Grid tied solar system.

[sensij]

If that would not fly you could pre wire everything with outlets and the UPS would not even be part of the house wiring.

You need an interlock on the breaker connecting the generator circuit into the main. Google "generator interlock kit" to see how it is done. You cannot satisfy NEC "procedurally", even with "double safety".

I would have thought that a smart Solar Inverter would not try to push all available power thereby exceeding overvoltage limits.

It is called a hybrid inverter. See Xantrex, Outback, Conext, etc. Expect to pay a few X more for it than for a dumb SMA grid tie inverter of equal power. You'll also need some kind of small battery for it to function properly, although it doesn't need much capacity if the rest of your contingency plan meets your needs.

I have posted both systems for comment; there is no need to revert.

The 10kW single inverter version is DOA... the only system worth discussing is the pair of 5 kW inverters in which you are getting some power from each SPS (1500 W max) in a way that does not require them to have a fixed phase difference, but otherwise running a genset and getting some surge support from the UPS. Again, this system is viable, in my opinion, if you believe that having this contingency plan is worth the money you are putting into it. I have no knowledge of how to ensure that the UPS system will be functional when you need it, but I would expect some sort of periodic test is necessary to make sure the battery capacity has not degraded.

[posplayr]

OK, if you want to have a chance of getting this to be code compliant (you seem to care about that with your tilt project), you're going to need to show that there is absolutely no way you can backfeed the grid. Simply saying "i'll remember to shut off the main breaker first" won't cut it... you'll need to plan for interlocked breakers or some other fool proof isolation technique.

Here is a install diagram for a Generac 6376 Generator Transfer switch. It is transferring Generator power behind a back driven breaker, not at the main 200A supply. Electrically, driving the transfer switch using the UPS as an alternate to the gen set is the same.

Procedurally to electrocute a lineman with the gen set, you would have to flip the transfer switch switch for the generator and start it up without pulling the main. Is that possible? Yes. Are you couplable of wanton disregard for life? Probably.

In the case of the UPS you would just need to flip the transfer switch on the UPS. Is that possible? Yes probably much more likely. Are you coupable of wanton disregard for life? Possibly but less likely so less likely to be approved.

Point is there is a double safety on the generator setup that would not exist in the UPS as presented, unless some other procedural safeguard was invoked.

Leave the UPS off and or disconnected. If the UPS is not even connected is there a sufficiently reduced threat? The procedure for startup of the contingency cycle would then be to turn off the main, start the generators and plug in the UPS's . From then on you cycle the transfer switch as required and the main remains open. IS there more of a threat now if I'm running on generator v.s. UPS and I throw the main back to connect to the grid?

If that would not fly you could pre wire everything with outlets and the UPS would not even be part of the house wiring.

The 2nd problem is that the novelty of your system revolves around the idea that you are using the UPS to "trick" the grid tie inverter into thinking the grid is present. That may work, but only as long as you have loads available to consume every single watt of power generated by the inverter (and no more). If you want to use the PV system to recharge the UPS, you'll have to deal with the fact that the power required to recharge the battery is not constant over time, so your other loads need to be variable. If those other loads are too high, the battery won't get recharged. If those loads is too low, the battery won't be able to take all the power, the system voltage will rise and the inverter will trip out, perhaps catastrophically if not managed well. As I see it, it may work, but you'll need some high quality monitoring and load shunting equipment to make sure you don't kill your inverter or your batteries.

I would have thought that a smart Solar Inverter would not try to push all available power thereby exceeding overvoltage limits. With the grid, the GT inverter can't lift the line voltage and so it (the grid) will absorb anything the GTI can offer and stay within the grids voltage limits. In island mode, the AC line voltage will rise sharply if the loads can not absorb the GTI output, so the GTI should (i.e. has to) regulate the power delivered to match the load else the voltage would run away. A shunt load regulator would work, but there is already a much more efficient series regulation device in the GTI.

What is SPS but a regulated output to meet whatever demand is applied?

This really is the crux of the issue. To make this work, the UPS only have to supply surge current for variations in line voltage (due to surge) and it does not have to absorb differential power between the solar output capability and the variable load demand. The UPS will only pull power as required to restore the charge on the LA batteries(that depleted in startup surge). The GTI manages the line load.

