Hello everyone! Haven't posted for a while but I'm happy to see all the activity and enthusiasm that this forum has always hosted.

I'm writing because I need some tips for a project.

I'm trying to design some sort of electromechanical switch that will allow me to commute between two different power supplies.

The device to be powered uses ~120VAC and considerable (but variable) power (several incandescent bulbs and an electronic controller).

I want to be able to switch between ~120VAC mains and ~120VAC fuel-powered generator ***without interrupting power to the device***.

The general considerations I've thought of (and what's mostly bothering my mind) are:

- I've thought I can use a group of capacitors to power the source while the sources are being switched, but I've only ever heard of this for DC, how does this work for AC? (Large capacitors are electrolytic and most importantly *have polarity*).

- If I do carge capacitors to supply energy during the commutation, a problem arises: Say, for example, that the mains is disconnected in the instant the wave is 120V+, the capacitor is charged to that voltage. I don't think it would be a great idea to reconnect the fuel-powered generator on the instant when it is 0V or -120V, causing a short circuit of somekind. And I definitely can't synchronise the sources, it's not an option.

- The device being powered has a regular power supply (transformer, rectifying diodes, capacitors, regulator), how would an irregular AC waveform affect it (when hopping from one wave to another un-synchronised one)?

- If it's not affected by the irregular waveform, then, maybe it's not affected by a sporadic disconnection of a few milliseconds? (meaning that I don't need to use those capacitors I mentioned above?)

Thanks for your feedback everyone!

__________________
"I have not failed. I've just found 10,000 ways that won't work."
Thomas A. Edison

 

How much power are you talking about ?

Light bulbs could care less about the transition as long as you do not produce some sort of spike that would blow them out.

The electronics have a power supply with caps on the DC side that will help it through the transition.

It is not difficult to turn one supply off when it crosses zero, and then wait for the second supply to cross zero and turn it on. Worst case your load would be without power for about 1/120th of a second. This assumes solid state switching/relays.

IIRC are some devices out there that are made to only switch on a zero crossing. Simple triacs will not shut off till they cross zero.

No need for caps.

Do not try to use 2 or more triacs or electronic relays in place of a larger one. They go bang! Smoke! Fire!

Did I get it right ???

3v0

 

Do you get to pick what time it will switch over at? Can you adjust the generator voltage at all?

If so, I'd use two fast-switching devices (maybe solid-state relays), and have your circuit do a make-before-break commutation right when both lines are at a peak. Most likely, the frequencies will be different enough that the phase between them rolls over every once in a while. If you can get the generator within a few volts of the line when the phases cross each other, then it's OK to short them for a few milliseconds while you transfer the load. The only restriction is to have them disconnected before the phases get too far from each other.

Again, I'd do the commutation when both lines are at a peak--not at zero. This is because the biggest hazard is inductive spikes, and inductive current is going to peak near the zero crossings.

 

Looks like I got it wrong...

3v0

 

I don't know about variating the voltage of the generator. It still hasn't been bought, is it common for fuel-powered generators to be variable?

I don't know if it's a good idea to vary the voltage of the generator. For once, the generator is supposed to be movable (a portable unit). And won't be attached to the switching circuit except for the power line (i.e. no control lines or additional circuits) so the voltage would have to be varied manually and the operators might not be very competent (they would have to quickly re-adjust the power to ~110 VAC quickly after switching because, after all, that's the voltage that is supposed to be applied to the system.

Also, I don't think the frequencies are going to be different enough for them to roll-over... They are both supposed to be 60Hz and I definitely can't vary the main one (comes from the public energy service). I don't reckon I'll be able to vay the generator's freq either... My institution won't want to buy an expensive generator because they'll have to buy several of them and also because it's non-technical people who will be operating them.

The system to be powered is mainly resistive, (light bulbs and/or LED lamps) and the controller's power supply which brings the 110VAC down to 5VDC. Is there much risk of inductive spikes?

I have a new concern, how bad is it that the frequency of the generator won't be in sync with mains power? Like: during the swap, I'll have skipped a crest or two of the wave and then I won't necessarily take over on whatever was next because they're out of sync and I don't reckon I'll be able to synchronise them.

__________________
"I have not failed. I've just found 10,000 ways that won't work."
Thomas A. Edison

 

What the *** am I talking about?

I'm not going to swap from mains to the generator. It's going to be the other way around!

The whole purpose of this thing is: in case of a power failure, someone will hop on a car and bring the generator to power the system. And then when mains power returns, we need to be able to hop from the generator back to mains without interrupting supply to the system. It will already have been interrupted once on mains failure, but one interruption is bad enough.

So, in short, it won't usually have to hop from mains to the auxiliary power of the generator. It will mostly have to hop from the generator to mains. (Cause we usually don't know in advance of power failures).

Thanks to everyone for your replies!

__________________
"I have not failed. I've just found 10,000 ways that won't work."
Thomas A. Edison

 

In this case, I think your biggest problem when switching from generator to mains supply is that the generator and the mains will not be synchronised.
That is they will not be running in phase, probably not even at the same frequency!

As mentioned elsewhere in this thread, the lamps won't care about the switching, the electronic controller will probably be OK if it has a large enough reservoir capacitor in its DC power supply. So, just use a big changeover relay which should change in about 10mS, and at a quick guess it will be OK.

JimB

__________________
Experience is directly proportional to the value of the equipment ruined.

 

Yes, no need for syncronisation, just a changeover relay to switch from mains to generator. If you use a mains AC coiled relay, you can connect the relay coil to the output of the generator - then when you start the generator it will automatically switch over - when you stop the generator it will switch back to mains.

This is how I switch my (still non-working) generator - I've tried starting it when the mains is still on (just for a test), and it switches so cleanly you don't notice it.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Ok, I guess this greatly simplifies everything...

Any idea as of some kind of test I should perform to make sure it's safe?

What happened to inductive spikes?

What about the life of the relay? Will it go for long? Can I tell the position it will assume in case of failure?

Wouldn't it be better to use solid state devices?

I reckon the devices will be exposed to variable temperatures (It will be installed in a metallic box that is exposed to the environment) It will have vents, so humidity will also be varying...

__________________
"I have not failed. I've just found 10,000 ways that won't work."
Thomas A. Edison

 

You would need to evaluate it, and decide what is safe and what isn't.

Why did you want some?.

Mine's been sat there 16 years so far, and never failed yet - just in case, I chose an octal based relay, so it's actually socketed, and I could change it in seconds.

Wouldn't have thought so, why complicate something that's cheap, reliable and simple?.

Relays are usually enclosed anyway - it shouldn't be a problem, and for what they cost it's certainly worth a try!.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Not that I wanted inductive spikes at all... But i_build_stuff said they were the risk.

__________________
"I have not failed. I've just found 10,000 ways that won't work."
Thomas A. Edison