Question for the Wizzards...why can’t Enphase micro inverters be connected in parallel?

ecat

10 W
Joined
Apr 8, 2009
Messages
88
I bought a dozen of these M215 grid tie inverters with the idea of being able to switch between feeding a charge controller and battery bank, and sending power back into the grid. The panels are all connected in parallel. One M215 works fine, it is self limiting, but hooking two up to the same DC source in parallel kills them both. (I now have four dead M215🙄.
Any idea why? And could I isolate the inputs with diodes or?

Thanks!
 
I think there are some inverters with CAN-Bus that can talk to each other to synchronize outputs.

If the sine waves aren't synchronized then one inverter will feed voltage into another one and it will go boom.
I'm not qualified to explain this really.
There are plenty of discussions about this though, you just need to look.
 
Thanks for the reply. These are grid tie inverters, so they automatically sinc with the grid. I am talking about the DC input side.
 
I'm pretty sure the DC side is irrelevant.
I could hook up a 1000 different kinds of inverters to one battery without a problem.
It becomes a problem when the inverters get connected in parallel on the outputs.

Or did they blow up without even having the outputs connected?
Maybe you connected them L-N, N-L instead of L-L, N-N? just a guess.
 
You are forgetting that these are grid tie micro inverters. They are meant to be hooked up to split phase 240V in parallel. Maximum 16 to a circuit I think. They only work when fed AC power, and shut down when AC from the grid is interrupted.
Each one mounts under its own dedicated panel as designed. I am try to use them in an unconventional way, mounting them back in the equipment room, and to more than one panel in parallel. It all works fine until I try to parallel another inverter to the same group of panels.
 
I guess it should be possible but the documentation is lacking quite a bit, to be honest.
You're probably better off contacting them: https://www.bluepacificsolar.com/home-solar/enphase-canadian-solar-6000.html
 
Actually, the M215 can have 17 micros in a single 240V string with a 20A double-pole circuit breaker.

The micros are for parallel use. They won't work in serial fashion. Maybe the upcoming IQ7A might where the DC source could be multiple solar panels or batteries. It might be 2 different sub-products, but it is on Enphase's radar. Reason for them to do this with the IQ7A is that at 450-Watt peak, it would pair great with some of their partner's panels — SunPower NGT, Solaria 72-cell, etc. But, it would also work with (2) "dumb" cheap mono 250W panels, too.

The IQ8 is going to have a built-in charge controller, so this is probably the product you are looking for. 2019 will be its debut.

Go Enphase!
 
ecat said:
I bought a dozen of these M215 grid tie inverters with the idea of being able to switch between feeding a charge controller and battery bank, and sending power back into the grid. The panels are all connected in parallel. One M215 works fine, it is self limiting, but hooking two up to the same DC source in parallel kills them both. (I now have four dead M215🙄.Any idea why? And could I isolate the inputs with diodes or?
1) You can of course parallel the outputs of the inverters. That's how they work.

2) You can't parallel lots of panels and feed them into one M215. Might work but the inverter isn't spec'd for it.

3) You can't use one panel and parallel the inputs into two inverters because both inverters will want to adjust the V/I curve to find the optimum power point, and they will drive each other nuts trying to do that.

4) You can't do what you want anyway because an M215 cannot drive a charge controller or battery bank.

Your options are:

a) use a regular array of M215's and a plain old AC charger. The M215's generate AC, the AC charger converts it back to DC.

b) switch the panel output between the M215 and a charge controller. Do this only when AC power is off, and use switches/relays that are rated for DC and the max voltage/current of the panel. This, of course, eliminates the benefit of having a microinverter.

c) get a completely separate array to run the battery bank. Panels are cheap nowadays.
 
A little back story...

This system has evolved to where it is now, with cobbled together parts on a budget. I picked up a dozen M215 for like 50.00 each on ebay, with the thought that they would be better that the cheap chinese grid tie inverters. They are a total steal at 50.00 each, and work great with one panel. They even work great in a battery discharge application, drawing about 215 watts for as long as the battery holds out with not much heat generated. Only problem is you have to have access to 240V split phase to hook them up to.
The system is mostly battery based, using a pair of 5.3 kw Tesla modules in parallel to charge my Nissan Leaf through an inverter. Thing is in the summer, 9 X 250 W panels produces more power than I can use for charging the leaf, so therefore the grid tie to use the excess power. The house has a grid tied system, so turning the meter back a bit more is not an issue with the utility.

Bill, thanks for your carefully considered response:

1. Yes got that.

2. Actually it works just fine. I have had 1000 watts of panels in parallel hooked up to one inverter. It just clips the power at 215 watts. It even works connected directly to the 22V the tesla module. Barely even gets warm.

3. OK, I didn't think of the MPPT of each inverter trying to adjust the operating voltage and fighting each other. I can see that being less than optimal, but why does it just kill both inverters instantly, and permanently? I'm trying to figure out why, and if anything can be done to isolate them from each other, short of a pair of isolated DC-DC converters.

