EEVblog 1500 - How an AC Automatic Transfer Switch Works

[EEVblog]

Automatic AC transfer switches are pretty cool devices. A look at what they do, a practical demo, teardown, and then reverse engineering to explain how it does it.
Bonus look at how my automated home battery backup system will work.

[johnwa]

Hi Dave, interesting to see inside one of these, I have seen ones where the switch contacts are actuated by a motor before, but not one with solenoids like this. It looks like some effort has been taken to prevent the mains and grid terminals ever being bridged, in order to guard against backfeeding.

One thing that wasn't clear from the video is what would happen if the mechanism jams, either due to an internal fault, or by the actuator being held externally. Will the microswitch still shut off the relevant solenoid? I very much doubt they are rated for continuous duty, I can see the magic smoke escaping (or worse) if they are left energised. Hopefully there are some thermal fuses at least?

I think those metal bars you saw are for arc quenching.

[PartialDischarge]

That braided cable in the "tranfee" seems uncoated, not good for humid or near the sea environments. The "transfer" switch seems to have a better solution.

[bsdphk]

Dave,

There is one problem with (many) freezers and fridges you should be aware of:

If the power disappears while the compressor runs and comes back immediately, the compressor may not be able to restart against the gas pressure, and stalls with power applied.

For reliability, you should put a "delayed-on" timer between your transfer-switch and the compressors, so that when power goes, it stays gone for at least a minute or two, before the compressors are restarted.

[JoeyG]

Does the auto switch meet the Australian standards and   will your house insurance cover you when installed?

[bitwelder]

From the 'reverse engineering' I gather that when idling in a stable status it consumes energy only when the primary supply is active, as the relay coil is energized. It doesn't seem to require energy to stay in backup mode. Correct?

[EEVblog]

From the 'reverse engineering' I gather that when idling in a stable status it consumes energy only when the primary supply is active, as the relay coil is energized. It doesn't seem to require energy to stay in backup mode. Correct?

Correct. Only the indicator LED.

[Dr Bob]

Dave, I have the same project planned for a chest freezer in my garage. One thing I plan on doing is modifying the freezer to bring the thermo-switch out to plug into an Arduino which will do the monitoring and control. I don't want to leave the inverter on the whole time drawing power from the batteries when power is not required for the freezer. The microcontroller will also light some LEDs to indicate if there have been certain issues. People have brought up some very good points about delay times for things that I haven't considered but will now put into the programming if something like the inverter shuts off due to low battery and the system will switch back to the grid. I'll also be adding a thermocouple and log the temp of the freezer. The Arduino will work from its own separate backed up power supply since it won't be drawing much power and I don't want it to reset if I lose the grid or the main battery.

The episode on the transfer switch was very timely and might be the better solution where you are running several fridges and turning an inverter on and off isn't likely to be a big savings. I also don't want to export power to the grid since the power company in California (Edison) is pushing to drop the feed-in rate AND charge customers a monthly connection fee based on the nameplate power of their PV system. I don't know if that will pass the State's very liberal government, but then again, it might. Edison could also insist on a periodic safety inspection that they charge customers for to allow the customers to keep their solar systems grid-tied.