Notes about problems with MCM 109

MCM-109 was originally tested 12-Jul-99. It failed because "2nd block of 32 channels does not seem to work -- they do not respond to Vcal input and show no `step` on the spy channels. The 4th group of channels has low gain, although they do respond to Vcal input. On the preamp spy channel, they (i.e. 4th group) seem to have normal gain. Serial readback usually did not work but other tests suggest the problem is in the readback ... MUX2 shows no response to Vcal in any chip and does not seem to work at all in chips 2,4,8." The block of dead channels is considered a fatal flaw. The weak response is not a serious problem.

Retest August 20, 1999
Sometime in the first half of August 1999, Jehanne discovered there was a short at one of the surface-mount capacitors for TGV chip #2 on this MCM -- which is the block of channels which did not work. This short was fixed and on August 20, 1999 the MCM was retested. The 2nd TGV chip worked, but there are numerous failures in the AMU/ADC section of the chip. Based on the spy channels, all 8 TGV's worked.
  • AMU/ADC chip # 2 -- AMU spy channel shows no response to Vcal. The ADC value is always about 128 (hex) for any Vcal setting. The same ADC value is seem with level 1 triggers.
  • AMU/ADC chip #4 -- AMU spy channel shows no response to Vcal. The ADC value is always about 385 (hex).
  • AMU/ADC chip #5 -- AMU spy channel shows no response to Vcal. The ADC value is always 3FF (hex).
  • AMU/ADC chip #6 -- AMU spy channel signal shape looks OK, but it does not respond to changes in Vcal and seems to be very noisy. ADC values are about 1F7 +- (about) 50 (hex).
  • ADC/ADC chip #7 -- AMU spy channel seems to work. ADC is always about 3E5 and is the same for all 32 channels in each event. Sometimes the ADC value changes a little from event-to-event, but it is always the same for each channel within an event.
  • Chips 1,3,8 seem OK, except the AMU spy line does not work in chip 8.

    New conclusion: multiple AMU/ADC problems. Only 3/8 of this MCM works.

    Details on earlier tests
    On August 9, 1999, this MCM was re-tested. I did not have a problem with the serial string download or readback. I will assume this was a problem with the (old) setup. I did not check all of the spy channels, but MUX1 and MUX2 both worked on the first chip. However, the problem with the 2nd block of 32 channels remained. It did not respond to bench-cal and the trace stayed constant at about 1.5V independent of changes in Vcal, Vfb, Vgate, and Vmid. I changed the ADC ramp (with Vref and Iref) and the ADC values changed as expected -- that is the ADC seems to see a constant input voltage and that is what it digitizes. I wrote some notes about this MCM in the logbook pp 56-57.

    Conclusion: The 2nd block of 32 channels still does not work. The problem looks the same as the problem with MCM serial #101. It looks like a TGV problem.

    updated 20-Aug-1999
    John Sullivan (sullivan@lanl.gov)