How to test a switch mode power supply with an analog oscilloscope

[j57H8I4]

I would like to test the PWM integrated circuit of a switch mode power supply with my analog oscilloscope, like it is tested from the minute 7:36 in this video: https://youtu.be/g4Jvm5Mh594 In that video that guy connects the PWM output and the mosfet gate of the switch mode power supply to a digital oscilloscope, I think working with a battery. When I tried to reproduce the same test using my analog oscilloscope the switch mode power supply fuse blew up, because the probe ground is connected to the mains ground. I would like not to use an isolation transformer because I am unfavorable to isolation transformers.

Is there a way to test a switch mode power supply PWM using an analog oscilloscope without an isolation transformer?

[Ian.M]

Not without spending hundreds of bucks for a high voltage isolated differential probe.

[j57H8I4]

So the question now is: is there any grounded testing point in a switch mode power supply that allows to test the power supply using an oscilloscope connected to the mains without an isolation transformer and an high voltage isolated differential probe?

[Ian.M]

NO

[mcinque]

Using both channels with probes on 10X, without GND leads connected and doing MATH A-B as a differential probe (with much less performance) should work.

EDIT: nevermind. Didn't read "analog oscilloscope". Do you have 2 channels and "Invert" on your analog oscilloscope? It should work like MATH A-B.

[Ian.M]

Not a good idea on the primary side of a SMPSU, unless you've got a pair of 1KV rated x100 probes.  With modern high frequency PSUs and >300V on the DC bus, its all too easy to exceed the probe's voltage limit derated for frequency.  Also, unless the scope's gain calibration is spot-on there will be a lot of 50Hz breakthrough.

[j57H8I4]

Yes, my oscilloscope has got two channels and the second channel has got invert function.

[j57H8I4]

No, my oscilloscope has got no calibration at all: I just bought it used and I received it yesterday! My oscilloscope is rated for 250V PK MAX, one probe is rated for 200V PK at 1x and 600V PK at 10x, I do not know yet about the voltage rate of my other probe because I have bought it today on eBay, it is the original probe for my oscilloscope but its voltage rate is not indicated in the user's manual. So, could I make an attempt to use it that way or the risks are too high?

[j57H8I4]

Do you think that my probe rated for 600V PK at 10x could be used the way mcinque indicated?

Please let me know that because I would buy another one of those probes because they are very cheap, it is this one: http://www.banggood.com/1-Set-P6100-DC-100MHz-Oscilloscope-Probe-100MHz-Scope-Clip-Probe-for-Tektronix-p-1025192.html?rmmds=myorder;

[Ian.M]

Chinese ratings are rarely trustworthy.  Using those probes on any circuit with more than 50V on it would be a gamble that could kill your scope, and also you, if you attempt to handle the probes with power on.

If you are probing higher voltage circuits, you need brand-name probes, manufactured and tested to real specs (i.e. backed by a 1st word standards  org), with an adequate voltage rating, at least 50% more than the max peak voltage you expect to find, and a minimum of a CAT II rating if working on the primary side of PSUs or on circuits fed by HV DC rails with large reservoir caps. If the probe doesn't have a manual (paper or online) with a curve for voltage derating with frequency, its almost certainly not reputable enough for this sort of usage.  Beware of fakes - if its far cheaper than brand-name's RRP and sold outside of official distribution channels, odds are it isn't genuine and wont meet its specs.

Use an isolating transformer and  hard-wire a ground to the negative side of the primary side reservoir capacitor, and your scope and probe will do just fine for probing around the chopper chip, provided you DO NOT attempt to probe the MOSFET drain (or BJT chopper collector), or any part of the primary snubber network.  With crappy Chinese probes I wouldn't try to look at ripple on the DC bus either, even though its within their nominal voltage rating.

However an isolating transformer removes any RCD/GFCI protection from circuits downstream of it so be very careful when probing the primary side of the PSU.  Its not a good idea to try to hold the probe on a pin - if you cant hook it on without shorting anything and work hands-free, switch off, take a 1" piece of tinned solid wire (e.g a cropped component lead), form a small loop in the end and twist to secure, then solder it on as a testpoint you can hook the scope probe to before switching the power back on.

[j57H8I4]

I have just ordered two of these oscilloscope probes: https://www.aliexpress.com/item/P4250-High-Voltage-Oscilloscope-Probe-2KV-100-1-250MHz-Alligator-Clip-Test-Probe-Electronic-Measuring-Instruments/32433214360.html?spm=2114.13010608.0.0.8hCoj5 and also this safety face shield: https://www.aliexpress.com/item/Adjustable-Electrowelding-Dustproof-Face-Mask-Protective-Shield-Full-Safety-Workwear-Eye-Protection-Welding-Helmet/32787027294.html?spm=2114.13010608.0.0.8hCoj5 I know, they are Chinese. Now I have to wait a couple of months for them to arrive. Thank you for all your very appreciated indications and advices.

