5 Sep
2025
5 Sep
'25
11:10 p.m.
When a power supply gives me trouble, I often pull *all* the electrolytic
capacitors, test them, and then install new ones if I have them (because if
you're going to solder a capacitor in, you might as well solder a new one
in). A couple times, it hasn't been the big filter capacitors that were the
problem. There's some risk in "shotgunning" component replacement like
I'm
suggesting in that you might make something worse by replacing a perfectly
good capacitor with a modern replacement that isn't quite the same. But it
has a good track record for me...
Also, old resistors can drift or open, so you might want to scrutinize
those too.
What value is your resistive load? Some switchers specify a minimum load.
Another option is just to see if you can find a meanwell open-frame power
supply or similar at mouser or digikey that's about the same size and has
the right specs and replace it. Had to do this with an Epson QX-10
switching supply. Even though I had the Sam's Notes for it, I gave up
trying to find the obsolete semiconductors that it needed. Smaller, more
efficient, cooler than the vintage power supply.
Scott
On Fri, Sep 5, 2025 at 6:56 PM Jon Elson via cctalk <cctalk(a)classiccmp.org>
wrote:
On 9/5/25 18:04, Rob Jarratt wrote:
future.
> -----Original Message-----
> From: Jon Elson <elson(a)pico-systems.com>
> Sent: 04 September 2025 15:32
> To: rob(a)jarratt.me.uk; 'General Discussion: On-Topic and Off-Topic
Posts'
> <cctalk(a)classiccmp.org>
> Subject: Re: [cctalk] Re: Repairing an Olivetti M24 PSU
>
> On 9/3/25 11:18, Rob Jarratt wrote:
>>> -----Original Message-----
>>> From: Jon Elson via cctalk <cctalk(a)classiccmp.org>
>>> Sent: 03 September 2025 15:39
>>> To: cctalk(a)classiccmp.org
>>> Cc: Jon Elson <elson(a)pico-systems.com>
>>> Subject: [cctalk] Re: Repairing an Olivetti M24 PSU
>>>
>>> Those are not real. They are conducted interference from the
>>> switching supply getting into the scope preamp via the ground lead.
>>>
>>> I have seen this MANY times, ignore it.
>>>
>>> Jon
>> Thanks Jon, obviously I don't have enough experience to know this. How
> can I recognise this in the future?
>
> Switching power supplies generally radiate a ton of electrical fields
at their
> switching frequency. If you see insanely
high frequencies in these
> measurements, you can usually assume they are radiated interference.
You
> can also turn on the scope's bandwidth
filter. I did see REAL ripple
in one of
> the traces, there were long straight lines
with slight tilt between the
noise
> pulses, those are the real ripple.
>
> Improving the ground connection at the scope probe also helps. The power
> supply injects currents into the ground terminal due to capacitance
between
> output transformer windings, and these
current flowing in the probe's
ground
> braid contaminates the measurement. Possibly
running a HEAVY copper
braid
between
the scope's ground terminal and the power supply ground will
reduce the effect.
Jon
Thanks Jon and Wayne for all the advice, I will try to remember for the
I thought that replacing two big output capacitors on the +5V output had
fixed the issue but it hasn't. The problem seems to be intermittent. I
tested it a couple of days ago with a simple resistive load on the +5V and
+12V outputs. The first two times I switched it on it operated normally.
The third time I got no output at all. It seems to start working again
after I leave it switched off for a few minutes. Over the last couple of
days I have tested it multiple times very briefly with the resistive load
and it has worked every time.
However, I am doubtful that it will always work. I don't think it can be
the SCR because I had similar symptoms with the SCR removed, so I am
wondering if there is anything else I could check for that might have this
kind of all or nothing intermittent behaviour?
Many switching supplies have VERY tricky startup circuits
that tickle the controller to start and then power the
controller from an auxiliary winding on the transformer.
Some of these circuits might require some capacitor to fully
drain before they can start properly again. Some of these
circuits may have a really high resistance value that
charges the controller's supply cap, and then a transistor
turns off that path when the controller is running to avoid
overheating that resistor. These circuits will be at mains
potential so be careful when testing them.
Jon