Re: Understanding the 8753ES frequency loop
Thank you.?
I¡¯m still recovering from a real nasty flu/cold so everything is going really slowly here.?
I¡¯m hoping that what I find is helpful to others as I have heard from another group member of a nearly identical problem. I¡¯d like to post sufficient information such that anyone else running into this will have specific accurate numbers to compare against and hopefully to lead to more repaired units.?
toggle quoted message
Show quoted text
On Mar 21, 2019, at 1:53 PM, Pete Manfre < pmanfre@...> wrote: New board should be to you tomorrow.?
Pete wa2odo On Thu, Mar 21, 2019, 1:32 PM Peter Gottlieb < hpnpilot@...> wrote: Thank you. All this helps me piece together how the setup works. Right now I am collecting data and once it starts forming a coherent picture I will post it.?
The firmware plays a big part in this as well. For example, after default pretune values are set, test 48 creates individualized pretune values for that specific instrument. There is only so far that can compensate, though, and that¡¯s apparently what I¡¯m running into.? What I don¡¯t yet know is whether the problem is on that board or if the YIG oscillator has drifted out of spec. It would be helpful to know just how far that firmware can compensate. I do see that my analog bus voltages are spot on, and seeing how your pretune voltage waveform is lower than mine, I still suspect a problem on my phase lock board. I have verified that the voltage to current converter is operating properly, and since the measurement resistor is verified well within its 1% tolerance, that the current through the YTO main coil should thus be within design tolerance as compared to the 1V/GHz signal. Thus any inability to be lower in voltage would have to be in the DAC, reference, voltage shift area that feeds the V/I converter. If however the replacement phase lock board has the exact same behavior I would have to look again at the YIG oscillator.?
Peter On Mar 20, 2019, at 11:17 PM, pianovt via Groups.Io < pianovt@...> wrote: Hi Peter,
The main and FM coil signals pass through the A3 ALC board like this:
<dummyfile.0.part>
The FM -? tap to the cavity oscillator is not part of the basic phase lock scheme, it's a spur avoidance fix. You don't need to waste time on studying that unless the fault is there.
Since you mentioned that the YTO is off (high?) by 200MHz, I think you should look at the main coil driver. It seems like something is preventing the main coil driver from delivering a low current. I am not aware of a PLL switch point at 1 GHz, but I have forgotten a lot about it so I may be wrong.
Vladan
|
Re: Understanding the 8753ES frequency loop
New board should be to you tomorrow.?
Pete wa2odo
On Thu, Mar 21, 2019, 1:32 PM Peter Gottlieb < hpnpilot@...> wrote: Thank you. All this helps me piece together how the setup works. Right now I am collecting data and once it starts forming a coherent picture I will post it.?
The firmware plays a big part in this as well. For example, after default pretune values are set, test 48 creates individualized pretune values for that specific instrument. There is only so far that can compensate, though, and that¡¯s apparently what I¡¯m running into.? What I don¡¯t yet know is whether the problem is on that board or if the YIG oscillator has drifted out of spec. It would be helpful to know just how far that firmware can compensate. I do see that my analog bus voltages are spot on, and seeing how your pretune voltage waveform is lower than mine, I still suspect a problem on my phase lock board. I have verified that the voltage to current converter is operating properly, and since the measurement resistor is verified well within its 1% tolerance, that the current through the YTO main coil should thus be within design tolerance as compared to the 1V/GHz signal. Thus any inability to be lower in voltage would have to be in the DAC, reference, voltage shift area that feeds the V/I converter. If however the replacement phase lock board has the exact same behavior I would have to look again at the YIG oscillator.?
Peter On Mar 20, 2019, at 11:17 PM, pianovt via Groups.Io < pianovt@...> wrote: Hi Peter,
The main and FM coil signals pass through the A3 ALC board like this:
<dummyfile.0.part>
The FM -? tap to the cavity oscillator is not part of the basic phase lock scheme, it's a spur avoidance fix. You don't need to waste time on studying that unless the fault is there.
