[SOLVED] Datron 4708 cant be calibrated error 4

[Samogon]

Hello volt gurus
I have datron 4708 calibrator and it is so off from displayed values that calibration of gain on DC volts cannot  be stored in memory.
For i stance on 10 volts it meaused 10.32 i can remove it with error -3.21%
And it shows pretty stable value with some fluctuations on 7th digid
What could  be wrong: precise reference or devider or even DC card?

Thank you

[pelule]

Sounds like there is an issue inside.
You may 1st check, if this issue is HW related.
I would start by checking some vaules in the different range and values for value and stability.
You mentioned, your measured value is stable with some fluctuations on 7th digid.
Did you checked it with an oscillioscop, if there is some high frequency.
If it's ok, it may a lost of calibration data.
If it's not ok, next step could check the internal supply volatges (value, noise)

[PA0PBZ]

I have a 1081 that was so f*cked up that it didn't even autorange anymore. Resetting the cal to default helped to get everything working again.
That's not a calibrator of course but just saying - if you can't get it to work and the hardware seems ok it could help to do a fresh start.

[lowimpedance]

 How clean is the air path from the air intake/filter to the exhaust ?. I have seen two 4800 series with dead electrolytics in the power supply assemblies and one of these resulted in the picture attached. Best assumption is overheating (perhaps long term!) and thus dead filter caps and poor power supplies leading to boom in this case.
 Just my experience only, not saying that's where yours is heading but its worth checking the quality of the power supplies first.

[Samogon]

This is Big badaboom! As I understand it is blown one of heatsinks on the picture?
No, unit is clean as it can be with air intake system. I definitely will check voltages and riples. Hope will have few hours on Monday.
Thank you for your support, will post results of voltages values and riple amplitudes asap.

[ManateeMafia]

Definitely check the supplies on the inguard supply. My 4700 had about 20 bad tantalums that all died one after the other. It took a while by measuring each rail on each board to get them all replaced.

My DC was off quite a bit before adjustment, but not as much as yours.

[Samogon]

So i have measured on output terminals AC component on different DC ranges it is the same
PtP average is around 17mV and AC RMS is around 4mV. I had to make low pass filter to 1Mhz. Is it big or not?
It appears i have very noisy environment in my lab.
Here is what i have got.
Next step would be is to check all power rails. +/- 15 +36 

[Samogon]

Lineriality if funny, on the 10 volts range it adds 33mV every 1 volt
For instance:
Set    Measured
1        1.033
2         2.066
3        3.099
So output function looks like f(x)=1.033x

[pelule]

You may first just collect/note you measures. A got ways doing this, is to do the "performance check".
This is a good ways to get an indication where the problem (if any exist) is caused.
Next step than you may check the power rails, if inside specs. If, you may check the reference voltage.
That both are the base of any follower circuits.

[Samogon]

All voltages except 36V measured 34.5V (which is in 5% to normal) are very close to normal with deviations not bigger than .1V
36 volts are supplied to precision reference. LTZ1000 output 7.2V

Also AC volts can be calibrated and offsets are stored in ram no problem with any offset values.

I am suspecting DC card now. Did not test DC current though.

[pelule]

Yes, I agree to your so far conclusion and targetting the DC card as potential root cause.
Never the less I would test at least one DC current vaule just to get an impression, if there may a problem.
It would be a real pitty, once you have finished your DC repair and than detect a problem there.

On DC card I would start a a mdidle value in the 10V range, say 6V. (measure volatge level, the noise, the load behavour and the current limit).

[Samogon]

Returning to it.
I did chech DC current and it exhibits exactly same deviation as DC voltage. And for me it is expected, because it id derived from voltage card.
So DC voltage card my poin of failure.
Will do some measurments. Looks like it is some kind of devider/multiplier issue.

[Samogon]

Reading my old thread.
Want to report that it was user error, i have skipped coarse gain and linearity pre-calibration procedure. Once i done it, i was able to fully calibrate it.
Special thanks to membres ManateeMafia and Pipelie for pointing it out.

[maginnovision]

Having just read it that's what I suspected. Coarse gain is very important. If you don't need anything but standard values linearity doesn't matter but it's simple enough no reason not to.