I have seen other GTI fakeouts posted but ones that use a separate set of panels and inverter to generate the reference AC signal. What distinguishes this method is that it does that with a dual conversion UPS that provides surge capacity and utilizes the same panels at full rated power. No additional panels required. In Normal operation the UPS is not even connected or might only be connected for battery maintenance.

If you revert back to just using the SPS to maintain the UPS, I think you'll have a system much more likely to succeed, which has already been proven to work. Really, I don't see how the SPS is useful except for recharging a critical loads UPS.

I have posted both systems for comment; there is no need to revert.

[sensij]

The AC-DC of the double conversion prevents battery overcharging and DC-AC output should supply whatever surge requirements you plan to try and drive in contingency mode. I put 80A breakers going into the UPS since I did not want the Sunny boy to kick off if it was trying to push too much current into the UPS but it might be irrelevant.

OK, if you want to have a chance of getting this to be code compliant (you seem to care about that with your tilt project), you're going to need to show that there is absolutely no way you can backfeed the grid. Simply saying "i'll remember to shut off the main breaker first" won't cut it... you'll need to plan for interlocked breakers or some other fool proof isolation technique.

The 2nd problem is that the novelty of your system revolves around the idea that you are using the UPS to "trick" the grid tie inverter into thinking the grid is present. That may work, but only as long as you have loads available to consume every single watt of power generated by the inverter (and no more). If you want to use the PV system to recharge the UPS, you'll have to deal with the fact that the power required to recharge the battery is not constant over time, so your other loads need to be variable. If those other loads are too high, the battery won't get recharged. If those loads is too low, the battery won't be able to take all the power, the system voltage will rise and the inverter will trip out, perhaps catastrophically if not managed well. As I see it, it may work, but you'll need some high quality monitoring and load shunting equipment to make sure you don't kill your inverter or your batteries.

If you revert back to just using the SPS to maintain the UPS, I think you'll have a system much more likely to succeed, which has already been proven to work. Really, I don't see how the SPS is useful except for recharging a critical loads UPS.

[posplayr]

Your plots then are based on using a tracker? I can't wait to see your 10 KW tracker, please send pictures.
However a tracker won't help you compensate for clouds. Extra panels facing east and west will, as clouds
disperse the sunlight. Bruce Roe

Yes a two axis tracker (still under development ); I have not derated it for 45 degree tilt limitations but it should be mostly there with only reductions of the skirts. Spacing will be unobstructed and no shading until 45deg sun angle. The plan is Qty 6 arrays of 2x3 panels each (6 total) on a 30"x45" shop that is 15" high with flat roof. Zoning limits max height above the roof to 8". It is a new invention, and I'm told by experts that it will be operational when pigs fly, but I have gotten some other more useful information here than that.

Arizona is clear low humidity most all year except for the monsoon season which is now. Humidity climbs to about 30-40%. Specular solar is most prevalent and provides for the the high harvesting yields.

[bcroe]

This is the PVwatts output. Based on a 50% probability I get 12 and 10.5 hours of
3K watt SPS daily for June and July respectively. We have had relatively cloudy monsoonal weather for most of July.

Your plots then are based on using a tracker? I can't wait to see your 10 KW tracker, please send pictures.
However a tracker won't help you compensate for clouds. Extra panels facing east and west will, as clouds
disperse the sunlight. Bruce Roe

[posplayr]

What if the grid goes down when nobody is home, or everyone is asleep? Without an automatic shut-off on the main breaker, you'll be feeding power into the grid, creating a safety hazard.

If you are referencing my diagram, IN GRID TIE mode: the transfer switch leaves both the UPS output and Generator output disconnected. The only thing the contingency is doing in GRID TIE mode is maintaining standby UPS battery power; with no discharging they will last a long time.

If the grid goes down, then the Solar inverter will shut down as required. If grid power returns before any contingency operation are invoked then power is restored and the solar inverter will simply resume operation.

The first step to invoking "CONTINGENCY MODE" is to shut off the main breaker. If you want you could use some alligator clips to attach a 120V buzzer to the grid side to let you know if power is coming back or you just look around the neighborhood to see if any other lights are on.