Since you mentioned that the YTO is off (high?) by 200MHz, I think you should look at the main coil driver. It seems like something is preventing the main coil driver from delivering a low current. I am not aware of a PLL switch point at 1 GHz, but I have forgotten a lot about it so I may be wrong.
Vladan
|
Re: Understanding the 8753ES frequency loop
Thank you. All this helps me piece together how the setup works. Right now I am collecting data and once it starts forming a coherent picture I will post it.?
The firmware plays a big part in this as well. For example, after default pretune values are set, test 48 creates individualized pretune values for that specific instrument. There is only so far that can compensate, though, and that¡¯s apparently what I¡¯m running into. ?What I don¡¯t yet know is whether the problem is on that board or if the YIG oscillator has drifted out of spec. It would be helpful to know just how far that firmware can compensate. I do see that my analog bus voltages are spot on, and seeing how your pretune voltage waveform is lower than mine, I still suspect a problem on my phase lock board. I have verified that the voltage to current converter is operating properly, and since the measurement resistor is verified well within its 1% tolerance, that the current through the YTO main coil should thus be within design tolerance as compared to the 1V/GHz signal. Thus any inability to be lower in voltage would have to be in the DAC, reference, voltage shift area that feeds the V/I converter. If however the replacement phase lock board has the exact same behavior I would have to look again at the YIG oscillator.?
toggle quoted message
Show quoted text
On Mar 20, 2019, at 11:17 PM, pianovt via Groups.Io < pianovt@...> wrote: Hi Peter,
The main and FM coil signals pass through the A3 ALC board like this:
<dummyfile.0.part>
The FM -? tap to the cavity oscillator is not part of the basic phase lock scheme, it's a spur avoidance fix. You don't need to waste time on studying that unless the fault is there.
Since you mentioned that the YTO is off (high?) by 200MHz, I think you should look at the main coil driver. It seems like something is preventing the main coil driver from delivering a low current. I am not aware of a PLL switch point at 1 GHz, but I have forgotten a lot about it so I may be wrong.
Vladan
|
Re: HP 3563A (3562A) A9 FFT
The 2nd FFT board arrived, sourced less than 1hr from here, and it is a good board.? All self tests passed.? The 3563A is working!??
On Tue, Mar 19, 2019 at 12:18 PM, morestainless via Groups.Io <lancelotburt@...> wrote: The signals and voltages appear to be ok.? Some of the IC pin signatures on the GI chips were correct, and they were all stable.? The board passes all fft tests aside from the global interface and rom.? The rom test is dependent upon the GI.
I just returned the 1st ebay fft board (cpu dead) and the 2nd is on the way. On Mon, Mar 18, 2019 at 2:23 PM, morestainless via Groups.Io <lancelotburt@...> wrote: Chuck,
Thanks for that bit of wisdom.? I had checked filter caps, clock, and voltages before.? All within spec.? The electrolytic and bypass caps on the board are good.? I'll check all the fft signal waveforms then.? Guess that's why they are described in the manual. On Mon, Mar 18, 2019 at 1:08 PM, Chuck Harris <cfharris@...> wrote: Two things:
1) signature analysis can fail if clocks and dividers
? aren't working right.
2) clocks and dividers can fail if power supply bypass
? capacitors aren't low enough ESR.
Before you condemn your board, do make sure that the
board's bulk bypass capacitors, and any filters that feed
them are good.
I have fixed several boards from this family of DSA's where
the only problem was bad electrolytic capacitors.
-Chuck Harris
morestainless via Groups.Io wrote:
> Anybody have a working A9 FFT board for this machine they would part with?? 3562A and 3563A? use the same board.? Signature analysis shows the global bus chips U511 - U518 are bad as are the two coefficient ROM's - U315 and U317.? Firmware images would also be a possibility as the interface chips are cheap enough.
