Re: on using Logic Analyzers to understand and reverse engineer systems
An additional comment about telling clocks from serial buses. The spectrum analyzer display will tell you if the signal is modulated.
I was just getting started reverse engineering an Instek GDS-2072E when I let the magic smoke out. I'm sure I fried the NAND flash, but won't know about the Zynq 7010 until I have learned how to remove the 1 Gb NAND flash chip. The reason for doing the reverse engineering is I'm working on a FOSS DSO FW stack for COTS Zynq and Cyclone V based DSOs.
On the bright side, it gave me a use for old PCs. I'm going to practice removing and replacing chips until I can do it reliably.
|
Harke, if that ROM is U32, that's the attenuator assembly cal constants ROM. Definitely make a backup of that & tag it with the serial number of your attenuator assembly if you ever end up with multiple machines and are swapping parts. That ROM is also only good for an 8664A (presuming it is original to the unit). It won't have the constants for an 8665A or 8665B.
That doesn't help with the FRAC-N issue, though....
Where are you located? I am in California, USA. I likely have a good FRAC-N module to test with if shipping won't kill us...though there is one on eBay right now for $75USD + shipping.
Once you get through the 3,504,311 error, the cal errors might go away on their own. Is the unit running firmware v4.2.2? If not, definitely grab that as well.
If you have error 5,030,xxx, there's a Python script I wrote that mimics the HP BASIC code in the service manual - It cleared the errors with a few of mine.
Colby
|
Re: HP5335A fan replacement
I didn¡¯t mean to suggest that but it¡¯s an intriguing possibility.?
toggle quoted message
Show quoted text
On Mar 25, 2019, at 10:52 AM, peter bunge < bunge.pjp@...> wrote: PeterG are you suggesting that the fan I used was a temperature sensing fan and this is what I was seeing where it began to roar at full speed after about 30 minutes? This would explain a lot. It was a baffling problem that was solved by a 'proper' replacement but I never really understood what was happening and blamed it on 'bad' very old fans. PeterB
On Mon, Mar 25, 2019 at 10:05 AM Peter Gottlieb < hpnpilot@...> wrote: I have used temperature sensing fans as some are available with remote sensors which can be mounted on a hot item like a heat sink. You can¡¯t always do this, for example the Tek 2465 scope with multiple critical hybrids. In that case I look for fans designed for ultra quiet operation of computers. Their curves are very good for their noise levels but can¡¯t match the highest volume fans used in some gear. They will work for at least 95% of the applications out there though and make an ENORMOUS difference.
Peter
> On Mar 25, 2019, at 9:56 AM, John Ackermann N8UR <jra@...> wrote:
>
> Very interesting!? An only slightly related learning:
>
> A few months ago I wanted to replace the very noisy and over-rated fan on a power supply with something quieter.? I got a nice unit that was a drop-in replacement.? I wired it in and... the fan didn't turn, and the power supply wouldn't turn on.
>
> After some fiddling around, I discovered that there was a current sensor on the fan line and if the draw was less than X ma, the power supply would shut down.? The new fan didn't pull enough current.? So, I had to wire a power resistor in parallel to fool the sensor.? Then everything worked fine.
>
> (Yes, I could have gone inside to change the sense circuit, but it was a densely packed PCB buried in a large and complex box, with zero documentation.? I mounted a metal-cased resistor to the outside of the cabinet, so its heat goes away naturally.? I'm wasting a watt, but since the supply is rated for 2.5 KW (50A @ 50V), I can afford it. :-) )
>
> John
> ----
>
>> On 3/25/19 9:33 AM, peter bunge wrote:
>> *_A lesson to be learned about fans._*
>> *_HP8753B Fan Repair_**__*
>> *_Problem_*
>> The analyzer is not working
>> *_Symptoms_*
>> The fan is not running. The analyzer will not work if the fan is not running.
>> *_Repair_*
>> Fan removed and bench tested, not working. New fan ordered and fan replaced.
>> The original fan was 4124KX. The new fan is a 4124GX ordered from Ebay however it appears to be dated 01.82? (Jan 1982).
>> *_Problems continue_*
>> 1.Fan is OK when first turned on but seems to speed up a bit after 5 minutes or so.
>> 2.When the analyzer was moved the fan suddenly took off and ran very fast and noisy (or appeared to do so).
>> 3.Shutting down briefly and starting again repeated the problem up to slightly noisy and when the analyzer was banged gently on the bench the fan took off again.
>> 4.Tapping on Post Regulator board A8 causes the fan to take off but not until it has run for a while and has sped up on its own.
>> *_Analysis_*
>> The fan is rated for 18 and 30 volts DC. It is speed controlled by voltage from 2 regulators. See page 17/18 FAN TROUBLESHOOTING in the Power Supply troubleshooting section of 08753-90156.pdf.? It seems that the power to it is suddenly jumping to max. It may be the reason that the original fan failed.
>> Find and check the power supply.
>> *_Troubleshooting_*
>> 1.Check regulators U2 & U3 on A8. The pins are easily accessed but there is confusion as to the pin identification and the regulator types. The schematic shows both to be LM337 which is possible but the part numbers on the devices do not match and neither do any of the pins agree with data sheets (NatSemi & Motorola). The parts list shows U2 as LM337T and U3 as LM317T which makes more sense but not one pin number is correct on the schematic. Correct the schematic.
>> 2.Measure all regulator voltages while forcing the fan runaway. None of them change by more than a volt. The fan voltages are -14.7 and about +4 or +5 (varies as to be expected because it is controlled by the temperature.
>> 3.This is a real surprise. It looks like it could be the fan is defective. Remove the back cover for access to the fan connections and check the voltages there directly. At startup 18.64 volts, when first speed-up 20 volts and climbing slowly, when racing 20.15 volts and climbing slowly (20.17 after a few seconds). The slight change in voltage is probably due to varying loads. The fan is highly suspect.
>> 4.Remove the fan for bench test. Remove top & bottom covers, remove 2 screws top and 2 screws bottom of fan cover, remove 3 screws that hold the cover to the power supply, pull cover out. Remove 4 fan screws (2.5 mm hex driver), disconnect fan. Note fan blows into the analyzer. Pull connector off motherboard, use three pin header for connections to a power supply (red is to back of analyzer, marked below motherboard).
>> 5.Fan run on the bench at 20 volts for ten minutes. No sign of trouble. Sound does not change. Banging on desk has no effect. Fan seems OK.
>> 6.Perhaps there is a bad filter capacitor and ripple is feeding through the regulators and upsetting the brushless motor control circuitry. Connect fan and check with ¡®scope.
>> 7.When the noise starts the negative supply is OK but the positive supply has 200 mv peaks showing through. When the runaway condition is forced the negative supply is still OK but the positive supply ripple increases to about 400 mv. Not much considering the dramatic increase in noise.
>> 8.Scope the input to the regulator while fault is there. It is clean, the noise is possibly being modulated on the adjust terminal of the regulator. ¡®Scope pin 8 of U1. Looks clean.
>> 9.This does not appear to be actual speeding up of the fan but rather a mechanical vibration that resonates and peaks at a certain frequency. As the heat sink warms up the fan speed increases due to the fan control circuitry. When cold the blips on the positive supply are at 71Hz and 200 mv. As the heat sink warms the frequency increases to 80 Hz but when the A8 board is tapped to force the fault the frequency actually drops to 69.4 Hz and 300 mv amplitude.
>> 10.Try tapping A8 to force the fault then loosen all four fan screws a few turns (loose) then re-tightened them and the howl was gone. With all four fan screws loosened 1/2 turn I can get the fault to happen. It seems to be something happening on A8 due to mechanical feedback.
>> 11.Tapping the middle of the fan removes the fault. I can make the fault come and go by tapping the fan. Perhaps it is the fan.
>> 12.Run the fan in place but from a separate power supply and with the analyzer off. The fault comes and goes with tapping the fan. It is better behaved with only two screws on opposite corners tight so it looks like it is very sensitive to the mounting. It makes various sounds when gently pushed or pulled from the corners. When left it makes ticking sounds as though the fan blades were hitting something, which they are not.
>> 13.Two grommets were slit in half and mounted between the fan and the mounts. A tiny dot of Loctite was used on the ends of the screws and they were tightened a turn past touching. After ten minutes of quiet running the roar returned.
>> *_Conclusion_*
>> It is the fan. A new one is ordered.
>> My memory is vague as to how the problem was resolved but this analyzer was used continuously for another two months. I think the 4124GX fan replacement was determined to be noisy and the vendor sent a similar replacement which was also noisy. One of them was taken apart and oiled and used. See below.
>> *_Problems continue_*
>> The fan noise problem.
>> When the front of the analyzer is lifted or the side slapped, while running, the fan noise returned.
>> ¡¤A 4124 GX fan, with the screws exposed, is in the analyzer. It is not the original which was a 4124KX.
>> ¡¤I remember taking a fan apart and lubricating it so this may be the first one bought that was repaired while waiting for a replacement.
>> ¡¤There is a definitely dead fan, with the label removed, taken apart and put aside so I don¡¯t really know which fan wound up in this analyzer.
>> ¡¤There is a second 4124GX, with an intact label, which is also noisy in the analyzer. It is probably the first replacement ordered.
>> ¡¤Exchange fans with the one in my own 8753B.
>> ¡¤The noise follows the fan to my analyzer while my fan works quietly in Ron¡¯s analyzer.
>> *_Conclusion:_*
>> The 4124GX fans are both defective.
>> *_Fan Replacement for HP 8753B_*
>> Contacted ebm-papst USA, Copyright ? 2007 ebm-papst Canada Inc. (905) 420-3533
>> *sales@...* <mailto:sales@...>1800 Ironstone Manor, Unit 2, Pickering, Ontario L1W 3J9
>> Is there a replacement for the Papst 4124KX fan?
>> Yes, this has been replaced by the 4184NX.? Please contact any of the below distributors to purchase this fan.
>> 4184NX
>>
>> *ARROW ELECTRONICS* <>
>>
>> 334
>> 4184NX
>>
>> *ALLIED ELECTRONICS, INC.* <>
>>
>> 259
>> 4184NX
>>
>> *ELECTRO SONIC * <>
>>
>> 105
>> 4184NX
>>
>> *MASTER ELECTRONICS * <>
>>
>> 100
>> 4184NX
>>
>> ** <>
>>
>> 100
>> 4184NX
>>
>> *NEWARK ELEMENT14* <>
>>
>> 300
>> 4184NX
>>
>> *SAGER ELECTRONICS* <>
>>
>> 231
>> It looks like Electrosonic is the best choice. Contacted them.
