Dana 4200 DMM repair

[GregDunn]

I've had this unit longer than any other piece of gear; I must have picked it up in the late 80s at a hamfest.  It's pretty nice for a manual ranging DMM and has current measurement too.  Unfortunately, when compared to my voltage/resistance references and my other DMMs it's definitely out of cal now.

I started by looking at the PSU voltages, which were on the low side.  New filter caps didn't help, so I looked at the series pass transistors (MPS-U05/U55).  One of them didn't measure quite right so I took it out and in the process saw that the leads had separated from the package.     Replaced both of the transistors just to be sure (used TIP31/32, though the pinout is different) and voltages look fine now.  Unfortunately, I suspect that cycling the power a few times during the checkout has resulted in something else failing, because now the display shows overload on all modes/ranges (it blinks 0.0000).

The schematic fragment shows essentially the dual-slope ADC; reference current comes in from the top and unknown voltage from the left.

Looking at Q1 as the likely culprit because it triggers display blanking during the measurement cycle, I stuck in a replacement with no effect.  Pulling the null detector output high with a resistor to +5V does switch off the overload condition and displays the last measured input, so it's definitely something in the null detector signal chain but after the ADC integrator.  It appears that the signal coming through Q29-Q30-Q31 is either lost or seriously attenuated so that Q1 isn't being turned off.  The voltage at W2 is exact (same as the input), and the ±20 and +5 points look good.  But replacing Q29-Q31 has no effect on the issue.  The resistors all seem OK (most of the resistors can be measured when the transistors are pulled) and the diodes give reasonable voltages in-circuit.  So I'm left with Q7 which gates the unknown voltage, or maybe the small caps C29-30, C38 which are ceramics and shouldn't be causing this?

I don't have a FET in my parts box which matches Q7 (Idss is higher than anything I have with a suitable Vgs cutoff) so I couldn't test it tonight.

[GregDunn]

Found a FET which was a good match for Q7 and replaced it too.  No difference.     The original Q29 measured suspiciously when out of circuit (it looked much like a BJT) whereas the original Q7 behaved like a JFET (and that's what my cheapie transistor tester saw too).  The replacement for Q29 tests as a JFET, plus its specs seem compatible, so I'm leaving it in circuit for now.

So now I've replaced all the suspect active devices in the ADC / null detector.  I guess I'll just have to start by checking voltages/waveforms again and see what else jumps to the front.

[magic]

You would perhaps have more luck if you posted schematic or some pics of the circuit you think is broken, no one knows what Q7 is.

[GregDunn]

   This is twice I've added an attachment and it hasn't posted properly.  It showed originally when I added the original msg and then disappeared when I hit "post".  Hopefully it's now here to stay...

[coromonadalix]

hit shift+alt+s to submit/post   does this work ?  and the total  combined image(s) size is not over 2000k

[GregDunn]

I just use the "Attachments and other options" panel and then the "post" button; I've never tried the key shortcuts.  Not sure why the web page very occasionally dumps my attachment when I post.  It's only happened a couple of times in the year or so I've been on here...

[bob91343]

I have one of these meters and it's not working.  I'd like to fix it or, worst case, sell it as a parts unit.  I like two things about it: the ac current ranges, and the fact that the dc current ranges go to such low currents that it's easy to measure diode leakage.

I have some documentation but it's not clear how I should proceed.

Bob