>
>
>
>
|
Re: HP 8664A signal generator repairability
Seconded,? I have gotten bit by the 2nd or 3rd harmonic even on a good meter.? Bird is considered a 5 or 10% meter at best great for tuning bad for accuracy ? Regards, ? Stephen Hanselman Datagate Systems, LLC 3107 North Deer Run Road #24 Carson City, Nevada, 89701 (775) 882-5117 office (775) 720-6020 mobile s.hanselman@... a Service Disabled, Veteran Owned Small Business DISCLAIMER: This e-mail and any attachments are intended only for use by the addressee(s) named herein and may contain legally privileged and/or proprietary information. If you are not the intended recipient, any dissemination, distribution or copying of this e-mail and any attachments is strictly prohibited. If you have received this e-mail in error, please immediately notify me and permanently delete the original and all copies and printouts of this e-mail and any attachments. ? ? ?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]> On Behalf Of alwyn.seeds1 Sent: Thursday, March 21, 2019 3:35 AM To: [email protected]Subject: Re: [HP-Agilent-Keysight-equipment] HP 8664A signal generator repairability ? Dear All, "From:?Bob Albert
Date: Wed, 20 Mar 2019 17:28:34 PDT? For a decent power meter you do NOT want a Bird.? They are pretty sloppy." Sorry for using the vague term ¡°decent power meter¡±. I meant something like an HP 435/436 with an 8481A measurement head. The Bird instruments are intended for measuring rather larger RF powers and, as Bob says, do not offer the accuracy required for testing a signal generator. While writing, a tip to avoid the mistake I once made. Power meters are, of course, broadband. Therefore if the signal to be measured has significant harmonic content an error in measuring the fundamental power will result. Forgetting about this, I was puzzling why the Cal. outputs of my 8560 series and 8590 series spectrum analysers measured differently. Then I realised that the 8590 Cal. source has strong harmonic output, whereas the 8560 series source does not. Similarly, some signal generators that use frequency doublers produce strong half-frequency outputs which would cause an error when checking output power on the highest frequency range with a power meter. ? _____________________________________________________ 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: Signal Generator Comparison Chart ?
I live in Ontario Canada and get heavy items shipped to the UPS Store in Ogdensburg NY, sometimes saving as much as $400 shipping! They charge $5 per package for 10 or 15 days, check. It used to be a week.?? It means 3 hours drive each way but my wife and I make a trip of it. PeterB
I have also been in the market for a 26GHz SG for a while with no luck. Although I have to say the biggest issue for me
has been the sellers not shipping to Canada. I came close to getting a 8673B for $150 but then the seller finally decided he wouldn't ship to Canada...
with regard to models:
8673B (has modulations) and 8673G (strip down version of B with no modulation) go from 2GHz-26.5GHz and at least for my purpose they are fine
both have excellent specs and they are synthesized signal generators. 8673D is also an option (0.05G-26.5G) but it is huge like a tank
the best overall in my opinion is 8340B or 8340A (no FM modulation) they go down to 10MHz and up to 26.5GHz with excellent
specs on accuracy and noise and they are standard calibration sig gens for spectrum analyzers.
remember 8350B is a sweep generator and the oscillator modules that come with it are NOT synthesized sig gens. they are generally poor on stability and accuracy
and that's why they have a low frequency resolution, too, I guess but their amplitude is stabilized in an ALC loop
I wouldn't go down that path unless you are in a hurry for a particular project
Next comes Wiltron synthesizers 67xx. The one I have been looking for is 6759B which is 10MH - 26.5GH synthesized with very good specs. very close to 8673B actually
their service manual is also available. 6769 goes up to 40GHz
and finally Gigatronics 1026 and 7100 and 7200. they all synthesize the frequency and have very good specs but not sure about service manual availability
On Wed, Mar 20, 2019 at 04:35 PM, Tim Tuck wrote:
Hi all,
I was wondering if there is a sig gen comparison chart somewhere.
Googling around I only found a pdf from Keysight of their latest generators.
Personally I'm in the market for something that goes to 26Ghz ( Id like
to go higher but the $'s start to get crazy for a hobbyist!) and I've
been looking at 8673's but they're not the only ones out there.
Scanning ebay for kit, I keep bumping into HP model numbers I've not
heard of before and then run down that rabbit hole investigating the
model. It'd be nice if there was a definitive list of all
HP/Agilent/Keysight sig gens to look at.
