Datron 4200 AC Calibrator -- FAIL 2 during selftest

[essele]

After the great success I had with my Datron 4000A DC Calibrator, I decided to grab a 4200 so I have an AC source as well.

There appears to be a couple of issues with this unit...

- Firstly, I get a FAIL 2 during self-test.
- During actual use everything is spot on except the 100V range, which is at about 85% of what you would expect.
- After 5 to 10 minutes it trips my house RCD (which the kids are starting to get annoyed about ;-))

I can't find a 4200 manual (of any kind) anywhere, so I'm looking at the 4708 manual which certainly seems to be pretty close, the schematics and board layouts look identical for the key boards, so I suspect they are the same modules.

I'm hoping that the FAIL 2 during self-test is because of the incorrect output on the 100V range, it's not entirely clear in the description of the self-test, it does say FAIL 2 is an over-voltage, and I'm seeing under-voltage.

I have also seen a couple of FAIL 5's when I've been adjusting the output, this is "Safety circuits watchdog" (if the 4708 is the same), but they don't seem to be frequent or easily reproducable.

It looks like the under-voltage is related to the 100V amplifier since this is only in the path in the 100V range, and everything else seems to be ok. I've had a quick look at the supply voltages for it (-400v) and they seem to be ok.

Does anyone have any experience with the 4200? Also, any clues as to why I would be tripping my RCD but only after a delay? I'm assuming it's heat related, but I'm struggling to understand why I'd get heat related earth leakage?

Thanks,

Lee.

[coromonadalix]

heve you done a visual inspection of the pcb,  you had a broken pcb trace on your 4000

Check all usual suspects,  caps  voltages test point, pcb contacts   etc ...  you know  lolll

[essele]

Hi coromonadalix ... yes, I've done a first pass quick check, and reseated all the boards, nothing obvious there.

However, in reading the schematics a bit more I realised that there is a positive and negative part to the 100V amp, so just had a quick check on a scope and that gives me rather a big clue!

(Apologies for the crudity of the photo, I didn't have time to make it to scale or to paint it!)

[essele]

Have determined that it's a problem in the gain or driver stage of the amp ... the drive to the external boards is also clipped, yet the input to the gain stage appears to be fine.

The +/- 400v and +15v all look ok.

There was a burnt resistor R53, which should be 61R9 1%, but was measuring 50R ... I've replaced it with a 68R for now. (I couldn't determine the colour code as it was burnt too badly, so I'm assuming the values are the same between the 4708 and the 4200. I have already found a few differences.)

I've run out of time for now ... sadly this will have to wait a week as I'm away!

[essele]

This is frustrating ... each time I leave this for a while something else goes wrong!

So I had pretty much narrowed down the problem to something with the gain/driver stage for the 100V range. Everything else was working pretty well (I still haven't look at current though.)

I was struggling to look at the OpAmp inputs/outputs with my Siglent ISFE as it's not the worlds most accurate (x200) and it's really noisy, so I ordered myself a Micsig to1104 while I was away so I could really get to the bottom of the issue.

Anyway, got back home and the 1000V range wasn't working .... oh well, not a major issue, I'll worry about that later.

I got everything setup on the main gain stage OpAmp and whilst I was watching the signals disappeared ... I was very careful in attaching test clips, so I'm pretty sure it wasn't me!

So at this point I have no output on any range!  After a bit more digging I can see that there is no input to the AC Amplifier, and some of the signals are not present on the Sine board. (I do have the Synth O/P and VCO O/P, but no OSC O/P.) All supplies still test ok.

Anyway ... another day passes, and now I have ErrorOL when I switch it on. Which (for the default power-on 10V range) seems to be either an issue with the 10V overload detection or an actual over-current issue.

(Helpfully the 4708 documentation and schematics are really helpful ... some of the schematics even mention the 4200)

I did have some spurious behaviour right at the beginning, so I hope all of this is related, but it looks like I am now three issues deep into it. I just hope it's not me breaking stuff as I go ... I do feel that I'm being pretty careful as I'm healthily terrified of the voltages in this thing.

... on to ErrorOL troubleshooting tonight!

[jfphp]

Hello,
I have sold a couple of weeks ago the complete paper-original manuals (with key, fuses...) of the 4200 to BNElecEng in UK through the EEVBLOG  forum. You can perhaps get in touch with him.
Regards
JFP

[BNElecEng]

That's correct, I now have a full set of manuals. I haven't gotten round to scanning them yet. It's only a few hundred pages...

