Grid tie inverter

[Nilao]

So it's approaching hurricane season here and I have 11 kw of solar on my roof with a grid tied inverter. It requires a sine wave present to turn on the inverter. Can I shut off the 200 Amp main and feed the panel with the 6kw inverter to provide it a sine wave so I could have my full 12 kw power if the grid went down?

[Sid Genetry Solar]

Possibly.  Several important questions first:

• What control board & firmware revision is your GS 6kw inverter?  (Rev. A.1 / B / C, and 1.1r??)
• What's the make/model on the grid-tie inverter?

I would be a little concerned about connecting 12kw of grid-tie to the output of a 6kw inverter...particularly as grid-tie inverters exist to blast as much power at the infinite grid as they can.  Yes, we've designed GS inverters to handle grid-tie systems operations (via automated frequency shift)--but if your grid-tie system doesn't support frequency shift (older ones generally don't), that might be a bit problematic.

Worth noting that the batteries can be charged through the GS inverter when grid-ties are connected to the output.  This actually is a side effect of the hardware design of an LF inverter--and the only thing the GS inverter can do to stop it is to frequency-shift "throttle" and/or shut down.  (This is why I'm a little concerned about connecting 12kw of power to a 6kw inverter.)

Honestly, while I've designed/written firmware for grid-tie support (GTM status = grid-tie master), nobody I personally know has a grid-tie system that we can test and tweak GS firmware on!  So if you don't mind being dev help with lots of diagnostics/testing (GS and grid-tie hardware permitting), we should be able to get this working.

[Nilao]

It's already tied to the meter and I am net metering. I understand the ramifications of backfiring the grid which is why I had the meter box upgraded to have a main disconnect to the meter.

 

The grid tied inverter is only about 6 months old and as was explained to me it sees the sine wave and sync up with it prior to connecting to the grid.

 

The make is Solar Edge, the model is SE10000H-US.

[Nilao]

Oh and I'm running Rev c 1.1r6

[Sid Genetry Solar]

I'll look up the grid-tie inverter later today, see if it supports linear frequency shift throttling.  I believe all newer grid-ties do--but if you got a NOS ("new old stock"), it might not.

 

33 minutes ago, Nilao said:

and as was explained to me it sees the sine wave and sync up with it prior to connecting to the grid.

Standard grid-tie inverter 😉.

[Nilao]

Got a fault this morning.  Had plugged it in to charge my semi batteries and run the fridge. Nothing is hooked to the output side. I do have extension cords plugged into the recepticals though. 

[Sid Genetry Solar]

3 hours ago, Nilao said:

Nothing is hooked to the output side. I do have extension cords plugged into the recepticals though. 

Output receptacles = output side, they're just one step beyond the output terminal strip.

If you can reliably reproduce this error, I will need to dig into it a bit further.  Basically, if the inverter measured AC voltage on the output at startup for ~0.5 seconds, it'll throw that error (instead of starting the FETs and blowing itself out).  A screenshot of the "DVLT" and "DAMP" oscilloscope channels (debug volt / debug amps) would be very handy for seeing what the inverter actually saw when it tripped out (all of the others continue refreshing regardless of whether it's running or not).

[Sid Genetry Solar]

5 hours ago, Nilao said:

The make is Solar Edge, the model is SE10000H-US.

As a company, they're as tight-lipped as could be--or whether or not they have lips should be the question.

The absolute most I could find was from this forum post:

https://teslamotorsclub.com/tmc/threads/hd-wave-with-frequency-shift-power-control.253395/

 

While the SE10000H is UL1741 SA and IEEE1547 compliant (which include frequency-shift throttling), said standards do not necessarily require that frequency-shift throttling is included to be certified.  The Solar Edge documentation provides absolutely no indication of whether or not frequency shift throttling is included--or if it is, what the specifications are.