Does such a thing exist ?
thanks
Tim
de VK2XAX
|
Thanks Tom, I will try it! Kevin
toggle quoted message
Show quoted text
----- Original Message ----- From: "Tom Holmes" <tholmes@...> To: < [email protected]> Sent: Thursday, March 21, 2019 1:18 AM Subject: Re: [HP-Agilent-Keysight-equipment] 8970b Use SP 1.3, since you have a more or less fixed IF frequency. Page 3-128 in
the 8970S manual is pretty clear about that. Also be sure to use SP 3.0,
which sets the IF frequency. Since you probably aren't doing a swept NF
measurement, ignore that part of the procedure. Since the output of the
noise source is within range of your 144 MHz input, and the IF at 30 MHz is
within the range of the 8970B, you can set this up just like doing a preamp,
except that mode 1.3 allows you to tell the system to use the ENR of the
noise source at 144 MHz, while listening at 30 MHz. During CAL, the 8970B
listens at 144 to get the source power levels, then during measurements,
listens at 30 MHz.
You might try, as an experiment, just using the 8970B in its SP 1.0 mode.
The input an output are not far apart and the source is pretty flat, as is
the receiver in the 8970B. You might find you get an almost identical result
to SP 1.3.
Yeah, the manual can be aggravating at times; it can take a lot of digging
to find what you need to know.
Hang in there, you'll get it working.
Tom Holmes, N8ZM
|
Re: HP 8664A signal generator repairability
Dear All,
"From:? Bob AlbertDate: Wed, 20 Mar 2019 17:28:34 PDT?
For a decent power meter you do NOT want a Bird.? They are pretty sloppy."
Sorry for using the vague term ¡°decent power meter¡±. I meant something like an HP 435/436 with an 8481A measurement head.
The Bird instruments are intended for measuring rather larger RF powers and, as Bob says, do not offer the accuracy required for testing a signal generator.
While writing, a tip to avoid the mistake I once made. Power meters are, of course, broadband. Therefore if the signal to be measured has significant harmonic content an error in measuring the fundamental power will result. Forgetting about this, I was puzzling why the Cal. outputs of my 8560 series and 8590 series spectrum analysers measured differently. Then I realised that the 8590 Cal. source has strong harmonic output, whereas the 8560 series source does not.
Similarly, some signal generators that use frequency doublers produce strong half-frequency outputs which would cause an error when checking output power on the highest frequency range with a power meter.
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: Signal Generator Comparison Chart ?
On Wed, Mar 20, 2019 at 10:06 PM, Tim Tuck wrote:
Next comes Wiltron synthesizers 67xx. The one I have been looking for is 6759B which is 10MH - 26.5GH synthesized with very good specs. very close to 8673B actually
their service manual is also available. 6769 goes up to 40GHz
They look nice, but I don't know anything about them - I've kept with HP for 99% of my lab because its reliable and I can service them but if there's reasonable documentation, parts etc. for those, I may consider it.
and finally Gigatronics 1026 and 7100 and 7200. they all synthesize the frequency and have very good specs but not sure about service manual availability,_._,_
Yeah... and that's the thing isn't it. I see no point expending $ on kit if there's no support.
Right, I'm finding this out with the Wavetek Pacific Measurements 1038-NS20 and NS201.? I'm still wondering about an Operation and Maintenance Manual with schematics for the N10. (Note the N10 & N20 aren't the signal source).?
I still haven't figured out exactly if the PM 1038 can handle up to 26GHz and if so... which modules.? Reads there are sensors/detectors that do rate up to 26GHz.?
|
Re: Signal Generator Comparison Chart ?
G'Day, On 21/03/2019 1:56 pm, amirb wrote: I have also been in the market for a 26GHz SG for a while with no luck. Although I have to say the biggest issue for me
has been the sellers not shipping to Canada. I came close to getting a 8673B for $150 but then the seller finally decided he wouldn't ship to Canada...