I don't think there is a way to screw up after that other than perhaps deep cycling your UPS batteries too far. You could get a Automatic generator transfer switch if you wanted that automatic.

The AC-DC of the double conversion prevents battery overcharging and DC-AC output should supply whatever surge requirements you plan to try and drive in contingency mode. I put 80A breakers going into the UPS since I did not want the Sunny boy to kick off if it was trying to push too much current into the UPS but it might be irrelevant.

For the cost differential of two SB5KTL SPS modules, SPS plugs and two 2.2K UPS's, the single SB10KTL and UPS6K provides full unlimited solar at close to a wash in initial cost outlay. Again the UPS6K handle surge loads or contingency overnight loads so it doe not have to be sized to full size of the SB10KTL.

[inetdog]

I was reserving dual conversion for a full solar capability not SPS. See previous figure. Why does that not work? There is no sync problem

When grid power is present, I do not see any reason to have the UPS except to ride through the transition from one source to another.
You could add one more manual transfer switch to allow the SPS to be used to help charge the battery of the dual conversion UPS when neither grid nor generator are available.

The conversion losses of the UPS in normal (not bypassed) operation are probably on the order of 5% or so, so that has to be taken into consideration when adding up the costs.

[sensij]

What if the grid goes down when nobody is home, or everyone is asleep? Without an automatic shut-off on the main breaker, you'll be feeding power into the grid, creating a safety hazard.

[posplayr]

Absolutely right. The only autostart is that once SPS mode has been initiated, an overload will trip the SPS off but it will try to restart after an interval (I believe 5 minutes) on the assumption that the insolation may be better or the surge load may not be present. There is no limit to the number of such restarts.

Turning the SPS switch off sets the unit up to validate a grid input when one becomes available and then enter GTI operation.

But it would not be difficult to wire up a relay to auto-start the SPS after a short delay whenever the grid goes down.
Note that you will also need a transfer switch to move the protected load from the grid connection to the SPS connection. And for a compressor load, that interruption while the SPS is staring up is almost certain to trigger a restart under pressure rather than just a normal starting surge. The five minute dead time should allow the pressure to bleed down before the SPS restarts.

I think with the contingency nature of this whole process, that the requirement for manual intervention is not such a burden. If it was normal operation, it is fully automatic grid tied. If the grid is down you have to manage your loads and power sources to get through the day. It still seems better than mothering a battery bank all the time and a lot cheaper.

[posplayr]

Since the SPS is designed for off-grid operation, it is not designed to synchronize to any other power sources. I do not think that you can parallel two SPS to get increased output. More likely both of them would shut down.
Unlike the SunnyIsland line, there is no provision for linking or synchronizing the SB inverters, since they are assumed to be synchronized to the power grid.
For the same reason, the SPS will not play together with a generator, except through a transfer switch.

Your only option that I know of would be to put a dual conversion UPS on the output of the SPS and use that to cover the surge.

The two SPS 10V outputs are isolated and at random relative angles(frequencies). Each serves a different leg of a 240 feed to serve 120V loads only. You could get lucky and the two legs are 120 deg apart but that would be as correct as a stopped clock.

I was reserving dual conversion for a full solar capability not SPS. See previous figure. Why does that not work? There is no sync problem

[posplayr]

EDIT 7/13/2015: Rather than just remove this post, I will detail what is wrong with it in case someone is asking about something similar.

First the transfer switch in this diagram probably has no practical way of providing an interlock between GRID power and the internal (solar, generator) power sources and so it is in violation the NEC.

The Grid Tied Inverter is of a special sort that with what is called "grid interactive" interface. What that means is that the GTI has to push current to the grid by just riding above grid voltage even through that grid voltage is moving up and down as the GTI meanders around with near and far load variations. This means that the GTI does not regulate on voltage but instead on current; it pushes as much current as the collected power from the solar and the grid voltage will allow.

If you hook this GTI up by faking it out that there is a Grid, you will have to do some fancy load balancing as the solar production and local "island" loads change during the day. If those loads and production are not balanced they would have only to go into battery storage and if the balance exceeds the charging capacity then the line voltage in the house will go up till you start getting shut downs or worse.

