Device for measuring mains power-on behavior within sinewave

[martin_nga]

Hi @all,

we're designing a product where we need to record the behavior when powered on at specific time during mains cycle (peak, zero crossing, in between). Recording the waveforms is not a problem, but switching precisely, safely and with as little as possible internal resistance is where my problem is.
We thought about making a box with a "high" voltage capacitor, charging / discharging series resistors and test switch.
But, does a tester like this, maybe a bit more sophisticated, exist as from-the-shelf unit?

Thanks!

[Kean]

ITech IT-M7722

Oh, that looks nice.  I have the older ITech IT7321 AC PSU and a Tonghui TH3321 Power Meter.  Both are very handy instruments, not that I do a lot of mains power stuff.

[martin_nga]

Thanks, I've had a look at those devices. I can't find info on my use case in the corresponding datasheets, so I think i need to give a few more info on my use case 

We're designing a product with bus communication interface. This usually outputs low voltage (16V), but needs to be tolerant when the electrician erroneously connects 230V AC to the bus terminals. For this, we've included a circuit that detects the voltage and isolates the bus. This happens within something like 1µs, where the worst case is applying the voltage at AC peak. This circuit now needs to be checked for switching time / failure current / destruction parameters, where only say 10µs after switching on is relevant. We've done some testing with just an isolation transformer and a switch, trying to hit the AC peak.
With that
- one needs many tries until successful
- we have switch bouncing
- we can see that the major current limiting element is the transformer internal resistance
- we can't set the internal resistance to a defined value (at least not reduce...)
- safety problems
We're now at the point that our time scale is so short that we don't need AC voltage at all, which finally brings me to the opening question:

We thought about making a box with a "high" voltage capacitor, charging / discharging series resistors and test switch.
But, does a tester like this, maybe a bit more sophisticated, exist as from-the-shelf unit?

[capt bullshot]

One could build a circuit using a Triac as the main switching element and some line frequency synchronized delay to turn on the mains voltage at defined phase angles. Don't use variacs or isolation transformers if you need low source impedance.

Otherwise, the kind of tests you describe are the "real world" kind of test, an electrician routing 230V to your comm port would cause serious bouncing and ringing, just as your tests do.

So, speaking from experience, I'd recommend to run both kinds of test - the well defined ones and the rather undefined real world tests. A device passing the well defined tests fails at the real world environment more often than one assumes.
Somewhere someone quoted the "file test" - using a rough file drawn across some metal electrode (one wire connected to the file blade, the other one to the piece of metal) to randomly interrupt the circuit producing all kinds of nastieness to test his circuits.

At a former workplace, we had the "Klapperkasten" - just a large contactor driven by an 555 circuit. This thing also produced all kinds of real world mess, used for testing instruments.

[Kean]

How about just using a TRIAC based leading edge dimmer module?
You'll get repeating edges if you leave it on - which may be useful for repetitive testing...

That's basically one of feature that the ITECH Programmable AC PSU's can simulate, but it can also just be set to start/end the AC waveform at a particular phase which I think is what blueskull was originally suggesting.

[martin_nga]

Just to prevent us wandering from the subject... my main question is, can I buy this device ready-made from a reputable supplier for a reasonably amount of money.

Testing directly on mains with random switching / noise / ... is of course necessary, but before doing that I would like to characterize the circuit in a controlled environment.

[martin_nga]

The AC source would be fine with me if it would state in the specifications that it behaves like mains within my timing requirement.

[nctnico]

I'd use a high voltage supply and a bank of capacitors. With more / less capacitors you can push more energy into the circuit if you like. I have made a capacitor bank with 6 2200uf capacitors in parallel. It is good enough for 1kA pulses when charged to 60V.

How complicated is your circuit and what other requirements are there? Nowadays you can get mains rated PTC fuses:
https://www.littelfuse.com/products/polyswitch-resettable-ptcs/telecom/trf600/trf600_150.aspx
A big fat TVS diode across the input and this fuse seems like a good start to me. You likely need to add some series resistance to keep the current through the PTC fuse in range though.

I'm also not sure whether over doing the protection is a good idea. In the end you are trying to prevent something which is very unlikely to happen and if it does it isn't your fault. Maybe a TVS and a sacrificial resistor is a better (and more reliable) idea. You have to take into account that any active protection circuit can fail and act on false positives. In the end that makes a device less reliable / more cumbersome to use.

[Kean]

The AC source would be fine with me if it would state in the specifications that it behaves like mains within my timing requirement.

Suggest you have a chat with Batronix.  They seem to sell the ITECH units, but only list the older models (like I have) on their site.
The newer models like blueskull mentioned are effecitively full AWGs for mains power with comprehensive measurement functions.
Batronix may also have some other suggestions...

Price will depend on maximum power rating but the smallest 300VA ones start at maybe $2,000, and they go up from there in leaps and bounds.
If that is within your budget, check the user manuals at www.itech.sh.  You may have other applications you can use it for to justify the purchase.
If your budget is more limited, you'll be hard pressed to find a cheaper option than a leading edge TRIAC dimmer module, or an appropriately size capacitor bank charged to the specified voltage.