Yep - I hear ya. At least it can be road freight from the US to you, I'm on a bloody big island! with regard to models:
8673B (has modulations) and 8673G (strip down version of B with no modulation) go from 2GHz-26.5GHz and at least for my purpose they are fine
both have excellent specs and they are synthesized signal generators. 8673D is also an option (0.05G-26.5G) but it is huge like a tank I missed out on a 8673G for $800 just recently :( - I do like those units since they can drive the millimeter-wave sources, and ... at some point... I'd be looking to get one or two of those two since I'm slowly working through the SHF bands building transverters and I intend to end up on 122Ghz at some point. I only need something to go from 3Ghz up since I have an ESG4432b to cover from 3Ghz down as well as a couple of other gens for down to a uHz.
the best overall in my opinion is 8340B or 8340A (no FM modulation) they go down to 10MHz and up to 26.5GHz with excellent
specs on accuracy and noise and they are standard calibration sig gens for spectrum analyzers.
Yes they are nice ! -? I have a saved search for them too.
remember 8350B is a sweep generator and the oscillator modules that come with it are NOT synthesized sig gens. they are generally poor on stability and accuracy
and that's why they have a low frequency resolution, too, I guess but their amplitude is stabilized in an ALC loop
I wouldn't go down that path unless you are in a hurry for a particular project
I'm in no hurry and they would be a last resort, one of my associates has one and a few plugins and a bunch of external attenuators, but its all too much kit and I only have a little lab that already has too many things in it, so what ever I get something has to leave to make room.
Next comes Wiltron synthesizers 67xx. The one I have been looking for is 6759B which is 10MH - 26.5GH synthesized with very good specs. very close to 8673B actually
their service manual is also available. 6769 goes up to 40GHz
They look nice, but I don't know anything about them - I've kept with HP for 99% of my lab because its reliable and I can service them but if there's reasonable documentation, parts etc. for those, I may consider it.
and finally Gigatronics 1026 and 7100 and 7200. they all synthesize the frequency and have very good specs but not sure about service manual availability,_._,_
Yeah... and that's the thing isn't it. I see no point expending $ on kit if there's no support. Thanks for the pointers, I'll do some research. regards Tim
|
Re: 16900A ballpark pricing and info
Glen,
Thanks for the previous two messages.
Good progress: The machine boots right into XP without problems. Needed to work around passwords on the unit, but that was easy enough using chntpw.
Self-tests pass on all modules, except for those that I believe require external hardware or configuration to pass.
The fitted 80GB IDE drive has only about 10GB in use. It's amazing what used to get us by. I ordered a SIL3114-based PCI-SATA card -- the best I can tell this should be supported natively. The USB ports are only USB 1.1 on the unit, which rules out using flash drives for anything "serious." I hope that gigabit ethernet will be faster!
Also ordered a cheap SSD which should perform at least better and more reliably than an old rusty spinner.
Keysight support has been great, and provided both BIOS update disc which I caught wind of, and a factory reset OS disk image. I haven't tried these yet, but will soon to at least ensure their integrity.
There are MacroVision lmtools .lic files in the Program Files directory. I think they get "applied" to the modules which I believe writes an EEPROM or some NVRAM on the module itself. According to support, no files need copied/persisted in order to maintain the options....They call the .lic files "Entitlement Certificates" which get emailed. They are small ~152 byte files.
Thanks,
Keith
toggle quoted message
Show quoted text
On 3/19/2019 3:11 PM, Glen Slick wrote: I still have to dig into this. Are the HDDs regular 3.5" SATA's? I might
replace it with an off-the-shelf cheap SSD. I'll probably image the
existing drive before I do anything.
The 16900A, 16902A, and 16903A drives are regular 3.5" SATA drives.
Small by today's standards. I just checked my 16903A, it is only 80GB
with about 65GB free. (The 16901A and 16902B switched to 2.5" drives,
either magnetic or SSD, maybe around 320-500GB).
oops, faulty memory. The 16900A, 16902A, and 16903A drives are regular
3.5" ATA drives, not SATA drives. There is no SATA interface on the
motherboard.
