HP 34970A self test fail - ideas needed

[WangQi]

Recently got an "as it is" HP 34970A from ebay. The original lithium battery is dead and I replaced it with a 3V CR2032. It did power on but failed in the self test reporting 6 ERRORs 609,610,611,617,618,621 which indicates some problem on the DMM board.
The DMM board seems intact and no obvious sign of failed components. All the voltage rails are OK, +/-10V ref are accurate, A/D converter seems to have perfect waveform when input is zero(see AD_CTRL) .
I also captured the waveform of the input buffer's (input scaling circuit) output during the self test, the output levels seem correct to me, at least for test items 609,610,611,617,618(I marked the test pulses with test item# - apparently the numbers are also used as the error code, see self test).
I don't have a MUX module yet so haven't done any real measurement. I'm completely lost in terms of the next step of debugging.
Anyone has some ideas?

The full schematic and error code description are also attached.
https://doc.xdevs.com/doc/HP_Agilent_Keysight/service/34970A/34970-90012_svc_Part_2.pdf

[plesa]

Did you check functionality of U101, it is shared between gain setting and ohm current source. Without relay card it can be difficult to test properly.

[WangQi]

According to the waveform of test items 609,610,611 seems U101 works normally - able to control the gain.

[simmconn]

The DMM board can be tested by rigging something up on its input connector. A multiplexer card is not absolutely necessary in debugging it. My 34970A came with 617 error (plus a couple of others that I can't remember). It turns out the AD706 was bad. The previous owner must have injected high voltage into the Ohm range. Despite the protection circuit sometimes the opamp does go bad. After replacing the opamp the unit passed self-test. That was long before I came across my first multiplexer card cheaper than what I paid for the switch unit.
Just follow the service guide and schematics to debug. There is enough information in there. And pray that the big analog ASIC didn't go bad.

[WangQi]

Thank you for your suggestions. But without the relay card I'm not able to get any readings from the device.
What's confusing is based on the self test waveform, the signal feeds to the multi-slope A/D is correct (e.g 609=0.7V,610=7V,611=18V) but it still reports error 609,610,611. The only reason seems to be the A/D is not working properly. While test item 612,613,615 pass which indicates the A/D is working properly...