HP gear had best-in-class specs and stability.? As parts inevitably change characteristics with temperature, the high airflow made the best of it by minimizing hotspots.? If you are always going to operate at 20C and maybe accept pushing the edges of the specs, you can go with less flow.
I have a load bank which had a fan louder than anything else I own. However, I rarely if ever need to run at full power.? I found a size compatible fan which is temperature controlled with an external sensor and mounted the sensor on the main heatsink.? The unit is far quieter but at high load the fan picks up speed a lot to keep things cool.? Some newer fans have blade designs meant to keep turbulence noise much lower.
Peter
toggle quoted message
Show quoted text
On 12/15/2024 9:46 AM, Wilko Bulte via groups.io wrote: /HPs cooling engineers must all have shared the same hobby: deafening rock music./
/
/
My bet is they were all working from the same, kinda horrible, environmental specs, e.g. 55 ambient temperature. That, more often than not combined with linear PSU, calls for Real Fans[tm] ????
Wilko
|
I have found that many "used" units have been run continuously in
racks or labs and the fans are often
very noisy. I recently got a Racal counter/times where the fan was
deafening.? I found a near replacement
"new" that was the same rating and cf/m fan on fleabay and the
difference was astounding, near silent.
A bench PSU I had was a bit too noisy, bearable but irritating,
it was new so changing the fan had no effect so I fitted a
temperature controlled? fan
controller that only cuts in as the heat-sink heats up. Another
possible solution.
Dave
On 15/12/2024 15:01, Adrian Godwin via
groups.io wrote:
toggle quoted message
Show quoted text
On Sat, Dec 14, 2024 at 10:16 PM, Martin wrote:
HPs cooling engineers must all have shared the same hobby:
deafening rock music.
?
Only the first one. After that the group was doomed by it's own
products.
Oddly enough, my desktop PC is an ex-business Hp because I've
found them much quieter than Dell and the like.
?
?
|
On Sat, Dec 14, 2024 at 10:16 PM, Martin wrote:
HPs cooling engineers must all have shared the same hobby: deafening rock music.
?
Only the first one. After that the group was doomed by it's own products.
Oddly enough, my desktop PC is an ex-business Hp because I've found them much quieter than Dell and the like.
?
?
|
HPs cooling engineers must all have shared the same hobby: deafening rock music.
My bet is they were all working from the same, kinda horrible, environmental specs, e.g. 55 ambient temperature. That, more often than not combined with linear PSU, calls for Real Fans[tm] ????
Wilko
|
Re: 3577A multiple problems - update
?
Having experienced that supply bypass components in modules and on PCB's
such as inductors or resistors (Allen Bradley carbon resistors in particular)
can go bad, I would recommend checking all supply voltages at the respective
module.
?
Ulf
SM6GXV
|
Re: 3577A multiple problems - update
Hi Peter.
Sounds a bit suspicious, multiple issues all at the same time.
As you had (and fixed) one one PSU issue, unless you have already
checked the other PSU output rails, I'd suggest you do that, and
the condition of power distribution network, interconnects +
decoupling caps elsewhere etc..
Older ROM's, EPROM's and RAM chips can "misbehave" when one or
more DC rail they need isn't quite up to snuff, as can (as I have
found) older CPU chips etc.
Moving RAM chips around (same type of course) does indicate one
or other is suspect, but again if the DC supplies are not clean
and stable, one might be "just on the edge" at the best of times
anyway.
The OCXO?? Leave it powered for a day and see if it can be pulled
back into spec.? Assuming nothing else goes "funny" while waiting.
Is the CRT subsystem OK, no random twitches that could indicate a
HV problem?? Perhaps also "spiking" other things that get upset.
Regards to All.
Dave 'KBV
|
Hi Hendrix,
I repaired one of those (actually the type with VFD display). Don't yours have a temperature controlled fan?
HPs cooling engineers must all have shared the same hobby: deafening rock music.
cheers
Martin
|
File /All HP, Agilent and Keysight instruments in folders by part numbers/3000 to 3999/3478A 5.5 digit Multimeter/HP 3478A Service Manual 03478-90004.pdf uploaded
#file-notice
The following items have been added to the Files area of the [email protected] group.