With these uncertainties in mind, I would be very hesitant to connect it to your GS inverter--ESPECIALLY as you have a 12v unit.  If the grid-tie inverter mercilessly dumps 10kw into the GS inverter, it's going to very quickly blow the GS inverter's FETs out with brute overcurrent (not to mention your batteries getting the 10kw of power dumped into them...until the FETs explode).  Which that is how grid-tie systems function: after a required "startup delay", they will usually put the pedal to the metal and dump as much power out as possible--and while the GS inverter is programmed to react and shift the frequency upwards (to throttle down/shut down) the grid-tie unit, that process takes time (1-2 sec to avoid oscillations).  And the one thing the FETs don't have in that case IS time.

 

If you knew for sure that the Solar Edge inverter supports frequency shift throttling AND if you could reduce the maximum output power down to <6kw--yes, we could get it working.

 

[NotMario]

In order to test frequency shifting support, could he just put a ~20A breaker in front of it?
Should trip the breaker if it starts shoving power through it. Not sure if that would be fast enough to save the fets, though...

[Sid Genetry Solar]

18 hours ago, NotMario said:

Not sure if that would be fast enough to save the fets, though...

Probably not...especially considering most breakers are thermal, not magnetic.

The biggest issue is that 12,000W / 12v = 1,000A (not including efficiencies).......and the FETs in the 6kw inverter are conservatively rated for 600A.  If he was at 24v or higher, they would probably survive.

[Nilao]

Would diodes work? I don't need the GS to do anything other than provide a sine wave signal so it thinks the grid is up.

[Sid Genetry Solar]

1 minute ago, Nilao said:

Would diodes work? I don't need the GS to do anything other than provide a sine wave signal so it thinks the grid is up.

No.  Two reasons:

• We're working with AC, not DC.  Diodes aren't an option here.
• The grid-tie inverter NEEDS a load on the output OR else it needs to be throttled down some way or another.

If the GS inverter did not dump power into the battery, the grid-tie inverter's output voltage will go through the roof and/or possibly blow itself up.  The reason the GS inverter would dump the power into the battery is because the GS inverter will throttle down to regulate the output voltage...and all that power has to go somewhere.

 

It's a lot like trying to use an engine that has the throttle stuck wide-open.  You either have to get the brick off the pedal, or load the output down.  If neither happens, the engine will over-rev and blow itself to pieces (after it's been "started" by the grid).

[dickson]

The biggest issue is that 12,000W / 12v = 1,000A (not including efficiencies).......and the FETs in the 6kw inverter are conservatively rated for 600A.  If he was at 24v higher, they would probably survive.

If the GS inverter is 36 v  then  using a fast blow 20 amp  glass fuse  the FETs   can be saved  ?         

[Sid Genetry Solar]

My bad, I missed a word in that post...if he was at 24v OR higher, the FETs would probably survive a 12kw blast from a grid-tie inverter.

[dickson]

My bad, I missed a word in that post...if he was at 24v OR higher, the FETs would probably survive a 12kw blast from a grid-tie inverter.

Your  48v  GS 6kw  with the rev C2 control board will  survive  a backfeed to the  grid  ?      Then if I   use  FETs rated for 600A  and 2  fast blow glass fuse instead of breaker to the output of my  15kw  Powerjack it will  also  survive a backfeed to the grid  ?        I  have not try to backfeed  to the output of my inverter   but Sean in the youtube say he did it by mistake  and the GS inverter survive .     

[Sid Genetry Solar]

Just now, dickson said:

Your  48v  GS 6kw  with the rev C2 control board will  survive  a backfeed to the  grid  ?      Then if I   use  FETs rated for 600A  and 2  fast blow glass fuse instead of breaker to the output of my  15kw  Powerjack it will  also  survive a backfeed to the grid  ?        I  have not try to backfeed  to the output of my inverter   but Sean in the youtube say he did it by mistake  and the GS inverter survive .     

