-
Fluke 8842A Repair
Posted by
TheDefpom
on 13 Oct, 2015 09:13
-
Hi All,
I am trying to source a vacuum fluorescent display for a Fluke 8842A, I picked the meter up cheap from eBay, stated as not working, all the internal voltages check out, but the VFD is cracked so the display is not working, I can hear relays etc. clicking as it boots up, so I am confident the only issue is the lack of display.
Anyone have a VFD or front panel laying around? this unit also has the rear plastic end trim broken too, if anyone has one of those.
If I cannot get a replacement VFD I might see about using an arduino or something to convert the display inputs and run an LCD instead, but wouldn't be as nice, so I would rather try and get this fixed up.
This one is dated 1983/1984 on the PCB's, and has the AC board installed too.
-
#1 Reply
Posted by
tautech
on 13 Oct, 2015 09:41
-
-
#2 Reply
Posted by
xwarp
on 13 Oct, 2015 11:11
-
I have a couple. Send me a message.
-
#3 Reply
Posted by
TheDefpom
on 13 Oct, 2015 18:49
-
Rob, I had seen those, I actually watched a VFD replacement video, I cannot remember who did it.
I have the service manual and diagrams for it, as well as a working unit for comparison if needed, the display is definitely toast, hopefully that is the only issue!
It looks like it was dropped, with the rear plastic bezel being broken and fairly smashed up.
-
#4 Reply
Posted by
TheDefpom
on 11 Nov, 2015 21:43
-
Hi All,
I am still wanting to find a VFD, and ideally a rear plastic bezel as well.
I just pulled this one apart, I was seriously considering replacing the display with a arduino powered LCD, (I already worked out how to convert the signals, at least in theory), but space is a bit tight, and I would need to cut out the front PCB to fit it, and would require me to bodge in jumpers for the front panel connector... I would rather get the right part, I don't want to butcher it just to fix it.
-
#5 Reply
Posted by
TheDefpom
on 19 Nov, 2015 17:22
-
Well it's getting there, Bill was able to supply me with a replacement VFD, so at least I can now see what it's doing.
Now the next issue is that it is behaving erratically, the rate indicators sometimes don't pulse and just stay lit, the readings are also very odd, and jump around wildly, never correct, and it is failing just about all self tests, but often passes no.1,2,3 sometimes 4, but always fails 5 and everything past that, almost every time it freezes on 10 and just sites there.
I disconnected the gpib and AC options to rule those out.
I hooked up my MSO to the ADC IC and its digital timings i checked looked fine, as per the manual.
I have verified that all the power supplies are within spec, and free of ripple.
I have checked the 8MHz crystal and 1MHz output from the IC, both fine and stable.
I have removed and reseated all of the IC's which are in sockets.
I still have a lot to work through.
The keyboard is working fine, the display updates with range changes, so I believe the display and keyboard section is ok.
The in putted voltage is getting to a test point (I cannot remember the no.) so it is passing through the front end.
-
#6 Reply
Posted by
TheDefpom
on 20 Nov, 2015 08:13
-
A bit of an update to this, I seem to have made some progress improving the issues, after re-soldering every single joint on the PCB it started to behave better and is actually giving some readings and not locking up when I change ranges or switch to Ohms, after soldering each section I would retry it, and noticed some self tests started to pass which previously would not (5,6), the locking up issue seemed to be related to the current source and resistance section, as that fault cleared once I resoldered that area.
Issues are still present though, it is still failing self tests 7,8,9,10, when it fails 10 the unit power cycles about 10 times by itself.
If I apply a POSITIVE voltage on DC it will read close to correct to a point (approx 50% full scale), and then suddenly jumps to a weird reading of 18.70749 or something very close to that when on the 20V range, other ranges also do similar things, HOWEVER, if I reverse the input to give a NEGATIVE reading, it reads normally to its range limit, so I suspect that there may be an issue with the calibration or ADC circuit, as the reading is not exactly inverse, so if I put in +3.5 volts I would get something close to that, maybe 3.4 but if I invert the input I get -3.6 so it is offset for some reason, there may be an issue with the way it is handling positive voltages.
