UPS switching time with oscilloscope

[drummerdimitri]

I want to measure how much time in ms it take my UPS to switch to battery power when mains fails.

How to set this up with an oscilloscope?

[Jeroen3]

You will need an high voltage attenuator.
https://www.eevblog.com/product/hvp70/

[drummerdimitri]

You will need an high voltage attenuator.
https://www.eevblog.com/product/hvp70/

no i just need to know how to set the trigger etc.

I can already get a reading for mains AC

[Jeroen3]

You could try timeout trigger or runt trigger, if you have those.

[donmr]

You might connect a small DC power supply and look for it's output to drop.  A cheap one with a load on it will drop the fastest.

When I wanted to measure this, I used a transformer that output ~6V AC to make is safe for the scope (and me).  I only wanted to see the switching time in multiples or fractions of a cycle.

[drummerdimitri]

You could try timeout trigger or runt trigger, if you have those.

I do but I have no idea how to set the trigger (noob)

[CM800]

You simply have to set the trigger to 'single' with a digital oscilliscope, then set the threshold voltage to below the upper bound of the voltage, and trigger on the falling edge. These settings should all be on it.

[tautech]

You could try timeout trigger or runt trigger, if you have those.

Yep, anything else will trigger on just the next mains sinewave cycle.

Interval trigger as it's known in some scopes.

drummerdimitri, I hope you're using something better than just 10:1 probes for your own and scope's safety.

[james_s]

A simple transformer ought to do the trick and should be completely safe. Something like a 9VAC output wall wart, the DC output type will usually have a capacitor that may hide the glitch you're looking for.

[ejeffrey]

Photo transistor taped over the "on battery" LED of your UPS?

Place a current clamp / current transformer around one of the battery leads and trigger off of that?

Use an electronically controlled power switch like this https://www.sparkfun.com/products/14236 to cut the input power to the UPS and trigger on the same logic signal you use to interrupt the power?

Tons of options.

[oldway]

Attention because simply switching off the mains AC power supply of the UPS is not representative of what is actually happening and the results may not be the same as those that occur during a power outage.

Indeed, the lack of AC voltage is often the result of a fault located upstream, far from the UPS .... the entire distribution network then appears as a short circuit and not an open circuit.

This may cause the UPS to operate quite differently during load transfer.

The transfer test must be done in a different way: the UPS is protected by a low-intensity external fuse and the AC input is short-circuited by a contactor.

The fuse blows and the AC input of the UPS feels a short circuit.

[sibeen]

The easiest is probably just to use the roll mode on your scope - if it has one - one channel on the input and the second on the output and then manually stop the cro.

[james_s]

I doubt sensing the LED is going to tell you much, it may well not come on in any precisely timed relation to what the UPS is doing.

[Mechatrommer]

I want to measure how much time in ms it take my UPS to switch to battery power when mains fails.
How to set this up with an oscilloscope?

are you capable of making some setup like a switch for power cut off trigger? do you have access to internal circuitry of the UPS for detection of internal mosfet activation? do the UPS output when power cut off is significantly different from mains supply signal, for example output will have slight increase of magnitude compared to mains?

[oldway]

If you are using a relay or contactor with dc coil to short-circuit the ac input of the UPS, simply use the relay coil signal to synchronize your oscilloscope.

[drummerdimitri]

I was able to trigger the event but I am not sure how to measure the transfer time.

Here's where I placed my two cursors is the 7.8 ms transfer time correct in this case? Where exactly do you have to place the cursors to get an accurate reading?

[james_s]

That looks reasonable, that has pretty clearly captured the transition from line to inverter, it's easy to tell by the way the waveform changes slightly.

[Jeroen3]

Looks good. You can measure it from the last good falling edge to the first new falling edge minus the period.

Max allowed 0% dip is 0.02 ms (one cycle). So this looks good!

[oldway]

It seems to be a low cost off line UPS with electromecanical transfer switch (relay).
Test is probably done without load.
Transfer time must be < 1/4 of cycle, so it is just acceptable.

I used to see far shorter transfer times in UPS with SCR's static switch. (AEG UPS)

There is also a phase lag between mains and inverter....not good at all !!!

NB: check the PLL of the inverter....