[Ian.M]

More dubious off-brand Chinese stuff.  It may be good or it may be totally dangerous shite.  Edit: See Tautech's reply below.     

When you get those HV probes, BEFORE USE, take a BNC socket, connect a 1Meg resistor from the center pin to the shell, plug the lead into it and HiPot test it, between the tip and the BNC shell at 1KV test voltage, and between the shell and all joints in the probe body.  Also check ground clip continuity to the shell.   100:1 probes should read an insulation resistance of 100Meg from the tip to the BNC shell.  After that, check the attenuation is as specified on a waveform <100V and then (and only then) I'd trust them up to 1KV if you keep them clean and undamaged.

N.B. the probe *MUST* be loaded by a 1Meg resistance at the BNC connector, same as a scope input, otherwise if you apply a high voltage to the tip there is a risk of insulation breakdown in the lead.  Never disconnect the BNC if the probe is connected to a HV circuit.

Get the right PPE for the job - that way you are far more likely to use it every time, all the time.

You are very unlikely to need the full face shield when working with electronics unless you are in the habit of fitting large electrolytics backwards or other destructive activities.  A pair of safety glasses with side shields (not fully enclosed goggles) would be more appropriate and far more comfortable, and can be ordered with prescription lenses if you need them.
 

[tautech]

I have just ordered two of these oscilloscope probes: https://www.aliexpress.com/item/P4250-High-Voltage-Oscilloscope-Probe-2KV-100-1-250MHz-Alligator-Clip-Test-Probe-Electronic-Measuring-Instruments/32433214360.html?spm=2114.13010608.0.0.8hCoj5 and also this safety face shield: https://www.aliexpress.com/item/Adjustable-Electrowelding-Dustproof-Face-Mask-Protective-Shield-Full-Safety-Workwear-Eye-Protection-Welding-Helmet/32787027294.html?spm=2114.13010608.0.0.8hCoj5 I know, they are Chinese. Now I have to wait a couple of months for them to arrive. Thank you for all your very appreciated indications and advices.

Those probes will be fine, they're made by one of the largest Asian probe manufactures that supply probes to several scope OEM's.
They are rebranded to Siglent, Rigol etc.

[innkeeper]

NOTE: doing this is highly dangerous as there are high voltage / current pathways to ground, and the chassis of the scope can be at high potential. I've added this warning good conscious - If you do not understand the dangers of doing this then DON'T EVEN TRY IT

I've done it... this goes back now circa 1985 ish.....never liked working on switching power supplies btw.. anyway

if you do this though and you hurt yourself or blow up your scope or worse .. don't come crying to me...

with that out of the way
what we used to do is use a plug to isolate the scope from ground so the chassis would be floating (either we tore out the ground pin on the power plug or used one of those 3-2 prong adapters that breaks out the ground.

then we would signal ground the scope to the switching power supply ground and go at it.

we would put the switching power supply under test on a variac. (not isolated) - this was a shop were we would get supplies someones already had a go at, so we would ramp up the voltage.

is it the right way to do it, no. does this assume a lot about how your scope is wired... yea... is it safe ... probably not.

[tautech]

I've done it... this goes back now circa 1985 ish.....never liked working on switching power supplies btw.. anyway

if you do this though and you hurt yourself or blow up your scope or worse .. don't come crying to me...

with that out of the way
what we used to do is use a plug to isolate the scope from ground so the chassis would be floating (either we tore out the ground pin on the power plug or used one of those 3-2 prong adapters that breaks out the ground.

then we would signal ground the scope to the switching power supply ground and go at it.

we would put the switching power supply under test on a variac. (not isolated) - this was a shop were we would get supplies someones already had a go at, so we would ramp up the voltage.

is it the right way to do it, no. does this assume a lot about how your scope is wired... yea... is it safe ... probably not.

Never is this method safe and should never be suggested on a public forum.
Users with limited knowledge of scope use and without the appropriate probes can kill themselves, I will not have that on my conscience. Get the proper equipment or give the repair to somebody that knows what they are doing.

[David Hess]

Not a good idea on the primary side of a SMPSU, unless you've got a pair of 1KV rated x100 probes.  With modern high frequency PSUs and >300V on the DC bus, its all too easy to exceed the probe's voltage limit derated for frequency.  Also, unless the scope's gain calibration is spot-on there will be a lot of 50Hz breakthrough.