>> They have 5 in stock in Toronto shipping is $10. UPS Ground, $62.78 each
>> *_Resolution:_*
>> It has been 5 years and the problem has not returned.
>> *_Lesson learned:_*
>> Don¡¯t go cheap on a repair!
>> On Sun, Mar 24, 2019 at 5:45 PM alwyn.seeds1 <a.seeds@... <mailto:a.seeds@...>> wrote:
>>? ? Dear All,
>>? ? I agree with Dave; flow at the required back pressure is what
>>? ? counts. If the fan design is not suitable, the hurricane of free air
>>? ? flow drops to a whisper when the fan is installed.
>>? ? Manufacturers' data sheets usually include graphs of air flow versus
>>? ? back pressure.
>>? ? Good manufacturers¡¯ fans are generally quieter than cheap computer
>>? ? fans; they also last longer.
>>? ? The Papst catalogue is enormous- most of the distributors only list
>>? ? a small part of the range.
>>? ? Regards,
>>? ? Alwyn
>>? ? _____________________________________________________
>>? ? Alwyn Seeds, Director
>>? ? SynOptika Ltd.,
>>? ? 114 Beaufort Street,
>>? ? London,
>>? ? SW3 6BU,
>>? ? England.
>>? ? Tel.: +44 (0) 20 7376 4110
>>? ? SynOptika Ltd., Registered in England and Wales: No. 04606737
>>? ? Registered Office: 114 Beaufort Street, London, SW3 6BU, United Kingdom.
>>? ? _____________________________________________________
>
>
>
|
Re: HP5335A fan replacement
PeterG are you suggesting that the fan I used was a temperature sensing fan and this is what I was seeing where it began to roar at full speed after about 30 minutes? This would explain a lot. It was a baffling problem that was solved by a 'proper' replacement but I never really understood what was happening and blamed it on 'bad' very old fans. PeterB
On Mon, Mar 25, 2019 at 10:05 AM Peter Gottlieb < hpnpilot@...> wrote: I have used temperature sensing fans as some are available with remote sensors which can be mounted on a hot item like a heat sink. You can¡¯t always do this, for example the Tek 2465 scope with multiple critical hybrids. In that case I look for fans designed for ultra quiet operation of computers. Their curves are very good for their noise levels but can¡¯t match the highest volume fans used in some gear. They will work for at least 95% of the applications out there though and make an ENORMOUS difference.
Peter
> On Mar 25, 2019, at 9:56 AM, John Ackermann N8UR <jra@...> wrote:
>
> Very interesting!? An only slightly related learning:
>
> A few months ago I wanted to replace the very noisy and over-rated fan on a power supply with something quieter.? I got a nice unit that was a drop-in replacement.? I wired it in and... the fan didn't turn, and the power supply wouldn't turn on.
>
> After some fiddling around, I discovered that there was a current sensor on the fan line and if the draw was less than X ma, the power supply would shut down.? The new fan didn't pull enough current.? So, I had to wire a power resistor in parallel to fool the sensor.? Then everything worked fine.
>
> (Yes, I could have gone inside to change the sense circuit, but it was a densely packed PCB buried in a large and complex box, with zero documentation.? I mounted a metal-cased resistor to the outside of the cabinet, so its heat goes away naturally.? I'm wasting a watt, but since the supply is rated for 2.5 KW (50A @ 50V), I can afford it. :-) )
>
> John
> ----
>
>> On 3/25/19 9:33 AM, peter bunge wrote:
>> *_A lesson to be learned about fans._*
>> *_HP8753B Fan Repair_**__*
>> *_Problem_*
>> The analyzer is not working
>> *_Symptoms_*
>> The fan is not running. The analyzer will not work if the fan is not running.
>> *_Repair_*
>> Fan removed and bench tested, not working. New fan ordered and fan replaced.
>> The original fan was 4124KX. The new fan is a 4124GX ordered from Ebay however it appears to be dated 01.82? (Jan 1982).
>> *_Problems continue_*
>> 1.Fan is OK when first turned on but seems to speed up a bit after 5 minutes or so.
>> 2.When the analyzer was moved the fan suddenly took off and ran very fast and noisy (or appeared to do so).
>> 3.Shutting down briefly and starting again repeated the problem up to slightly noisy and when the analyzer was banged gently on the bench the fan took off again.
>> 4.Tapping on Post Regulator board A8 causes the fan to take off but not until it has run for a while and has sped up on its own.
>> *_Analysis_*
>> The fan is rated for 18 and 30 volts DC. It is speed controlled by voltage from 2 regulators. See page 17/18 FAN TROUBLESHOOTING in the Power Supply troubleshooting section of 08753-90156.pdf.? It seems that the power to it is suddenly jumping to max. It may be the reason that the original fan failed.
>> Find and check the power supply.
>> *_Troubleshooting_*
>> 1.Check regulators U2 & U3 on A8. The pins are easily accessed but there is confusion as to the pin identification and the regulator types. The schematic shows both to be LM337 which is possible but the part numbers on the devices do not match and neither do any of the pins agree with data sheets (NatSemi & Motorola). The parts list shows U2 as LM337T and U3 as LM317T which makes more sense but not one pin number is correct on the schematic. Correct the schematic.
>> 2.Measure all regulator voltages while forcing the fan runaway. None of them change by more than a volt. The fan voltages are -14.7 and about +4 or +5 (varies as to be expected because it is controlled by the temperature.
>> 3.This is a real surprise. It looks like it could be the fan is defective. Remove the back cover for access to the fan connections and check the voltages there directly. At startup 18.64 volts, when first speed-up 20 volts and climbing slowly, when racing 20.15 volts and climbing slowly (20.17 after a few seconds). The slight change in voltage is probably due to varying loads. The fan is highly suspect.
>> 4.Remove the fan for bench test. Remove top & bottom covers, remove 2 screws top and 2 screws bottom of fan cover, remove 3 screws that hold the cover to the power supply, pull cover out. Remove 4 fan screws (2.5 mm hex driver), disconnect fan. Note fan blows into the analyzer. Pull connector off motherboard, use three pin header for connections to a power supply (red is to back of analyzer, marked below motherboard).
>> 5.Fan run on the bench at 20 volts for ten minutes. No sign of trouble. Sound does not change. Banging on desk has no effect. Fan seems OK.
>> 6.Perhaps there is a bad filter capacitor and ripple is feeding through the regulators and upsetting the brushless motor control circuitry. Connect fan and check with ¡®scope.
>> 7.When the noise starts the negative supply is OK but the positive supply has 200 mv peaks showing through. When the runaway condition is forced the negative supply is still OK but the positive supply ripple increases to about 400 mv. Not much considering the dramatic increase in noise.
>> 8.Scope the input to the regulator while fault is there. It is clean, the noise is possibly being modulated on the adjust terminal of the regulator. ¡®Scope pin 8 of U1. Looks clean.
>> 9.This does not appear to be actual speeding up of the fan but rather a mechanical vibration that resonates and peaks at a certain frequency. As the heat sink warms up the fan speed increases due to the fan control circuitry. When cold the blips on the positive supply are at 71Hz and 200 mv. As the heat sink warms the frequency increases to 80 Hz but when the A8 board is tapped to force the fault the frequency actually drops to 69.4 Hz and 300 mv amplitude.
>> 10.Try tapping A8 to force the fault then loosen all four fan screws a few turns (loose) then re-tightened them and the howl was gone. With all four fan screws loosened 1/2 turn I can get the fault to happen. It seems to be something happening on A8 due to mechanical feedback.
>> 11.Tapping the middle of the fan removes the fault. I can make the fault come and go by tapping the fan. Perhaps it is the fan.
>> 12.Run the fan in place but from a separate power supply and with the analyzer off. The fault comes and goes with tapping the fan. It is better behaved with only two screws on opposite corners tight so it looks like it is very sensitive to the mounting. It makes various sounds when gently pushed or pulled from the corners. When left it makes ticking sounds as though the fan blades were hitting something, which they are not.
>> 13.Two grommets were slit in half and mounted between the fan and the mounts. A tiny dot of Loctite was used on the ends of the screws and they were tightened a turn past touching. After ten minutes of quiet running the roar returned.
>> *_Conclusion_*
>> It is the fan. A new one is ordered.
>> My memory is vague as to how the problem was resolved but this analyzer was used continuously for another two months. I think the 4124GX fan replacement was determined to be noisy and the vendor sent a similar replacement which was also noisy. One of them was taken apart and oiled and used. See below.
>> *_Problems continue_*
>> The fan noise problem.
>> When the front of the analyzer is lifted or the side slapped, while running, the fan noise returned.
>> ¡¤A 4124 GX fan, with the screws exposed, is in the analyzer. It is not the original which was a 4124KX.
>> ¡¤I remember taking a fan apart and lubricating it so this may be the first one bought that was repaired while waiting for a replacement.
>> ¡¤There is a definitely dead fan, with the label removed, taken apart and put aside so I don¡¯t really know which fan wound up in this analyzer.
>> ¡¤There is a second 4124GX, with an intact label, which is also noisy in the analyzer. It is probably the first replacement ordered.
>> ¡¤Exchange fans with the one in my own 8753B.
>> ¡¤The noise follows the fan to my analyzer while my fan works quietly in Ron¡¯s analyzer.
>> *_Conclusion:_*
>> The 4124GX fans are both defective.
>> *_Fan Replacement for HP 8753B_*
>> Contacted ebm-papst USA, Copyright ? 2007 ebm-papst Canada Inc. (905) 420-3533
>> *sales@...* <mailto:sales@...>1800 Ironstone Manor, Unit 2, Pickering, Ontario L1W 3J9
>> Is there a replacement for the Papst 4124KX fan?
>> Yes, this has been replaced by the 4184NX.? Please contact any of the below distributors to purchase this fan.
>> 4184NX
>>
>> *ARROW ELECTRONICS* <>
>>
>> 334
>> 4184NX
>>
>> *ALLIED ELECTRONICS, INC.* <>
>>
>> 259
>> 4184NX
>>
>> *ELECTRO SONIC * <>
>>
>> 105
>> 4184NX
>>
>> *MASTER ELECTRONICS * <>
>>
>> 100
>> 4184NX
>>
>> ** <>
>>
>> 100
>> 4184NX
>>
>> *NEWARK ELEMENT14* <>
>>
>> 300
>> 4184NX
>>
>> *SAGER ELECTRONICS* <>
>>
>> 231
>> It looks like Electrosonic is the best choice. Contacted them.