No, no, no, no, don't confuse three completely separate scenarios.......

a) @Nilaois wanting to "backfeed" a 12kw grid-tied solar inverter to the GS inverter...which with all of the unknowns, is likely not a good idea especially on a 12v GS.

b) Sean accidentally SHORTED the output of the 12kw GS and it self-protected.  That was not backfeeding it from grid power.  If there's another incident where he did accidentally backfeed the output without smoke, that means the inverter happened to somehow be in sync with the grid--otherwise instant destruction would result.  (If you put limitless power up against limited FETs, the FETs are going to lose.)

c) Direct connecting the grid to the output of ANY off-grid (i.e. not grid-tied) inverter is simply asking for trouble.  Try multiplying the strain of a short circuit by 15x for starters--it's pretty well guaranteed to blow something out.

[dickson]

That was not backfeeding it from grid power.  If there's another incident where he did accidentally backfeed the output without smoke, that means the inverter happened to somehow be in sync with the grid--otherwise instant destruction would result.  (If you put limitless power up against limited FETs, the FETs are going to lose.)

Then I  better NOT  try  to backfeed the  output of my Powerjack to the grid  or  short circuit the output .   I  thought  you was able to design a control board that  will  detect  a  backfeed  and  shutdown the inverter to save the  FETs .   Designing a  control board  that  can save the  FETs  cause by user  mistake  will save a lots of money  as  if that   happen  then the whole  GS 12kw inverter  is destroy  but only  the transformer and shell is  save .      Thank you  for explaining  the difference  .      

[Sid Genetry Solar]

2 minutes ago, dickson said:

Then I  better NOT  try  to backfeed the  output of my Powerjack to the grid  or  short circuit the output .   I  thought  you was able to design a control board that  will  detect  a  backfeed  and  shutdown the inverter to save the  FETs .   Designing a  control board  that  can save the  FETs  cause by user  mistake  will save a lots of money  as  if that   happen  then the whole  GS 12kw inverter  is destroy  but only  the transformer and shell is  save .      Thank you  for explaining  the difference  .      

A PJ probably won't survive either case.

Yes, I have the firmware to TRY to detect a backfeed and shut down--but the FETs have to survive a full half-cycle of being blasted upside down before that will trigger.  It's nearly impossible to truly protect against a backfeed event.  The issue is that if the backfeed is 180 degrees out of phase, the power will go through the blocking diodes in half the FETs, and through the "turned on" FETs in the other half.  Instant destruction will result.

[JIT]

On 6/1/2022 at 3:53 AM, Sid Genetry Solar said:

Worth noting that the batteries can be charged through the GS inverter when grid-ties are connected to the output.  This actually is a side effect of the hardware design of an LF inverter--and the only thing the GS inverter can do to stop it is to frequency-shift "throttle" and/or shut down.  (This is why I'm a little concerned about connecting 12kw of power to a 6kw inverter.)

Frequency modulation is not the only mechanism to control backfeed power from GTI.  All UL1741/IEEE1547 compliant GTI's (i.e. all GTI's legally connected to the grid in the US) will shutoff or curtail output power if either frequency or voltage gets too high/low.  Depending on the inverter hardware design, you can modulate voltage to control backfeed power from GTI.  Also, modern GTI's have sophisticated anti-islanding detection so they will often not stay activated if the microgrid voltage is unstable or has too much harmonic content.  So, GS inverter output voltage control loop needs to have not only fast and accurate voltage response at the fundamental frequency but also compensation at harmonic frequencies.  You may want to study how other established LF inverters that support this capability works.

 

On 6/1/2022 at 3:53 AM, Sid Genetry Solar said:

Honestly, while I've designed/written firmware for grid-tie support (GTM status = grid-tie master), nobody I personally know has a grid-tie system that we can test and tweak GS firmware on!  So if you don't mind being dev help with lots of diagnostics/testing (GS and grid-tie hardware permitting), we should be able to get this working.

Honesty is good but frankly a bit surprised to see this since the GS inverter manual on the web site makes it look like a working feature.  Perhaps you should mark it as experimental until you have successfully tested it?

