Parallelling Power

I'd start by learning about voltage and frequency and how each one might affect phasing.

Then I'd have another drink, and plug 'em in.

The isolation transformer could screw things up, on account of phase delay between primary and secondary. (See "vector group")

Stick with plan A, but be sure to check first that all corresponding items are the same phase/polarity.

This is what I was seeing running around 3 vac. I hooked up the scope for the trace display and a power quality analyzer which easily displays vectors to get a look at the phases and found about a 3.3 degree difference between the two sources. I'll do more math later.

My real question is this. What effect will this really have on the wiring and receptacles and ultimately the equipment which is on the branch for a momentary parallelling of the busses? I know I will have localized heating.

But is there a real world example of how this would be done? I am thinking critical loads and generator synchronization and such. For example generators can be synchronized with incoming power before switching over. They are parallelled for a preset time and then the incoming power is cut off for the generators to take the whole load.

If there are real world examples, what are they? Is this only done on large-scale systems which can absorb these losses and gains easily?

<...What effect will this really have on the wiring and receptacles and ultimately the equipment which is on the branch for a momentary parallelling of the busses?...>

With a voltage and phase difference between the two supplies, at the first attempt to make them parallel and assuming they are both working correctly, at least one of the two circuit protective devices will operate, disconnecting its/their supply/supplies from the parallel point, preventing damage.

If there is no phase or voltage difference between the supplies, the fact that the cables are being supplied from two circuit protective devices effectively in parallel during the changeover makes the cables downstream of them rather susceptible to overload and thermal damage should a fault occur during the changeover event; any fault may appear in the cables rather than being disconnected safely by the circuit protective device(s) beforehand.

The problem is that in the prospective mission in the original post, the equipment is being asked to doing something other than what it was designed to do. It's a bit like the Apollo 13 recovery: the people involved may survive, or they may not.

When you are 3.3 degrees out of phase, the voltage potential between the two traces will be negligible when at the peak.

Create for yourself a solid state SPDT (or if both legs are 'hot', a DPDT) switch to do your switching at the peak like this: Rig up some Random switchable SSRs that will switch between the two sources 16.59ms after zero-cross. Use a timing circuit to create the delay, triggered from a zero cross detector (attached probably to the original source, whichever crosses zero first) along with a pushbutton switch as two inputs to a logical AND circuit. Pass the output signal through a one-shot multivibrator and then drive the SPDT or DPDT circuit. (Note: the switch will need to have an invertor so the two switches will operate -one goes off, and the other goes on. If there is a significant propagation delay in the invertor, that must be subtracted from the 16.59ms delay, and must be less than 76 microseconds.) One leg, without the delay, and normally connected should be a male plug for the receptacle, and marked primary or original source source, and the second normally open should be also a male plug for its receptacle, and marked secondary source. The output of your circuit will have a female receptacle.

If you can understand and build this circuit, (test it first with maybe a 57W lamp load, then) plug your circuit's primary plug into the original source outlet circuit, plug the secondary plug into the secondary outlet source. You will need to fashion a bypass jumper cable possibly with clip leads on one end, (if you do not have a receptacle on the 'keep alive circuit' load side of the circuit) and with a male plug on the other end to directly attach your load that you do not want to interrupt the power to, to a spare outlet on the live primary source (be careful if you will be using the clip leads, lethal voltage here!), and hook it up.

Now you can unplug the power cord normally used to power the load you want to keep live (it will stay live because of your jumper cable). Then plug this cord into your female receptacle on the switching circuit you built, and then unplug the jumper cable from the original source, leaving the other end attached, but in a safe place (It will be hot, still!)

Go ahead and trigger the circuit with your pushbutton. It will trigger the one-shot and switch the load over to the new power source, just before the peak, splitting the difference of the 3.3 degrees phase shift over the top of the waveform. (I'm not sure this splitting hairs of degrees helps, but my brain is tired, and this is a freebie. )

Now plug your jumper cable into another outlet on your secondary source. Now you can remove the plug that normally powers your important load, and plug it into another receptacle on your secondary source.