By: Jared Cabot <jaredcabot@...>
Description:
A clean, high resolution, bookmarked version of the 3478A Service Manual.
|
I purchased this HP6645A power supply from Radu Dicher about a month ago and have been a bit slow getting it set up for use due to my other obligations. It's exactly as advertised, appears to work just fine, and I even added banana jacks in the rack mounting holes (which I'm not otherwise using) so I can use standard test leads with it. Looks like a really sweet unit. My only issue is fan noise - not horrible, but a little annoying for continuous use. Has anyone else replaced an HP 6645A fan with some quieter type?
Steve Hendrix
|
HP 8648B RF Sig Gen. Repair and Output Attenuator 333222-600111 Woes
S/N of the signal generator is 3623A01683 circa late 90's, LCD display.? Options Rev Pwr.? Worked swell for a decade.
S/N of attenuator is 06102
?
Original symptom was output dropping 40 dB (with an attendant clicking flutter.? Originally thought problem was power dropping out- sounded like logic fluttering around but eliminated this.? Zeroed in on output attenuator misbehaving.? Didn't understand the intermittent part.? Attenuator is marked 24V on the outside and without any service data,? I assumed it was a simple 24V open collector drive type thing initially.? I identified the +24V and ground pins and applied that from a bench supply, no smoke, current low, all made sense.? Applied gnd to all the pins assuming that I'd be pulling down relay coils etc.? No luck.? Inserted back in SG and attenuator wasn't functioning even as well as before- must have mis applied some volts to the attenuator.
?
Took apart the attenuator, all the O'Rings were intact but I will replace them anyway.? What I was surprised to find was a little surface mount PCB with 4x dual monostables (74HC123), 2x 7 channel relay drivers- some appear paralleled and some are unused (ULN2003A), power comes in at 24V directly to dual coil solenoids (4x), 24v also goes through a Zener dropper (8v) and into a little linear (78L05) that powers the logic.? ?There are bunch of pull ups etc as I would expect.? All this stuff was so cheap that I just stripped the board of IC's and replaced them all- its $5 and all easily available.? No problems, some SMT rework practice.
?
Powered up and started tracing circuits, it basically is a dual coil magnet latching kind of deal. When the input changes, it hits the monostable that hits solenoid with a 24V pulse to set and another to reset.? There are little perm magnetic hats that go over solenoids for latching.? Its pretty cool, power is super low- no solenoids energized in static state.? The monostable takes in logic signals from the main board (from one of those blasted SCX6206TQM/V2 gate arrays- its obtainable and might be next).? Trying to get attenuator to operate on the bench with no luck.? The logic has high value input series resistors and lots of pull ups and downs, etc.? I'm grinding through documenting it but its tedious and my hearts not in it.? Has anyone does this and have the info?
?
My question is, does anyone have experience with these "smart" attenuators and with this logic in them.? Any help would be appreciated.? I will document and share what I find in any case.
?
I have a feeling that the original fault might have been that gate array and rest of this is just shop wear.? Regards and Happy Holidays.
?
John
|
Re: 3577A multiple problems - update
The unlock is due to the A7 module.? The problem seems to be in the 300-500 MHz oscillator.? Using the W1 jumper to set it to the low and high frequencies is supposed to give a 300 MHz total range (say, 250-550 MHz) but I get 363 to 529 MHz.? So something's wrong in the oscillator circuit that prevents it from going to the lowest frequencies.? As a test, I tried to set the analyzer to 190 MHz where I would expect the synthesizer to put out 490.250 MHz, and it did, however after a short time it went unlocked again and went up to 529 MHz.? So something additional seems to be wrong.? I checked the internal +12 volt supplies on that board and they are all in spec.? Yikes, might be a hard to find varactor diode.
?
I still get a CPU invalid instruction error after a while.? I think I'm going to take a break from the VCO problems and see what I can find in the CPU.? Perhaps a bad ROM which is temperature sensitive?
?
Anyone have similar problems with this instrument?
?
I can't believe how messed up it is.
?
Peter
?
|
Re: 3577A multiple problems
Update:
The +15 issue turned out to be a bad adjustment pot. Replaced and it looks good now.?
