8116A Code 42

[Mick B]

Thanks Mark, I understand. I regret wasting any of your time, like I said just learning.
I have changed all the results to reflect, Volt's from U32 ~ U34 Starting at « Reply #70
leaving all the text the same for people reading these posts to understand that I was using the term "mv" when it should have been "V" very misleading, details matter.

[MarkL]

Ok, now with the updated numbers...

Both the old and new U32 I agree look ok.  All the outputs are complements of each other and at good ECL logic levels.

Is this a new U37?

U37.15 is too low for an output, and may be the reason for the heating.  It may be getting dragged down.  As a test, you could unplug U37 and place a 100 ohm resistor to ground on U37.15.  The only other things on that node would be two inputs U36.6, U36.7 and pull-down R62.  ECL inputs are high impedance, so the inputs wouldn't matter and this test would form a voltage divider made up of 100 ohms and 470 ohms (R62).  You should get a reading of around 0.93V.  If not, one of the inputs on U36 is probably bad.

Almost all of the logic levels on U33 and U34 appear ok, meaning the outputs are consistent with the AND function, except for the output U33.9 which is not correct.

Also, the -0.7V logic HIGH tie point is measuring -0.967V, which is right on the edge of what's considered a logic HIGH.  That's something that should be investigated.

One thing at a time.  I would remove U33 and U34 and let's focus on why U37.15 is not right, as per above.  I think it's ok to leave U32 in for now.

[Mick B]

OK!

Is this a new U37?

Yes All chips are new.
Setting up a voltage divider brilliant.
The voltage divider  with U33 and U34 removed produced -0.96V
   

Also, the -0.7V logic HIGH tie point is measuring -0.967V, which is right on the edge of what's considered a logic HIGH.  That's something that should be investigated.

Is this the common input Starting at R64?
Or is it U33.12 that's suppose to be -0.7 V but is -0.948 V with chips removed
OR I don't know what you mean. I'm sorry i'm tired a long day in the sun.

[MarkL]

...
The voltage divider  with U33 and U34 removed produced -0.96V

Well, that value is correct, which says it's not U36.6 or U36.7 loading the output of U37.15, and it also verified the pull-down R62 is ok.

You could try the old U37 in that socket and see if you still get -2.60V on U37.15.

For comparison, my U37 is at 48.7C.
   

Also, the -0.7V logic HIGH tie point is measuring -0.967V, which is right on the edge of what's considered a logic HIGH.  That's something that should be investigated.

Is this the common input Starting at R64?
Or is it U33.12 that's suppose to be -0.7 V but is -0.948 V with chips removed
OR I don't know what you mean. I'm sorry i'm tired a long day in the sun.

It's supposed to be a fixed -0.7V provided by the voltage divider R58/R65, if we can believe the schematic.

However, calculating the resulting voltage using the resistor values on the schematic, it turns out it actually results in -0.92V (or -0.94V in your case with a -5.3V supply).  Why they labeled it as -0.7V I don't know, but is probably a red herring at this point.  Placing the logic HIGH tie point so close to the threshold was not a good design choice in any case.

So let's leave that one alone for now, but keep it in mind.

[Mick B]

OK Good news, I could not find the old U37. But I did find another chip from a different vendor and the results are
U37.15
Power on -1.77V
pushing the MAN button once -1.28V
pushing the 1 cycle button once -1.77V
pushing 1 cycle one more time -0.785V
and one more time and back to-1.77V
I pushed these buttons in this order right after power on. I believe I had a bad chip, It is mildly warm

I have enough 1% resistors to get -.7V  if you think that might be something to look into when this comes to and end. or leave well enough alone
Tomorrow, I will post the results of the whole chip, Put U33 & U34 back in and see  what's what

[MarkL]

OK Good news, I could not find the old U37. But I did find another chip from a different vendor and the results are
U37.15
Power on -1.77V
pushing the MAN button once -1.28V
pushing the 1 cycle button once -1.77V
pushing 1 cycle one more time -0.785V
and one more time and back to-1.77V
I pushed these buttons in this order right after power on. I believe I had a bad chip, It is mildly warm

The replacement is mildly warm?  If so, that sounds a lot better.

Those voltage levels are also good, except the -1.28V.  That's in between HIGH and LOW.  Are you still measuring these values with scope?  Maybe it's pulsing?