>> They have 5 in stock in Toronto shipping is $10. UPS Ground, $62.78 each
>> *_Resolution:_*
>> It has been 5 years and the problem has not returned.
>> *_Lesson learned:_*
>> Don¡¯t go cheap on a repair!
>> On Sun, Mar 24, 2019 at 5:45 PM alwyn.seeds1 <a.seeds@... <mailto:a.seeds@...>> wrote:
>>? ? Dear All,
>>? ? I agree with Dave; flow at the required back pressure is what
>>? ? counts. If the fan design is not suitable, the hurricane of free air
>>? ? flow drops to a whisper when the fan is installed.
>>? ? Manufacturers' data sheets usually include graphs of air flow versus
>>? ? back pressure.
>>? ? Good manufacturers¡¯ fans are generally quieter than cheap computer
>>? ? fans; they also last longer.
>>? ? The Papst catalogue is enormous- most of the distributors only list
>>? ? a small part of the range.
>>? ? Regards,
>>? ? Alwyn
>>? ? _____________________________________________________
>>? ? Alwyn Seeds, Director
>>? ? SynOptika Ltd.,
>>? ? 114 Beaufort Street,
>>? ? London,
>>? ? SW3 6BU,
>>? ? England.
>>? ? Tel.: +44 (0) 20 7376 4110
>>? ? SynOptika Ltd., Registered in England and Wales: No. 04606737
>>? ? Registered Office: 114 Beaufort Street, London, SW3 6BU, United Kingdom.
>>? ? _____________________________________________________
>
>
>
|
Re: on using Logic Analyzers to understand and reverse engineer systems
On Mon, 25 Mar 2019 05:16:46 -0700, you wrote: I can't speak to doing this in HW, but I spent many years fixing large (>50KLOC) programs which had *no* comments other than the author's name. In the aggregate about 2 million lines. Ouch. This is when you shoot the programmer, and all of his managment that allowed any of this to happen.
I relied on my knowledge of what the programs did and all the ways you could do that. I've actually had to pull out professional papers and read the equations to figure out if a DSP operation was in the time or frequency domain.
General strategy was to look at the overall structure, identify key events in the operation and then guess at tests cases which would isolate the fault. Most importantly, I never spent more than 3-4 hours without taking a long break. I came to refer to the crest of an overpass about 5 minutes from work as "the point of epiphany" because of the number of times I left work early after a fruitless afternoon hunting for the problem and when I reached the point of epiphany it would suddenly become clear. I'd go in the next morning and have it fixed in 15-20 minutes.
Yep, with any microprocessor design, you have to know the processor family, then look at the interrupt vectors, and then start to disassemble that code. If you have table driven systems, then it's experimentally disassemble some of the addresses you think are in the table. Early stuff is full of "clever" schemes that have long outlived their usefulness. If it's disassembling a higher level language generated program, you're probably lucky if you can figure out the routines. All of what I've tried is assembly language programmed.
I'd love to see a checklist of questions related to reverse engineering HW designs. With the shift to OEMs not providing component level data, most current T&M gear will become unrepairable except by reverse engineering. I just got a Tek 11801 going after a couple of days struggling with the lack of service data. And it was a *simple* problem! A dead NVRAM battery. I don't know that I could isolate a fault on a board with 120 TTL parts and several undocumented ASICs with multiple ROMs and over 100 control signals on ribbon cables. If I did succeed I'm sure it would take many days to accomplish. So it would have to be *very* valuable.
I'd be willing to see if I could help on that. Depends on what you want to do in reverse engineering, though. Fault find is one thing, re-engineer is another....
I came across someone who repaired industrial process control equipment for a living for whom repairing undocumented systems from companies that no longer existed was a common assignment. He made a handsome living doing it.
Not that we'd know any companies like that..... Harvey
|
Re: on using Logic Analyzers to understand and reverse engineer systems
On Mon, 25 Mar 2019 00:49:55 -0400, you wrote: In addition to debugging my own designs, one of my common use cases for
using logic analyzers is to peer into designs of existing hardware
systems. These systems are sometimes poorly documented with schematics
and theory of operation info rarely available or reliable. They
sometimes employ custom ASICs, and just having a pinout with names of
signals is all I have to go on. Now this can be nasty....
Where there are datasheets available for the chips, then it's not too
bad to understand the signals, although I might not have direct
experience with how those chips work, or have a real appreciation for
the data format.
Well, there are only so many formats to consider. In terms of parallel, you look for lines changing, then some sort of control strobe. In terms of serial, you look for known data formats, SPI (a method, not a format) is common for fast data. I2C generally is used to connect at a subsystem level and is slower. SPI is 3 lines + a select, I2C is only two. Some of my struggles are related to multiple clock domains, sometimes to
bit-ordering(is D0 the MSB or LSB?), sometimes a pin turns out to be
serial data, sometimes there are time-multiplexed systems using slots,
and so on. All of which can happen. It does sound as if you're looking at some complicated stuff.
With busses, I often don't realize "who's doing the talking?"
True enough. Unless you know the chips, you'll have a problem. it's possible that the HP current tracer probe would help on TTL stuff, where there's enough current to maybe take a look and trigger the probe (1 ma is most sensitive). For CMOS? also possible, but I don't know how much. Otherwise, you're trying to look at voltage levels and guess.
It's not that I don't eventually get there, but it seems that I stare
endlessly at logic analyzer traces until I finally have the Eureka
moment and see what I'm doing wrong.
That may be about it. I swear I spend more time trying to understand whether I have everything
configured properly than I do actually putting the data available to
good use. But without proper configuration, is the data any good?
What do you do? Have some orderly organizational method? Look at the pin
on a 'scope and first determine all the clocks in play, which chips are
in which domains? Determine if they are serial vs just being a fast chip
enable?
Well, I'll admit that I cheat a bit. I debug things I've designed, so I've got good information on all the chips. Most of my problems are in the category of "is it doing what I want it to or is it doing what I told it to do?" I'd start looking at the pins on a scope, which gives me both logic level and an idea of data rates, but little interrelationship of signals. It's often more useful to start at the outside and go in. If you have some peripheral chips that you can identify, if only by function, (memory, I/O, oscillators, data transceivers, that kind of thing), then you at least have an idea of what they should do and what some of the expected signals are. Sometimes you need to identify the black box by throwing tennis balls in a dark room and seing what says "ouch".
I usually lean towards using Timing mode because the change-in-data
sometimes gives me information about whether there are intermittent
timing violations occurring and because I don't have an appreciation for
the significance of the multiple clocks in the design. This of course
results in many data duplicates which compounds the "too much data" problem.
You may be trying to do it all at once. Analyzing an entire screen of timing mode traces is going to be nastier than analyzing just a few. I'd tend to the subsystem approach, and see what that tells me. As an example, take a common (inexpensive, SPI display with a touch screen). You know what it is, and can hopefully get close on the pin connections if not exact. In this case, you don't care what the data is, but you do care where it comes from. While SPI has three lines (clock, MISO, MOSI (master out slave in)), it also requires a chip select line for that individual chip. The three lines can be common to any number of chips, but the select line is unique, and is driven by only one thing. Ditto for things like SPI memory. Naturally, I'm a bit biased towards microprocessor driven systems, since that'a a lot of what I do. A tentative block diagram is a good start, not functional as much as simply chips....
Many designs seem to employ "tricks" where taking advantage of different
chip propagation delays based on the technology used help prevent race
conditions. Or say, knowing that the data isn't actually consumed until
later, so having data valid in the first half of the clock cycle isn't
necessary....and so on.
That's typical 6800, 6502 style microprocessor interfacing. It's also typical that data is allowed to change, say, when the clock is low, and must be stable when the clock is high. That's for writing data since the clock high is the write cycle clock. For read cycles, typical memory allows the address to set up during the low clock period, then supplies data during the clock high, where the data is guaranteed to be stable n nanoseconds before the clock low to high transition. These are pretty standard conventions.
Do others struggle with this type of problem? Have you found some online
resources or books that address this type of thing?
Lots of experience can help. Knowing digital hardware design helps, too. For what I'm doing (currently, one of the projects is reverse engineering the CPU board on a Tektronix DM5010), I've gotten the firmware ROMS, and the service manual, so I know where things are supposed to be. Disassembling at least parts of the ROM code shows me where some of the interface code is, but there's only so much that I (in theory) need. Since I'm using a C driven microprocessor with an FPGA to fake the hardware interfaces (front panel and logic to go to the actual DMM circuitry), all I need to know is how it should work, and then design the FPGA to handle that. The logic analyzer will be used to verify the circuit behavior. I've designed the board so that there are logic pod connectors on it, thus allowing easy access to circuits. This is somewhat of reverse engineering, but not quite at the level you're heading for. Any of this help? Harvey
Thanks,
Keith
|
Re: HP5335A fan replacement
In this case, the power supply is for a 500W 2M amplifier. Maximum draw is about 20 amps @ 50 volts, and that's never more than 50% duty cycle for <5 minute periods, so the 50A continuous supply is vastly over-rated. With the original noisy fan, I swear the output air felt *colder* than ambient! It could *almost* work with no fan, but I'm not prepared to go that far.
John
----
toggle quoted message
Show quoted text
On 3/25/19 10:05 AM, Peter Gottlieb wrote: I have used temperature sensing fans as some are available with remote sensors which can be mounted on a hot item like a heat sink. You can¡¯t always do this, for example the Tek 2465 scope with multiple critical hybrids. In that case I look for fans designed for ultra quiet operation of computers. Their curves are very good for their noise levels but can¡¯t match the highest volume fans used in some gear. They will work for at least 95% of the applications out there though and make an ENORMOUS difference.
Peter
On Mar 25, 2019, at 9:56 AM, John Ackermann N8UR <jra@...> wrote:
Very interesting! An only slightly related learning:
A few months ago I wanted to replace the very noisy and over-rated fan on a power supply with something quieter. I got a nice unit that was a drop-in replacement. I wired it in and... the fan didn't turn, and the power supply wouldn't turn on.
After some fiddling around, I discovered that there was a current sensor on the fan line and if the draw was less than X ma, the power supply would shut down. The new fan didn't pull enough current. So, I had to wire a power resistor in parallel to fool the sensor. Then everything worked fine.
