Power supplies for rod and petal production

From Sandor Czellar, November 2002

The scheme proposed for the power supply unit to be used in rod and petal production is given in pdf format, gif format and jpeg format.

Basic features 

1) The components are housed in a 3U VME crate.

2) The unit is based on a switching power supply recuperated from the DELPHI HPC, modified for the Tracker needs.

3) Two linear regulators are used to issue the voltages needed. In this version, the power supply yields a fixed voltage, and the sensing is used to obtain fast feedback at the level of the regulators. 

4) Current limitation is implemented on each of the two lines.

5) Up to four humidity measurements and four temperature measurements can be used to interlock the power supply. The thresholds can be adjusted from the front panel of the PS unit. An additional temperature probe is mounted on the heat sink carrying the linear regulators, with hardwired threshold. 

6) An additional input is available for an external logic signal (e.g. ON/OFF signal from the cooling unit).

7) Measurements of humidity and temperatures, as well as currents and voltages, are made available for monitoring. 

8) The device can be controlled from the PC. 

The communication between the PS unit and the PC could be made with a commercial card (e.g. http://www.labjack.com/labjack_u12.html). For use in large systems, several such cards should be multiplexed into one PC.

Items 6-8 need optical decoupling to avoid ground loops between the PC (or the cooling plant) and the PS. The optical isolation is under development.

Current status

A pre-prototype has been realized and tested. It is now in use in the TOB system test.

It implements the basic features 2-4. The performance is summarized in this document (doc file, pdf file). 

The most important aspects already implemented and verified are listed below.
No overshoot at startup: slow startup.
Current limitation: implemented and working correctly on both 1.25V and 2.5V lines.
Very low noise: much less than 1 mV at full load.
Fast reaction: overvoltage smaller than 200 mV for a variation of the load on the 2.5V line from 10A to 0A. The test was made with a long cable and also using the IC BUS. For the test with the IC BUS, the load was placed at the end, which is much worse than the real case, where the load is distributed and larger in the first section.

Next step

The device as described above is rather complicated. It seems unlikely that all the developments needed can be completed in time for the start of the rod/petal production. For this reason it has been envisaged to produce first a simplified version with the features existing in the present pre-prototype, plus an automatic interlock system based on two temperature measurements and two humidity measurements. These features are considered sufficient to start the production with minimized risks, while the development of the complete device is carried on. (NB: in the simplified version there will no monitoring, and the device will not be controllable from a PC). The number of simplified units to be built will have to be decided taking into account the schedules of the different labs involved, and the time needed to complete the development.

Duccio