I have enough 1% resistors to get -.7V  if you think that might be something to look into when this comes to and end. or leave well enough alone
Tomorrow, I will post the results of the whole chip, Put U33 & U34 back in and see  what's what

Looking back on the measurements from my unit, it's also not -0.7V (it's -0.93V).  I think there was an error in calculating the resistor values in the design, and/or that node should not have been labeled "-0.7V".  But obviously it's working.  We should leave this alone.

Interested to see what you get with U33 and U34 back in circuit.  With the prior measurement reporting errors, it's possible the old ones were ok.

[Mick B]

Sorry Mark to take so long to get back to you, I have been very sick. I caught the flu & got a sinus infection at the same time. I still feel like shit, but I couldn't lay in the bed another day, thank God for antibiotics.
Here is the newest reading of U37, U33, U34 with all chips in. Settings:  Mode, EBUR 14. Frequency 1K, Wave NONE, Amp 1V, offset NONE, output ON. ALL voltages checked with my Rigol 5354 Scope. AND on U37 pins 11. & 14 my HP 34401A multimeter these voltages seemed low hopefully these are close enough to evaluate.  I have done a self-calibration on my scope It takes almost an hour. If you think I need to redue these I will. Right now, I'm going back to bed apparently, I don't feel as well as I thought I did     
   U37
  4. (I) -2.17 V
  5. (I) -2.15 V
  2. (O)-1.17 V

  6. (I) -1.77 V
  7. (I) -1.76 V
  3. (O)-1.70 V

10. (I) -5.22
11. (I) -0.460 V ...NOTE: HP Meter 0.513 V   
14. (O)-0.671 V ...NOTE: HP Meter 0.739 V   

12. (I) -0.735 V
13. (I) -1.71 V
15. (O)-0.722 V
  9. (O)-0 V

    U33
  4. (I) -1.70 V
  5. (I) -0.715 V
  2. (O)-1.77 V

  6. (I) -1.64 V
  7. (I) -1.69 V
  3. (O)-1.78 V

10. (I) -1.69 V
11. (I) -1.79 V
14. (O)-1.78 V

12. (I) -0.904 V
13. (I) -0.710 V
15. (O)-0 V
  9. (O)-1.76 V

    U34
  4. (I) -1.69 V
  5. (I) -1.78 V
  2. (O)-1.77 V

  6. (I) -0.735 V
  7. (I) -1.69 V
  3. (O)-1.78 V
 
10. (I) -1.69 V
11. (I) -0.715 V
14. (O)-1.77 V

12. (I) -0.903 V
13. (I) -1.68 V
15. (O)-0 V
  9. (O)-0.731 V

[MarkL]

No problem; I hope you are feeling better!

Except for U37.2, between all the other U34 U37, U33, and U34 readings you posted, there aren't any readings that are out of range of ECL logic levels, or inconsistent with the gate functions.  That doesn't mean they're necessarily working, but perhaps we're almost passed the point of locating shorts and similar issues using this static measurement method.

U37.11 is coming from a level shifter, so I'll give them a pass on being a little outside a normal ECL HIGH.

Is anything still getting hot?

On U37.2, you are reporting it as -1.17V, which is between an ECL HIGH and LOW, but on the other seven pins attached to this node you are reporting it as -1.70V.  Was U37.2 a mistake?  If a mistake, then I would say all the nodes are consistent.

It might be interesting to see if anything is working now on the burst function by setting the unit to internal burst mode (I.BUR).

EDIT: Fix typo.

[Mick B]

Feeling a little better today, no fever. Apparently, that U37.2 reading was an error, I wrote down -1.71 but typed -1.17 dyslexia strikes again, I try very hard to catch these. The hottest chip on the board U34 @ 46.2C So all is well......? almost I'm still having issues with the BUR light intermittently working and I wonder about the error light, not exactly sure when it is supposed to come on I read it somewhere but don't remember where. as I adjust the timing depending on other settings, I can make the error light flash.
I did the performance test for Burst mode, it made it up to 816 in EBUR and I looked at EBUR & IBER on my scope and it looked good, if I understand IBER, it just keeps repeating what you have it set at, Correct? I have never used this mode before. The sweep mode also works.
This adventure so far has been just Great. 

