Fluke 8842A 5.5-digit meter repair and calibration.

[TiN]

Another flukey was saved recently from the dumpster fate.

Caveat was dodgy rework job here:



And broken voodoo resistor here:





Full article with process and troubleshooting steps.
Enjoy the read. 

Pity no GPIB option to capture some stability data...

[Kleinstein]

Having so much trouble with the ADC / amplifier, it might have been easier to swap the display to the parts donor unit. 

Having a non heated reference is a 2-sided thing when it comes to fast turn on.  Even with careful TC compensation it might take some time to stabilize, as it takes time to get the final temperature differences.  The heated references LM399 / LTZ1000 usually are rather fast in getting there temperature. It is more the rest of the circuit that takes it's time. Here the added power of the reference may make settling of the parts around the heated reference a little more important.

[TiN]

Even with careful TC compensation it might take some time to stabilize, as it takes time to get the final temperature differences.

I very much doubt it matters in 5.5 digits meter, those effects completely dwarfed by 10ppm resolution of the A/D in here.

Having so much trouble with the ADC / amplifier, it might have been easier to swap the display to the parts donor unit.

But then we will not have this article and thread, eh? Also my meter have shattered high-voltage divider network, and who knows what other problems.

The heated references LM399 / LTZ1000 usually are rather fast in getting there temperature.

Yep, getting heat to reach equilibrium with everything around them is still hours. Hence I never turn off my 8.5d meters, except for service/relocation. And LTZ can take days to reach <0.5ppm from previous value, if stayed cold for long.