


September 16, 1986


TO: N. Kelly


FROM: R. Briody


SUBJECT: Status report


1) Several power supply vendors were given our power requirements for the 
diskless node. To date, none have given a positive response. Since there
doesn't appear to be an "off the shelf" solution to our power supply need,
I propose that we modify a Todd unit to suit our requirements. The following 
modifications would be required:

	A) According to an engineer at Todd, the +5 VDC supply output is 
	current limited by the schottky rectifier and not the transformer.
	The MBR3045PT schottky rectifier's maximum forward current of 30
	amps is a die limitation NOT a heat limitation. It will be necessary
	to parallel a second rectifier across the existing rectifier. The 
	application engineer at Motorola said the forward voltage drop of 
	the rectifier varies approximately 50 mV between devices. A Vf of 50 mV
	will mean a minor imbalance in forward current shared by the two
	rectifiers (current hogging).
	B) A +12 VDC to -5.2 VDC inverter would have to be added to the supply.
	This could be easily done on a separate PCB and mounted to the Todd 
	supply.

BOTH of the modifications outlined above can be done WITHOUT drilling holes
in the Todd power supply frame or disassembling the supply.

Modifying the Todd supply should be an acceptable solution to our needs. For
the short term, the supply is cost effective ($154 in 100 pcs.) and readily
available. For the long term, we can continue searching for "off the shelf"
supplies which will meet our requirements without modifications.

2) I'm currently looking at two possible vendors for the V24 system power
switch. Both have long lead times for volume orders. The problem is compounded
by the illuminating lamp. 120 VAC and 240 VAC switches are readily available
yet switches with 24 VAC lamps aren't. Lead times are being quoted at 8-12
weeks for illuminated rocker switches with 24 VAC lamps. I am continuing to
search for other vendors.

3) Testing of the 500 watt MPC supply has been temporarily suspended as 
Ken Brunskill's group is using the V24 chassis as a test bed for the memory
board parity error problem. When the parity problem is solved, we will resume
testing the supply.