[MarkL]

...
The hottest chip on the board U34 @ 46.2C So all is well......?

ECL dissipates heat even if nothing is switching, and that temperature is fine.  I don't think you have any internal shorts now.  So that's good.

I'm still having issues with the BUR light intermittently working and I wonder about the error light, not exactly sure when it is supposed to come on I read it somewhere but don't remember where. as I adjust the timing depending on other settings, I can make the error light flash.

The error light comes on when the settings don't allow the unit to complete a triggered event before the next trigger comes along.  For example, in burst mode it would be an error if the repetition rate was too short and did not allow the set number of burst pulses to occur.  An error could also be casued if the repetition clock and/or pulse clock were not in spec (i.e., could cause the current event to get overrun by the next event).

Keep in mind all the time related circuitry on this unit is analog, so the exact settings where the error starts to occur is going to vary a little.

I did the performance test for Burst mode, it made it up to 816 in EBUR and I looked at EBUR & IBER on my scope and it looked good, if I understand IBER, it just keeps repeating what you have it set at, Correct? I have never used this mode before. The sweep mode also works.
This adventure so far has been just Great. 

If it passes the performance verification, that's good.  That's correct on I.BUR.  It will keep repeating the set number of pulses, and will have a gap at the end if the repetition time permits.  E.BUR waits for a manual or external trigger before emitting the set number of pulses.

Is there anything that's not working now?

Does the error light come on unexpectedly?  If so, under what conditions and what is the resulting output?

EDIT: typo.

[Mick B]

Hello Mark, At the moment it would seem the Burst problem is healed, but it might be interesting to see how this would play out (Page 10.6-7)   I was also hoping you might help me go through this, so I can get a working knowledge of my (HP 5005A Signature Multimeter) while verifying the work is done. My son bought me this piece of test gear, I have never used it, I have the manual and the basic bootup and functions all seem to be ok with no errors.
 I believe the Error light is working as intended. leaving the intermittent BER light I am going to attempt to find what's up with that, it's off more than its on, and looking at U20 as on Pg. 10.6-7 with the Address decoder and my Rigol 16 channel Logic Probe on my scope would go a long way in my understanding of how to use this equipment and the confidence to use it in the future. I realize this is unnecessary, look at how far you got with a voltmeter! I really want the understanding you have, to be able to look at a circuit understand how it's supposed to work, like understanding ECL logic and knowing what the inputs should be and what to expect at the outputs and where to look if it's wrong.
You have been great at explaining things to me, very easy to understand and pointing me at the right literature to supplement that. Please note, your time is not wasted on me.     
   

[MarkL]

Great!  Sounds like everything is working now.

The HP 5005A is many instruments in one, so I would recommend a thorough read through the user manual.  I have a 5006A, which is just a signature anayzer.

If you want to try out the signature capabilities on your 5005A, there are very explicit step-by-step instructions on how to set up the 8116A for signature analysis, and then tables of expected probe results.  All starting on page 10.7-17.

One useful thing to do is verify your ROM(s) on page 10.7-19.  ROMs can go bad, which usually leads to outright failure, but it's still not a bad thing to verify them.  There are several versions of firmware and you may have to look through the service manual change information to find the right signature table for your version.  Each firmware version has it's own ROM part number and you'll need to cross reference the ROM part number(s) to the change information to find the right table.

And finally, if you want to learn more about signature analysis, there's a whole HP Journal dedicated to the topic:

  http://hparchive.com/Journals/HPJ-1977-06.pdf


For more general learning, my favorite book is the Art of Electronics by Horowitz and Hill (two volumes):

  https://www.amazon.com/Art-Electronics-Paul-Horowitz/dp/0521809266

  https://www.amazon.com/Art-Electronics-x-Chapters/dp/1108499945

There's also a course book and a student manual, neither of which I've seen.  I don't know how applicable they would be without a professor and appropriately equipped lab to go with them.

[Mick B]

Well, wouldn't you know it my 5005A is having an issue with one of the pickups the Red Qual/Stop one. resister quick test shows open. don't think it's a big deal.
thanks for the links, I have the 2nd edition The Art of Electronics. and the service manual for the HP 5005A such as it is.
also, the first link you sent is all about a power supply, I read it all. And from there I was able to find lots of stuff about Signature Analysis Here is a good one. Application Notes 222-4

https://www.hpmemoryproject.org/an/pdf/an_222-4.pdf

Guidelines for Signature Analysis, Understanding the Signature Measurement

and this one is for all the Application Notes 222-xx

https://www.hpmemoryproject.org/ressources/resrc_an_03.htm

Thank you that website is a gold mine.

