Fluke 332D DC Voltage Standard Restoration

[TERRA Operative]

I found this link that looks a bit interesting, I'd give it a go (once my unit is working) if I could find some spare header pins...
It might be a a way to upgrade the chopper amp if yours is faulty anyway maybe.

http://www.friedrich-messtechnik.de/index.php/bauanleitungen/87-fluke335-332

[TERRA Operative]

Time to stick this one back on the bench... It's been in hibernation for a little while.

So the latest is that the unit would trip out as soon as you turned the power dial to the 'operate' position. But if I took the lid off and left the safety interlock switch open, turned the dial to 'operate' and then pressed the interlock to activate the unit, it would come alive and kick out a super stable voltage.

So, after looking at the schematics it seems to be something in the Current Limiter PCB A5A6.
Luckily I have a spare (with completely trashed capacitors), and if I swap it in and let the caps reform a little, the unit starts to work.
Ok, I have my culprit...

After squinting at the PCB, and with the spare next to me, I soon found some problems....

It seems I am an idiot....

I installed a 22K resistor in place of the old 2.2K carbon composite (durrrrrrr), and somehow I used 470uF capacitors in place of the original three 250uF parts.... 

Ok, well the resistor is easy to get in Akihabara, and I'll add the capacitors to my next Digikey order.




More to come...

(I also now have a much older 335A to look at next too )

[TERRA Operative]

Well, looks like the rabbit hole does far deeper than initially expected, as usual..

So it turns out that Current limiter board is actually ok. When I actually paid attention to the schematics, the capacitors are just bulk smoothing caps after the bridge rectifiers and vreg section etc, the DC coming out is perfectly stable and all is well.
I replaced the 22K resistor with a spare one for now (It's actually too big, I only have a 1W, where the spot is for a 1/2W, new correct resistor coming soon) and that PCB is working fine as far as I can tell.

Now I can consistently make the unit fail by swapping out the Series Pass Driver PCB A5A2.
I'm not convinced it is actually at fault as such though. I have checked all the semiconductors on my Peak Atlas component testers and they all test ok, capacitors and resistors are good (no swapped values this time!) and everything seems fine...
My thinking is that the old board I am testing with the old components may have drifted enough to get a bit loosey-goosey and more tolerant of an actual fault somewhere else.

Although, I did notice that the schematic calls for two 12V zener diodes, but my PCB has two 10V zeners instead as per the parts list... The part markings in my PCB match for 10V parts, and were original parts, and they test bang on in the Zener diode tester too.
I've noticed a number of errors and ommissions from the schematics so far, so I've been updating it in the PDF file as I go. I'll release the file once I have been through it enough to catch what errors I see (and generally tidy up, bookmark, etc, the PDF file too)


I tested the range switch (all 1 million decks of that 3 position rotary switch!) just to make sure it is all ok, and after a small blast of Deoxit to clear up one contact that was reading a couple ohms, all contacts are now reading around 0.1 ohms (not accounting for test lead resistance) so it is now perfect.


I did notice that if I try to turn the power knob from Standby to Operate, the current meter on the front panel shoots up then the unit trips out. Like the unit is hitting over-current.
But if I leave the internal interlock switch open to inhibit the output when I turn the power knob to Operate, then press the interlock switch to enable the output after a second or two, the unit works perfectly.

Hmmm, could be an inrush current problem with the new capacitors somewhere maybe?

[TheDefpom]

Now I can consistently make the unit fail by swapping out the Series Pass Driver PCB A5A2.
I'm not convinced it is actually at fault as such though. I have checked all the semiconductors on my Peak Atlas component testers and they all test ok, capacitors and resistors are good (no swapped values this time!) and everything seems fine...

As good as the Peak testers are they don't pick up high voltage breakdown, I had an issue on my Fluke 5200A where one of the high voltage output transistors tested fine but when high voltage was put on the part it would collapse and go to low resistance...(I showed it in the video series I did), check it on a curve tracer that can make sure it meets its spec for voltage breakdowns.

Don't rule out that transistor, my money is on that :-)

[TERRA Operative]

Yeah, true that.

All the smaller transistors are socketed, so I was able to swap them between the old spare good board and bad board from this 332D, still the bad board was bad. Although there is a small TO-66 packaged transistor that is soldered in so I didn't try swapping it yet.

[TERRA Operative]

And the Defpom called it.

Turns out that Q3 on the Series Pass driver is actually bad, even though the Peak DCA75 Pro thinks it looks ok.
I swapped in the transistor from the Series Pass Driver board in my 335A that is next up on the bench and now the 332D is now mostly working!