(Yes, I could have gone inside to change the sense circuit, but it was a densely packed PCB buried in a large and complex box, with zero documentation. I mounted a metal-cased resistor to the outside of the cabinet, so its heat goes away naturally. I'm wasting a watt, but since the supply is rated for 2.5 KW (50A @ 50V), I can afford it. :-) )
John
----
On 3/25/19 9:33 AM, peter bunge wrote:
*_A lesson to be learned about fans._*
*_HP8753B Fan Repair_**__*
*_Problem_*
The analyzer is not working
*_Symptoms_*
The fan is not running. The analyzer will not work if the fan is not running.
*_Repair_*
Fan removed and bench tested, not working. New fan ordered and fan replaced.
The original fan was 4124KX. The new fan is a 4124GX ordered from Ebay however it appears to be dated 01.82 (Jan 1982).
*_Problems continue_*
1.Fan is OK when first turned on but seems to speed up a bit after 5 minutes or so.
2.When the analyzer was moved the fan suddenly took off and ran very fast and noisy (or appeared to do so).
3.Shutting down briefly and starting again repeated the problem up to slightly noisy and when the analyzer was banged gently on the bench the fan took off again.
4.Tapping on Post Regulator board A8 causes the fan to take off but not until it has run for a while and has sped up on its own.
*_Analysis_*
The fan is rated for 18 and 30 volts DC. It is speed controlled by voltage from 2 regulators. See page 17/18 FAN TROUBLESHOOTING in the Power Supply troubleshooting section of 08753-90156.pdf. It seems that the power to it is suddenly jumping to max. It may be the reason that the original fan failed.
Find and check the power supply.
*_Troubleshooting_*
1.Check regulators U2 & U3 on A8. The pins are easily accessed but there is confusion as to the pin identification and the regulator types. The schematic shows both to be LM337 which is possible but the part numbers on the devices do not match and neither do any of the pins agree with data sheets (NatSemi & Motorola). The parts list shows U2 as LM337T and U3 as LM317T which makes more sense but not one pin number is correct on the schematic. Correct the schematic.
2.Measure all regulator voltages while forcing the fan runaway. None of them change by more than a volt. The fan voltages are -14.7 and about +4 or +5 (varies as to be expected because it is controlled by the temperature.
3.This is a real surprise. It looks like it could be the fan is defective. Remove the back cover for access to the fan connections and check the voltages there directly. At startup 18.64 volts, when first speed-up 20 volts and climbing slowly, when racing 20.15 volts and climbing slowly (20.17 after a few seconds). The slight change in voltage is probably due to varying loads. The fan is highly suspect.
4.Remove the fan for bench test. Remove top & bottom covers, remove 2 screws top and 2 screws bottom of fan cover, remove 3 screws that hold the cover to the power supply, pull cover out. Remove 4 fan screws (2.5 mm hex driver), disconnect fan. Note fan blows into the analyzer. Pull connector off motherboard, use three pin header for connections to a power supply (red is to back of analyzer, marked below motherboard).
5.Fan run on the bench at 20 volts for ten minutes. No sign of trouble. Sound does not change. Banging on desk has no effect. Fan seems OK.
6.Perhaps there is a bad filter capacitor and ripple is feeding through the regulators and upsetting the brushless motor control circuitry. Connect fan and check with ¡®scope.
7.When the noise starts the negative supply is OK but the positive supply has 200 mv peaks showing through. When the runaway condition is forced the negative supply is still OK but the positive supply ripple increases to about 400 mv. Not much considering the dramatic increase in noise.
8.Scope the input to the regulator while fault is there. It is clean, the noise is possibly being modulated on the adjust terminal of the regulator. ¡®Scope pin 8 of U1. Looks clean.
9.This does not appear to be actual speeding up of the fan but rather a mechanical vibration that resonates and peaks at a certain frequency. As the heat sink warms up the fan speed increases due to the fan control circuitry. When cold the blips on the positive supply are at 71Hz and 200 mv. As the heat sink warms the frequency increases to 80 Hz but when the A8 board is tapped to force the fault the frequency actually drops to 69.4 Hz and 300 mv amplitude.
10.Try tapping A8 to force the fault then loosen all four fan screws a few turns (loose) then re-tightened them and the howl was gone. With all four fan screws loosened 1/2 turn I can get the fault to happen. It seems to be something happening on A8 due to mechanical feedback.
11.Tapping the middle of the fan removes the fault. I can make the fault come and go by tapping the fan. Perhaps it is the fan.
12.Run the fan in place but from a separate power supply and with the analyzer off. The fault comes and goes with tapping the fan. It is better behaved with only two screws on opposite corners tight so it looks like it is very sensitive to the mounting. It makes various sounds when gently pushed or pulled from the corners. When left it makes ticking sounds as though the fan blades were hitting something, which they are not.
13.Two grommets were slit in half and mounted between the fan and the mounts. A tiny dot of Loctite was used on the ends of the screws and they were tightened a turn past touching. After ten minutes of quiet running the roar returned.
*_Conclusion_*
It is the fan. A new one is ordered.
My memory is vague as to how the problem was resolved but this analyzer was used continuously for another two months. I think the 4124GX fan replacement was determined to be noisy and the vendor sent a similar replacement which was also noisy. One of them was taken apart and oiled and used. See below.
*_Problems continue_*
The fan noise problem.
When the front of the analyzer is lifted or the side slapped, while running, the fan noise returned.
¡¤A 4124 GX fan, with the screws exposed, is in the analyzer. It is not the original which was a 4124KX.
¡¤I remember taking a fan apart and lubricating it so this may be the first one bought that was repaired while waiting for a replacement.
¡¤There is a definitely dead fan, with the label removed, taken apart and put aside so I don¡¯t really know which fan wound up in this analyzer.
¡¤There is a second 4124GX, with an intact label, which is also noisy in the analyzer. It is probably the first replacement ordered.
¡¤Exchange fans with the one in my own 8753B.
¡¤The noise follows the fan to my analyzer while my fan works quietly in Ron¡¯s analyzer.
*_Conclusion:_*
The 4124GX fans are both defective.
*_Fan Replacement for HP 8753B_*
Contacted ebm-papst USA, Copyright ? 2007 ebm-papst Canada Inc. (905) 420-3533
*sales@...* <mailto:sales@...>1800 Ironstone Manor, Unit 2, Pickering, Ontario L1W 3J9
Is there a replacement for the Papst 4124KX fan?
Yes, this has been replaced by the 4184NX. Please contact any of the below distributors to purchase this fan.
4184NX
*ARROW ELECTRONICS* <>
334
4184NX
*ALLIED ELECTRONICS, INC.* <>
259
4184NX
*ELECTRO SONIC * <>
105
4184NX
*MASTER ELECTRONICS * <>
100
4184NX
*ONLINECOMPONENTS.COM* <>
100
4184NX
*NEWARK ELEMENT14* <>
300
4184NX
*SAGER ELECTRONICS* <>
231
It looks like Electrosonic is the best choice. Contacted them.
They have 5 in stock in Toronto shipping is $10. UPS Ground, $62.78 each
*_Resolution:_*
It has been 5 years and the problem has not returned.
*_Lesson learned:_*
Don¡¯t go cheap on a repair!
On Sun, Mar 24, 2019 at 5:45 PM alwyn.seeds1 <a.seeds@... <mailto:a.seeds@...>> wrote:
Dear All,
I agree with Dave; flow at the required back pressure is what
counts. If the fan design is not suitable, the hurricane of free air
flow drops to a whisper when the fan is installed.
Manufacturers' data sheets usually include graphs of air flow versus
back pressure.
Good manufacturers¡¯ fans are generally quieter than cheap computer
fans; they also last longer.
The Papst catalogue is enormous- most of the distributors only list
a small part of the range.
Regards,
Alwyn
_____________________________________________________
Alwyn Seeds, Director
SynOptika Ltd.,
114 Beaufort Street,
London,
SW3 6BU,
England.
Tel.: +44 (0) 20 7376 4110
SynOptika Ltd., Registered in England and Wales: No. 04606737
Registered Office: 114 Beaufort Street, London, SW3 6BU, United Kingdom.
_____________________________________________________
|
Re: HP5335A fan replacement
I have used temperature sensing fans as some are available with remote sensors which can be mounted on a hot item like a heat sink. You can¡¯t always do this, for example the Tek 2465 scope with multiple critical hybrids. In that case I look for fans designed for ultra quiet operation of computers. Their curves are very good for their noise levels but can¡¯t match the highest volume fans used in some gear. They will work for at least 95% of the applications out there though and make an ENORMOUS difference.
Peter
toggle quoted message
Show quoted text
On Mar 25, 2019, at 9:56 AM, John Ackermann N8UR <jra@...> wrote:
Very interesting! An only slightly related learning:
A few months ago I wanted to replace the very noisy and over-rated fan on a power supply with something quieter. I got a nice unit that was a drop-in replacement. I wired it in and... the fan didn't turn, and the power supply wouldn't turn on.
After some fiddling around, I discovered that there was a current sensor on the fan line and if the draw was less than X ma, the power supply would shut down. The new fan didn't pull enough current. So, I had to wire a power resistor in parallel to fool the sensor. Then everything worked fine.
(Yes, I could have gone inside to change the sense circuit, but it was a densely packed PCB buried in a large and complex box, with zero documentation. I mounted a metal-cased resistor to the outside of the cabinet, so its heat goes away naturally. I'm wasting a watt, but since the supply is rated for 2.5 KW (50A @ 50V), I can afford it. :-) )
John
----
On 3/25/19 9:33 AM, peter bunge wrote:
*_A lesson to be learned about fans._*
*_HP8753B Fan Repair_**__*
*_Problem_*
The analyzer is not working
*_Symptoms_*
The fan is not running. The analyzer will not work if the fan is not running.
*_Repair_*
Fan removed and bench tested, not working. New fan ordered and fan replaced.
The original fan was 4124KX. The new fan is a 4124GX ordered from Ebay however it appears to be dated 01.82 (Jan 1982).
*_Problems continue_*
1.Fan is OK when first turned on but seems to speed up a bit after 5 minutes or so.
2.When the analyzer was moved the fan suddenly took off and ran very fast and noisy (or appeared to do so).
3.Shutting down briefly and starting again repeated the problem up to slightly noisy and when the analyzer was banged gently on the bench the fan took off again.
4.Tapping on Post Regulator board A8 causes the fan to take off but not until it has run for a while and has sped up on its own.
