9 May
2020
9 May
'20
11:14 p.m.
On 2020-May-09, at 9:32 AM, Rob Jarratt via cctalk wrote:
I have recently been trying to improve the ripple on
the output of my
MicroVAX 3100 Model 95 PSU because occasionally it would fry a memory
module. I replaced a bunch of capacitors, some of which had started to leak.
However, the ripple does not seem much better. There is a scope trace here:
https://rjarratt.files.wordpress.com/2020/05/microvax-3100-model-95-psu-ripp
le-after-re-capping.png
Ch1 is the 12V output and Ch2 is the 5V output. I had an old RD53 connected
as a dummy load. It is possible that the memory was breaking because of
occasional spikes that are worse, but I don't know. Does that seem OK?
Two guesses come to mind, but a lot depends on, or greater insight might be gained from,
knowing the design configuration of the power supply.
The spikes are suggestive of ringing from a fast and undamped switching operation. Might
be from switching of the primary-side driver, or from the switching of the secondary-side
rectifiers in the conduction/non-conduction cycle.
More scope observation and knowledge of the design might sort out whether the spikes are
synchronous to the primary driver vs. the sec rectifiers.
Guesses:
- The snubbing/damping C/L/diode networks around the primary winding of the transformer.
- The snubbing caps (usually) paralleling secondary-side rectifiers.
As the spikes look to be heavier on the 5V circuitry, it may be those on the 12V output
are 'just' induced reflections through the transformer from what's happening
on the 5V circuitry.