So, time to go find a '40250' TO-66 transistor, or equivalent thereof.


Now, the next issue.

I can consistently get the machine to operate when in 100V or 1000V mode, but if I start in 10V mode it trips out.
BUT, if I start in 100V or 1000V mode, then switch to 100V mode it operates flawlessly besides not being able to regulate output current in 10V mode so if I try to run the current limit adjustments (put an ammeter directly across the output) the dial on the inbuilt gauge immediately shoots right up and then the unit trips out. The 100V and 1000V ranges operate correctly and flawlessly.

For a laugh, I ran through the adjustment procedure, just to make sure all the dials and levers were in the right place, to make sure I hadn't screwed up an adjustment somewhere so bad that it was the problem, but nope.
I was able to fully adjust the unit and it's stable as bedrock, besides the two issues on the 10V setting.

So, next problem to look at is the 10V setting. I'll swap a couple more boards back and forth to see if I can narrow it down a bit further.

[TERRA Operative]

Ok, looks like the Series Pass Element Board is the one playing games now.

I swapped in a spare board and it operates fine when switching from standby directly to the 10V range.
Current limiting is still a bit blegh, but that may be just down to adjustment to take into account of the old Series Pass Element board I'm testing with, I'll deal with that if needed once the original board starts on the 10V range setting..
I also swapped across all the socketed transistors with my fingers crossed it would be that easy, but the fault remains on the original board, so I gotta look deeper at the eight 2N3739 transistors on that board...


I gotta get my Tek 577 curve tracer working (damn tantalums!). Too much stuff to fix!

[TheDefpom]

Excellent, I hope you are recording a video for the repair.

Don’t forget to re-cap, and check all the carbon composite resistor values have not drifted too far from their original values.

[Tj138waterboy]

I second that statement for a video.

[TERRA Operative]

Yep, I think I have some video from years back, will finish it off.

All electrolytic capacitors have been replaced, and all carbon composite resistors have also been replaced with 1%, 100ppm-ish metal film resistors back when I started the restoration, so they are a known good element to my troubleshooting.

[TERRA Operative]

Ok, so after poking away and testing this and that, I'm down to testing the pass element transistors (seven of 2N3739) and all eight 200V zener diodes.

The Peak Zener tester only goes to 50V, so I'll have to dig out a power supply capable of 200V and some resistors for current limiting to test the zeners.
My only power supply is a huge heavy rack-mount HP thing... yaaaay...

As for the transistors, not sure. I'll have to figure out an easy way to test them. I have no space to pull out the Tek 577 curve tracer to fix it.

[DavidKo]

For zener testing you can use cascade voltage multiplier (behind some transformer) + resistor to limit the current if needed. You can also use 2 multi-voltage transformers with same output voltage and connect secondary windings and with Japanese transformers you will have plenty options how to connect the outputs :-). What I have seen they are 2x isolated 120V with 100 V and 110 V taps.

[TERRA Operative]

I ended up using the 332D itself.

I stuck the working series pass board in and set the output to 210V. The zener diodes are 1N3051A rated to 200V, 1W, so I strung two 22K resistors in series to limit the power to about a watt and gave them a test.
They were all within their 10% tolerance, the worst being -7.5-ish%, most hovering around -5%, so they are all good.

Onto testing the transistors under close to full working voltage and load, gotta figure out a rig for that now.



[EDIT]
So, I an the transistors on the Peak DCA75 Pro again just for fun, and took more notice of the numbers this time.
Turns out there's one that has a HFE of half of the others, and half of the rated specs... I think that is most likely to be the bad one.

Some new transistors have been ordered from Europe via ebay... Now to play the waiting game...

[factory]

For higher voltage Zeners & some diodes, I've used an insulation tester (aka Megger) with a DVM connected in parallel to the output, this is how I found some incorrect diodes fitted in place of 100V Zeners at work, breakdown voltage of the diodes was at least double. 
As well as being useful for checking capacitors for leakage, the Megger has also been used to find failing neons & number tubes, the failed ones require higher voltage to make them light.



Since upgraded to this Sullivan Megohmmeter, not sure what the output current is, but it's more than the wee Megger can manage, shown here testing a suspect batch of tubes (most were cracked), I added a decade resistor and the DMM is measuring the volt drop across the tube.



David

[TERRA Operative]

Oh! I didn't even think of that!
I have a Fluke megger here even.

It outputs from 50V to 1000V, not sure on the current, but it would probably work for a zener.