Beginning to lose faith in old HP gear.?
The oven oscillator first had no output. Dirty power connector to board. Now, after it warms up, I can¡¯t get the adjustment to bring it to frequency. Best is 10 Hz low. ?I can use with an external reference, but it looks like it needs to be replaced.?
I get A5, A6 or A7 unlock error message on internal or external reference. That¡¯s next?
But I am also getting intermittent bus errors and illegal instruction errors. Wow.?
Not sure what ISN¡¯T wrong with this thing! ?And it was working fine before. Yeah I reseated the boards and used DeOxit Gold on the connectors. Power supply seems solid and I didn¡¯t see anything on a scope when getting the CPU errors.?
toggle quoted message
Show quoted text
On Dec 14, 2024, at 1:24?PM, Peter Gottlieb via groups.io <hpnpilot@...> wrote:
? ? Thanks.?
I¡¯m pretty confident two RAM chips are bad as when I move them the reported fault moves with them.?
I just discovered that I cannot adjust the +15 supply and it is at the high end of acceptable at 15.5 volts. Also there is excessive ripple. Pretty sure it is the LM338K so I ordered a replacement. Also if you watch long enough you see the voltage jump around. The -15 supply is adjustable and very close. I am not confident I have found all power supply issues. Sometimes it takes a few minutes of watching to see something going wonky so it is a slow process.?
I am also seeing oscillator unlock messages as well as bus errors but won¡¯t go after those until I am sure the power supply is perfect.? On Dec 14, 2024, at 12:53?PM, dan.meeks222@... via groups.io <dan.meeks222@...> wrote:
? Peter -
No help here, I just wanted to give you a data point: I have had one of these for a long time, and it sits unused for weeks or months between uses, and I have not had any issues.?
If there is something I can do to help you troubleshoot, let me know. I probably don't want to dig deep into the guts or solder anything but if there is something less invasive that might help I am happy to do that.
Dan
|
Re: 3577A multiple problems
? Thanks.?
I¡¯m pretty confident two RAM chips are bad as when I move them the reported fault moves with them.?
I just discovered that I cannot adjust the +15 supply and it is at the high end of acceptable at 15.5 volts. Also there is excessive ripple. Pretty sure it is the LM338K so I ordered a replacement. Also if you watch long enough you see the voltage jump around. The -15 supply is adjustable and very close. I am not confident I have found all power supply issues. Sometimes it takes a few minutes of watching to see something going wonky so it is a slow process.?
I am also seeing oscillator unlock messages as well as bus errors but won¡¯t go after those until I am sure the power supply is perfect.?
toggle quoted message
Show quoted text
On Dec 14, 2024, at 12:53?PM, dan.meeks222@... via groups.io <dan.meeks222@...> wrote:
? Peter -
No help here, I just wanted to give you a data point: I have had one of these for a long time, and it sits unused for weeks or months between uses, and I have not had any issues.?
If there is something I can do to help you troubleshoot, let me know. I probably don't want to dig deep into the guts or solder anything but if there is something less invasive that might help I am happy to do that.
Dan
|
Re: 3577A multiple problems
Peter -
No help here, I just wanted to give you a data point: I have had one of these for a long time, and it sits unused for weeks or months between uses, and I have not had any issues.?
If there is something I can do to help you troubleshoot, let me know. I probably don't want to dig deep into the guts or solder anything but if there is something less invasive that might help I am happy to do that.
Dan
|
Re: 3577A multiple problems
The connectors and PC B edge connectors are first suspect In these 1970s..1980s HP meters.
Opening up and removing, reseating all the PC Cards , May fix it.
Then see the connectors especially any multi pin ribbon, etc.
Second suspect is any switches and pots that are scratchy.
Jon
|
Re: HP8594E repair with "ADC-2V FAILL"
What a coincidence!
I have just switched on my trusted HP8593E and it shows exactly the same problem.
It had been running without problems for the last 10 years in my lab.
?
Many thanks for the detailled description of your work, Vincent.
I'll see if it is the same problem. A good starting point anyway...
?
First, I have to fix my Tek scope that died earlier this week. Not a good week. :-(
?