On an analog oscilloscope, the variable controls can be used to trim the combined gain of the probes and oscilloscope to match for good low frequency common mode rejection.  Analog oscilloscopes actually work better for this operation than almost all digital oscilloscopes (1) because they can subtract one channel from another without adding quantization noise.

If you watch carefully when doing this, you can *see* the noise level fall when the vertical gain adjustment is correct.

(1) A weird exception is the old Tektronix 2232 series of DSOs which do the subtraction in the analog domain before the digitizer.

[tggzzz]

I've done it... this goes back now circa 1985 ish.....never liked working on switching power supplies btw.. anyway

if you do this though and you hurt yourself or blow up your scope or worse .. don't come crying to me...

with that out of the way
what we used to do is use a plug to isolate the scope from ground so the chassis would be floating (either we tore out the ground pin on the power plug or used one of those 3-2 prong adapters that breaks out the ground.

then we would signal ground the scope to the switching power supply ground and go at it.

we would put the switching power supply under test on a variac. (not isolated) - this was a shop were we would get supplies someones already had a go at, so we would ramp up the voltage.

is it the right way to do it, no. does this assume a lot about how your scope is wired... yea... is it safe ... probably not.

There is ONE AND ONLY ONE WAY TO SAFELY FLOAT A SCOPE; see the photo.

It is very dangerous to encourage beginners to do dangerous things they don't understand.

[mcinque]

With modern high frequency PSUs and >300V on the DC bus, its all too easy to exceed the probe's voltage limit derated for frequency.

I've done it with my DSO with 10X probes but your're right: I made the assumption he has better input specs on his oscilloscope.

Chinese ratings are rarely trustworthy.

This is true, you cannot trust specs from a not respectable manufacturer.

[sibeen]

There is ONE AND ONLY ONE WAY TO SAFELY FLOAT A SCOPE; see the photo.

Love that photo. Back in the day I used to use a bit of electrical tape around the exposed edges of the cro and across the buttons most used. Wouldn't do it nowdays, but I was once young and stupid.

[dave_k]

If you just want to test the PWM function of the SMPS controller IC, try powering it using a low-voltage bench supply .. no need to dick about with floating scopes or differential probes. Stay away from high-voltage stuff unless you know what you're doing...

[guenthert]

I would like to test the PWM integrated circuit of a switch mode power supply with my analog oscilloscope, like it is tested from the minute 7:36 in this video: https://youtu.be/g4Jvm5Mh594 In that video that guy connects the PWM output and the mosfet gate of the switch mode power supply to a digital oscilloscope, I think working with a battery. When I tried to reproduce the same test using my analog oscilloscope the switch mode power supply fuse blew up, because the probe ground is connected to the mains ground. I would like not to use an isolation transformer because I am unfavorable to isolation transformers.

Is there a way to test a switch mode power supply PWM using an analog oscilloscope without an isolation transformer?

You had me confused there with the 'digital' vs. 'analogue'.  I watched now the video (sadly I don't speak a word of Italian, love the sound of it though) -- the fellow there is using a (digital) handheld oscilloscope, which is (implicitly) floating off ground.  'digital' vs. 'analogue' is the wrong dichotomy, as most digital oscilloscopes are not floating off ground.  The common (GND) on digital bench oscilloscopes is connected to chassis and mains earth, exactly like that of their analogue counterparts. Ian and others already explained how to work around that, I just thought above was worth clarifying.

[james_s]

Why on earth would you try so hard to avoid an isolation transformer? They exist specifically for situations like this, and they work very well.

You're lucky you didn't seriously damage your scope and the probe, and the power supply you were trying to test.

[innkeeper]

Never is this method safe and should never be suggested on a public forum.
Users with limited knowledge of scope use and without the appropriate probes can kill themselves, I will not have that on my conscience. Get the proper equipment or give the repair to somebody that knows what they are doing.

Point taken, note added to the post

[Ian.M]

Your note is wrong. 

• The hazard is the live chassis of the floated scope, which makes all its controls a touch hazard, and any exposed metal parts a deathtrap.
  .
• Also, if you float it by lifting the ground wire, it is far too easy to blow up the the scope as its PSU transformer insulation could be stressed with double the supply voltage between the primary and secondary.
  .
• Lastly, again with the ground lifted, an unsafe condition  could exist even with nothing connected to the scope as it almost certainly wont meet the safety standards for a Class II (double insulated) appliance.

[j57H8I4]

Why on earth would you try so hard to avoid an isolation transformer? They exist specifically for situations like this, and they work very well.

You're lucky you didn't seriously damage your scope and the probe, and the power supply you were trying to test.

I have not found in any book or user's manual the indication to use an isolation transformer. It is very hard to find isolation transformers on the market, there must be a reason for that.