MDC-Upgrade: Cooperation with Yanyu Wang



Introduction

In the HADES-DAQ-Upgrade project the MDC-Upgrade is an essential part. The expected data rates for heavy ion systems can not be handled any more with the original MDC-DAQ.

A description of the concept of the DAQ-upgrade and the new DAQ-system can be found at:

http://hades-wiki.gsi.de/cgi-bin/view/DaqSlowControl/DaqUpgradeOverview

The idea behind the MDC-upgrade is to simplify everything (get rid of VME,PPC,SAM,ROC,DTU) and put everything on a common platform for the whole HADES-DAQ system.

A short overview of the old and new MDC-DAQ can be found here: http://hades-wiki.gsi.de/pub/DaqSlowControl/DaqUpgradeMDCOverview/MDC-setup.pdf

The TRBv2 (optionally without TDCs when only digital communication is needed, like in the case of the MDC-DAQ) will serve as this platform. To test the new readout-scheme one needs to connect the TRBv2 to the MDC-FEE (driver cards on the motherboard). For this purpose a MDC-ADDONv1 card was developped and is currently (2007-02-02) being programmed by Attilio Tarantola.

As this board is just using RS484 drivers for the communication the known problems will stay:
  • crosstalk and "ringing" of the MDC-FEE due to copper cables squeezed between motherboards
  • heat disspation
  • missing possibility to reprogramm the motherboard-CPLD
  • hassle with the chains of motherboards

Project definition

Therefore, the project was started to evaluate and build a prototype board for the TRBv2 -> MDC-Driver-Card -> Motherboard communication. The idea is to use only optical transmission for the data and the clock (only point-to-point connections). The data should be tranmitted via the optical transceivers and SERDES bidirectionally. Also the common or-signal can be transported via the data-stream (with high priority which is forseen in the TRB-Net protocol). Only the "Common-Stop" signal needs a good time resolution, so a dedicated optical transmitter has to be put on the driver-card and a receiver on the TRBv2-MDC-Addon_v2. In total we need around 400 new driver cards (with FPGA). The following picture illustrates the project. Each TRBv2-MDC-AddonV2 will have 32 optical transceivers as this is the maximum number of motherbaords for one MDC sector.

2008-05-19 It turned out in October 2007, that it is not possible to transmit a good timing signal via the cheap optical trasceivers.

Optical Transmission

More details about the project progess can be found here.

-- AttilioTarantola - 16 May 2008

MB and DC Current measurement:

-- AttilioTarantola - 22 Apr 2008

Today Yanyu and me measured the MB+DC currents. We measured the following currents for long and short MB with all threshold fixed to 10. The calibration trigger was enabled such that we had one normal trigger and one calibration trigger. The Driver card were plugged onto the MB(as in normal operation): this means the TDCs, the CPLD and the 2 transceivers have the same common voltage +5VD.

Motherboard Voltage Current(readout stopped) Current(readout running)
short +1V(green) 150 mA 160 mA
short -3.3V(yellow) -360 mA -360 mA
short +3.3V(orange) 360 mA 360 mA
short +5V(red) 440 mA 850 mA
long +1V(green) 220 mA 230 mA
long -3.3V(yellow) -520 mA -520 mA
long +3.3V(orange) 510 mA 510 mA
long +5V(red) 540 mA 700 mA

Here the same measurement changing thresholds:

Motherboard Threshold Dataword number(calibration) Dataword number(normal trigger-noise) Current relative +5V
short 60 (all DBs) 67 1 500mA
short 10 (all DBs) 67 52 850mA
long 10 (all DBs) 577 50 970mA
long 60 (all DBs) not enabled 2 970mA
long 10 (all DBs) not enabled 60 1,03A
long 60 (all DBs) not enabled 2 650mA
long 10 (all DBs) 530 60 1,03A

We powered the MB(+5V,+3V,-3V,+1V) and the driver card (+5V) individually. Here the current relative +5V for the DC:

