Fluke 8800A/AF issues

[GregDunn]

It arrived today.  It's a little nicked up on the outside (handle, edges, etc.) but otherwise pretty clean.  It's the /AF model with the optional external ref, and most important to me, a later model with the updated Ohms board - no AA batteries!  Internal date codes seem to indicate the latter half of 1980.

I put the shorting bars on the correct binding posts, and did a quick check.  ACV looks pretty good from < 0.1V up to about 12V - the highest AC output I had readily available.  It tracked my semi-calibrated Heath IG5218 meter (the signal source) pretty nicely over 3 decades.    Auto/manual ranging seems to be OK.

DCV is a bit weird - it's like the measurement is divided by 2 and offset by a volt or so.  0V input reads about -1V and 10V reads about 6V.  Other than that it seems to track input voltage monotonically.  No discontinuities or random garbage, just the wrong values.  Seems to do the same on auto or manual ranging.  The zero cal pot has no effect.

Ohms is just wacky: Open circuit seems to give the all-segments-on overload indication reliably after I worked the range switches a few times (cleaning probably needed).  Putting a 2.2K resistor or higher gives that same 5K reading; connecting a 220 gives me about 90 or 9K, depending on how it feels at the moment.  Short circuit gives OL on the 200 range, and a steady 0.95, or 9.5, or 950, as you move UP toward higher ranges - yes, the dp moves to the right instead of the left.  Auto gives the 0.95 reading.  Almost like it's not scaling, but thinks it is.

I checked the PSU and the ±18V and 13V supplies were correct (haven't looked at the waveform to see if they're clean DC).  The +35 reads +25 and the -35 reads -37.  Someone has clearly replaced C20-21 with caps of unknown provenance, but C18-19 are the dreaded original Spragues.  C22 had the brand "Brasileira" written on it but if it was replaced, it certainly was done well - the joints look professional.

So, I think all the PSU electrolytics probably need to go, and the Tantalums as well - correct?  Is it worth looking at anything else until that's taken care of?  I haven't been able to locate a schematic of the newer Ohms board - all the manuals I've seen show the old one with the AA batteries.  That board seems to have something broken, though it has no obvious visible defects.

Edit - found the 8810A manual which appears to have the new ohms board schematic.

[GregDunn]

OK, late followup - the +35 VDC looks terrible on the scope - clearly C18 is toast and I'm sure that is wreaking havoc with the VDC stage - it depends on the 35V supplies for balance and offset, whereas the VAC hardware does not.  This is enough to convince me to hold off on further troubleshooting of the VDC issue till the PSU is fixed.

Next, after scratching my head for a while I decided to pull the ohms converter and see if there was anything visually at fault.  In a nutshell, yes there is.

Obviously someone has replaced Q4 and not done a very good soldering job - that's not helping.  Additionally, someone replaced Q6 and it looks like they splattered flux all over, in addition to causing that nasty-looking joint when putting the yellow wire (OHMS OUT) back in. 

In the schematic, Q4, Q6, and Q8 are all drawn as FETs, but in the BOM, only Q6 is listed as a FET (N-channel, 343830 / NSSF50024).  Not sure if that'll be a problem to source if it turns out to be dead.  Looks like it's only being used as a switch to pass OHMS OUT to the next stage, so hopefully not.  The other two are listed as garden variety 2N3904.

If anyone has an 8810A or a later 8800A with this ohms converter, I'd be glad of any critiquing of my analysis.

[med6753]

The +35V supply at +25V will definitely be the source of your DCV issues. When I purchased my 8800A the -35V was down around -20V and resulted in no DCV readings. Re-cap of the PSU is definitely in order.

My 8800A is also the later vintage without the batteries. But I've had no issues with the ohms section so I really can't offer much advice. Once the PSU issues were resolved mine was in spec and required no adjustments. But I must say that is some crappy work on that board. Someone needs to step away from the soldering iron. 

[GregDunn]

Haven't had a chance to look at it again today, but I'm going to fix those joints and clean the board properly while waiting for PSU parts (or maybe I can find a cap to stick in there temporarily).  With luck maybe that mess is the only cause of the ohms board issue and I can avoid having to replace the FET.

This meter apparently resided in the repair shop at Nellis AFB; I suspect someone was distracted or in a hurry.  It does happen; I worked at a Navy depot and we had some great techs as well as a couple of not-so-great ones.

[GregDunn]

Ohhhhhhh yeah...  Bodged in a decent cap in parallel with the bad one and clipped one of its leads so it was out of circuit.  ±37V supplies are now right on the money.  DCV measurement with the leads shorted will now zero, and my AD584 voltage reference measures within .0005V of the written value (sanity checked by my 8600 and Keithley 197a).

