Electronics > Projects, Designs, and Technical Stuff
Scoping a mains powered switched inductive load
Truglodite:
Hello all,
I hope I don't sound like the guy who's smart enough to get in trouble, and too dumb to avoid it...
I am troubleshooting a problem I'm having with a microcontroller operated relay that switches a small mains powered shaded pole motor on/off (110VAC... I'm in the US). In particular, when the uC turns the relay off, transients disturb the DC side enough to where I am seeing false button presses executed. I've researched a bit on fixing the problem using an RC snubber to absorb the transients. I could measure the motor current with low bandwidth using a clamp, but don't think I have the correct equipment to measure the inductance of the motor (suggestions how to do this with an average hobbyist's tools are welcome).
So to optimize the RC values I figured I could capture the voltage across the relay with my scope (1054z w/ 2150 probes) as I adjust R to minimize spikes when the relay is switched (w/ sort of 'standard' C values for this application of 0.1uF and 0.2uF). The probes can handle the 110VAC sure, but I worry exactly what voltages the probe would have to deal with across that relay when the motor is switched off. It's just a small shaded pole motor in a food dehydrator... much like a desk fan motor... probably 50W or less. So I am guessing transients are not going to hit ~10kV/1kA, but I just got my scope and destroying it would be a disaster for me (I'd rather guess and test over the next few months than do that lol).
So if I was to do this, would I need 100x probes to be safe? ...or given the specs would it be ok to probe it with my stock 10x probes?
Since I mentioned mains work, and probably sound like a total noob... I also will mention I've worked in the construction industry doing electrical for a while, and have plenty of experience (30+yrs) with lower voltage measurements and design... just want to avoid having to spend on 100x probes since I probably wouldn't use them often... and the motor seems too small to have transients that really bite... but that's where my experience falls short. I could probably like I said just guess and test, but my plan is to share the snubber specs with others that have the same dehydrator. I figured one of those glass passivated paper impregnated RC snubber units would be perfect, but I don't want to tell a bunch of ppl to get a part that is far from optimal, and end up having issues with it.
Thanks in advance,
Kevin
T3sl4co1l:
300V 10x probes are fine. Don't leave them connected forever, a surge could take them out (has to do with the CAT rating), but that's unlikely just for a look.
Don't use the ground clip, however -- you have some chance of clipping it to the wrong line, which will obviously not be pretty*. The scope is already grounded (assuming the mains outlet is), so you have some reference to the line already. (The neutral may also have a few volts on it, even though it's [supposed to be] grounded at the panel.)
*The scope may survive, maybe even the probe. It's a lot of noise and smoke (and a blown fuse/breaker) for a simple mistake though. :)
You will likely see a nice spike coincident with turn-off, and maybe some rapid-fire spikes as the contacts open up and the motor inductance discharges across the gap. Don't discount the rising edge at turn-on either -- it comes with a very small spark as the wires are charged up; this happens in mere nanoseconds, and can disrupt susceptible circuits just the same.
Note you'll have to flip the switch a few times to get a good trigger. Turn-off current is maximum near zero-crossing. Turn-on rise is worst at peak voltage.
Products are normally tested for EFT (electrical fast transients) and ESD, and filtered and shielded to prevent upsets due to these, which can come from within (like your relay contacts) or from anything nearby. Best practice is to use a circuit board with ground plane, filtering on external connections, and preferably a metal (shielded) enclosure.
Tim
bdunham7:
I have one of these and it is actually pretty good. It is not a huge investment and you may find it useful in the future--I sure have, and not just for high voltage work. Just read the instructions and pay attention to the V-Hz derating curve.
https://www.ebay.com/itm/300MHZ-Oscilloscope-clip-probes-100X-up-to-5KV-for-Tektronix-HP/183013764876?hash=item2a9c786f0c:g:69YAAOSwWEZaWFJz
You are right to address your RFI issues in the mains with X and Y capacitors, but I don't know that you need to do that much research. Just do your best to capture the biggest spikes you can, put some X and Y caps in and then see if they go away. I don't think fine tuning is all that important. However, I don't think you should overlook the frailty of your low voltage control circuits--they should have their own EMI filtering and shouldn't freak out so easily. You need to spend some time with the scope on the low voltage side. I have some experience with this, and my only advice would be to not be surprised when you see very powerful but very short spikes--you may need to use the full bandwidth of your scope and the peak detect function.
Truglodite:
Thank you both for the very useful replies. I had a feeling it would be OK since the probes are CAT2 rated and this is technically speaking a CAT2 situation. I won't be hooking up a ground strap since the scope is referenced to mains earth; didn't plan to but thanks for the healthy reminder. I do plan to sit down and play with it for a while to get some good triggers and a chance of capturing max Vpp for both on and off. That said, I wouldn't leave it hooked up unless I'm there using it... and I'll make sure this testing isn't done on a day with thunderstorms hehe. I like the mention of zero crossing vs peak in the on vs off situations. It will likely take a many tries to get close to the full picture. While I'm at it I'll be sure to have a look at the low volt side as well.
I understand that researching this is a bit overkill, but that is my nature LOL. This is partly to satisfy my (perhaps unhealthy) drive for perfection, and another part for me is academic (observing EM theory in action). It feels like it would be at least a minor failure on my part to not at least try and characterize the transients with various RC snubber values while I'm in there having fun... and I've already got the parts to run a small test matrix on it.
bdunham7, thanks for the link to the Tek HV probes! Heck $40 isn't much I'll grab a set for that. I was looking at some Caltest stuff that was more in the $100 range. :-+ :-+
Now that I'm thinking about it, I'd like to probe the low volt side at the same time if possible. The 5V is supplied right now by a wall wart... and of course that makes me pause for a second with the probes. I'm not sure how the wallwart was designed, but I'm guessing it's isolated as it doesn't have an earth pin. I'm overthinking that one I think... should be fine if I just verify the 5V gnd is either isolated or tied to earth (using perhaps low impedance measurement), correct?
Thanks,
Kevin
bdunham7:
Just to be clear, those are generic probes, not Tek branded. Still, my example seems pretty good--and I've used it on a 4kVp-p 20kHz power supply (o-scope CRT HV driver). Hands-off, of course--you should always be nervous at those levels! :bullshit:
As for your wall wart, while there may be some leakage current, you can still use the ground lead on either side and the leakage current will just go to ground. All mains-powered systems have at least some such leakage current--as long as it is very low, there are no safety or ground current issues. You would need to test it in various configurations, grounded and ungrounded, to find your interference.
Navigation
[0] Message Index
[#] Next page
Go to full version