Apparent ECONDEventPacket internal non-rogue decoding issues

[joshgreaves332]

This has only been internal non rogue decoding.
There seems to be inconsistency in what the decoder is decoding for multiSampleECONDEventPackets. Sometimes the decoding shows zeros in random channels, sometimes it show zeroes only in ch>36 (corresponding to 1 i_link), sometimes it works perfectly fine for a while.
Multi vs Single packets: the .adc() value returns 0 for the ch>36, in some instances.

For instance taking pedestals goes from:

• Pedestals showing apparent elink dependence (although BOTH elinks have zero ADC values at different points) and also exhibit some repitition: there are repeated values for the same orbit/bx/event
• Then a few minutes later with no changes made, PEDESTALS produced output which exhibits randomly appearing zeros, but with seemingly no repetition.
• Then a few minutes later we get complete output with no zeros

Steps to reproduce the behavior:
Hard to say... haven't been able to consistently affect which state the outputs are coming in.

Expected behaviour:
Consistent ADC values for both ELinks, and all channels, for multisampleECONDpackets.

Terminal Print Out
Decoding comes with these warnings"

(ldmx-env) jgreav@rdsrv436:/u1/greaves/pflib/build$ ./econd-decoder 2 pedestal_20251213_171017_mini.raw 
(2025-12-13 17:10:38) [econd-decoder]  warn: The ECOND decoding is not well tested. Inspect the results carefully and please contribute!
(2025-12-13 17:10:38) [econd-decoder]  info: popping 0 event from stream
(2025-12-13 17:10:38) [decoding]  warn: Bad header marker from ECOND 0x0a + 0b1
(2025-12-13 17:10:38) [decoding]  warn: Incomplete event packet, stored payload length 458 is larger than the packet length 160
(2025-12-13 17:10:38) [decoding]  warn: Event header 8b CRC does not match trasmitted value: 01111000 != 00000000
(2025-12-13 17:10:38) [decoding]  warn: bad subpacket checks 000
(2025-12-13 17:10:38) [decoding]  warn: bad subpacket checks 000
(2025-12-13 17:10:38) [decoding]  warn: CRC over all link sub-packets does not match transmitted value 0x218db6e1 != 0x08721400

Clearing the buffer in EXPERT->FC->BUFFER_CLEAR seemed to effect change temporarily, but that was not consistent either.

[tomeichlersmith]

This warning

(2025-12-13 17:10:38) [decoding]  warn: Incomplete event packet, stored payload length 458 is larger than the packet length 160

is telling us that we did not read out enough words and there is nothing the decoding can do.

Since the rest of the printouts that the decoding would see are all zero's, I'm guessing this is more of an issue with how we are capturing the data than the decoding.

#349 has one patch to the data capturing that could resolve this issue, but that could only be confirmed if you are able to reproduce this error.