Re-flowed the nasty joints on the ohms converter board, cleaned carefully with flux remover, allowed to dry thoroughly.  Ohms now works a treat in auto and manual modes, again hitting my sample resistors within tiny error margins (also sanity checked).

From  to  to  in 5 minutes.      Now to wait for my replacement caps to come in and do a proper fix.

While cleaning up the 8800A, I discovered that the power button has apparently been broken off and epoxied back together rather professionally.  Is that a good excuse to buy a "parts special"?     Seriously, I love this meter already and I've hardly had it for a day.

[vindoline]

Congratulations! Another Fluke meter brought back from the brink.

[particleman]

Give the piano buttons a good cleaning too. I love those meters also, mV, 4 wire ohm, 1200V, selectable range + autoranging. They seem to hold their calibration very well. Not much to not like about them. I have repaired a lot of things with that meter. Glad you like yours.

[GregDunn]

Yup, once I get the new PSU caps I'll need to pull the case apart anyway; that'll be a good time to clean switches and connectors!  I'm impressed I was able to get this far with only the metal shield removed and nothing physically taken out of the case.

I've only had about 3 hours of burn-in since the bodge fix, but I'm seeing just less than 1 LSD drift so far.  Quite impressive.

My only quandary now is what to do about feedback to the seller.  It was described as "working - very good to excellent condition" but I'm not sure if 2 of the 3 modes being broken disqualifies it.  I'm inclined to be lenient but don't want to be dishonest when responding.

[lowimpedance]

In the schematic, Q4, Q6, and Q8 are all drawn as FETs, but in the BOM, only Q6 is listed as a FET (N-channel, 343830 / NSSF50024).  Not sure if that'll be a problem to source if it turns out to be dead.  Looks like it's only being used as a switch to pass OHMS OUT to the next stage, so hopefully not.  The other two are listed as garden variety 2N3904.

If anyone has an 8810A or a later 8800A with this ohms converter, I'd be glad of any critiquing of my analysis.

 The 8810a is the later version I think , (Without checking  ) , as I have both the 8800a and 8810a. The 8800a has the old blue color case and the poxy battery ohms pcb where as the 8810a has the later color scheme like the 8XXX series.
Also the plugins were options and not standard for the 8810a. As who ever originally purchased the 8810a I have thought VDC was all they ever needed  .  , the function select switch had a blocking strip installed which prevented the
non installed options from being selected. Shouldn't complain to much though as it was a dumpster job.

BTW The 2n3904 is a NPN BJT

[GregDunn]

Yes, I have no idea what serial number the 8800A started being shipped with the 8810 style ohms board, but it was no later than 1980.  I fully expected mine would have batteries but needless to say was happy it didn't!

I have tons of 3904s in my parts box; I was just surprised that they list it in the BOM when the schematic shows a JFET.  Apparently they decided, after drawing it, that a less exotic part would do the job.  That's OK by me; one fewer weird transistor to track down if it ever fails.

[GregDunn]

All old caps replaced with shiny new ones having 105C ratings.  The two 220µF caps measured a bit low and lossy after pulling them, so I'm sure they were ticking time bombs as well.  The Brasileira 4700µF cap has been replaced as well, even though it measured within spec; I want a modern name brand on every cap in the power supply.

As expected, everything still works and the calibration seems unaffected (not that I would expect it to be; however, we all know what flux residue may do to high impedance circuits, and though I made every effort to clean it off, one never knows).  I'm returning this one to service and looking forward to many years of good use.

[particleman]

Great news. Im glad you finished it up. On to the next project.? It will be nice to use it to fix some stuff   

[GregDunn]

It's a definite winner on the bench.  Believe it or not, I actually have a valid use for multiple DMMs on many of my projects - when repairing an amplifier recently, the long term stability of the PSU under load and the state of both channels was pretty much unknown (it turned out I needed to replace several damaged components on both channels) and I needed to monitor:

1. AC out of the variac
2. Amp DC PSU voltage
3. Amp offset voltage
4. Amp bias current
5. at least one more test point as needed for troubleshooting
6. then, the scope channels after I get the DC operating points sorted.

So having 4 bench units and a couple of portables is very very handy.  I don't always need 0.1% accuracy, but knowing I can trust the meters implicitly is just one less worry as I look for issues.  A quick scan of the displays tells me if the PSU voltage is dropping, a component is drawing too much current, etc. etc. long before the smoke test fails.  None of this delicately moving probes around while the amp is powered up and risking an accident. 

I have another couple of small amps to work on, but before I do that, I need to get to my HP 6920B repaired and reassembled.  It probably won't get as much use as the DMMs, but having something to sanity check my meters and scope with will be good.