Now you can remove both ends of the jumper cable, and put your circuit up on eBay, hopefully you will never need to use it again because you have decided to use a BACK-UPS source for such important loads.

Good luck!

SAFE ANSWER: don't do it

IF YOU ARE GOING TO DO IT: learn as much as possible first then wear SAFETY GLASSES and gloves that are both electrically and thermally insulating.

If you need someone to list on your insurance policy I'm available.

The fact that you are asking questions like this tells me you have very limited knowledge about electricity....This might seem like it's no big deal, but this kind of tinkering is dangerous....Not only might you cause harm to yourself, you may harm somebody else unintentionally...There is also the possibility that cross phasing or direct shorting your electricity could damage the components or wiring in your system....Do you even have a multimeter to test these circuits with? Do you even know how to use a multimeter? No electrician would touch this without a meter to test the voltage and phase with...If you had a test meter and you knew how to use it, you wouldn't be asking these questions...So you are clearly not qualified to be doing what it is you want to do...

What exactly are the devices and the application, as some have built in energy storage which could allow you to do rapid switching while maintaining continuous operation.

Alternatively, what about using a uninterpretable power supply (or UPS)? They are fairly cheap now-days?

What's the application?

Jack has a great idea. Use an UPS unit for the switch over instead of a power strip. Plug your appliance into the UPS, plug the UPS into outlet 'A'. Turn on the appliance. Unplug the UPS from outlet 'A' and plug it into outlet 'B'. During the swap the UPS carries the appliance with no interruption!

This is safe, and it doesn't matter if the two outlets are on different circuits, or have different voltages. The UPS takes care of all that.

Skip the UPS. Once you unplug the devices to plug them into the UPS, you can just as well plug them directly into their intended destination.

Obviously the device would need to be plugged into the UPS first with the supply to the UPS itself being the thing that gets transfered over (either the primary supply or some backup secondary supply), hence a continuous supply to a set of devices off a power strip connected to the outlet of the UPS.

Hopefully The.Tinkerer can supply some more detailed information as to the application to help clarify things.

If you are switching mains power, the safe and sensible way is to have a clean break from one and then switch in the other.
The big question is ...
What equipment are you powering?
If the actual equipment is DC powered (like most stuff e.g a PC) then it's own power supply should hold it up long enough for the switch over.
If it's big AC motors and stuff wot goes roundy roundy then mechanical inertia should keep it going for the 50milliseconds or so necessary to switch over.
Del

Here is what you need to do. Create a power cord that has 2 input male input plugs with standard plugs on them, say 20 feet long, so they span 20 feet. (not a male and a female - which is a common extension cord) Then locate 2 wall sockets on the same line, or fed from the same side of an AC circuit. What this means is there will be no voltage difference between the ground, hot or neutral of the 2 wall plugs. Test to be sure. I assume the machine is plugged into one power bar with spare power sockets. Then plug into the power bar with the machine in it and plug the other end into a wall socket. You will now have a power bar fed by two in phase AC lines and your machine plugged into that same power bar. You can now unplug one ac source, with the machine getting power from the other wall socket, and move the machine, and move the power cords, plug them in and out as you go, making sure you are in phase with the same hot leg and neutral. If you plug one into one hot leg and the other into the other hot leg = bang + fuse and you do not want to do that.

Only do this if you know what you are doing

When you parallel the supplies there will be a circulating current. I would guess this would be small, depending on the size of the transformer. Reduce the size of your fuses and give it a try. If all is OK replace original fuses and use as you like.

From the info you have presented you can do it with no problem as long as everything is in phase. The biggest hazzard will be the hot male plugs during this operation and it would be advisable to have a helper with an understanding of what is going on if only to hold them away from everything until the changeover is complete. The 3volt and 3deg. difference is mainly from transformer loading and is not an issue.

I feel I could do this (but I find it totally unnecessary), but I would need a dual beam 'scope to make sure I get my phasing correct.

If you need to feed equipment continually with mains, depending on the size of the load, a correctly sized UPS would allow you plenty of time to change the source with no danger to you or anyone else.

Also then you don't need a 'scope either!!!

