Safely measuring balanced audio / bridged amplifiers

[jared555]

I am looking into getting a rigol 1054z for various projects (audio signal measurements, arduino, raspberry pi, multicopters, possibly dmx lighting projects, etc.)

Will do my best to avoid triggering any wars for now (I was looking at the 2072a but I can hopefully save that expense, 25mhz digital should be the upper boundary)

If I am feeding a balanced signal into an amplifier configured for bridge mono output (two amplifier channels of opposite polarity) and want to measure both the input and output, would this be the correct method without differential probes?  Obviously differential probes would make this easier but I doubt I would use them enough to benefit, at least for now.

Channel 1: XLR + / XLR shield
Channel 2: XLR - / XLR shield
Channel 3: Power amp channel 1 / Common or chassis
Channel 4: Power amp channel 2 / Common or chassis

Then either set up math add / subtract between CH 1/2 or CH 3/4 to either measure total voltage?  I wouldn't always need to measure everything at once, just using this as an example. 

My bigger amplifier would be approximately 73V RMS on each channel with the possibility of working with an amplifier at about 98 V RMS.  The big amplifiers have built in meters but it would still be nice to be able to measure them.  If I am calculating things correctly, that would be approximately 135V peak which should be within the limits of the 10x probes, although that would be 270V (27V as seen by the scope?) difference between the two inputs worst case.

The short term goal is just to find the exact point where I am starting to clip both the ADC and DAC/power amp stages to set limiters.  I know there are plenty of ways to do this but I figured it would be a fun way to learn the scope.

[2N3055]

You absolutely have to use differential probes... Those voltages are dangerous, and accuracy of measurement done that way wouldn't be very accurate anyways....

Even hand made transformer would be better than that..

Do you need to see waveforms or just measure voltages? If you only need voltages, isolated DMM with audio bandwidth TRMS would give much more precise measurement of power and levels..

You could monitor just one side of the bridge output to the ground, to see clipping.. But not on a factory 10x probe.. You would need 100x probe with at least 500V rating... Those 1x/10X probes are too easy to move switch inadvertently, and have voltage ratings of max 300V despite division... And on higher frequencies even less..

You wouldn't expensive high frequency diff probes, 100-200 kHz would be more than enough, and those could be had for not so much...

[jared555]

I need to see where clipping starts since often amplifier specs are... creatively written.

I had a feeling the 'don't go near those voltages' warning would come up.  I would mostly be working with under 100W systems (30V RMS or so) and only one channel so I will wait to do any measurements on the bigger stuff.  Everything has banana plugs on the outputs so I can definitely avoid touching anything energized.

Which of the ~$400 differential probes would you recommend?  I see the B&K (a different rebadge) often and rigol has a couple.

[2N3055]

You can literally go with simplest one..

You could even use audio trafo, ones used for 100V installations, and put it in reverse, watch secondary low voltage for distortion, and then just use good TRMS multimeter (would need to have 100kHz bandwidth) to measure exactly output and input of amplifier... And if you use good AC TRMS voltmeter, it would be very accurate...

Also 100x passive probe would allow you to check one and then other side of bridge amps to see if they clip symmetrically...
You couldn't measure across the speaker with it, of course, but one side and then other ...

And I know why you're asking, you were very polite saying creatively written specs... they simply lie with the specs 

[jared555]

This is mostly justification for getting the equipment and learning it.  Once I have it I can use it for more advanced projects which was literally the only reason I was researching the 2072a for a while.

> You could even use audio trafo, ones used for 100V installations, and put it in reverse, watch secondary low voltage for distortion, and then just use good TRMS multimeter (would need to have 100kHz bandwidth) to measure exactly output and input of amplifier... And if you use good AC TRMS voltmeter, it would be very accurate...

I had actually considered just getting a DI rated for speaker level input and feeding it into one of my audio interfaces.

> Also 100x passive probe would allow you to check one and then other side of bridge amps to see if they clip symmetrically...
> You couldn't measure across the speaker with it, of course, but one side and then other ...

One thing I was also curious about was testing to see if phase was 100% in line across all channels.  Obviously it should be and it wouldn't be the kind of phase plot you get with SMAART but it would be enough to see if the amp/receiver had any weird processing going on. 

> And I know why you're asking, you were very polite saying creatively written specs... they simply lie with the specs 

I am not too concerned about my professional level equipment since most of it is tested to a real standard that you can check, but the consumer stuff I am integrating with...  The sad thing with some companies is the engineers put in a bunch of work in getting accurate measurements and then the marketing department gets their most creative people involved.

I have also heard of some amplifiers that actually have the positive / negative output circuitry balanced between the positive and negative to get away with lower rated components so I will be sure to verify that isn't the case too.  If it is then either a transformer or differential probes are the only option for that amp.

[mzacharias]

I would simply use the x10 probes with the invert - add function enabled, view the result on the scope screen while also measuring with (preferably Fluke) multimeter to measure the True RMS voltage just prior to clipping. Done it many times with no problem, although I do go "off cal" in the 5 volt/division range to reduce the waveform size on screen. I use analog scopes, a Tek 2246, sometimes a 465, and I would also point out that in the U.S. at least, one can pick up a used B&K 1472 or 1477 type scope pretty cheap on eBay (often in the $50.00 or so range), and they can go to 20 VOLTS per division even before applying a x10 probe setting.

[2N3055]

Love the "preferably Fluke" comment...  I respect Fluke , have Fluke that's been serving me well for 15+ years...

But for audio work you need good frequency response, and want TRMS voltmeter with cca 100 kHz bandwidth...
 
So Fluke 287/289 is a good choice for audio. Fluke 87 is not despite being darn good meter... Most of the Fluke are made for industrial use and not for audio work.. they go to few kHz, F87 goes to 20kHz..

For audio work UNI-T 71 is better than F87... And Brymen BM869 even better, more precise than  UNI-T 71 (and F87), they both go to 100kHz...

And for low voltage the fact that UNI-T is not very electrically robust doesn't matter.. And Brymen is that too... So no, Fluke is excellent brand, but it's no holly cow...
It's tool for the job...

[2N3055]

You absolutely have to use differential probes... Those voltages are dangerous, and accuracy of measurement done that way wouldn't be very accurate anyways....

Why? I use my scope to measure mains. As long as you know what you are doing and you acknowledge the fact that a scope shares the same ground over all channels, I don't see a problem.
Of course, to work on such voltage, you do need a pair of very steady hand.

You could monitor just one side of the bridge output to the ground, to see clipping.. But not on a factory 10x probe.. You would need 100x probe with at least 500V rating... Those 1x/10X probes are too easy to move switch inadvertently, and have voltage ratings of max 300V despite division... And on higher frequencies even less..

Many scopes are rated for 300V peak wrt ground. At least I know Rigol DS1054z supports non destructive 300V input, so does my Keysight MSOX6004A and my previous MSOX3104A.

Amplifiers in question are bridged, no obvious  common ground..  And differential inputs too..
It is sometimes hard to find safe ground...  It's not max voltage that will damage scope, it's wrong ground, like when people work on primary side of switchers and BOOM....

So that's why I said differential... I also said he can measure single ended, I do it all the time as you do.. But it's easy in a lab, with open amp and schematic...

Also, even mediocre differential probe will have more than  order of magnitude better accuracy that two probes, two channels, their isolation  and 8 bit resolution with mathematical subtraction of channels..

So you are 100% correct it can be done, but it's not a best way to do it if you consider all factors... At least to the extent that I wouldn't recommend to somebody to do it...
These are 1000W+ amplifiers, plenty of current and voltage to do real damage, both to you and equipment....