*_Analysis_*
The fan is rated for 18 and 30 volts DC. It is speed controlled by voltage from 2 regulators. See page 17/18 FAN TROUBLESHOOTING in the Power Supply troubleshooting section of 08753-90156.pdf. It seems that the power to it is suddenly jumping to max. It may be the reason that the original fan failed.
Find and check the power supply.
*_Troubleshooting_*
1.Check regulators U2 & U3 on A8. The pins are easily accessed but there is confusion as to the pin identification and the regulator types. The schematic shows both to be LM337 which is possible but the part numbers on the devices do not match and neither do any of the pins agree with data sheets (NatSemi & Motorola). The parts list shows U2 as LM337T and U3 as LM317T which makes more sense but not one pin number is correct on the schematic. Correct the schematic.
2.Measure all regulator voltages while forcing the fan runaway. None of them change by more than a volt. The fan voltages are -14.7 and about +4 or +5 (varies as to be expected because it is controlled by the temperature.
3.This is a real surprise. It looks like it could be the fan is defective. Remove the back cover for access to the fan connections and check the voltages there directly. At startup 18.64 volts, when first speed-up 20 volts and climbing slowly, when racing 20.15 volts and climbing slowly (20.17 after a few seconds). The slight change in voltage is probably due to varying loads. The fan is highly suspect.
4.Remove the fan for bench test. Remove top & bottom covers, remove 2 screws top and 2 screws bottom of fan cover, remove 3 screws that hold the cover to the power supply, pull cover out. Remove 4 fan screws (2.5 mm hex driver), disconnect fan. Note fan blows into the analyzer. Pull connector off motherboard, use three pin header for connections to a power supply (red is to back of analyzer, marked below motherboard).
5.Fan run on the bench at 20 volts for ten minutes. No sign of trouble. Sound does not change. Banging on desk has no effect. Fan seems OK.
6.Perhaps there is a bad filter capacitor and ripple is feeding through the regulators and upsetting the brushless motor control circuitry. Connect fan and check with ¡®scope.
7.When the noise starts the negative supply is OK but the positive supply has 200 mv peaks showing through. When the runaway condition is forced the negative supply is still OK but the positive supply ripple increases to about 400 mv. Not much considering the dramatic increase in noise.
8.Scope the input to the regulator while fault is there. It is clean, the noise is possibly being modulated on the adjust terminal of the regulator. ¡®Scope pin 8 of U1. Looks clean.
9.This does not appear to be actual speeding up of the fan but rather a mechanical vibration that resonates and peaks at a certain frequency. As the heat sink warms up the fan speed increases due to the fan control circuitry. When cold the blips on the positive supply are at 71Hz and 200 mv. As the heat sink warms the frequency increases to 80 Hz but when the A8 board is tapped to force the fault the frequency actually drops to 69.4 Hz and 300 mv amplitude.
10.Try tapping A8 to force the fault then loosen all four fan screws a few turns (loose) then re-tightened them and the howl was gone. With all four fan screws loosened 1/2 turn I can get the fault to happen. It seems to be something happening on A8 due to mechanical feedback.
11.Tapping the middle of the fan removes the fault. I can make the fault come and go by tapping the fan. Perhaps it is the fan.
12.Run the fan in place but from a separate power supply and with the analyzer off. The fault comes and goes with tapping the fan. It is better behaved with only two screws on opposite corners tight so it looks like it is very sensitive to the mounting. It makes various sounds when gently pushed or pulled from the corners. When left it makes ticking sounds as though the fan blades were hitting something, which they are not.
13.Two grommets were slit in half and mounted between the fan and the mounts. A tiny dot of Loctite was used on the ends of the screws and they were tightened a turn past touching. After ten minutes of quiet running the roar returned.
*_Conclusion_*
It is the fan. A new one is ordered.
My memory is vague as to how the problem was resolved but this analyzer was used continuously for another two months. I think the 4124GX fan replacement was determined to be noisy and the vendor sent a similar replacement which was also noisy. One of them was taken apart and oiled and used. See below.
*_Problems continue_*
The fan noise problem.
When the front of the analyzer is lifted or the side slapped, while running, the fan noise returned.
¡¤A 4124 GX fan, with the screws exposed, is in the analyzer. It is not the original which was a 4124KX.
¡¤I remember taking a fan apart and lubricating it so this may be the first one bought that was repaired while waiting for a replacement.
¡¤There is a definitely dead fan, with the label removed, taken apart and put aside so I don¡¯t really know which fan wound up in this analyzer.
¡¤There is a second 4124GX, with an intact label, which is also noisy in the analyzer. It is probably the first replacement ordered.
¡¤Exchange fans with the one in my own 8753B.
¡¤The noise follows the fan to my analyzer while my fan works quietly in Ron¡¯s analyzer.
*_Conclusion:_*
The 4124GX fans are both defective.
*_Fan Replacement for HP 8753B_*
Contacted ebm-papst USA, Copyright ? 2007 ebm-papst Canada Inc. (905) 420-3533
*sales@...* <mailto:sales@...>1800 Ironstone Manor, Unit 2, Pickering, Ontario L1W 3J9
Is there a replacement for the Papst 4124KX fan?
Yes, this has been replaced by the 4184NX. Please contact any of the below distributors to purchase this fan.
4184NX
*ARROW ELECTRONICS* <>
334
4184NX
*ALLIED ELECTRONICS, INC.* <>
259
4184NX
*ELECTRO SONIC * <>
105
4184NX
*MASTER ELECTRONICS * <>
100
4184NX
*ONLINECOMPONENTS.COM* <>
100
4184NX
*NEWARK ELEMENT14* <>
300
4184NX
*SAGER ELECTRONICS* <>
231
It looks like Electrosonic is the best choice. Contacted them.
They have 5 in stock in Toronto shipping is $10. UPS Ground, $62.78 each
*_Resolution:_*
It has been 5 years and the problem has not returned.
*_Lesson learned:_*
Don¡¯t go cheap on a repair!
On Sun, Mar 24, 2019 at 5:45 PM alwyn.seeds1 <a.seeds@... <mailto:a.seeds@...>> wrote:
Dear All,
I agree with Dave; flow at the required back pressure is what
counts. If the fan design is not suitable, the hurricane of free air
flow drops to a whisper when the fan is installed.
Manufacturers' data sheets usually include graphs of air flow versus
back pressure.
Good manufacturers¡¯ fans are generally quieter than cheap computer
fans; they also last longer.
The Papst catalogue is enormous- most of the distributors only list
a small part of the range.
Regards,
Alwyn
_____________________________________________________
Alwyn Seeds, Director
SynOptika Ltd.,
114 Beaufort Street,
London,
SW3 6BU,
England.
Tel.: +44 (0) 20 7376 4110
SynOptika Ltd., Registered in England and Wales: No. 04606737
Registered Office: 114 Beaufort Street, London, SW3 6BU, United Kingdom.
_____________________________________________________
|
Hello Group,
I have obtained a very cheap HP8664A, but, of course, with issues. Ser no starts with 3744 but internal date codes show well into 2000. It has options 010, 004 and an oven. It has a special high shielding case. Each time upon power-on it will calibrate itself, a long operation. Then it will give the Result code: 3,504,311. MSSG gives: Hardware Failures; 5 and 1 and Calibration Errors: 2 and 1. The repair manual I have suggests to look at A10, the fractional N module. I opened it and found a beautiful pcb without any visual or smelling anomalies. The power supply is fine. On my SA (HP8562A): between about 400 MHz and 2.9 GHz everything is fine. Nice clean stable signal, frequency and amplitude spot-on. Can be modulated as well. Outside this range the signal is clearly out of lock=useless. I checked the Eeprom (28HC64) at the interconnection board and it reads fine on my Eprom programmer.
ILD test 320 gives the code; +3,504,311, quite similar to the calibration routine. The HW Failures and Calibration Errors given above can not be found in the manuals I have.
Any help in bringing this beautiful instrument back into service in my shack is highly appreciated.
Kind regards,
Harke
|
Re: HP5335A fan replacement
Very interesting! An only slightly related learning:
A few months ago I wanted to replace the very noisy and over-rated fan on a power supply with something quieter. I got a nice unit that was a drop-in replacement. I wired it in and... the fan didn't turn, and the power supply wouldn't turn on.
After some fiddling around, I discovered that there was a current sensor on the fan line and if the draw was less than X ma, the power supply would shut down. The new fan didn't pull enough current. So, I had to wire a power resistor in parallel to fool the sensor. Then everything worked fine.
(Yes, I could have gone inside to change the sense circuit, but it was a densely packed PCB buried in a large and complex box, with zero documentation. I mounted a metal-cased resistor to the outside of the cabinet, so its heat goes away naturally. I'm wasting a watt, but since the supply is rated for 2.5 KW (50A @ 50V), I can afford it. :-) )
John
----
toggle quoted message
Show quoted text
On 3/25/19 9:33 AM, peter bunge wrote: *_A lesson to be learned about fans._*
*_HP8753B Fan Repair_**__*
*_Problem_*
The analyzer is not working
*_Symptoms_*
The fan is not running. The analyzer will not work if the fan is not running.
*_Repair_*
Fan removed and bench tested, not working. New fan ordered and fan replaced.
The original fan was 4124KX. The new fan is a 4124GX ordered from Ebay however it appears to be dated 01.82? (Jan 1982).
*_Problems continue_*
1.Fan is OK when first turned on but seems to speed up a bit after 5 minutes or so.
2.When the analyzer was moved the fan suddenly took off and ran very fast and noisy (or appeared to do so).
3.Shutting down briefly and starting again repeated the problem up to slightly noisy and when the analyzer was banged gently on the bench the fan took off again.
4.Tapping on Post Regulator board A8 causes the fan to take off but not until it has run for a while and has sped up on its own.
*_Analysis_*
The fan is rated for 18 and 30 volts DC. It is speed controlled by voltage from 2 regulators. See page 17/18 FAN TROUBLESHOOTING in the Power Supply troubleshooting section of 08753-90156.pdf.? It seems that the power to it is suddenly jumping to max. It may be the reason that the original fan failed.
Find and check the power supply.
*_Troubleshooting_*
1.Check regulators U2 & U3 on A8. The pins are easily accessed but there is confusion as to the pin identification and the regulator types. The schematic shows both to be LM337 which is possible but the part numbers on the devices do not match and neither do any of the pins agree with data sheets (NatSemi & Motorola). The parts list shows U2 as LM337T and U3 as LM317T which makes more sense but not one pin number is correct on the schematic. Correct the schematic.
