[hpsdr] Metis & dual Mercury boards

[K1rqg at aol.com]

Folks
 
For some time now I have been experimenting with K5SO diversity software  
with a great deal
of success. I have been following the progress of others on this reflector, 
 and would like to
mention something to be aware of. 
 
When Metis arrived, I immediately changed over from Magister (OZY as well)  
and was thrilled
with the results. Over a period of about two weeks, and during many  
software changes, I
began to see degradation in performance of diversity operation. The  noise 
floor would rise,
and on occasion it appeared that one of the Mercury cards would "take off". 
 This got progressively
worse, so I switched back to OZY and all seemed to play well again. 
 
Also with Metis, all was well with just one Mercury on the back plane. I  
saw Georg, DL2KP, 
comments and it sounded a bit like the things I had been seeing. 
 
What I found in my case was the LPU was failing after being on for a few  
minutes with
Peny, Metis, Excalibur, and two Mercury boards. I monitored the  regulated 
12v and that
stayed stable, however, the regulated 5 v started at 4.98 v and continued  
to drop down
below 4.2 v and this is where the noise floor started rising. I measured  
excessive drop 
across R6 (2.2 ohm 10 W) resistor and also notice a physical split  
longitudinally down
the center. 
 
I switched back to my old ATX power supply and all problems vanished  
(except for the
occasional noise hump associated with switching P/S). I fashioned 4 x 2 ohm 
 5 W 
resistors (series/parallel now 20 W dissipation) and the LPU seems to be  
holding up for 
over an hour. I will test it longer tomorrow.
 
I thought there must have been some software issues, but so far it has  
turned out to be
LPU. Check the +12 v and +5 v on the Atlas board (the other voltages are  
not needed)
for stability and ripple.
 
73 de Joe
K1RQG
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openhpsdr.org/pipermail/hpsdr-openhpsdr.org/attachments/20110314/48c4b392/attachment-0003.htm>