Maynuo M8813 PSU testing - power on spikes

[Circuitous]

After having a microcontroller get damaged on a breadboard, I stopped using my Maynuo M8813 power supply until I could get around to testing it.

Here's a video showing some nice capabilities of the power supply, along with some disappointments.


This video just tests a few aspects of the PSU that Martin didn't cover in his big review of the M8800 series.

Spoiler Alert:  If you have one of these power supplies, follow the often given advice to disconnect your DUT before switching the PSU on or off (at the hard switch).

I felt this video got a bit long, so I trimmed it as much as I could.  Making this gave me more even appreciation for Dave's rapid-fire approach to reviews... I'll have to work on my speed!  I was also recovering from a cold, so the audio changes a bit from clip to clip.

[H.O]

Hi,
I'm definitely no expert on these things but since I've got a M8811 I thought I'd try it at my end, this is what I've found so far.

Using a pair of ~500mm long banana to aligator clip cables, connecting a 10k resistor (yes, I know you used a 1k but I don't thing it matters much) to the M8811 and then a Rigol RP3500 1:10 probe across the resistor I can pretty much replicate your results - very large, like +/-60V, voltage spikes appears to appear across the resistor when turning the M8811 on/off using the mechanical power switch (output OFF in all cases).

Then I tried placing the 10k resistor right at the bananajacks on the M8811, used banana to BNC adapter and a BNC-BNC coax to the scope (changing the scope readout to reflect the 1:1 input ratio). At this point the voltage seen by the scope reduced by a factor of 10 or thereabouts, I never saw "more" than +/-5V - which of course still isn't good. But.....

Then, just for kicks I went back to the first setup. However this time I pressed the power switch on my M9712 Electronic load which sits on top of the M8811 - somehow the M8811 still generated some pretty severe voltage spikes - even though it was off and I didn't even touch it. I can easily get +/-40V spikes to appear across the resistor connected to M8811 by power cycling the M9712 or another instrument on the shelf.....

Attached is screenshot of the waveform picked up across the resistor connected to the M8811 whille turning off the M9712. Obviously the duration of the spikes aren't as severe as when powercycling the M8811 but the fact that they appears to be there as well as the fact that they appear to greatly reduce in amplitude when not using the long aligator clip cables and the scope probe makes me wonder if they ACTUALLY appears on the M8811 output when you power cycle the instrument or if, in reallity they are generated and picked up by some other mechanism.

Try it out, same setup as in your video but then powercycle another instrument that uses a mechanical power switch.

This makes me wonder if Rigol perhaps is doing something right with their dicky soft power switching.

Your uC board still fried though so.....

/H.

[sync]

These turn-on spikes look like common mode noise to me. Do you get them too when you connect the scopes ground lead and tip to the same terminal of the PSU?

[Circuitous]

These turn-on spikes look like common mode noise to me. Do you get them too when you connect the scopes ground lead and tip to the same terminal of the PSU?

If I connect both the probe and ground to the same output jack, I get a tiny spike of about 3.4V peak-to-peak.

@H.O  I tried your tests out.  With a 10K resistor, I show a ringing of 142V p-p !!
  If I connected my Array 3723A electronic load to the PSU, with the PSU off, then turned-on the e-Load, I do see a spike in the 30V range, but it's very very short, and no ringing. 
  So I modified that test... I took the PSU out of the circuit and connected the scope to the Array load, and then turned on the load... a short 39V spike.  So, I then tested the Array load for common mode noise by connecting both the probe and ground to the same terminal on the load, on power-up of the load it generated a short 2.48V spike.
 
To just check for something about the input of the DS4024 being an issue, I also dragged out my old DS1052E... same results, just a lot less resolution (although it does save the image a lot quicker).

[sync]

Do you really believe it's possible that the PSU can output -70V and a 140Vpp signal with several MHz?

[H.O]

Hi,

If I connected my Array 3723A electronic load to the PSU, with the PSU off, then turned-on the e-Load, I do see a spike in the 30V range, but it's very very short, and no ringing.

I'm not sure but I think you might have partly missed my point. I can get the spikes to appear across the M8811 output by power cycling my M9712 (or my DM3068) which is not even connected to the M8811 in any way except thru the (properly earthed) mains power. The voltage spikes "are there" but the ringing is greatly reduced compared to when power cycling the M8811 - just as you say.

I guess what I'm trying to say is that if the M8811 was behaving bad on power up/down why would similar spikes appear across its outputs when another, completely different, instrument not even connected to the M8811 is power cycled?

I'm not saying there isn't an issue with the M8811, all I'm saying is that I don't think it's as bad as it looks initially - if it's actually there at all. If the spikes was really generated by the M8811 then why is the amplitude reduced by a factor of 10 or more when the output is probed with a banana to BNC adapter and then to scope as opposed to using long lab cables and a scope probe?

Not trying to start an argument, just trying to understand.

/H.

[sync]

If the spikes was really generated by the M8811 then why is the amplitude reduced by a factor of 10 or more when the output is probed with a banana to BNC adapter and then to scope as opposed to using long lab cables and a scope probe?

Had you switched the scope to 10x probe setting?

[H.O]

I'll quote myself

Then I tried placing the 10k resistor right at the bananajacks on the M8811, used banana to BNC adapter and a BNC-BNC coax to the scope (changing the scope readout to reflect the 1:1 input ratio). At this point the voltage seen by the scope reduced by a factor of 10 or thereabouts, I never saw "more" than +/-5V - which of course still isn't good.

In other words, when I used the 10:1 probe the scope readout was set to reflect that. When I used the straight coax cable the scope readout was set to reflect that (1:1).
/H.

[sync]

So the peak at the input of the scope is basically the same. Try it again with the probe and scope set to x1. I think it will similar to the banana cable.

[H.O]

Yes, pretty much.
The RP3500 probes are fixed 10x , I've got a couple P6137's but they're 10x too - don't have an 1x probes around atm.

[bendzo]

Hi,
I have second hand M8813, DVM works only up to ~60V, is it normal? (specs are quite vague, DVM  precision is in fact specified only up to 58V... but full DVM input range is not specified)
Could somebody with this "SMU" confirm DVM input range?
Thanks!
Jacob