Tom
|
Today I fired up my 3577A to begin to do some tests.? It has been sitting in my lab, fully environmentally controlled, for several years unpowered.? It was working perfectly when it was put away.
?
Immediately I got a trace memory failure, shows as banks 2 and 6.
?
I ran further tests and got failures:
Log sweep signal level test
Log sweep flatness test
Linear sweep signal level test
Linear sweep flatness test
Synthesizer and LO test
?
So I checked the power supplies and reference, all look great.? Hmmm, what could be a common cause then?
?
I started with the trace memory.? I put in a bunch of sockets and moved around chips.? Indeed two of the 2k x 8 Toshiba RAM chips appear to be bad as the failure follows them to whatever bank I plug them into.
?
Anyone else just have these chips die while sitting around unused?
?
As I need to order replacements I will have some time to probe around.? Apparently something else has failed just sitting around.
?
Any thoughts as to where I should start?
?
|
Re: HP 8593A Display repair
thank you for help.
picture added.
Am 13.12.2024 um 00:11 schrieb Yves Tardif via groups.io:
toggle quoted message
Show quoted text
Hi,
When the Spectrum loses all its calibration constants, there is a chance that you will lose the synchronization of the display on the screen. You should try to do a "DEFAULT CAL DATA" and "CRT HORZ" for the display.
But do in order if you don't see the softkey menu:
PRESET
CAL
MORE (softkey #6)
CRT HORZ (softkey #5) and CRT VERT (softkey #4)
TURN KNOB TO TRY SYNCHRONIZE DISPLAY
------------------
PRESET
CAL
MORE (softkey #6)
MORE (softkey #6)
DEFAULT CAL DATA (softkey #3)
If it's possible, send us a picture of your screen.
Yves
-----Message d'origine-----
De : [email protected] <[email protected]> De la part de Martin M
Envoy¨¦ : 12 d¨¦cembre 2024 14:36
? : [email protected]
Objet : [HP-Agilent-Keysight-equipment] HP 8593A Display repair
I have a repair of HP8593A with the following problems:
The PSU was fail. now complete recapping, it works.
The lithium battery was rip and the memory empty. Now it have a fresh battery, but the shp have to learn what is in the memory.
The display is fail: I see only a part of the screen, like a zoom in, and vertical triple.
How I can find the problem? It will be in the monitor box, inside.
thank you.
|
HP8594E repair with "ADC-2V FAILL"
A friend sent me a failed HP8594E with the error message "ADC-2V FAILL" over the screen. I searched the network and most of the info tell me the power supply is failed. I opened the machine and saw the +15V LED is dark to say the power for +15V is abnormal.
1. disassemble the PSU and there is no obvious burn-out device found.?
2. Should I focus on the PSU or check other part of the machine ? I took the 2nd way to isolate the other boards.
3. By pull out the board over the shielding compartment one-by-one and check the PSU +15V LED, it still turn dark.
4. In the final, I pull out the analog interface board, then the PSU +15V LED turn light and the the HP8594E to boot properly.
5. After checking the resistance over the bypass capacitor with multimeter, the +15VF show a low impedance to ground(below 1 Ohm)
6. My first guess is the large electrolytic capacitor over +15VF. Check the CLIP manual, I remove C317(470uF, 25V). No luck, short-circuit keep over +15VF to ground
7. In the next day, borrow a thermal image camera from friend. Using a power supply to provide 0.5V over +15VF and ground, the current consumpion is over 500mA. In the same time to check
the board with thermal image camrea. Yes, the C19 is highlighted over the camrea. With visual inspect, there is nothing wrong with it.?
8. After remove C19, the short-circuit over +15VF and ground is disappeared. Replace C19 with a new one and assembly everything back. Now, the HP8594E is worked again
9. This SA is not to work for a long time and it seems need to further calibration to keep its accuracy.
?
73 de bv3ue, Vincent
?
|
That's interesting that yours is the same. Yes I also found if I reduced the source power slightly everything was OK. I have the IF bandwidth set at auto rather than 4khz?
?
I've not noticed this with the one we have at work. Next time I'm in I'll check the settings on it.?
|
Martin
Froggie the Gremlin
bv3ue