PS. I was unable to deduce what you need to feed from the mains, or its load current, that may make my answer incorrect....

Would an UPS solve this problem. Connect the device to a UPS. This will give you at least 5 minutes to unplug the UPS from the first outlet and plug it into the second.

"If I "reverse" feed power to a power strip, not a surge suppressor, from another receptacle on the same branch or phase, whatever you want to call it, momentarily then unplug the original source I will have succeeded. This is of course all hypothetical with all rules and regulations aside."

Not hypothetical, it will work, basically, you are just doubling the wire size of your original circuit. However, keep in mind that you will end up handling male power plugs that are 'hot', so be very careful.

"Now, let's say that the second source I am using to reverse feed the power strip has gone through an isolation transformer and there is a 3 vac difference between the two "sources" so it's not quite in phase, but is derived from the same phase before the transformer, which is now a separately derived service."

Since the 'new' source is after an isolation transformer, it is 3.3 degrees behind the original, a relatively short time interval. Because of these time and voltage differences you don't want to just tie them together. But I think you can accomplish what you want to do by following these points:

1. Build the entire switching assembly before inserting it into the circuit.
2. Keeping in mind your backfeed power strip idea, build an assembly to insert into the primary circuit, a double pole double throw, zero cross switching, solid state relay. This gives you the ability to switch on and off the present power source at will, and have it happen at a time when the voltage is zero.
3. Connect your 'new' power supply to another double pole double throw, zero cross switching, solid state relay. Again, this gives you the ability to switch on and off the 'new' power source at will, and have it happen at a time when the voltage is zero.
4. Tie the outputs of the two solid state relays in parallel. Make sure that you match the polarities of the circuits, hot to hot, neutral to neutral.
5. Tie the two solid state relays to the same control voltage. This will switch both sources at the same time.
6. Wire the primary source through the normally closed contacts of it's solid state relay. Because the relay is off, the circuit is maintained.
7. Wire the secondary source through the normally open contacts of it's solid state relay. Keep in mind that because the relay is off this circuit will be open, and therefore not yet connected.
8. Use your backfeed idea to insert the switching assembly into the circuit on the primary side.
9. Connect the secondary source to the switching assembly.
10. Disconnect the original power strip supply plug from the original outlet. Careful it's HOT.
11. Switch on the solid state relays, this will switch you to the new source.
12. Plug the original outlet strip supply plug into the new source.
13. Disconnect the switching assembly.

Here's the timing, when you turn off the control voltage, the original source will turn off at the next zero voltage crossing. The secondary source will not yet be at the zero cross, and won't turn on, until it's next zero cross happens 3.3 degrees later. All your switching happens when the voltage is zero, therefore the current is also zero during the switching.

Waayyy too complicated.

Hello The.Tinkerer are you there?

Can you please supply some more information to help clarify your application to keep this thread on track (we are now just guessing).

To original poster.

and WHY are you doing this?

This is something we did routinely at the power utility I worked at but unless you are building a table top demo project to show how a big public utility handles feeder switching it makes no sense. As several posters have already suggested A UPS will do the same task much more effectively.

A UPS would work after the devices are plugged into it, but not before--which the OP needs. [Revisiting #8.]

?????

Seriously, what are you talking about?

If the starting state is "plugged into the UPS", what is the problem then?

All that he may need to do is plug the UPS in a different outlet....as long as he doesn't take longer to replug that the UPS can supply mains = no problem.

Tell me what I am missing please?

That it is not plugged into the UPS in the first place. (That's what the before pertains to.)

Based on the information given in the original request (specifically paragraph 2 sentence one - I want to switch some devices from one receptacle to another without losing power to the device) and in post#9 (among others) the device would be expected to be plugged in to the UPS before the UPS is connected up to the primary source, not after, to create a no-break power supply system that could be feed from multiple independently selectable supplies (which power the UPS and hence the customers load(s)).

This is how I read the limited information provided in the request.

Sorry. If the device is already plugged into a UPS, then the OP would have no problem in the first place; he would simply unplug and replug the UPS. The only reason for a problem is that his device(s) are not already plugged into a UPS.