There was an E5862A external boot drive option, which was a basically
just a Buslink external SATA drive enclosure and a Buslink SPCI2P PCI
SATA interface card. That card uses a Silicon Image Sil3112A PCI-SATA
chip.
External Boot Drive for Keysight Logic Analysis Systems Installation Guide
I picked up a couple of Buslink SPCI2P PCI SATA interface cards cheap
on eBay a while ago and installed one in a 16900 series analyzer to
use a SATA drive instead of an ATA drive. I don't remember there being
any real difficulty in getting that to work, although I might have had
to update the BIOS flash chip on the SPCI2P card. There are different
images for regular vs. raid operation if I remember correctly.
|
Re: HP5335A fan replacement
On Tue, Mar 19, 2019 at 07:22 PM, Joe / KN5U wrote:
By the way, the original fan was 120 VAC, but repair note changed the fan to 24 VDC. This was precipitated by the power supply relay and socket overheating. The repair note had a description of a circuit mod using a diode and also replacing the fan and socket. I have done a few of these mods and this a fairly simple fix. Let me know if you need a copy of the change.
Joe, KN5U
The fan on my HP5335A is 24V, so I guess the modification you mention has already been applied.
|
Re: HP5335A fan replacement
On Tue, Mar 19, 2019 at 07:17 PM, Joe / KN5U wrote:
Dan,
This is a great question. My experience is that the critical parts for cooling consideration are the input amplifier U3 and U4 (Schmitt amplifier). These are highly susceptible to shortened life due to heating. Of the 200+ units I have refurbished, interestingly, those that had underpowered fan replacements also had one or both Schmitt amplifier chips not working. So my suggestion is to replace the fan with one of the proper air flow. Especially considering that the Schmitt amplifiers are no longer available.
Regards,
Joe, KN5U
Thanks Joe. I will look for another fan with better airflow characteristics. Of course, I could always go for a Papst fan with better noise characteristics.
|
Re: My $25 "In Poor Shape" 3468A has arrived
On Tue, Mar 19, 2019 at 08:04 AM, Alex wrote:
BTW wanted to direct your attention to that black resistor towards the front of the board and just to the right of one of the red relays next to the lithium battery, that I assume might be related to the Amps readings (which I have not tested yet). Is this the normal appearance for this part? I really does not look too abnormal, but it seems like it has a tiny bit of some brown-reddish stuff on the sides where the wires meet the body of this resistor. I just put up a more detailed picture in the gallery as well. Thanks!
Jeff, That is the 0.1 Ohm wire wound current sensing resistor.? It looks kind of ratty because of the coating but that is normal.? The reddish-brown stuff I do not know -- could be oxidation from flux. The negative readings seem suspiciously like thermal voltages from all the contacts you are making in shorting the input.? The best way to apply a "short" is to NOT use these fancy "gold plated" speaker lug/clips.? ?Make you own banana plug SHORT out of soft copper tubing --? type " L" and just bend the tube into a "U" so you have no solder connections or other interfaces other than the tube connecting with the input banana jacks -- two contacts? Hi and Lo.? Choose about 2.5" overall length to start with.? Another way to make a very good SHORT is to use? AWG #6 solid copper wire - again start with about 2.75" overall length.? The input jacks have an inside diameter of about 3.8mm and #6 AWG solid wire has a OD diameter of??
