Juggling PV & grid proving difficult

[vtpeaknik]

Hi all.  I'm on the grid but added a backup power system based on PV.  I'm trying to make use of the PV power while at the same time avoiding deep discharging of the batteries, both to lengthen their life and to avoid being caught with half-full batteries if or when the power goes out.  I now have roughly 800 watts peak PV, but this is in the cloudy Northeast US so does not amount to a lot.  Nevertheless at this time of year (longest days) I should be able to run the refrigerator and freezer on it, I thought.  That's roughly 50-60 watts average.  And I can, but to keep the batteries mostly full I've set up the following arrangement:

During sunny weather I disconnect the system from the grid and let the batteries do it all.  They get charged up full again by mid-morning.  But during not-fully-sunny weather (most of the time) I've plugged the inverter into the grid at night.  The inverter (Xantrex Prosine 2.0) has a built in transfer switch, so it passes the grid power through to the frig & freezer.  To automate the nighttime grid connection, I've added a relay that is activated by the "aux load" output of the PV charge controller: it thus connects to the grid whenever the battery voltage is below 13V.

I expected this will shave about 1 KWH/day off my grid electricity usage, which totals about 5 or 6 KWH/day so it would be noticable.  After 2 months of doing this I've now seen two electricity bills.  The drop was a lot smaller.  Scratching my head over it, I've finally thought of putting the Kill-A-Watt meter between the grid and the inverter.  Turns out that, with no load, and with grid power applied, the inverter is using 14 watts constantly.  (And 93 volt-amps!)  This is disappointing.  The same inverter uses only about 2 watts of battery power in standby mode when the grid is disconnected.  Apparently its designers thought that when "shore power" is available one is not concerned with the phantom draw.

Any ideas how I could arrange the system to do what I really want it to do?  I can think of a complicated timer-based system but it's not appetizing.  The timer would be needed to force the frig to not run most of the night, and only run during a shorter period (or two) when grid power would be applied.  Simply disconnecting the grid power for most of the night would just result with the frig running on battery power whenever its thermostat feels like.

I also wonder how safe it is to make my own "transfer switch", can I count on a relay to never somhow fail and connect the grid to the inverter output.  And does it have to be a DPDT relay or can it be SPDT (with the "neutral" side connected to both grid and inverter)?

[bob g]

i can maybe offer an option for your last question

look on ebay for a "reversing contactor" they are more commonly 3 phase but don't let that concern you, you simply use 2 of the three contact sets for single phase

they are about bullet proof for reliability, and are failsafe when it comes to not interconnecting both sources.

they are basically two contactor/relay units, that are mounted on a common plate, they have a mechanical interlock and an electrical interlock on the coils (typically)

what the interlocks do is guarantee that both cannot be closed at the same time, only one can close and the other must remain open circuit.

used they are quite reasonable, just remember to check the coil voltage requirements, although the coils are generally replaceable why go the added expense when the proper voltage coils are offered on the units sold on ebay. typical voltages are 24vac, 120vac and 240vac
although it is possible to find other less common voltages as well.
also keep in mind that a small control transformer can provide the needed coil voltage if need be as long as you can provide either 120 or 240vac.

don't be too concerned about used contactors, because they generally have replaceable contacts and come in a kit to rebuild them. doing a rebuild is pretty straight forward and most folks would have no problem doing so if need be. in any case i would take apart a used unit to inspect the contacts and clean them if need be.

the oem's will tell you to never file the contacts, while it probably is not a good idea, they can be dressed up just fine with a fine point file
just keep in mind you want to keep the proper profile, generally a bit of a dome or convex shape to the contact point. they are designed to be self cleaning with a wiping action when activated, so keeping the profile is key to refiling and having a good result.

i use them in my switch gear, and have several sizes

they come in at least five common sizes, type 1, 2, 3, 4 and type 5
type 5 is a rather large piece of equipment able to handle something in
excess of 50 amps per pole iirc, (it might be higher than that i don't recall)

a type 2 reversing contactor would probably handle your needs quite well.

bob g

[vtpeaknik]

Interesting idea, bobg.  Would those interlocked reversing contactors work OK if driven by an SPDT relay (or switch), so that as soon as one coil in the contactor is released the other one is energized, or do they need some delay for the interlock to get out of the way before the other coil is energized?

A disadvantage of using such a double-coil contactor is that nothing conducts unless one coil or the other is energized.  I.e., need a bit of power (whether grid or PV) at all times.  And I cannot think of a way to set it up so that if the grid power is lost and the inverter is on standby it would keep the load (frig) running on inverter power.  Only if I'm around to move the frig plug to a different outlet then OK, but not unattended.  Unless the coil itself (or with some dummy load added in parallel) is enough of a load to get the inverter out of standby, in which case it'll use 25W of battery power constantly just to run the inverter, plus the coil and dummy loads, even when the frig is not running.