 

[Sid Genetry Solar]

1 hour ago, JIT said:

All UL1741/IEEE1547 compliant GTI's (i.e. all GTI's legally connected to the grid in the US) will shutoff or curtail output power if either frequency or voltage gets too high/low.  Depending on the inverter hardware design, you can modulate voltage to control backfeed power from GTI.

And that's precisely the issue: grid-tie inverters often don't indicate whether they support frequency-shift throttling.  AFAIK the spec doesn't require it.

Yes, I'm fully aware that grid-tie inverters have to shut down if frequency (or voltage) go out of range--but an on/off throttle is considerably more common (especially on older units).  The uncertainty coupled with the oversizing of the grid-tie inverter in this case...is what really comprises the issue.

 

1 hour ago, JIT said:

Honesty is good but frankly a bit surprised to see this since the GS inverter manual on the web site makes it look like a working feature.  Perhaps you should mark it as experimental until you have successfully tested it?

So I have tested the GS inverters' response to backfed power--yes, it will throttle the frequency upwards until either it shuts down ("GTM Regulate Fail") due to exceeding 62Hz, or until the backfed power goes down/away.  And if the load increases beyond the backfed power (i.e. power is being pulled from the battery), the GS inverter will edge the output frequency down to 60Hz until either the backfed power comes back up--or until it reaches normal 60Hz operation.

On paper--and with bench experiments--the function is tested and known to work.

What hasn't yet been tested is how the GS inverter would interact with an actual grid-tie system--and this may be highly dependent on the type of grid-tie-system.

[Sid Genetry Solar]

1 hour ago, JIT said:

You may want to study how other established LF inverters that support this capability works.

Without getting into the super-high-trim big-name-brand inverters (Outback, Schneider, etc., etc.), the only LF inverter I'm aware of that even acknowledges that this sort of functionality exists are certain Sigineer inverters with a "frequency shift dry contact input" that you have to externally control.  Without an external smart controller that really knows what's going on, if you connect a grid-tie inverter to the output of one of these Sigineers and forget about it, you'll probably boil your batteries to pieces (or launch the next lithium-fueled rocket, depending on the chemistry).

https://www.sigineer.com/features/dry-contacts-providing-ac-output-frequency-switch-between-60hz-62-5hz/

GS inverters already have an automatic frequency shift shutoff.  Again, I don't know how this will play with each individual different grid-tie system.  And in the case of the OP, the unknowns and potential power scale/amperages just don't play well with the GS inverter.

[JIT]

21 hours ago, Sid Genetry Solar said:

And that's precisely the issue: grid-tie inverters often don't indicate whether they support frequency-shift throttling.  AFAIK the spec doesn't require it.

It's not hard to tell.  Perhaps you don't know the difference between UL1741 and UL1741SA (SA is short for Supplement A).  For UL1741 compliant GTI, frequency-watt (i.e. frequency shift) curtailment is not required.  For CA Rule 21/UL1741SA compliant GTI, frequency-watt curtailment is required.  Similar requirements for voltage-watt curtailment. 

 

21 hours ago, Sid Genetry Solar said:

Yes, I'm fully aware that grid-tie inverters have to shut down if frequency (or voltage) go out of range--but an on/off throttle is considerably more common (especially on older units).  The uncertainty coupled with the oversizing of the grid-tie inverter in this case...is what really comprises the issue.

Inverters supporting GTI backfeed power 2x the nominal inverter output power capacity are available.  So, it's definitely doable depending on the inverter HW/SW design.  BTW, OP's GTI can only put out 10kVA max so the max GTI backfeed power is most likely well under 10kW probably closer to 1.5x the nominal inverter output power capacity.

 

21 hours ago, Sid Genetry Solar said:

 

So I have tested the GS inverters' response to backfed power--yes, it will throttle the frequency upwards until either it shuts down ("GTM Regulate Fail") due to exceeding 62Hz, or until the backfed power goes down/away.  And if the load increases beyond the backfed power (i.e. power is being pulled from the battery), the GS inverter will edge the output frequency down to 60Hz until either the backfed power comes back up--or until it reaches normal 60Hz operation.

