FMC Projects » Simple PCIe FMC carrier (SPEC)

First field tests of two SPEC boards running a White Rabbit timing link between them have been made in Siberia. These tests are made by DESY to verify the usability of the board at the Gamma-Ray and Cosmic-Ray experiment HiSCORE-EA at the Tunka site. Full results of the tests will be made available soon. Two kilometres of fibre where outside, while the SPEC boards were nicely warm inside the counting room.

In the year 2011 the company Creotech has produced and sold fifty SPEC boards. With the twenty cards from the pre-production of the seventy that CERN has ordered from the company Seven Solutions, in total seventy SPEC boards have been produced in 2011.

The pre-series of the first commercial series production of SPEC cards has succeeded. After repairing some assembly problem with the FMC connector, all twenty cards passed the test on the test bench that was prepared by CERN. The cards will be sent to CERN for a final inspection before giving the OK to produce another fifty.

CERN has placed an order for EMC testing according to EN 55022(2006) /A1 + EN 55024(1998)/A1 (2001)/A2. Once these tests are done, it will allow the board to carry the CE approval.

The modification for a better mechanical fit has been made and halogen-PCBs will be ordered to start the first series production of SPEC boards.

A serious piece of automated test software has been made that will check all soldered connections on the board and that will measure the different supply voltages. This is all possible because of the groundwork done by the Production Test Suite project software logically called PTS and the FMC Carrier tester test card that measures all connections to the FMC connector. We have used these to create the "PTS for SPEC" production software. The user manual gives a good overview of the ease of use.

We are now ready to send a complete test bench (PC, SPEC reference card, FMC test card, cables etc.) to the company that will produce and test 70 boards. First 10 though, just to be sure...

The V3 of the design was ready for production on 22 July. We'd like to produce 70 of those cards of this version, apart from the fact that a user at GSI just found a mechanical problem that could appear with a certain combination of SFP plugin and PC type. Being just in time to correct this by simply moving the SFP 2.5 mm further out on the PCB, we decided to make a V4 of the PCB.
Why a V4 while the V3 never was produced? As this will ease the tracibility as possibly copies of this Open Design design may already be around!

Three V2 boards have been received. One has already passed the full production suite, two others have only received a very short test (and passed). This means that this version can be used to start a series production.
Actually there is still a not-understood issue that we don't manage to get communication between the two SATA connectors working, while from each one to the FMC test board works OK. We are currently looking into that and believe it is only a firmware problem.

For the production of the SPEC cards specific production software has been written and demonstrated. This software provides a simple go/no-go interface with logging facilities to easily find production errors. The software is written in Python and uses a CERN-written test environment called Bakara. To test the connectivity FMC connector on the SPEC card, a specific FMC carrier test card is developed that is also able to measure the different power supplies and test the different gigabit links.

The layout of the V2 of the SPEC board is ready. It is an improvement of version "1.1". We will build 3 of these pre-production boards before we will order 70 boards. Two production test benches are being set up.

Five engineers have reviewed version 2 of the SPEC card. Many small changes have been made to make the board better producable. E.g. instead of 23 resistor values only 10 will be needed. The production files will soon be generated by CERN's design office.