Thanks again for all your help, perhaps we can do it again in the future. 
Mick.

[MarkL]

Oops, wrong month.  Here is the correct one:

  http://hparchive.com/Journals/HPJ-1977-05.pdf

That's a good app note, thanks.  Very detailed and practical.  The HP Journal focuses more on the theory and design.

Best of luck, and hope you enjoy your working 8116A and learning the 5005A!

[Mick B]

WOW, the 5005A is not your daddies VOM. Looking into open in the Qual/Stop lead doing the quick test. I changed the leads on the A6 timing pod when I first got this as there were no grabbers and messed up ends, then performed Q test and all was good, put it on a shelf until yesterday, then this happened. HP has a very convoluted was to test resistance in this thing. PS I downloaded the journal, I will read it.
On the outside chance, you might want to partake in the festivities IF I run into trouble, here is the service manual. Enjoy your weekend.

 https://arcarc.xmission.com/Test%20Equipment/HP/HP%205005A.pdf

Just in case, I DL-ed this file no virous 
                   

[Mick B]

I found the problem..... it's the wire harness. I might have found a new one complete with A6 pod, will see I'm waiting for them to get back to me. I'm also looking into making a new one its only 3 wires between the A6 board and A3 board Red, White, & Black. In the A6 Pod Green, yellow and red go through several RC circuits and come out as 2 wires and a ground.

[Mick B]

Hello Mark, while trying to calibrate this HP 8116A I've run into a problem. Perhaps you can steer me in the right direction. 
when setting up the voltages IE: Square High Amplitude & Offset Pg. 9.4-5 It would seem I have a LARGE Over-Under shoot issue
In the picture the AMP is set to +10V. at +16V it is off the screen. At +10V you can see the P-P voltage is +13.44
the brown lines are indicators for the Setting VMAX & VMIN the overshoot is in the 30s% and not <4% making it impossible to set.
This is one of the first adjustments. This was not like this when I started, I made it all the way through the calibration and checked to see if the frequency and voltage levels were good, and I found this.
Do you see a place to start?

[MarkL]

First, I would try a quick check to make sure it's not a measurement issue.  Move your cable to the trigger output and make sure it's a clean square wave on your scope.  Levels will be at 0V and 2.5V.  Assuming that's ok...

What do those spikes look like if you reduce the time/div?  Is it a simple overshoot or is it ringing?  Is it the same in complement mode?

Are the "Overshoot and Transition Time Adjustments" starting on page 9-6 still within acceptable limits?

If everything there appears to be ok, I would verify any adjustments having to do with frequency response, like flatness or distortion.  Anything timing related like frequency or width are probably not involved in the overshoot.

Failing all that, I would take a look at TP4 and see if it has the same shape of waveform as the output.  If it does, the output amplifier is being true to the shaper output and the problem adjustment is probably in the shaper area.  Otherwise, it's an adjustment further downstream in the output amplifier.  (Assuming the current issue is introduced by a change in adjustment.)


As a future hint, one thing to do is carefully mark all adjustments with a fine Sharpie before changing anything.  That way you can at least get back to where you started, and you are more free to experiment with any adjustments.  The marks can be removed with isopropyl and a q-tip, if needed/desired.  High-res photo(s) can also work, as long as they show everything.

[Mick B]

Starting over all pots at center, as per the instructions. The previous owner did this. I have to go with it,4 chips and the main board were replaced.

Resistance measured on an HP 34401A

Picture's 1~4
 
1. This a 50.3 ohm terminated BNC TEE. With the 3ft cable the resistance is 50.65 through a banana adaptor
 
2. This a picture of a new BNC cable with a 50.1-ohm 2W pass through resistor on the trigger output (perfect)

3. This is the something but with the cable I was using for the main output.

4. This is the same cable as picture 3 but with the 50 ohm terminator resistor from picture 1

This does not do this all the time. It could be doing this with the pass through I just can't get it to do it now
adjustment on page 9-4.5 to 9-4.15 came out perfectly.
What could make this do this? I have been messing with this for hour-----days before I stumbled onto this

I'm going to do the calibration to the end again, using just pass through on the main output and a 20 dBm on the trigger output. on the scope, for adjustments that use the trigger output on my counter HP 5334A has a 50 ohm / Z button do you see a problem with this idea?