The 200V/2000V ranges are out of wack too, but I am just pleased to (seemingly) get it functional without locking up.
I haven't tried clearing the existing calibration,or doing a new one, as I am still unsure if that is the issue, and the existing may actually be OK, also I don't have the required equipment to calibrate over 30VDC, or an AC source.
-
#7 Reply
Posted by
xwarp
on 20 Nov, 2015 08:47
-
Since you have seen improvements after reflowing some joints, reflow the joints for the power transformer. Those joints are known to create issues.
-
#8 Reply
Posted by
TheDefpom
on 20 Nov, 2015 09:43
-
Yep, indeed, I have already done those ones on the transformer, I did the entire PCB just to be sure that nothing like that is the cause, even though most of the soldering looked absolutely fine, with very few suspect joints to see, only 2 or 3, but obviously even though they looked OK, they were not, as resoldering the entire PCB made a difference and fixed some of the issues.
-
#9 Reply
Posted by
xwarp
on 20 Nov, 2015 10:51
-
What are those tests that it's failing?
-
#10 Reply
Posted by
TheDefpom
on 22 Nov, 2015 21:18
-
I spent some more time on this yesterday, and have narrowed it down to an issue with either the ADC (U101) or the op-amps that are tied to it (the comparator U102, and the AD Amp U103).
I can see the voltage level changing correctly at TP103 (-2V through to +2V) which is the feed into the ADC (I also injected my own voltages there to confirm), but the digital waveform at TP101 when viewed on a scope freezes between certain voltage inputs, and correlates to the incorrect voltage reading on the display (it reads correctly to about 1.2V, then reads 1.26V when between 1.20V and about 4V, and then jumps UP to a fixed 18.707V until the voltage input is above that level, then it reads correctly again), when the digital waveform is changing with the changing input voltage the display is correct, when the waveform freezes, so does the display reading.
Interestingly, when using a negative voltage input it works correctly about below about -1.2V input.
I tried injecting the -2V through to +2V directly at TP103 to bypass all the input circuitry to rule that out completely, it didn't change the outcome, so I am sure all of that is OK, as I pretty much confirmed operation of all of that as OK already.
I found some replacement LF411CN and LF412CN and ordered those, they were not that expensive, so I might just replace then anyway to rule those out, which would leave the ADC IC, a Fluke 715680 device... does anyone know if this is a rebranded off the shelf part, or is it really a custom part for Fluke? I found some parts which are in the same package called 715680 online, so I ordered a couple, they might be the same... maybe!
I also tried using a 10K resistor on the resistance setting, it resulted in an output of 1V at the ADC as it should, and gave the same 18.707 reading as the voltage input does, so that looks fine too.
-
-
How clean is the PCB ?. maybe a good IPA scrub around the ADC etc. (The PCB needs to be very clean!).
Since it appears to have been dropped there may be a hair line fracture on the main PCB, does lightly flexing the board change the test's outcome or affect the readings?.
-
#12 Reply
Posted by
TheDefpom
on 22 Nov, 2015 22:51
-
After resoldering the PCB I have cleaned it as well, so no flux etc is on the board.
Flexing the PCB doesn't change anything, no signs of PCB damage.
I have tried replacing a few of the caps this morning, C101,C102,C103 in case they were off value or ESR and effecting it, they where fine, measured OK when removed, no change in testing, so I put them back again afterwards (probably better quality than the replacements I have laying around).
-
#13 Reply
Posted by
xwarp
on 23 Nov, 2015 14:13
-
I still have more parts that I can pull from the spare if you need.
-
#14 Reply
Posted by
TheDefpom
on 24 Nov, 2015 00:26
-
If you have the 715680 ADC IC I think that is what is wrong with this unit.
This morning I used my MSO to track the digital outputs of the ADC (pins 27, 29, 30, 32, 33, 36 (interestingly it seems the manual might be wrong, it states pin 34 is 1/8th bit and pin 36 is 1/16th bit, but it seems that is wrong, as the 1/8th bit appears to be on pin 36, and the 1/16th bit is on pin 33!), with both scope inputs also monitoring the voltage pass through output of the IC (pin 23) and the input returned from the AD amp (TP101), I could see the digital was not changing (display not changing either), but the passthrough bit voltage was changing, as was the bit voltage at TP101 which is fed back into the ADC.