Well, that was a bit silly on the part of the original poster in not using an UPS in the first place.

There's still no answer to the questions about what it is that needs to be kept running in this way and why. What happens when the facility experiences a power cut/blackout upstream?

Make sure you have suitable fuses in the transformer secondary circuit feeding the second receptacle, make sure the two supplies are in phase and not 180 out of phase.

Connect a second plug to the load with a view to disconnecting the first one later, all live, then go ahead, plug in the secondary circuit.

At this point you still have the main supply and the transformer primary and secondary in parallel, acting like a choke with two sets of windings, the out of phase and 3vac difference will cause some circulating current around the two windings which will depend on the size of the transformer the impedance and resistance of the windings. But at this time the load is supplied from the primary side because of the mains low impedance. So I feel the transformer will not come under any load until you unplug the primary supply, circulating current will be relatively small.

Then as someone else has warned you have a live plug connected to the load wich needs disconnecting or making safe.

Sorry for not replying sooner...I had to work on some other issues related to the power swap.

Our entire building has two UPS' of its own. We are doing some maintenance on the static switch which will make UPS power unavailable. However, we have generator power for the building and the "normal" power panels for lighting and wall power will still have power. The generators also feed the input to the UPS during a hazard condition. Here is the desired sequence of events:

1) Building goes on generators. Both regular and UPS power is derived from the generators...I parallel/swap power for our equipment...the UPS equipment goes down for maintenance...when UPS power is available again we swap back to UPS power for the equipment (this time with small UPS' in line to prevent future issues, we'll just be able to move the small UPS plug from source to source.)

Note: I had already thought of this and have the UPS' charged and ready, but I greatly appreciate the collective's same suggestion.

2) The normal power closest to the equipment is phase A. The equipment is also phase A.

Note:The 480 steps down to 208 at the main distribution transformer. This is the first phase delay. The panel that feeds the equipment normally is also on an isolation transformer and adds more phase delay. Both transformers are separately derived services, not that it matters as much.

3) I have a Fujitsu high bandwidth router with dual power supplies in parallel so swapping one power supply and then the next is not a problem. The Fujitsu will stay online. My problem is the time code generator and the coax to fiber converter which service the router. They are single source.

4) An outage on the router is not advisable nor desirable. It is an extremely odd event where we lose UPS power for the building...it hasn't happened for almost 9 years. That was the time which a power quality survey engineer dropped a shipping bar into one of the 480 transformers. That was fun...

5) I will be putting a small UPS into this cabinet in the future for just such an event...but right now that thought is too little too late.

My ultimate question to the collective was basically if anyone had ever tried anything like this before. If they had or hadn't, why or why not? What is the ultimate or real harm for a parallel feed of the same phase from two sources when the phase difference is so small?

Again the simplest, cheapest, quickest and safest method would be to include more smaller UPSs powering loads that can keep the router and other critical low current drain devices up and running during a failure of the power supply.

Having small UPSs downstream of a larger UPS is not that uncommon and makes sense for critical loads performing a vital role that you cannot afford to have go down. This is what I would do if I were you, and likely what your IT department would advise also.

Sure you could implement a ring main system powered from two independent supplies with a source transfer switch in the middle, but is it really economical in this instance compared to, say a few good quality (say) $100 2kW UPSs?

Just to tease you, Jack, but you will have to move the decimal points on that UPS cost estimate. 2KW of UPS power is going to cost like five hundred bucks. For a hundred bucks you get three or four hundred watts.

I get mine at the dump, and buy new batteries. Got a sweet 900 watt APC, the batteries were less than a hundred, delivered.

I give the dump manager a cold drink and he lets me pick over the electronics pile.

Good times, good times.

I love the price.

Yes, for the future that would be a good idea and I plan on implementing something like that.

However, right now we don't have this arrangement. I do not want to schedule an authorized service interruption for these lines of communication. A lot of people depend on these normally highly available circuits.

I will be running some tests (experiments) with similar loads. I will post back with my results. I was hoping one of the members here would have viable real world experience on this subject.