2.Measure all regulator voltages while forcing the fan runaway. None of them change by more than a volt. The fan voltages are -14.7 and about +4 or +5 (varies as to be expected because it is controlled by the temperature.
3.This is a real surprise. It looks like it could be the fan is defective. Remove the back cover for access to the fan connections and check the voltages there directly. At startup 18.64 volts, when first speed-up 20 volts and climbing slowly, when racing 20.15 volts and climbing slowly (20.17 after a few seconds). The slight change in voltage is probably due to varying loads. The fan is highly suspect.
4.Remove the fan for bench test. Remove top & bottom covers, remove 2 screws top and 2 screws bottom of fan cover, remove 3 screws that hold the cover to the power supply, pull cover out. Remove 4 fan screws (2.5 mm hex driver), disconnect fan. Note fan blows into the analyzer. Pull connector off motherboard, use three pin header for connections to a power supply (red is to back of analyzer, marked below motherboard).
5.Fan run on the bench at 20 volts for ten minutes. No sign of trouble. Sound does not change. Banging on desk has no effect. Fan seems OK.
6.Perhaps there is a bad filter capacitor and ripple is feeding through the regulators and upsetting the brushless motor control circuitry. Connect fan and check with ¡®scope.
7.When the noise starts the negative supply is OK but the positive supply has 200 mv peaks showing through. When the runaway condition is forced the negative supply is still OK but the positive supply ripple increases to about 400 mv. Not much considering the dramatic increase in noise.
8.Scope the input to the regulator while fault is there. It is clean, the noise is possibly being modulated on the adjust terminal of the regulator. ¡®Scope pin 8 of U1. Looks clean.
9.This does not appear to be actual speeding up of the fan but rather a mechanical vibration that resonates and peaks at a certain frequency. As the heat sink warms up the fan speed increases due to the fan control circuitry. When cold the blips on the positive supply are at 71Hz and 200 mv. As the heat sink warms the frequency increases to 80 Hz but when the A8 board is tapped to force the fault the frequency actually drops to 69.4 Hz and 300 mv amplitude.
10.Try tapping A8 to force the fault then loosen all four fan screws a few turns (loose) then re-tightened them and the howl was?gone.?With all four fan screws loosened 1/2 turn I can get the fault to happen. It seems to be something happening on A8 due to mechanical feedback.
11.Tapping the middle of the fan removes the fault. I can make the fault come and go by tapping the fan. Perhaps it is the fan.
12.Run the fan in place but from a separate power supply and with the analyzer off. The fault comes and goes with tapping the fan. It is better behaved with only two screws on opposite corners tight so it looks like it is very sensitive to the mounting. It makes various sounds when gently pushed or pulled from the corners. When left it makes ticking sounds as though the fan blades were hitting something, which they are not.
13.Two grommets were slit in half and mounted between the fan and the mounts. A tiny dot of Loctite was used on the ends of the screws and they were tightened a turn past touching. After ten minutes of quiet running the roar returned.
*_Conclusion_*
It is the fan. A new one is ordered.
My memory is vague as to how the problem was resolved but this analyzer was used continuously for another two months. I think the 4124GX fan replacement was determined to be noisy and the vendor sent a similar replacement which was also noisy. One of them was taken apart and oiled and used. See below.
*_Problems continue_*
The fan noise problem.
When the front of the analyzer is lifted or the side slapped, while running, the fan noise returned.
¡¤A 4124 GX fan, with the screws exposed, is in the analyzer. It is not the original which was a 4124KX.
¡¤I remember taking a fan apart and lubricating it so this may be the first one bought that was repaired while waiting for a replacement.
¡¤There is a definitely dead fan, with the label removed, taken apart and put aside so I don¡¯t really know which fan wound up in this analyzer.
¡¤There is a second 4124GX, with an intact label, which is also noisy in the analyzer. It is probably the first replacement ordered.
¡¤Exchange fans with the one in my own 8753B.
¡¤The noise follows the fan to my analyzer while my fan works quietly in Ron¡¯s analyzer.
*_Conclusion:_*
The 4124GX fans are both defective.
*_Fan Replacement for HP 8753B_*
Contacted ebm-papst USA, Copyright ? 2007 ebm-papst Canada Inc. (905) 420-3533
*sales@...* <mailto:sales@...>1800 Ironstone Manor, Unit 2, Pickering, Ontario L1W 3J9
Is there a replacement for the Papst 4124KX fan?
Yes, this has been replaced by the 4184NX.? Please contact any of the below distributors to purchase this fan.
4184NX
*ARROW ELECTRONICS* <>
334
4184NX
*ALLIED ELECTRONICS, INC.* <>
259
4184NX
*ELECTRO SONIC * <>
105
4184NX
*MASTER ELECTRONICS * <>
100
4184NX
*ONLINECOMPONENTS.COM* <>
100
4184NX
*NEWARK ELEMENT14* <>
300
4184NX
*SAGER ELECTRONICS* <>
231
It looks like Electrosonic is the best choice. Contacted them.
They have 5 in stock in Toronto shipping is $10. UPS Ground, $62.78 each
*_Resolution:_*
It has been 5 years and the problem has not returned.
*_Lesson learned:_*
Don¡¯t go cheap on a repair!
On Sun, Mar 24, 2019 at 5:45 PM alwyn.seeds1 <a.seeds@... <mailto:a.seeds@...>> wrote:
Dear All,
I agree with Dave; flow at the required back pressure is what
counts. If the fan design is not suitable, the hurricane of free air
flow drops to a whisper when the fan is installed.
Manufacturers' data sheets usually include graphs of air flow versus
back pressure.
Good manufacturers¡¯ fans are generally quieter than cheap computer
fans; they also last longer.
The Papst catalogue is enormous- most of the distributors only list
a small part of the range.
Regards,
Alwyn
_____________________________________________________
Alwyn Seeds, Director
SynOptika Ltd.,
114 Beaufort Street,
London,
SW3 6BU,
England.
Tel.: +44 (0) 20 7376 4110
SynOptika Ltd., Registered in England and Wales: No. 04606737
Registered Office: 114 Beaufort Street, London, SW3 6BU, United Kingdom.
_____________________________________________________
|
Re: HP5335A fan replacement
A lesson to be learned about fans. HP8753B Fan Repair
?
Problem
The analyzer is
not working
Symptoms
The fan is not
running. The analyzer will not work if the fan is not running.
Repair
Fan removed and
bench tested, not working. New fan ordered and fan replaced.
The original
fan was 4124KX. The new fan is a 4124GX ordered from Ebay however it appears to
be dated 01.82? (Jan 1982).
?
Problems continue
1.??????
Fan is
OK when first turned on but seems to speed up a bit after 5 minutes or so.
2.??????
When
the analyzer was moved the fan suddenly took off and ran very fast and noisy
(or appeared to do so).
3.??????
Shutting
down briefly and starting again repeated the problem up to slightly noisy and
when the analyzer was banged gently on the bench the fan took off again.
4.??????
Tapping
on Post Regulator board A8 causes the fan to take off but not until it has run
for a while and has sped up on its own.
?
Analysis
The fan is
rated for 18 and 30 volts DC. It is speed controlled by voltage from 2
regulators. See page 17/18 FAN TROUBLESHOOTING in the Power Supply
troubleshooting section of 08753-90156.pdf.? It seems that the power to it is suddenly
jumping to max. It may be the reason that the original fan failed.
Find and check
the power supply.
Troubleshooting
1.??????
Check
regulators U2 & U3 on A8. The pins are easily accessed but there is
confusion as to the pin identification and the regulator types. The schematic
shows both to be LM337 which is possible but the part numbers on the devices do
not match and neither do any of the pins agree with data sheets (NatSemi &
Motorola). The parts list shows U2 as LM337T and U3 as LM317T which makes more
sense but not one pin number is correct on the schematic. Correct the
schematic.
2.??????
Measure
all regulator voltages while forcing the fan runaway. None of them change by
more than a volt. The fan voltages are -14.7 and about +4 or +5 (varies as to
be expected because it is controlled by the temperature.
3.??????
This
is a real surprise. It looks like it could be the fan is defective. Remove the
back cover for access to the fan connections and check the voltages there
directly. At startup 18.64 volts, when first speed-up 20 volts and climbing
slowly, when racing 20.15 volts and climbing slowly (20.17 after a few
seconds). The slight change in voltage is probably due to varying loads. The
fan is highly suspect.
4.??????
Remove
the fan for bench test. Remove top & bottom covers, remove 2 screws top and
2 screws bottom of fan cover, remove 3 screws that hold the cover to the power
supply, pull cover out. Remove 4 fan screws (2.5 mm hex driver), disconnect fan.
Note fan blows into the analyzer. Pull connector off motherboard, use three pin
header for connections to a power supply (red is to back of analyzer, marked
below motherboard).
5.??????
Fan
run on the bench at 20 volts for ten minutes. No sign of trouble. Sound does
not change. Banging on desk has no effect. Fan seems OK.
6.??????
Perhaps
there is a bad filter capacitor and ripple is feeding through the regulators
and upsetting the brushless motor control circuitry. Connect fan and check with
¡®scope.
7.??????
When
the noise starts the negative supply is OK but the positive supply has 200 mv
peaks showing through. When the runaway condition is forced the negative supply
is still OK but the positive supply ripple increases to about 400 mv. Not much
considering the dramatic increase in noise.
8.??????
Scope
the input to the regulator while fault is there. It is clean, the noise is
possibly being modulated on the adjust terminal of the regulator. ¡®Scope pin 8
of U1. Looks clean.
9.??????
This does
not appear to be actual speeding up of the fan but rather a mechanical
vibration that resonates and peaks at a certain frequency. As the heat sink
warms up the fan speed increases due to the fan control circuitry. When cold
the blips on the positive supply are at 71Hz and 200 mv. As the heat sink warms
the frequency increases to 80 Hz but when the A8 board is tapped to force the
fault the frequency actually drops to 69.4 Hz and 300 mv amplitude.
10.??
Try
tapping A8 to force the fault then loosen all four fan screws a few turns
(loose) then re-tightened them and the howl was?gone.?With all four
fan screws loosened 1/2 turn I can get the fault to happen. It seems to be
something happening on A8 due to mechanical feedback.
11.??
Tapping
the middle of the fan removes the fault. I can make the fault come and go by
tapping the fan. Perhaps it is the fan.
12.??