[essele]

Fantastic ... @BNElecEng happy to donate if I can grab a copy!  Some of the external scanning services don't seem to be that pricey if you'd like save personal effort!

[essele]

Right, I think I've finally found the cause of the "Error OL" ... it looks like I have a faulty DAC on the Sine board, it seems to be drawing enough current through the FB pin that it trips the over-current detection. This is M43. M42 seems ok, and if I swap them then the "Error OL" goes away.

Obviously swapping them is only a test since I suspect the device is faulty. So two new ones on order (just to be safe), then hopefully I'll be back onto the "missing signal" problem ... and fingers crossed it's also related to the DAC.

I just hope I haven't buggered anything else in the pursuit of this one ... I think I've pretty much desoldered half the bloody device by now ;-)

[essele]

So the “Error OL” is now fixed, however this hasn’t fixed the missing sine wave problem. Lots more digging and countless test wires later it seems a safe bet that the problem is with the sine oscillator.

The AC board seems ok, the quasi sine is being generated and I can probe various things that match the frequency setting on the front panel.

The main oscillator on the sine board seems to start at power on, at about 500kHz, it builds in amplitude and frequency (to around 650kHz) and then damps very quickly over about 6 cycles to zero, then swings to the positive rail slowly.

I’m not sure if it’s one of the control circuits (amplitude or frequency) that’s not behaving or whether the power on sequence switches a different feedback loop into play and that’s the problem?

I am so out of my depth here it’s mad, but at least the Micsig is getting a good workout!

[essele]

This really is the gift that keeps giving! Just had a “Fail 8” error, which is the 38v supply (however can often be the 15v supply) ... further testing showed both 15v supplies not working.

Aaaarrrgggghhhh!

Luckily during further analysis I had my first ever live tantalum failure ... exciting stuff ... and it seemed to be the cause of the problem, replaced it and it’s back to the previous situation.

It really is one thing after another!

[maginnovision]

At least you won't get bored with it. 

[essele]

OK ... second 4200 arrived today, primarily for spares, however it really does seem salvageable ... it's been through a FAIL 9 and then a FAIL 8, but after replacing a few tants (power supply and AC board) it appears to be all ok, however the front panel doesn't respond to any keypress (and the B symbol isn't lit), so it may be a keyboard controller type issue .. will get to that one shortly.

Someone has also broken the power input bit (Potter 7100-0002) and cut the fan wires ... I've soldered the fan wires back and it seems fine ... I may have to get creative with Epoxy unless someone has a spare power input unit??

However, if I put the Sine board from that one into my existing unit I get reasonable output ... so the issue is definitely with that board.

So far I've replaced every IC apart from the LH0032 and the uA714 ... I don't really want to mess with the working board unsoldering these bits, so I think I'll wait for the spares to arrive.

[essele]

More progress...

On the second unit it seemed to be a stuck key causing the problems, once the front panel was off everything was fine.

The 1kV range didn't work, but that turned out to be the "disable 1kV range" switch (sw1) on the Power Amp board ... so apart from the broken power input, this unit now seems to be totally fine (it doesn't have the current module unfortunately.)

I also took the opportunity to pinch the LH0032 from the Power Amp board and try it in my Sine board on the broken unit ... still no joy. Really not sure what to do here now other than start changing transistors! Will rethink overnight!

[essele]

It turned out to be a broken resistor network (AN14), one of the resistors was measuring over 100k (possibly open, I haven't removed it yet). I soldered a 10k temporarily across it and all is working!

(Not sure how I missed it before, I'm pretty sure I checked all of these!)

Anyway, I will post separately on this, but it does seem that you can't get DIP8 resistor networks any more ... the best I think I can do is go for a surface mount version and solder it to an adapter board, otherwise I'm into separate 10K resistors which presumably wouldn't be so good as I suspect they wanted the TC matching.

The old part is a BECKMAN 694-3-R10kD ... it's a 4 x 10K at 0.5% resistor network in a DIP8 package ... Mouser and DigiKey show them as non-stocked, obsolete in 2013.

Any ideas on where I can get something similar?

Thanks.

[maginnovision]

Glad to hear you figured it out. Maybe you can search for some NOS outside of major suppliers.