[Roman oh]

Hi Terra Operative.
Interested in your repair. I had a catastrophic failure with me series pass element - don't know what happened, but came in in the morning after the unit had been on overnight, and it was stuck with no output. Turns out two of the 200V zeners ( the first two) and five of the 2N2739s (all except the first two!) had gone to meet their maker. What excited me in your post was "I have ordered some from Europe via ebay". Can you tell me where? I have spent hours searching for suitable replacements. Thought it would be simple, as there are zillions of HV transistors around in power supplies, CFLs, etc. But these all (leaving aside the package differences, which I was willing to try to work with, if needs must) have very low hFE in the 10-30 range or lower, so are probably unsuitable.
I shall watch your repair story with interest, and would really appreciate a pointer to a 2N3739 replacement source...
Roman

[factory]

Probably this listing in Europe; https://www.ebay.co.uk/itm/221751361830

David

[TERRA Operative]

I'm going through the return process with them at the moment, bought 7 of them and half had hfe lower than the lowest spec in the datasheet, they were a mixed bag of differently stamped parts (some didn't match the others). They were also shipped stuck straight into plain foam in a plastic bag, no ESD protection at all....

[factory]

I bet they would be 10x the price if they had tested them & chucked the duds, that's the trouble with third party suppliers, never know what you are going to get.

Is there a modern part that could be used instead?

David

[TERRA Operative]

That's my next step if I don't decide to cannibalise my spare Series Pass PCB...

[Roman oh]

David,
Thanks for the link. First thing I notice is "Doesn't post to Australia". Then TerraOperative's experience...Hmmm...might just keep looking.
Thanks again
Roman

[Roman oh]

Hi  again Terra Operative,
Fyi, I have made it a project over several years to "improve" this beast mainly with a view to reducing its power consumption, so I wouldn't have to wait hours for it to stabilise. (No load power consumption after ref heater warmup is now around 11W, instead of the 30-40 that is was). As part of that, some years back, I "improved" the series pass by making the series string of 2N3739s into a darlington configuration using some 2N6517s in a TO92 package. This enabled me to increase those hot 22K resistors(R28-R34) feeding the bases to 220K and save a few of those watts. So the dependence on hFE of the 2N3739s was reduced. Worked reliably for a number of years, so probably not a factor in my recent mishap.

Out of curiosity (since you have yours opened up) I notice that my board has, for for R6,8,10...18 and R21-26, uniformly 150K 1/2W CC resistors fitted whereas the schematic shows a series of stepped values from 56K -100K for the former, and 100K for R21-26. The board does not appear to have been modded, but hard to tell. Spice model of my configuration (both with and without the darlington mod) suggests a pretty good distribution of voltage up to 1200V in across the string, nowhere approaching the 300Vceo limit of the transistors, so I'm not too concerned. But I would be interested to know whether your boards are similarly configured.
Roman

[TERRA Operative]

I am definitely interested in your mods!
I have some PCB's made way back when I started looking at this thing to replace the chopper amp with modern silicon which will be installed once I get it all working in it's original factory configuration, and I was thinking to replace the zener regulators on the right-most board (Current Limit board) with good quality linear regulators too for another easy upgrade.

I have a couple versions of the series pass board, so I'll check tonight and let you know. Will take a few photos roo.

[kawal]

Ok, so after poking away and testing this and that, I'm down to testing the pass element transistors (seven of 2N3739) and all eight 200V zener diodes.

The Peak Zener tester only goes to 50V, so I'll have to dig out a power supply capable of 200V and some resistors for current limiting to test the zeners.
My only power supply is a huge heavy rack-mount HP thing... yaaaay...

As for the transistors, not sure. I'll have to figure out an easy way to test them. I have no space to pull out the Tek 577 curve tracer to fix it.

I had a lot of issues with the same unit. The old transistors would test good in diod mode but had gain close to 1.  Check the gain of the transistors!

[TERRA Operative]

Ok, updates.

I mixed and matched the pass transistors with what I had on hand to select the best ones.

Now the fault is still there! It will start and work fine in the 100V and 1000V modes, but it doesn't work in the 10V mode.... 
When switching from standby to operate in the 10V mode, it trips out, and if starting in 100V or 1000V mode then switching to 10V mode, it doesn't trip but it gives no output.....

I really need to find a modern equivalent for these 2N3739 pass transistors so I can buy brand new parts that I know and can trust are 100% good...


Ugh, I'm getting tired of this thing. I'll poke it some more before I throw it on the 'for sale as-is' pile.....