4. I am just trying to feed excess power back into the grid once the batteries are charged.

a) The way the system is set up for battery charging is 9 panels, 3S, 3P, so this would require a total re do, and would not work when the grid is down, and I use it for backup power.

b) This could work, but I would have to switch between parallel/series wiring the panels to get the voltages right, as well as switching between the charge controller and grid tie, so really complex switching.
Right now I am switching one M215 using the charge controller's "waste not" setting, it switches the AC on to the M215 when the batteries are floating. The DC side is permanently connected to the batteries, this works fine, but will not be enough in the summer.

c) No room left on the shed roof for more panels.

The simple thing would be to purchase a hybrid inverter, but that would require big $$$
The other way would be to purchase a single larger grid tie inverter, but the good ones need a minimum of 150V so would require series parallel switching again, as the charge controller maxes out at 150V.
 
On my side.. Yes they work in parallel... well at least the M190 that we today can find for about 10$ each used and still 100% working!

I tried 2 in parallel from a batch that i bought from Ontario,Canada.. and YES it work !!!

I got some 300W 60 cell panel and i tought it was a big sacrifice to let go 110Watts unused instead of full 300W..

So i connected tweo i tarallel and tried on my power supply to simulate the solar panel.. and i varied the voltage to see how the MPPT handle that.. It work WELL !!!

I now have a equivalent Enphase 380W AC for 20$!

I confirm. It worked for me when i tried!!!

Paralleling M190.png

Doc
 
I consider similar configuration. From Ecat/Doc's posts we have below:
1. two M215 with input parallel to PVs, and putput to grid 240VAC, burnt immediately. Outputs I believe is via trunk cable, and input polarity also correct.
2. two M190 with input parallel to a power supply, and output to grid 240VAC, works fine.
if this is the case, then the impedance of PV would be the cause? power supply impedance in case 2 may be larger? a logical derivation would be that limited parallel of PVs would still work with two paralleled M215.
Could Doc clarify a bit? Did you tried with real PVs instead of power supply? what kind of power supply(V/A) you use and how can you used to test MPPT? - I guess, when you increase power supply voltage the inverter output will go up or it won't find MPP point as the function is linear no peak.
Could Ecat update a bit? did you tried to add inductor at each inverter input to increase its input impedance? or slow fuse?
thanks for any comment
 
blacowjoe said:
I consider similar configuration. From Ecat/Doc's posts we have below:
1. two M215 with input parallel to PVs, and putput to grid 240VAC, burnt immediately. Outputs I believe is via trunk cable, and input polarity also correct.
2. two M190 with input parallel to a power supply, and output to grid 240VAC, works fine.
if this is the case, then the impedance of PV would be the cause?
What may have happened:

The 300 watt panel puts out about 210 watts in local sunlight. The M215's start up and they both try to juggle MPPT. Inverter 1 increases impedance and pulls down the voltage, the other one sees that and reduces impedance. All the load goes toward inverter 1, and it draws so much that the panel voltage drops. Then inverter 2 pulls a little more while doing its search algorithm. Inverter 1 starts backing off and inverter 2 rapidly ramps to full power. They slam on and off (0 % to almost 100%) until they fail.

The M190's start up - but neither, on its own, can pull enough power to get the voltage to dip appreciably. So the slamming on and off is eliminated.
 
"Inverter 1 increases impedance and pulls down the voltage, the other one sees that and reduces impedance. All the load goes toward inverter 1, and it draws so much that the panel voltage drops. Then inverter 2 pulls a little more while doing its search algorithm. Inverter 1 starts backing off and inverter 2 rapidly ramps to full power. They slam on and off (0 % to almost 100%) until they fail."
if Inverter 1 increases impedance, PV voltage would go higher. Yet I agree MPPT may be the issue, MPPT messed up due to interfering from the other, it might be designed w/o overload protection so it burns.
M190 case, I think it uses power supply not PV to emulate, impedance of them are different. Yet this needs clarification from Doc.Or may be M190 has protection circuit.
 
It might help to put a big capacitor across the input lines. Switching noise would be reduced.
 
Anyone made this work successfully? I know the m190s are not too expensive, but still, a shame to blow them up. I have a 400 watt new (well, broken glass to be resealed) panel, that I'd like to try with 2 m190s, or any other older enphase inverters that work with the older ENV-120 Envoy.
Thanks! Dale
 
dalek said:
Anyone made this work successfully? I know the m190s are not too expensive, but still, a shame to blow them up. I have a 400 watt new (well, broken glass to be resealed) panel, that I'd like to try with 2 m190s.
Use one inverter per panel. Simple and foolproof.
 
Cool, I didn't realize that this thread was still going. Could it maybe be something simpler and more fundamental with the input design of the M215?

Like it has some sort of tie to ground on one of the inputs, or one leg floated so that when they are tied together it just blows up the input transistor. They die instantly, not sure if it is when they are first connected to live panels, or when the AC power is turned on. when the system.
 
I did connect up 2 and 3 m190s in parallel to one panel. I made Y adapters out of extra mc4 connectors from dead microinverters. So far I used inverters that were working but put out less power, 140 watt peak, vs the 200 that they should do. Nothing bad happened that I could tell, but I didn't exhaustively test them individually again afterward.

Together, they never made more than what just one would do. It appears that one does mppt tracking while the others make zero or some small amount. Sometimes the one that's doing the most switches to a different inverter.

I might try leaving the AC side disconnected on two, so the won't start up at all. That may add extra capacitance on the input side, and let 2 micros work together, out of 4 total. I might even try good working ones
 
Back
Top