 

[Mick B]

Hello all, well it made it through the calibration as before, everything is working as it should. It sucks a faulty 50-ohm termination that breaks down under load caused this.

On to making a cable for my HP 50005A I can't find one, it looks like three RG-316 coax in a sheath.
Anyone seen such a thing?

[MarkL]

Glad you figured it out.

The terminator could have been marginal or constructed with a cheap resistor.  At its maximum, the 8116A can output +/-8Vpp square wave which would be 2W 1.28W into 50 ohms.  A good 2W terminator should be able to handle this.

I'd recommend keeping around a few brand name terminators for calibration tasks, like from Minicircuits or Tektronix.  The feedthrough terminators are even better for high frequency work since there's not a stub, e.g. Tektronix 011-0049-01.

Sorry, no idea on obtaining tri-coax cable.  I would give it your best shot to salvage the existing cable by stripping back past the bad section(s).  Failing that, you could try to replace it by bundling three individual coaxes, but you would need to determine the impedance of the old cable to select the right replacement.  This can be done with an LCR meter or TDR.

EDIT: fix 1.28W

[Mick B]

Hi Mark, after some investigating, I think that the terminator was 1 watt. I ended up getting 2watt attenuators for calibrating this and I had a 2watt passthrough. Bottom line I just didn't think about the wattage of the stubby one. I won't do that again. 
The Tektronix 011-0049-01. are very expensive I found some used ones on Ebay $49.00 but he's on vacation and has 10 I think, I can get a few of these.
I kept cutting back a few inches and still have an open, now the cable is short. And I will still have to overcome (see picture # 3 crimped on). The braids were screwed, crimped, and peened it took me about 45min to get the braids out of these holding devices.  
Nobody has tri-coax cable everyone's impedance is either 50 or 75 ohms I haven't measured HP's yet but hopefully its 50ohm, if it is I am leaning towards one of these
I have 50ft of the RG-316 and then there's Pasternack's LMR-100A-UF. either way I will have to come up with a sheath of some kind or someway to keep the 3 somewhat together. OR PUTTING IT IN THE CLOSET UNTEL A PART SHOWS UP!

[Mick B]

Hay all, this is a P.S. to the above post speaking about Attenuators, I should have added that (quit using attenuators to adjust this HP 8116A after your (MarkL) explanation as to why HP were using them in the first place) This can be found in an earlier post in this thread. What a time saver.

[MarkL]

...
The Tektronix 011-0049-01. are very expensive I found some used ones on Ebay $49.00 but he's on vacation and has 10 I think, I can get a few of these.

That's a pretty high price.  There's a couple of sellers in the $25 to $30 range. As long as it's ebay category "Used", you can return it if damaged.

Also, I discovered that the 5005A data pod,  #05005-60006, is the same one that's used in the 5006A.  It may help broaden your search to look for a dead 5006A, but it's probably just as scarce.  The pod itself has been superseded (later rev) by #05005-60102, according to Keysight Find-A-Part, so that could work too.

[Mick B]

Hay Mark, thanks for the info, in the meantime I'm going to make one until one comes up. I just got two 011-0049-01 for $39.00 2 for $50.00.
I'm now looking for some better-quality Tee's/splitters, the ones I have the male connector are crimped and are starting to fail at the crimp. Do you have a Brand to recommend for this part?

Alos I am going to start on the other HP 8116A without option 001, it has the same problem high overshoot at 16V@100kHz square wave. I'm going to look at the recommendations you gave me for the other one with this problem.
 I have been wondering how this unit produces a clean square wave.
 and was thinking about the circuit were the caps 528 and 530 are. The manual says they effect the overshoot. Pg.10.4-5. (see picture) can you explain this circuit?
 I don't have any very small PF caps but I'm going to get some, the problem is they are 1pf 200V is this voltage necessary? is there a spike potential? I can get 50V all day long. I will have to confirm this, but I think I saw this overshoot on the trigger output also TP4 looks good
Well that's another project.   Thanks.