9 May
2020
9 May
'20
5:32 p.m.
Hello,
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?
Thanks
Rob
9 May
9 May
5:44 p.m.
On 05/09/2020 11:32 AM, Rob Jarratt via cctalk wrote:
Hello,
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?
These spikes are very likely not real, but a result of probe
ground impedance. There likely ARE
some spikes and ripple, but not to the extent of the giant,
sharp vertical spikes. If you are using a X10
probe, does the 200 mV / div on the scope trace reflect that
attenuation?
In other words, are the spikes 400 mV or 4 V?
Jon
6:13 p.m.
The spikes are 400mV. The probes are set to x10 and the scope is set for
them to be x10.
Regards
Rob
-----Original Message-----
From: Jon Elson <elson at pico-systems.com>
Sent: 09 May 2020 17:45
To: rob at jarratt.me.uk; Rob Jarratt <robert.jarratt at ntlworld.com>; General
Discussion: On-Topic and Off-Topic Posts <cctalk at classiccmp.org>
Subject: Re: MicroVAX 3100 Model 95 Ripple
On 05/09/2020 11:32 AM, Rob Jarratt via cctalk wrote:
> Hello,
>
>
>
> 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-ps
> u-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?
These spikes are very likely not real, but a result of probe ground
impedance.
There likely ARE some spikes and ripple, but not to
the extent of the
giant,
sharp vertical spikes. If you are using a X10 probe,
does the 200 mV /
div on
the scope trace reflect that attenuation?
In other words, are the spikes 400 mV or 4 V?
Jon
10 May
10 May
12:14 a.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.
1472
days inactive
1472
days old
3 comments
3 participants
participants (3)
-
bhilpert＠shaw.ca -
elson＠pico-systems.com -
robert.jarratt＠ntlworld.com