On paper--and with bench experiments--the function is tested and known to work.

What hasn't yet been tested is how the GS inverter would interact with an actual grid-tie system--and this may be highly dependent on the type of grid-tie-system.

Bench testing is a good start but frankly not adequate for quality power electronics engineering.  I would suggest you do at least similar level of testing with real target loads/systems as you did for the 12kW inverter (i.e. testing with a real central A/C) before release as a generally available function.  I like what you guys are doing and I am fan of agile development in general but not so much for power electronics.  Apps crashing on a phone is a bit different from power electronics failures.  Perhaps you don't have enough customers yet but sooner or later you could have a serious liability problem.

Also, real GTI behave quite differently due to their built-in anti-islanding detection mechanisms that differ across vendors and even across models.

 

[JIT]

22 hours ago, Sid Genetry Solar said:

Without getting into the super-high-trim big-name-brand inverters (Outback, Schneider, etc., etc.), the only LF inverter I'm aware of that even acknowledges that this sort of functionality exists are certain Sigineer inverters with a "frequency shift dry contact input" that you have to externally control.  Without an external smart controller that really knows what's going on, if you connect a grid-tie inverter to the output of one of these Sigineers and forget about it, you'll probably boil your batteries to pieces (or launch the next lithium-fueled rocket, depending on the chemistry).

https://www.sigineer.com/features/dry-contacts-providing-ac-output-frequency-switch-between-60hz-62-5hz/

GS inverters already have an automatic frequency shift shutoff.  Again, I don't know how this will play with each individual different grid-tie system.  And in the case of the OP, the unknowns and potential power scale/amperages just don't play well with the GS inverter.

Sigineer has a crude hack they put in for Jack Ricard/EVTV.  It's not a good reference.  I understand Outabck and similar are high end but not sure why you would exclude them.  If anything, you should study what they do (instead of hacks like PJ) and provide similar functions with better cost/benefit tradeoffs for your customers.  I would suggest you study Victron inverters (e.g. Multiplus, Quattro).  Or even the OzInverter.

Edited June 4, 2022 by JIT

[Sid Genetry Solar]

43 minutes ago, JIT said:

Sigineer has a crude hack they put in for Jack Ricard/EVTV.  It's not a good reference.  I understand Outabck and similar are high end but not sure why you would exclude them.  If anything, you should study what they do (instead of hacks like PJ) and provide similar functions with better cost/benefit tradeoffs for your customers.  I would suggest you study Victron inverters (e.g. Multiplus, Quattro).  Or even the OzInverter.

Best I could find on the OzInverter was their usage of a voltage rise to shut off grid-ties--but this is not a very common specification for grid-tie inverters.  (Yes, they should shut down if the voltage goes out of range--but if I understood some briefs of the latest certification specs, some of the updated specifications increased the "out of voltage range timeouts" to try to not brown out as easily.)  Adding to that is that common AC line voltages can be 220v--and easily past 260vAC.  We've got several customers that have 260v line AC--either a grid-tie won't run, or if it does, how high does the input voltage have to go to cause it to shut down?

For that matter, I have never heard of grid-tie inverter throttling based on line voltage--as line voltage is not close to much of a "defined standard" as line frequency is.  Without an in-depth examination, I would presume that the OzInverter is an on/off system.  (Batteries full, raise the voltage and kick out the grid-ties.)  That's not the linear throttling I am hoping to achieve.

 

All of the others inverters are too expensive for me to buy.

Besides, there's no joy in "copying someone else"--that's an old trick that China exists on.  I would rather learn how to do it the best way and be original.

[TheButcher]

IIRC the inverter limit in my state of QLD Australia, it's 255VAC.  The pretty loose spec considers from about 216V up to 253V to be valid.  Grid tied inverters must progressively throttle as line voltage rises up towards 253V and beyond and completely shutdown at 255V.  This allows better supply stability rather than having a bucket of inverters abruptly turn off and the voltage drop etc.