Run the
fan in place but from a separate power supply and with the analyzer off. The
fault comes and goes with tapping the fan. It is better behaved with only two
screws on opposite corners tight so it looks like it is very sensitive to the
mounting. It makes various sounds when gently pushed or pulled from the
corners. When left it makes ticking sounds as though the fan blades were
hitting something, which they are not.
13.??
Two
grommets were slit in half and mounted between the fan and the mounts. A tiny
dot of Loctite was used on the ends of the screws and they were tightened a
turn past touching. After ten minutes of quiet running the roar returned.
?
Conclusion
It is the fan.
A new one is ordered.
My memory is
vague as to how the problem was resolved but this analyzer was used
continuously for another two months. I think the 4124GX fan replacement was
determined to be noisy and the vendor sent a similar replacement which was also
noisy. One of them was taken apart and oiled and used. See below.
?
Problems continue
The fan noise
problem.
When the front
of the analyzer is lifted or the side slapped, while running, the fan noise
returned.
¡¤????????
A 4124
GX fan, with the screws exposed, is in the analyzer. It is not the original
which was a 4124KX.
¡¤????????
I
remember taking a fan apart and lubricating it so this may be the first one
bought that was repaired while waiting for a replacement.
¡¤????????
There
is a definitely dead fan, with the label removed, taken apart and put aside so
I don¡¯t really know which fan wound up in this analyzer.
¡¤????????
There
is a second 4124GX, with an intact label, which is also noisy in the analyzer.
It is probably the first replacement ordered.
¡¤????????
Exchange
fans with the one in my own 8753B.
¡¤????????
The
noise follows the fan to my analyzer while my fan works quietly in Ron¡¯s
analyzer.
Conclusion:
The
4124GX fans are both defective.
?
Fan Replacement for HP 8753B
Contacted ebm-papst USA, Copyright ? 2007 ebm-papst Canada
Inc.?? (905)
420-3533
sales@... ??1800 Ironstone Manor, Unit 2, Pickering, Ontario L1W 3J9
?
Is there a replacement for the Papst
4124KX fan?
?
Yes, this has been replaced by the
4184NX.? Please contact any of the below distributors to purchase this
fan.
|
4184NX
|
|
334
|
|
4184NX
|
|
259
|
|
4184NX
|
|
105
|
|
4184NX
|
|
100
|
|
4184NX
|
|
100
|
|
4184NX
|
|
300
|
|
4184NX
|
|
231
|
?
It looks like Electrosonic is the best
choice. Contacted them.
They have 5 in stock in Toronto shipping
is $10. UPS Ground, $62.78 each
?
Resolution:
It has been 5 years and the problem has not
returned.
?
Lesson learned:
Don¡¯t go cheap on a repair!
?
On Sun, Mar 24, 2019 at 5:45 PM alwyn.seeds1 < a.seeds@...> wrote: Dear All,
I agree with Dave; flow at the required back pressure is what counts. If the fan design is not suitable, the hurricane of free air flow drops to a whisper when the fan is installed.
Manufacturers' data sheets usually include graphs of air flow versus back pressure.
Good manufacturers¡¯ fans are generally quieter than cheap computer fans; they also last longer.
The Papst catalogue is enormous- most of the distributors only list a small part of the range.
Regards,
Alwyn
?
_____________________________________________________
Alwyn Seeds, Director
SynOptika Ltd.,
114 Beaufort Street,
London,
SW3 6BU,
England.
Tel.: +44 (0) 20 7376 4110
SynOptika Ltd., Registered in England and Wales: No. 04606737
Registered Office: 114 Beaufort Street, London, SW3 6BU, United Kingdom. _____________________________________________________
|
At 2019-03-24 12:53 PM, [email protected] wrote:
Does anyone know where I can
find a line cord to fit it? I found the operating and service manual, and
I'd like to get it back to operating condition. I'll have to save to get
at least a coax thermistor mount.
I know I've seen those somewhere, but I can't think where. Maybe on the
back of one of my instruments, in whihc case I'm not going to be too
willing to sell it!
If you don't already have a manual for it, I ran across this one while
looking to see what kind of line cord it needs. Three-prong, circular
sockets for pins in the back panel, with the center ground prong slightly
offset. Smaller than the standard NEMA connectors of today, of course.
And it was a refreshing trip down memory lane, seeing a "real"
user manual for an instrument, much like ones I remember seeing when my
dad was into that sort of thing.
Steve Hendrix
|
Re: on using Logic Analyzers to understand and reverse engineer systems
BTW For finding clocks, a spectrum analyzer and a small H field probe is much better than a scope. Less risk of disturbing the circuit operation and it's easy to locate the clock lines even if they are buried.
I've used a simple loop of 26 AWG wire on a scope probe tip with the SA app for the Instek MDO for this. Strictly speaking it's both E&H, but H dominates. I also made a heatshrink condom to insulate a scope probe tip for use as an E field sensor. The 1/4" antenna localizes signals extremely well. Both ideas came from an old article by Bob Pease.
Don't use your probe tip as a form for winding the loops to fit the probe tip. Measure them and use drill bits for forms.
|
Re: on using Logic Analyzers to understand and reverse engineer systems
I can't speak to doing this in HW, but I spent many years fixing large (>50KLOC) programs which had *no* comments other than the author's name. In the aggregate about 2 million lines.
I relied on my knowledge of what the programs did and all the ways you could do that. I've actually had to pull out professional papers and read the equations to figure out if a DSP operation was in the time or frequency domain.
General strategy was to look at the overall structure, identify key events in the operation and then guess at tests cases which would isolate the fault. Most importantly, I never spent more than 3-4 hours without taking a long break. I came to refer to the crest of an overpass about 5 minutes from work as "the point of epiphany" because of the number of times I left work early after a fruitless afternoon hunting for the problem and when I reached the point of epiphany it would suddenly become clear. I'd go in the next morning and have it fixed in 15-20 minutes.
I'd love to see a checklist of questions related to reverse engineering HW designs. With the shift to OEMs not providing component level data, most current T&M gear will become unrepairable except by reverse engineering. I just got a Tek 11801 going after a couple of days struggling with the lack of service data. And it was a *simple* problem! A dead NVRAM battery. I don't know that I could isolate a fault on a board with 120 TTL parts and several undocumented ASICs with multiple ROMs and over 100 control signals on ribbon cables. If I did succeed I'm sure it would take many days to accomplish. So it would have to be *very* valuable.
I came across someone who repaired industrial process control equipment for a living for whom repairing undocumented systems from companies that no longer existed was a common assignment. He made a handsome living doing it.
|
Re: 8753ES YTO freq vs. Main Coil mA
Hi Peter
I don't know this instrument, but here a couple of things I would check..
I cant remember if you already checked this, but can you connect a coupler to the 3.8GHz oscillator, to tap of a small bit and check that this is always stable?? I guess it will be, but will be good to get it verified.
The YTO unlock instability, you should see on the FM coil drive.
Next I would follow the signal all the way to the PLL circuit. Check that it don't get attenuated to much at lower frequencies so the PLL can't lock to it.? ??
Good luck with this.?
Br, Askild
?
On Sun, Mar 24, 2019 at 10:29 PM Peter Gottlieb < hpnpilot@...> wrote: I checked the FM coil drive and it is right about at zero so that seems like
it's not the issue.
Nobody seems to know how the firmware works so we are left to guess and try
things.? There is some range which the firmware can correct for and I heard from
someone who needed to nudge his phase lock board to get his system into that
range.? I tried the same thing and can get some measure of success as well,
although apparently I have another problem causing me grief.? At this point I
have this temporary fix in which permits the unit to run on default pretune
values but won't pass test 48 which sets more precise pretune values.
Below about 50 MHz I have a PLL instability which gets worse as the frequency
decreases.? The PLL begins to lose lock below about 25 MHz.? I took two photos
of the unit output when set to 25 MHz in CW mode showing the instability, then I
put it into source tune mode which opens the phase lock loop.? See attached.?
Notice how the output is nice and stable as soon as the loop is opened.
Not sure what to make of this.? The YTO main coil drive doesn't look any noisier
at 25 MHz than it does at higher frequencies.
On 3/24/2019 3:28 PM, Ed Breya via Groups.Io wrote:
> I don't know if this will help, but here goes.
>
> First, don't worry about the apparent offset in coil tuning current. YTOs are
> only specified to operate over a certain range, like an octave or two, but can
> run a ways beyond at either end. It doesn't help (other than mathematically to
> get the tuning chart numbers) to extrapolate down to zero coil current - the
> YTO will stall out somewhere a little below its lowest spec frequency and coil
> current. Any slope variations or offsets extrapolated at zero just represent
> the character of the magnetic structure in the YTO.
>
> The proper range, slope, and offset needed for tuning a particular YTO are
> typically handled by the YTO driver circuit, which can make it so the net
> result is a nice convenient GHz/V type system, at its input. There are usually
> sufficient adjustments (pots and/or DAC settings) in the driver circuitry to
> accommodate the tolerances of the YTO, the 3.8 GHz oscillator in this case,
> and the rest of the parts, to make everything come out right, during calibration.
>
> There may be circuitry included to improve the tuning linearity (due to
> magnetic saturation), and compensate for the dynamic response (magnetic delay
> and hysteresis). Ideally, the YTO frequency is perfectly and instantly
> proportional to the magnetic field, but in reality the field depends on an
> iron core, coil inductance, AC core losses, and so forth.
>
> The FM coil may be used in a PLL, for fine tuning. Its effect is very small
> (but much faster), compared to the main coil. It too has to be properly
> driven, of course.
>
> In a down-converting generator made for reaching low frequency, it would
> normally be set up to avoid spectral inversion - that's why there's a minimum
> output frequency that can be expected. If the 3.8 GHz LO is free-running, its
> drift may be enough to limit how far down you can go at the low end,
> regardless of the YTO's tuning precision. For sources that can get way down
> toward zero (DC even), both the LO and YTO need to be precisely controlled
> (synthesized).
>
> The coarse and fine tuning, compensation, and calibration all contribute to
> successful operation, so there may be quite a few possible suspects.
>
> If you have schematics, you should be able to find all of these elements, and
> track down the proper conditions for diagnostics and repair.
>
> Good luck.
>
> Ed
>
|
on using Logic Analyzers to understand and reverse engineer systems
In addition to debugging my own designs, one of my common use cases for using logic analyzers is to peer into designs of existing hardware systems. These systems are sometimes poorly documented with schematics and theory of operation info rarely available or reliable. They sometimes employ custom ASICs, and just having a pinout with names of signals is all I have to go on.