[bob g]

the reversing contractor is not without it fantom load, yes one or the other coil must be energized to keep the contactor locked on and the other locked out.

another option is the double pole double throw relay, while there is the possibility of an interconnect of at least one leg should a contact stick
it requires no power to the coil to be connected to one side.

i bought some open frame style, dpdt relays off of ebay about two years ago, they are rated at 30amps per pole, nicely made first world units, were brand new unused... i got them for about 2 bucks each when i bought 2 dozen of them.

probably for a 2kva inverter setup, one of these dpdt relays would be the ticket so long as one accepts the remote possibility of an interconnect should the contacts stick or malfunction which is probably fairly remote.

thinking about it, ups systems routinely use dpdt relays to switch back and forth from inverter to mains when supplying the load. i don't think they have a failure rate worth mentioning and they are of much lower quality than what is available in my opinion.

bob g

[vtpeaknik]

Another option is two relays in series, to make the always-on failure possibility quite remote.  (Their contacts in series, not the coils...)  They would need to be more complicated than DPDT though.  More like 4PDT, at least one of them.  Either that or use the secondary relay on the "hot" side only, then both can be DPDT.  And, if one can get 3PDT or 4PDT relays, can use the extra contacts to create an electrical interlock by running the second relay's coil via the first relay's contacts.

Another advantage of using plain relays is that one can easily get them with 12VDC coils, and run them directly off the (in my case) 12VDC battery and controls system.

[vtpeaknik]

After further thinking: if I want to achieve the following features:
* reasonably safe overall
* no battery power used at night (while using grid power)
* reverts to battery power automatically if grid power fails
* controlled by a 12V signal
then I think a way to go is a combination of two relays.  One, a 12V SPDT relay that is energized when solar power is available.  When it is NOT energized (i.e. via its NC contacts) it lets grid power (115VAC) through.  That's what I have now, but that AC now goes to the inverter's internal transfer relay.  Instead it would then go to the coil of a second relay, with AC coil and 4PDT contacts (for example perhaps fleabay item #300696944497 which I have no affiliation with!).  (Another option is two DPDT relays instead of the one 4PDT relay.)  Thus, if the grid power fails then the second relay reverts to its "off" state which is to feed the inverter power to the frig.  The 4PDT contacts of the second relay would be wired as follows:
* One pole (SPDT section) connecting the "hot" wire for the frig to either grid power (*) or inverter power (**) - the latter when there is no current in the coil.
* Another pole doing the same but for the neutral side.
* A third SPDT section wired with its NO contacts between the actual grid "hot" wire and the point (*) mentioned above.
* A fourth SPDT section wired with its NC contacts between the actual inverter "hot" wire and the point (**) mentioned above.
The latter two items are for safety in case of the first SPDT section failing in a "sticky" way.  Thus any single relay contact failure will not cause an unsafe situation.  The Neutral side of the AC power does not get the same safety treatment, but in reality both neutrals and the ground are all shorted together in the system anyway.

Does this make sense?  This does all the AC switcharoo inside one relay.  Do I need to worry about other failure modes, e.g., arcing between the hot and neutral sides inside the relay?

[vtpeaknik]

So I went ahead and built a transfer circuit as described above.  For a 4PDT relay I used an Omron LY4, the contacts of which are rated 10A.  Found it on fleabay complete with a matching socket that has screw terminals for the wiring.  Works now, no smoke thus far!  This is the version of the relay that has a 120VAC coil.  The coil makes an audible buzz the whole time it is energized (all night) but I suppose that's normal?  The relay gets a bit warm but not much.  The coil wattage is about 1 watt.

Question though: when it switches from grid power to the inverter there can be up to 2 seconds with no power output, since the inverter is in standby mode where it checks for a load every 2 seconds.  If that happens while the refrigerator is running, the compressor will lose power for up to 2 seconds.  I know generally it's not a good thing to disconnect a compressor from power and then reconnect without waiting several minutes.  Is a 1-2 second break a problem in that sense?  Or does it take longer than that for the refrigeration fluid to lock up the compressor?

[OperaHouse]

Anything like a compressor I would consider putting on a ON DELAY timer.  Any power drop out would reset the timer for say 10-15 minutes.  That would protect a small inverter from stall currents.  Franky, all these things just draw current for not much reason and fridges are designed for these power outages.

[vtpeaknik]

My inverter can handle it (Prosine 2.0, 2000 watts continuous, 4000 watts peak).  I've seen it report an 87 amps draw (on the 12VDC side) for a few seconds as the compressor spools up.  I just hope an occasional 1-2 seconds dropout won't hurt the compressor.

[vtpeaknik]

I've had that transfer circuit based on a 4PDT relay in operation for 2 months now.  Mostly works as intended.  But, 3 times now I've come home to find the inverter (Xantrex Prosine 2.0) has stopped inverting, displaying an error message along the lines of "ac power feed back".  I turn it off and back on and it works again for several weeks.

So, is the transfer relay not doing its job to isolate the two power sources?  Despite the two relay switches in series on each line?  What's going on?  I have two conjectures, I wonder what y'all think of them:

(1) If the relay switches from one AC source to the other while the refrigerator (my only load through this circuit, usually, sometimes the chest freezer too) is running, there may be enough inductive effect in the motor to cause a spark in the relay across its switch, between the NO and NC contacts.  During the time the spark lasts, it allows conduction, causing the quick-acting protection circuitry in the inverter to notice and react.

(2) If the frig is running on grid power at the time the relay switches to inverter power, perhaps the motor, still spinning for some milliseconds after the switchover, creates enough voltage via its own inductive inertia to confuse the inverter into thinking that it's connected to a live grid circuit?