| AWG 6 |
OD dia? 0.1620 in |
OD dia? 4.115 mm |
? |
? |
? |
? ? ? ? ?micoOhms? ? per inch? ? ? ? 32.93? ? ? ? ?? |
milliOhms per foot
0.3951 |
milliOhms
per meter: 1.296 |
You can whittle down the diameter of the #6 wire to fit the input jacks of the 3468/3478 multimeter by putting the ends of the wire in the jaws of a pair of adjustable lock pliers that has gripping grooves in the jaws and rotating the wire to shave the diameter a bit.? Then bend the wire into a "U" with proper spacing between the ends. This wire short will be good to 66 microOhms as measured on my Hp3458A using offset compensated Ohms which nulls out the thermal offset voltages?(1 count on the low Ohms range =? 10 microOhms).? Length of wire short when inserted in the meter is about 2"? so the inherent resistance of the short = ?2" x 33 microOhm/inch? =? 66 microOhm as measured on the HP3458A If you make two such shorts for the INPUT and SENSE input jacks and jumper a wire between them for 4-wire Ohms, this type of short is almost as good as the massive thermal block short I have for calibrating meters.??The difference is in thermal voltages generated and thermal stability of the short which is only a difference of 1 count?on the Hp3458A (10 microOhms). Now if you calibrate the Ohms ranges in 4-wire ohms and switch to 2-wire Ohms, it is not unreasonable for the meter to display as much as 0.5 Ohm with the short applied due to the current from the Ohms converter passing thru trace resistances, wiring to the input terminals and the return ground circuit to the reference ground of U101 along with?thermal voltage offsets generated internal to the meter between the meter input and the input to the DCV amplifier U101.? Shorting the meter inputs with a short in 2-wire Ohms is meaningless if the meter was caled in 4-wire Ohms.? Doing such can give you all sorts of strange readings positive and negative which depends on the nature of thermal voltages? and contact resistances generated by attaching test leads to the meter. If you are going to ZERO cal in 2-wire Ohms, it makes no sense to do so with a short applied to the meter input between Hi and Lo for the current from the Ohms converter passes from the Hi terminal to the outside world and will produce a voltage drop along your test leads which will register as an offset equal to the resistance of your test leads and after calibrating will be added to the resistance you are trying to measure.? To ZERO cal the meter in 2-wire Ohms, do so with your test leads attached and shorted at the probe tips of the test leads and not at the meter input.? Then use the test leads you caled with for making your measurements in 2-wire Ohms.? This will give you the most accurate readings. Good metering to you all and happy equinox. George? ? March 20, 2019? 10:50pm
|
Re: Understanding the 8753ES frequency loop
Hi Peter, The main and FM coil signals pass through the A3 ALC board like this:  The FM -? tap to the cavity oscillator is not part of the basic phase lock scheme, it's a spur avoidance fix. You don't need to waste time on studying that unless the fault is there. Since you mentioned that the YTO is off (high?) by 200MHz, I think you should look at the main coil driver. It seems like something is preventing the main coil driver from delivering a low current. I am not aware of a PLL switch point at 1 GHz, but I have forgotten a lot about it so I may be wrong. Vladan
|
Re: Signal Generator Comparison Chart ?
I have also been in the market for a 26GHz SG for a while with no luck. Although I have to say the biggest issue for me
has been the sellers not shipping to Canada. I came close to getting a 8673B for $150 but then the seller finally decided he wouldn't ship to Canada...
with regard to models:
8673B (has modulations) and 8673G (strip down version of B with no modulation) go from 2GHz-26.5GHz and at least for my purpose they are fine
both have excellent specs and they are synthesized signal generators. 8673D is also an option (0.05G-26.5G) but it is huge like a tank
the best overall in my opinion is 8340B or 8340A (no FM modulation) they go down to 10MHz and up to 26.5GHz with excellent
specs on accuracy and noise and they are standard calibration sig gens for spectrum analyzers.
remember 8350B is a sweep generator and the oscillator modules that come with it are NOT synthesized sig gens. they are generally poor on stability and accuracy
and that's why they have a low frequency resolution, too, I guess but their amplitude is stabilized in an ALC loop
I wouldn't go down that path unless you are in a hurry for a particular project
Next comes Wiltron synthesizers 67xx. The one I have been looking for is 6759B which is 10MH - 26.5GH synthesized with very good specs. very close to 8673B actually
their service manual is also available. 6769 goes up to 40GHz
and finally Gigatronics 1026 and 7100 and 7200. they all synthesize the frequency and have very good specs but not sure about service manual availability
toggle quoted message
Show quoted text
On Wed, Mar 20, 2019 at 04:35 PM, Tim Tuck wrote:
Hi all,
I was wondering if there is a sig gen comparison chart somewhere.
Googling around I only found a pdf from Keysight of their latest generators.