As elnav has said parallelling systems is something power supply companies do as a matter of routine. You have carried out the same process as they do. There will be circulating current. The value will depend on the voltage and phase differance across the break and the impedance from each side of the break to the point of common coupling ie where they join. If there is any doubt a reduced value of fuse or relay setting would be used to prove the parallel. From the figures you have given I doubt that you would have any problem, as I said some time ago.

Your going to have to break the circuit some time to either implement a separate UPS system or a source transfer system. If it were me and the load is that critical I would not go with an unproven system that had the chance of not only not working properly but also potentially damaging the equipment it was supposed to protect. Keep it simple.

Can you not schedule some down time during a low usage time (a few minutes for each circuit should do to just add a second UPS), perhaps at night?

My ultimate question to the collective was basically if anyone had ever tried anything like this before. If they had or hadn't, why or why not? What is the ultimate or real harm for a parallel feed of the same phase from two sources when the phase difference is so small?

I likely wouldn't bother with parallel backup using the same phase as the redundancy in this case, I would try and run the main unbacked-up circuits back to the main UPS supply so everything is either powered or dead, and/or add separate UPSs to critical unbacked-up loads using maintenance switches to allow taking the UPSs off-line for battery replacement while keeping the critical loads going.

Additionally you may find there are problems with the way you are thinking about implementing the parallel phase feed (and potential backfeed), so don't forget to check your local electrical standards and regulations.

Downtime isn't much of an option...it is better to come up with an alternate solution. There are too many users to schedule with at this time, though it may indeed come to that. It's a 24 hour enterprise.

I am not trying to set up a parallel backup...just a momentary transfer of power until the main UPS maintenance is complete. When I switch back over a small personal UPS will be in place on the the system to prevent the need to do this in the future.

I honestly think it would be safer just to schedule some down time with your IT department or send an email notification saying there will be a momentary interruption in the connection at such time on such date for such period of time. Could you do the transfer and reset the router in about a minute or so, surely this is possible?

Perhaps it is time to have a talk with your IT department to see if perhaps you are trying to make things harder than they need to be.

I work for the "IT" department. We aren't talking about a little business here with a few thousand computers. This is a backbone node for a major ring network. We have hundreds of thousands of clients all over the world. We have had service interruptions before...but as I said it's been about 9 years.

I am not looking for options to either not turning the network off or asking for help in scheduling downtime. An authorized service interruption is my last option.

I was looking for real world experience from a learned person who may know what the limitations were for parallelling the same phase with same branch phase lag induced by separately derived services through isolation transformers on a 208Y system.

Imagine that someone in the ICU depends on equipment that cannot go down...a heart and lung machine...that is almost the type of reliability that I need...a bit dramatic, I know, but one would get the impression it was that important.

This puts a new light on things. From your initial comment I had a smaller task in mind.

Usually systems like this have methods to shunt the tasks to various boxes in the site, or even to other sites, so boxes whose tasks have been offloaded can be powered down for maintenance, moving etc.

Is this sort of redundancy not available? Is this a sole site?

If you can shift the load as suggested by Aurizon, that will probably be the best solution.

If not, I suggest that you re-consider my suggestion, using solid state relays to switch the load at the zero cross. If you modify the circuit to turn on the relay control voltage when the incoming supply is at it's peak, both relays will be 'energized' waiting for the next zero cross to switch. One path switches off, and the other switches on 3.3 degrees later.

The power supply source will be off for a very short time, and there are no eddy currents to worry about. After all we are talking about 3.3 out of 21,600 degrees per second (60 Hz system) only 1.528 X 10-4 seconds, the electrolytic caps in the power supplies should carry that without any trouble.

Please let us know what you decide to do and how it works out.

Good Luck.

Everything worked out okay with my little experiment. I just fashioned a power cord with male parts on both ends, donned my gloves and made the connections. I experimented first with some old equipment and measured the results. There was no effect as far as localized heating and a .1 volt variation as measure with a simple fluke dmm. I just connected the cord, then opened the breaker for the normal feed. Voila! No service interruption...everyone's happy.