-
#15 Reply
Posted by
xwarp
on 24 Nov, 2015 01:00
-
I have two of those. Let me swap them out with my working 8842 and verify them.
Would a bad calibration have an affect?
-
#16 Reply
Posted by
TheDefpom
on 24 Nov, 2015 02:08
-
It works at certain voltages, and when it does it is very close to correct, so I think the calibrations are probably ok, if I inject a negative voltage it behaves better, but still has voltages where it does not change the reading or the digital outputs to the precision resistor ladder, despite the other points I mentioned changing in line with the voltage applied.
I am 90% sure it's the ADC which is faulty. I ruled out lots of other items and today's testing showed the digital not changing when it's input was changing, in the dead voltage ranges.
Another point I didn't mention before, prior to resoldering the pcb a couple of times the display changed to show just "05" and at that time I noticed the ADC was hot, so I suspect it had an internal short, or something.
-
#17 Reply
Posted by
TheDefpom
on 24 Nov, 2015 21:18
-
Here is a photo of the probe setup, and a video of the scope screen as I run through the voltage range, it shows where the digital and the display stop changing, but the tp101 test point shows the voltage as still changing, I couple of times I run up and down at a voltage to show it changing without the display or digital changing.
The digital outputs are the ones which feed the resistor ladder, it is probably the only way to try and work out what the IC itself is actually trying to do, apart from the data bus to the display, which is shown anyway on the display.
-
#18 Reply
Posted by
mlefe
on 22 Dec, 2017 14:12
-
Hi! I'm sorry to resurrect such an old thread: was it finally de A/D? Where did you find a replacement?
Thanks!
Sent from my SM-G950F using Tapatalk
-
#19 Reply
Posted by
TheDefpom
on 22 Dec, 2017 19:18
-
Hi, yes the adc ic had failed, I was lucky and someone from this forum sent me one from a broken unit, this fixed it.
-
#20 Reply
Posted by
xwarp
on 22 Dec, 2017 19:40
-
Hi! I'm sorry to resurrect such an old thread: was it finally de A/D? Where did you find a replacement?
Thanks!
Sent from my SM-G950F using Tapatalk
Let me look and see if I have any more of those.
-
#21 Reply
Posted by
mlefe
on 22 Dec, 2017 21:23
-
Hi! I'm sorry to resurrect such an old thread: was it finally de A/D? Where did you find a replacement?
Thanks!
Sent from my SM-G950F using Tapatalk
Let me look and see if I have any more of those.
Thank you!
Sent from my Galaxy S8
-
#22 Reply
Posted by
Velund
on 28 Feb, 2018 22:01
-
It looks like it was dropped, with the rear plastic bezel being broken and fairly smashed up.
I currently working on restoring 8842A with the same problem (only few small pieces left out of rear bezel, probably dropped unit).
Maybe someone can make detailed photos and make measurements of critical dimensions - to try making a 3D model and print quick'n'dirty bezel replacement from gray ABS on a 3D printer?
Anyway looks like I will need to find some broken joints, as my unit is working fine sometimes and sometimes can start to malfunction.
-
#23 Reply
Posted by
TheDefpom
on 04 Mar, 2018 03:58
-
Check the transformer connections carefully, on one of the units I had they looked fine, but were actually poor.
-
#24 Reply
Posted by
Velund
on 10 Mar, 2018 22:02
-
Transformer was the first thing I checked and resoldered (just to be sure).
My multimeter is working with some intermittent faults. Sometimes everything is fine and all self tests pass. Sometimes it start to show some strange results (like relatively large current with zero actual current) and start to fail some tests.
Looks like I should check and resolder everything, especially heavier parts, as it was subjected to a stress (rear bezel is destroyed completely, just a couple of pieces around screws remains - looks like it was dropped).
Still unable to find photos of rear bezel "from inside". I'm planning to try to make rear bezel replacement on a 3D printer, but have to make a 3D model first.