The only way to power your house (as an island) is if you have an inverter with some type of voltage control and a standard GTI just does not do that. There are solutions where you create a shunt regulator with some type of a shunt load so that you can pull the excess power available away from the line through this load. This will keep the voltage in check and you can run what ever you want but the shunt has to be sized for the largest excess you would expect to see between production over your demand.

The other figure in this thread is also nonNEC compliant as the transfer switch does not have a mutually exclusive interlock with the GRID, but SPS does provide a voltage regulated output that would not over drive you circuits; it would provide what ever current is demanded with compliant voltage levels assuming the solar production could support it.

A full output solar system for Grid tied disconnect contingency.

Modes of operation:

NORMAL GRID:TIE

• Transfer switch is off and Solar Invertor backfeeds the transfer switch==> Sub Panel==> Main Panel ==> Net Meter ; all loads supported

GRID DOWN MODES: 200A MAIN IS OFF

• No Solar: Transfer switch selects either UPS or Gen Set (SB10KTL Main thrown off to prevent potential backdriving Transfer switch).
• Solar:Transfer Switch selects UPS to Power the Transfer Switch SB10kTL Turns on after verifying "grid" power.

Note that the UPS is doing basically what it is designed for which is fast surge protection for starting loads. Battery sizing for extended amp-hr pure drain operation is optional and offset by the generator.

I dont know if anybody has done this. With pure double conversion (as opposed to line interactive), I don't see why it would not work.

[inetdog]

One last shortcoming of the SPS is that I do not believe it will automatically start to produce power if the grid goes down. I believe you have to manually initiate the sequence to go from grid production to emergency production and then back again.

Absolutely right. The only autostart is that once SPS mode has been initiated, an overload will trip the SPS off but it will try to restart after an interval (I believe 5 minutes) on the assumption that the insolation may be better or the surge load may not be present. There is no limit to the number of such restarts.

Turning the SPS switch off sets the unit up to validate a grid input when one becomes available and then enter GTI operation.

But it would not be difficult to wire up a relay to auto-start the SPS after a short delay whenever the grid goes down.
Note that you will also need a transfer switch to move the protected load from the grid connection to the SPS connection. And for a compressor load, that interruption while the SPS is staring up is almost certain to trigger a restart under pressure rather than just a normal starting surge. The five minute dead time should allow the pressure to bleed down before the SPS restarts.

[inetdog]

Since the SPS is designed for off-grid operation, it is not designed to synchronize to any other power sources. I do not think that you can parallel two SPS to get increased output. More likely both of them would shut down.
Unlike the SunnyIsland line, there is no provision for linking or synchronizing the SB inverters, since they are assumed to be synchronized to the power grid.
For the same reason, the SPS will not play together with a generator, except through a transfer switch.

Your only option that I know of would be to put a dual conversion UPS on the output of the SPS and use that to cover the surge.

[posplayr]

I understand your design. The gen set is the main power source until the sun comes up to do it's job. And you are correct being in Tucson you have a much better chance of long cloudless days.

I just didn't want others the think that they could "always" get 8 full hours of the full 3kw output of those SPS inverters.

One last shortcoming of the SPS is that I do not believe it will automatically start to produce power if the grid goes down. I believe you have to manually initiate the sequence to go from grid production to emergency production and then back again.

In Tucson , you don't need an alarm to tell you the AC is off even at 9:00 am. As stated this is contingency only.

Dec 8.5 hours

[SunEagle]

Well that is the stressing case, however you overlooked that this is Tucson and I would have a tracker with 45 degree tilt capability.

Further, perhaps the diagrams are confusing, but there is no reliance on the SPS. The Gen set provides for a flexible switch over capability.

I understand your design. The gen set is the main power source until the sun comes up to do it's job. And you are correct being in Tucson you have a much better chance of long cloudless days.

I just didn't want others the think that they could "always" get 8 full hours of the full 3kw output of those SPS inverters.

One last shortcoming of the SPS is that I do not believe it will automatically start to produce power if the grid goes down. I believe you have to manually initiate the sequence to go from grid production to emergency production and then back again.