Connecting a 'scope to an SMPS

[Radio Wrangler]

Scopes' volts/div knobs often change only the attenuation in front of the amplifier input stage, though in some scopes they do decade steps before the front end and do the 1-2-5 ratio steps after the front end.

Fit a x10 probe and your wanted signal shrinks by a factor of 10, so you change the volts/div switches by a factor of 10 and voilá you are right back were you were. a factor of 10 of your input attenuation simply got exported to your probes, the overall levels hitting the input stage are back to the same and it's a zero-sum game winning you nothing.

I've had the job of developing SMPS, some rather odd ones (320W sine wave output at 40kHz) where there wasn't much prior art in the public domain. There had to be a lot of looking at live waveforms in running equipment. It's nice to avoid having to probe such stuff, but you don't always have the option.

I used to solder on a wire to a few points I wanted to look at and use a 4-channel scope with probes connected to the wites and safely insulted. The equipment was on an isolating transformer with a carefully chosen and securely connected earth, and only when I was happy and well clear was the thing brought up on power. Tests were done, traces recorded and then it was powered down, disconnected from mains and checked that things had discharged before I opened it again.

If anyone's interested, it was the power supply in the HP70000B modular measurement system mainframe, getting around some obsolete chopper transistors and upping the power, and doing a 275v DC input version. They were for test suites on the US Navy aircraft carriers. Had to get it right!

David

[cmjones01]

To add to the differential probing discussion, in my youth I tried the "A-B" scope thing for debugging a switch mode power supply primary side, using a good quality scope (a Tektronix 1A1 plugin) and 10x probes. It didn't work. The half-mains common-mode signal was just far too much for the vertical amplifiers and it was impossible to see anything useful.

Since then I've always used an isolating transformer or a proper differential probe.

Incidentally, the common-mode problem can be even worse on digital scopes. They mostly do the A-B maths digitally, which is fine as long as the common-mode signal doesn't exceed one screen height. That means that trying to use the A-B mode to see a small signal riding on top of a DC offset, for example, is completely pointless. You can't gain any resolution by using the differential mode. I tried this on a budget digital scope (I think it was a Hantek or similar) to see the signal across a current-sense resistor in a low-voltage circuit and was puzzled by the lack of results until I discovered what was going wrong.

There are good reasons why people pay money for good differential probes and isolating transformer.

Chris

[Keith956]

Originally Posted by cmjones01

There are good reasons why people pay money for good differential probes and isolating transformer.
Chris

I do agree, I use both. They are however expensive, even the cheapest diff probe will set you back a couple of hundred.

I also tend to agree with McMurdo that scope probing around the high side of a SMPS with a scope is to be avoided if at all possible. Even if you don't electrocute yourself, accidentally shorting that high voltage smoothing cap will cause an almighty bang.

[Welsh Anorak]

Back to the original question, I agree that it's not a good idea to be advising someone at arm's length when dealing with this sort of circuit.

However David makes a good point to get the circuit running with a bench PSU. There have been no bangs and the Vcc to the control IC is present and correct. There are no downsteam shorts. I think I'll suggest replacing the IC (it's a Sharp 16-pin DIL) and see what happens, then if no joy he'll throw in the towel.

Oh, it's an Iiyama monitor, not the Philips.

[kalee20]

Check the data sheet of the IC if at all possible - there may be a low-voltage inhibit circuit which stops drive if the mains is low - it makes it impossible to wind-up the supply (with an auxiliary supply to the IC present) from zero unless the low-voltage cut-out is disabled.

It's almost impossible to fault-find an SMPS without a circuit diagram (though, it's possible, if tedious, to trace one out).

[Tim]

To simply check if the oscillator transistor /mosfet is running you can quite often simply hold the end of a scope probe near the switching transformer. Alternatively you can wind a few turns of insulated wire, connect each end between the probe tip and probe ground and go “ sniffing”.
I happen to have an early Philips 3217, which is not “ grounded “ in the conventional sense. I could use that , and while it wouldn’t blow up, I would need to bear in mind the case would be at “ half mains”. Rubber gloves job!
However an isolation transformer is always best. Yes they are costly, but what is the value of your life?

[trh01uk]

There is a third alternative to transformers and differential probes, that no-one appears to have mentioned so far. That is an isolating probe. Essentially the signal is fed through opto-isolators for safety and high pass filters to remove the mains component.

Its a bit academic though because I have never seen these isolating probes for sale. The closest I have come to one is a design for one in Elektor magazine in recent years. Sorry don't have the article any more so can't give more details. Might set some one thinking though....


Richard

[Keith956]

Originally Posted by trh01uk

Its a bit academic though because I have never seen these isolating probes for sale.

Here you go, a snip at the price!

https://www.tek.com/en/products/oscilloscopes/oscilloscope-probes/isovu-isolated-voltage-probes

[kevinaston1]

Years ago, when my techs had their first taste of SMPS service, I wrote the attached article for them.

Might be of use to someone.

[Radio Wrangler]

The headline price on that linked page is only the starting price for the lowest bandwidth model, while the performance numbers are those of the most expensive one! Look further down you'll see the price goes up like a rocket.

Impressive, though.

£34,900.... what's that in equivalent RS isolation transformers?

David

[Julesomega]

Originally Posted by kevinaston1

Years ago, when my techs had their first taste of SMPS service, I wrote the attached article for them

Your general approach sounds sensible, and with a full schematic and parts list it can be applied more generally.
Was the 'ZED' series and OEM module used in consumer equipment?

[McMurdo]

allen and heath audio mixer I will say

[trh01uk]

Originally Posted by Keith956

Here you go, a snip at the price!

https://www.tek.com/en/products/oscilloscopes/oscilloscope-probes/isovu-isolated-voltage-probes

Keith,

thanks for turning up absolutely absurd numbers. My recollection of the Elektor design was that the build cost was in the tens of pounds, so what this Tek product must be offering at that price boggles the mind!


Richard

[MotorBikeLes]

I have one of Peter Waugh's isolation amplifiers with its Xs1 and Xs 10 probes.
Rarely used but there if needed.
Les.

[G6fylneil]

All this for a monitor? I would have thought that if "we" had more information some one will say " it's the 4.7uF on the ..." or something similar.