Where there are datasheets available for the chips, then it's not too bad to understand the signals, although I might not have direct experience with how those chips work, or have a real appreciation for the data format.
Some of my struggles are related to multiple clock domains, sometimes to bit-ordering(is D0 the MSB or LSB?), sometimes a pin turns out to be serial data, sometimes there are time-multiplexed systems using slots, and so on.
With busses, I often don't realize "who's doing the talking?"
It's not that I don't eventually get there, but it seems that I stare endlessly at logic analyzer traces until I finally have the Eureka moment and see what I'm doing wrong.
I swear I spend more time trying to understand whether I have everything configured properly than I do actually putting the data available to good use.
What do you do? Have some orderly organizational method? Look at the pin on a 'scope and first determine all the clocks in play, which chips are in which domains? Determine if they are serial vs just being a fast chip enable?
I usually lean towards using Timing mode because the change-in-data sometimes gives me information about whether there are intermittent timing violations occurring and because I don't have an appreciation for the significance of the multiple clocks in the design. This of course results in many data duplicates which compounds the "too much data" problem.
Many designs seem to employ "tricks" where taking advantage of different chip propagation delays based on the technology used help prevent race conditions. Or say, knowing that the data isn't actually consumed until later, so having data valid in the first half of the clock cycle isn't necessary....and so on.
Do others struggle with this type of problem? Have you found some online resources or books that address this type of thing?
Thanks,
Keith
|
Re: 16900A cleanup and modernization experience including installation notes!
On Mon, 25 Mar 2019 00:06:43 -0400, you wrote: Responses in-line
On 3/24/2019 11:37 PM, Harvey White wrote:
On Sun, 24 Mar 2019 23:00:42 -0400, you wrote:
Thanks Harvey!
Now what should I do with the other 554 channels? <big grin> <sound of bits chasing each other.... a lot like crickets.....> Heh. I was mostly bragging because I'm excited about the possibilities!
:) So forgive me! Not a problem. I sometimes wonder what I'd do with all the channels.
What I've found on the lower (16700 series) LA's is that the deep
capture is not as useful as you'd like. 2M samples X 16 or more
channels on the display would give a snail a run for its money. I've got a fully-loaded 16700 too. Yeah scrolling through those is a bit
painful. I've done upwards of about 32-channels at a time. Most of the
time, because the triggering is pretty decent, you don't need so many
samples, you just need the RIGHT samples.
Mostly what I'm doing is looking at the outputs of either an ARM microcontroller or an FPGA for a moderately long pulse sequence, PWM for LEDS, serial to an ESP8266 (whoever picked a variable length serial data response for those things deserves to have to put them in a system. A system with an operating system running them and NO flow control).
I had a 32-channel 500mhz USB model that only had 2000 sample memory
depth. 2000 TRANSITIONS --- so long times between data was ok, like you
find in serial data. I really thought it was going to be limited, but it
wasn't horribly so. I did eventually "upgrade" to a 24mhz, but super
deep (billions of samples), usb one..... and that was good for some
things too.
I've got a 10 dollar 8 channel LA that does a remarkably good job of decoding I2C and serial data. For debugging communications, that did a good job, surprisingly better than my 56K baud maximum HP communications analyzer. (I'm working at 115K baud to the ESP8266 WIFI chip). However, the triggering isn't all that sophisticated, and it can't take apart the I2C messages.
They used to say, fast, cheap, deep, pick two.
Yep.
Sometimes, though, I don't know exactly what I'm looking for.....so I'm
collecting samples to look through and determine the
possibilities....what common values might I be seeing.
sometimes I just don't know what's going on, so paying attention to SPI data going in, data going out, that kind of thing helps a lot.
You might want to use different blades for different tasks, I've only
got one blade in my 16702B, and that does it for most of what I need,
two pods for I2C and the rest for whatever I need. Monitoring the
result of an I2C transmission in another processor board could be
interesting. Which blade do you have? 715A, 717A, 752A?
717A I think. The 16555A that I can pirate from my 16500 is also available, but I really don't quite need it unless I'm looking at two systems. Ironic that the 16510 has about the capability of the one 1661 LA that I still keep if I need something more portable.
How do you decode the I2C? On the unit itself? I forget, was serial
decoding an option?
Serial decoding is an option to an extent, but I cheat. I use a CPLD to go from I2C serial to a state machine. I number the bytes in the CPLD, flag ack/nak for each byte, tag start and stop, etc. That was the basis for a standalone I2C decoder that I built. Uses LM339 comparators for SCL, SDA and an interrupt line, uses the CPLD as a oneshot for the activity led, shows me high and low stuck lines. That's going to go into a more complicated FPGA as a logic analyzer plugin to an existing project, so I will be putting that, two USARTS, two FIFOs (1K*8) and likely a FIFO for the I2C stuff (24 bits * 1K). It's up to the ARM processor to display data in state form. Of course, I can also do timing as well, I guess. The LM339's are in there to protect the FPGA, seriously. I never saw an I2C decoding option in any of the 16700 series or below.
I was exporting via FTP the flat data files from captures, and then
using the latest Logic Analyzer 5.90 software on a faster machine to
interpret them. Just much nicer using a modern PC...... I'm happy that
the modern PC can now control and connect to the other LA remotely!
Ah, I may end up doing that. I do PC programming in Pascal (Lazarus), and depending on whether or not I get WIFI going on these projects, I may end up with all the 488 bus stuff in the lab on WIFI. Likely will be able to do that with I2C. I've got one design with a small (320 * 240) color touchscreen, and an add on board (when debugged) will allow me to go ahead and use an 8 inch VGA touchscreen. Add in the processor, FPGA and the like and you've got a reasonably nice debugging tool.
I may end up doing that, but I don't think I need quite so many
channels...
<grin> Thanks! ;)
Quite welcome. Harvey Keith
|
Re: 16900A cleanup and modernization experience including installation notes!
Responses in-line On 3/24/2019 11:37 PM, Harvey White wrote: On Sun, 24 Mar 2019 23:00:42 -0400, you wrote:
Thanks Harvey!
Now what should I do with the other 554 channels? <big grin> <sound of bits chasing each other.... a lot like crickets.....> Heh. I was mostly bragging because I'm excited about the possibilities! :) So forgive me! What I've found on the lower (16700 series) LA's is that the deep
capture is not as useful as you'd like. 2M samples X 16 or more
channels on the display would give a snail a run for its money. I've got a fully-loaded 16700 too. Yeah scrolling through those is a bit painful. I've done upwards of about 32-channels at a time. Most of the time, because the triggering is pretty decent, you don't need so many samples, you just need the RIGHT samples. I had a 32-channel 500mhz USB model that only had 2000 sample memory depth. 2000 TRANSITIONS --- so long times between data was ok, like you find in serial data. I really thought it was going to be limited, but it wasn't horribly so. I did eventually "upgrade" to a 24mhz, but super deep (billions of samples), usb one..... and that was good for some things too. They used to say, fast, cheap, deep, pick two. Sometimes, though, I don't know exactly what I'm looking for.....so I'm collecting samples to look through and determine the possibilities....what common values might I be seeing. You might want to use different blades for different tasks, I've only
got one blade in my 16702B, and that does it for most of what I need,
two pods for I2C and the rest for whatever I need. Monitoring the
result of an I2C transmission in another processor board could be
interesting. Which blade do you have? 715A, 717A, 752A? How do you decode the I2C? On the unit itself? I forget, was serial decoding an option? I was exporting via FTP the flat data files from captures, and then using the latest Logic Analyzer 5.90 software on a faster machine to interpret them. Just much nicer using a modern PC...... I'm happy that the modern PC can now control and connect to the other LA remotely! I may end up doing that, but I don't think I need quite so many
channels...
<grin> Thanks! ;) Keith
|
Re: 8753ES YIG oscillator circuit WCA
Also, it appears if you have manual control of your preset DAC¡¯s then you can manual tune the DAC 0 to max (it sounds like you have 2040 bit DAC¡¯s if 1019 is your center). Monitor the YIG freq. and change the DAC tune no¡¯s. You should see binary stepped YIG frequencies. The 2 LSB¡¯s should show no change since they are not used due to DAC noise issues. You can calculate the delta freq./DAC step and it should be the same across the entire DAC tune. What I have seen are stuck bits that cause a nonlinear stair step, resulting in tune or unlock problems. Hunting between an upper and lower DAC tune that is above and below the correct DAC number it cannot get because it is stuck will cause FM on the YIG at the freq. of the FM lock loop slew rate.
Once you know the freq. delta/bit change then if you have the freq. FM or offset then you will know what bit is stuck.
Don Bitters
|
Re: 16900A cleanup and modernization experience including installation notes!
On Sun, 24 Mar 2019 23:00:42 -0400, you wrote: Thanks Harvey!
Now what should I do with the other 554 channels? <big grin> <sound of bits chasing each other.... a lot like crickets.....> What I've found on the lower (16700 series) LA's is that the deep capture is not as useful as you'd like. 2M samples X 16 or more channels on the display would give a snail a run for its money. You might want to use different blades for different tasks, I've only got one blade in my 16702B, and that does it for most of what I need, two pods for I2C and the rest for whatever I need. Monitoring the result of an I2C transmission in another processor board could be interesting. And if your system is interrupt/keyboard driven, you could monitor a complete event/message sequence/result and so on.... I may end up doing that, but I don't think I need quite so many channels... <grin> Harvey
Keith
On 3/24/2019 8:57 PM, Harvey White wrote:
On Sun, 24 Mar 2019 14:45:44 -0400, you wrote:
I purchased a fully loaded 16900A for quite a bargain recently. It has
(1) 16950B, 16M, 667MHZ state mode, and 4ghz timing zoom module. It also
has (5) 16910A's, 16M option, 500MHZ state mode option. The 16950B is
68-channels, and the 16910A's are 102-channels. I wasn't a math major,
but that's almost 600-channels of simultaneous capture.
A couple friends asked, what are you going to do with all those
channels? While truthfully I don't know, I can certainly do whatever the
heck I'd want to do with this beast! Debugging I2C communications between microprocessors:
8 channels for data,
roughly 8 channels for status (start/stop/first byte, ack/nak, etc per
byte)
8 channels (roughly) for message byte count, resetting after a stop:
24 channels
Then you monitor anything that the I2C message was supposed to
trigger....
Now the 24 channels worth of data is parallel, state machine, and you
have to figure out how to take it apart.
Harvey
<snip>
|
Keith Monahan