##################################################################################
Date: Wed, 27 Apr 2005 00:43:30 +0100 (IST)
From: mikolaj cwiok
Subject: Re: L1 to L2
Hi everybody,
There was a partial success in getting data out of the TAB to L2beta test site in DAB2 this afternoon.
Mike and me tried to send the famous AAAA BBBB pattern to the ch#0 of the VTM+FIC cards, which are in
turn connected to ch#2 of the MBT card.
Upon manual L1ACCEPT issued from TAB/GAB GUI the MBT card reported each time that some data was arrived
to the buffer. Unfortunately, when we listed the contents of this buffer it was filled essentially with zeroes...
We've also checked that it wasn't simple issue of wrong numbering of 4 optical cables delivered to DAB2
(well, just in case :-). Only particular MBT channel received something upon manual L1ACCEPT. MBT card
is working for sure. It receives valid "fake event" data pattern on ch#1 as it should, because it's still
connected to the black box event generator triggered by a pulser.
Cheers,
Mikolaj.
##################################################################################
Date: Sat, 30 Apr 2005 08:05:51 +0100 (IST)
From: mikolaj cwiok
Subject: L1Cal to L2 SUCCESSFUL, too !!!
Hello,
James, Linda, Md and Mikolaj were trying to understand why Level-2 stuff (VTM+FIC+MB2+L2Beta) was receiving
only zeros from L1Cal TAB card.
It was not possible today to do the planned simultaneous test with L3 in the Sidewalk (DAB1). Instead, we
tried to read test events send by TAB (manual L1ACCEPT from VisTabsoft.tcl program) directly from L3 VRB
card memory, but without any luck. VRB responded correctly to the basic VME commands described in the manual
(e.g. it's firmware date-->4/25/2005). After using a hack to mask L1&L2 accept we tried to read VRB's fifos,
but it contained only 32-byte header.
However we were able to successfully read optical power from two VTM cards (one connected to FIC, one to VRB)
using some sets of VME commands send from BTQCHECK program to FIC and VRB, respectively. Tests were made with
and w/o optical splitter, as well.
Moreover, later today we succeeded to read AAAA BBBB test pattern sent from TAB to VTM+FIC installed in the Sidewalk
using BTQCHECK program (see attachment #1).
Then James and I tried to inspect what's going on in DAB2 test stand using another set of VTM+FIC connected to MBT
and d0l2beta9. Using 2 different debugging programs (for FIC and MBT cards, respectively) James and me verified
that both FIC and MBT receive AAAA BBBB pattern from TAB. However MBT needed some initialization work first (see
attachments #2 and #3 for FIC and MBT cards, respectively).
After modifying my test program to dump arbitrary number of words from MBT buffers I was able to see AAAA BBBB pattern
in L2Beta, as well at the end of the day (see attachment #4). Previous versions of L2Beta test program dumped only
first 84 words of MBT's memory which was exactly the size of the "event" generated by "fake event generator".
The problem I encountered was that when one connects FIC's output to MBT's ch2 (ch0 is restricted, ch1 was connected to
"fake event generator" at that time) the data was shifted by 84 words from the expected position after sending Event #1.
I was dumping 300 words and even 1000 words for each event, despite that I was expecting only 192 words from the TAB.
After event #2 position of the data was shifted to 188 and so on. When I disconnected "fake event generator" from MBT's ch1
and read MBT's ch2 again this memory shift persisted! Than I've checked that for MBT's channels no. 3..6 such shift was
even higher. The good news is that when one sends data to MBT's ch1 it always appears at the beginning of the dumped buffer.
What still have to be understand:
1. Why James and I couldn't read good events from FIC's fifos at all yesterday? I.e. before talking to MBT and L2Beta.
- Could it be a wrong pattern loaded to TAB?
- However we dumped TAB's memory and it did contain some AA BB words.
2. This evening there were no transmission errors reported by the FIC's front panel LEDs. Yesterday James and me noticed
that they appeared periodically. The FIC card was not moved from the crate so, certainly, it was not a mechanical issue.
3. Why the data appears at different MBT's memory address in each event for channels > 1 ?
- Need MBT card tech.specs, rather than treat is as a black box.
I will put this info in the e-logbook later.
Cheers,
Mikolaj.
Attachment #1: l1_to_l2_first_transmission_20050429.JPG
Attachment #2: l1_to_l2_seen_in_DAB2_seen_by_FIC_20050429.JPG
Attachment #3: l1_to_l2_seen_in_DAB2_seen_by_MBT_20050429.JPG
Attachment #4: l1_to_l2_transmission_seen_by_d0l2beta9_10events_FICch0_T0_MBTch2_fake_event_box_disconnected_300words_20050429.txt
##################################################################################