Personally I'm in the market for something that goes to 26Ghz ( Id like
to go higher but the $'s start to get crazy for a hobbyist!) and I've
been looking at 8673's but they're not the only ones out there.
Scanning ebay for kit, I keep bumping into HP model numbers I've not
heard of before and then run down that rabbit hole investigating the
model. It'd be nice if there was a definitive list of all
HP/Agilent/Keysight sig gens to look at.
Does such a thing exist ?
thanks
Tim
de VK2XAX
|
Use SP 1.3, since you have a more or less fixed IF frequency. Page 3-128 in
the 8970S manual is pretty clear about that. Also be sure to use SP 3.0,
which sets the IF frequency. Since you probably aren't doing a swept NF
measurement, ignore that part of the procedure. Since the output of the
noise source is within range of your 144 MHz input, and the IF at 30 MHz is
within the range of the 8970B, you can set this up just like doing a preamp,
except that mode 1.3 allows you to tell the system to use the ENR of the
noise source at 144 MHz, while listening at 30 MHz. During CAL, the 8970B
listens at 144 to get the source power levels, then during measurements,
listens at 30 MHz.
You might try, as an experiment, just using the 8970B in its SP 1.0 mode.
The input an output are not far apart and the source is pretty flat, as is
the receiver in the 8970B. You might find you get an almost identical result
to SP 1.3.
Yeah, the manual can be aggravating at times; it can take a lot of digging
to find what you need to know.
Hang in there, you'll get it working.
Tom Holmes, N8ZM
|
Re: HP 8664A signal generator repairability
On Mon, Mar 18, 2019 at 07:41 PM, Colby Burkett wrote:
Sean, I have a lot of parts for the 8664/8665 series sig gens. Let me know what you need and I'll see if I have it.
?
Colby
Would you happen to have the bit of hardline coax that connects the front panel RF out to the inside? I'm worried that might have been damaged. Sean
|
Re: HP 8664A signal generator repairability
On Wed, Mar 20, 2019 at 05:38 PM, Colby Burkett wrote:
Agreed on the Bird reference. They're great inline watt meters for high power needs, but not for fine performance verification. I use the E4418B and E4419B w/8482A and E9300A heads, or a U2001H USB Power Meter - depending on where I am....
?
Colby
Thanks as well...good stuff to start watching for. Sean
|
Re: HP 8664A signal generator repairability
On Wed, Mar 20, 2019 at 05:28 PM, Bob Albert wrote:
For a decent power meter you do NOT want a Bird.? They are pretty sloppy.
?
I use my HP410B for example.? I calibrate it with my HP 3456A in the audio range and then it's good up to at least 500 MHz.? My HP 8657B isn't all that accurate and I have to use the Offset to get more accurate readings.? The spectrum analyzers are, as you say, not too accurate.
?
On the other hand, I don't need precision amplitude.? All I do is calibrate S meters and you can barely see 1 dB on them, and most aren't too accurate over the range.? So as long as I have 50 microvolts I am happy.? Flatness is another issue but then the SA does a good job.
?
Bob
?
?
?
On Wed, Mar 20, 2019 at 03:03 PM, alwyn.seeds1 wrote:
Dear All,
?
Measuring signal generator output accuracy with a spectrum analyser, even as nice a one as the 8566B, does not make a lot of sense since the amplitude accuracy is limited. The 8566B flatness is +/- 2.2 dB to which has to be added the calibrator accuracy of +/- 0.3 dB. A good signal generator has much better accuracy.?
?
What you need for the job is a decent RF power meter, which will give you a total measurement uncertainty of about +/- 0.4 dB.
?
Regards,
?
Alwyn ?
Hi, I am looking into getting one. The purpose of the above exercise is simply to make sure there are no gross problems with the 8664. Of that I am quite sure now, so I am keeping the instrument and getting a bit of a refund from the seller. Sean
Thanks for the pointers on power meters. Needed one eventually anyway (isn't that the best excuse for buying more test gear!) Sean
|
jafinch78 .
Keith Monahan