Motherboard Threshold Current for DC(+5V)(readout stopped) Current for DC(+5V)(readout running) Dataword number(calibration) Dataword number(normal trigger-noise)
short 10 (all DBs) 250mA 370mA 67 40
short 60 (all DBs)   260mA not enabled 2

Here the current relative +5V for needed by the TDCs+CPLD:

Motherboard Threshold Current for TDCs(+5V)(readout stopped) Current for TDCs(+5V)(readout running) Dataword number(calibration) Dataword number(normal trigger-noise)
short 10 (all DBs) 250mA 370mA 67 45
short 60 (all DBs) 250mA 250mA not enabled 1

+5Volts study:

-- AttilioTarantola - 16 May 2008

14 May 2008 Yanyu and me tried to see the working mode of one short MB changing the +5V. The +5V(normal operating voltage for one short MB) is needed by the 8 TDCs, the CPLD and 2 transceivers. The results are:

* Operating voltage (5V-4.7V): the readout is not affected by the voltage drop. This means we were able to get normal data (60 datawords per event) and calibration data (all channels calibrated). The MB worked as expected:no changes in DST (data strobe), AOD(address/data) and RDY(token back to Addon) signals.

* Operating voltage (4.4V): the event size for normal data does not change (60 datawords per event) but the data is always 0 and the calibration still is working fine (all channels calibrated). (Speaking with Joern, he suggested this behavior is due to the fact that the CMS does not reach the TDCs. At this operating voltage the TDCs should work without any problem and this we see in the calibration events where the common stop is not needed by TDCs). No changes in DST(data strobe), AOD(address/data) and RDY(token back to Addon) signals.

Conclusion: the LDO, which Yanyu is looking for, might drop down to 4.85V. In this case the MB(TDC+CPLD) will not be affected.

-- AttilioTarantola - 04 Aug 2009 Jan and me: additional current measurement. Setup: 1 long MB, 2 DB stack, connected to 10 cm cable to a power distributor board (v1 bridged). We read the current on the voltages power suppllies:

The voltages measured on the OEPB(input connector): 5.59V, 3.66V, 1.93V, +3.130V, -3.190V

OEPB(oputput LDO): 1.198V,3.342V, 1,000V, 5.007V,

On MB(TDC side): -3.17V,+3.10V

Long MB:

Th *Nr.dataword I(+5V) I(3.5V) I(1.5V) I(3.0V) I(-3.0V) comment
before init - 144 176 468 560 582  
x"00" - 1440 176 472 568 580   LDO limit to 4.45V
x"05" - 1440 @ 4.45V 176 472 568 580   LDO limits reached, voltage dropping with time and temperature
x"10" 100-120 1440 176 472 568 580   LDO limit to 4.45V
x"20" - 668 176 472 568 558  
x"70" ~10 550 176 472 568 580  

Short MB:

Th *Nr.dataword I(+5V) I(3.5V) I(1.5V) I(3.0V) I(-3.0V) comment
x"00" - 1255 175 384 392 391  
x"10" - 980 175 394 392 391  
x"20" - 460 175 394 392 391  
x"30" - 385 175 394 392 391  
x"40" - 385 176 396 378 389  
I Attachment Action Size Date Who Comment
GBit_POF_Transceiver_tcm278-76277.pdfpdf GBit_POF_Transceiver_tcm278-76277.pdf manage 134 K 2007-03-16 - 10:31 YanyuWang GBit_POF_Transceiver_tcm278-76277.pdf
Optical_Transmission.jpgjpg Optical_Transmission.jpg manage 29 K 2007-02-05 - 22:30 MichaelTraxler Optical Transmission
driver-card-real1.jpgjpg driver-card-real1.jpg manage 65 K 2007-03-16 - 09:34 YanyuWang Driver-Card-real
driver-card1.jpgjpg driver-card1.jpg manage 80 K 2007-03-16 - 04:20 YanyuWang Driver-card
Topic revision: r17 - 2009-10-27, JanMichel
Copyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki Send feedback | Imprint | Privacy Policy (in German)