Using solar micro inverters with batteries instead of panels

[jimbob232]

I have not purchased inductors yet so perhaps jmbob32 could answer that question (along with a link to his source for 100uH inductors .

Check this out! https://amzn.eu/d/3dHCsBG

These are what I'm using, 3 in series. Resistance is under 0.1 ohm but not zero so I think it can be considered as a LCR filter along with the input capacitors. They will take 15a but no idea what the saturation current is..

 

[fafrd]

Check this out! https://amzn.eu/d/3dHCsBG

These are what I'm using, 3 in series. Resistance is under 0.1 ohm but not zero so I think it can be considered as a LCR filter along with the input capacitors. They will take 15a but no idea what the saturation current is..

100uH 15A. So your earlier reference to ‘1mH’ was a typo, I assume.

Do you have any way to measure AC current at 60Hz (50Hz in your case)?

The Microinverter should be switching between 2 operating points @ 60/50Hz and seeing the two voltages involved is pretty much impossible without a scope, but you might be able to at least measure the 50/60Hz AC current which would represent the ‘average’ DC current on top of which the 25kHz ripple current would ride…

 

[Hedges]

It is an LC circuit. If the resonant frequency and under-damping allows, it may ring wildly. If a huge bulk capacitance, then my approximation of DC voltage applied to L alone might be OK.

If you use a transistor to pull an inductor up to +24VDC, current starts ramping up. When transistor turns off, inductor snaps to negative infinite volts. Or to the point of breaking down transistor. Parasitic capacitance (of inductor windings or circuit) may slow the rise (negative edge) time and limit voltage somewhat.

A reverse-biased diode across inductor while applying +24V will clip that negative infinity down to about -0.7V (not above +24V). Inductor will drive current through it for a while. That's the basic operation of a Buck regulator. A MOSFET in place of (or in parallel with) diode reduces dissipation and wasted power.

If what you have is a solenoid or relay you want to turn off fast, the freewheeling diode may keep current conducting too long. A resistor and diode in series is a snubber, and voltage is limited to I x R, can be sized to remain within transistor's voltage handling ability.

The orthogonal L R impedances might be useful for calculating current through transformer with secondary open, especially in saturation where inductive reactance is reduced. Conjugate impedance matching involves adding a "C" which resonates with "L" at operating frequency, so series LC presents zero ohm impedance. That leaves just "R", with lower impedance magnitude so higher current and power than what SQRT(R^2 + XL^2) let through.

At elevated frequencies, "skin effect" means resistive component of inductor is not same as DC resistance.

Bunch of EE stuff, some of which I occasionally apply. Mostly, I used Ohm's Law and Geometry.

"If we assume inductor current can drop to 0A before the end of OFF time"
That is "discontinuous conduction mode". If inductor is big enough for a given load, you get a DC current with sawtooth ripple riding on it. There are various SMPS modes of operation, producing various noise that may matter in sensitive applications.

I think you've got the first-order idea, which is about what I've got. I did go through the different switching regulator architectures in a college class. Someone who specializes in it would be more on top of the analysis. I've bounced around from microprocessor design to missiles to multi-chip modules to atomic mass spectrometers to vacuum applications involving electrons, and EMI/EMC. It keeps me confused and I have to dig up examples of how to do things.

 

[twrlp]

I had some time last night to play around with the M250's; here's what I've discovered:

• They will power up when connected directly to a battery, HOWEVER:
  • These models (and iq series) perform a DC resistance check at startup, and if too low they mark themselves offline (solid red LED) which persists across reboots. Per docs this can be rectified using an Envoy; I've ordered one from ebay. Scratch 2 m250's for now.
• Inserting the 4 Ohm power resistor I use for precharging my off-grid AIO permitted the m250 to boot up and begin producing power. m250’s MPPT settled down to the minimum 16V supported by the m250, leaving a ~10.6V drop through the resistor at 10V/4R=2.5A=25W burned as heat, 2.5A*16V=40W output to AC. Not particularly efficient
• Bypassing the resistor after MPPT stabilized didn't trigger the "DC Resistance Low" error. (yes risked 1 more for SCIENCE), but no power was produced - my ammeter suggests m250’s MPPT was repeatedly hunting and sleeping.
• I tracked down a couple of 0.1 Ohm power resistors in my junk drawer.
  • Adding one to the DC input generated full power (9.5A, 248W AC, 10W heat loss on the 0.1 resistor) when 4W was removed
  • So I tried putting two in parallel (0.05 Ohm) which also worked, and as a bonus only generated 2.5W each.
  • Still probably need a heat sink for these and for the 4 Ohm bootstrap resistor.

My current thinking is that I'll use an 80A SSR to switch DC to the inverter(s), in parallel with a relay and in-line 4 Ohm resistor. The ardunio/esp32/rpi (haven't decided) would sequence the relay for bootup, then the SSR, then turn off the now-bypassed relay.

Before leaving this unattended I want to put snap-disc thermostats on at least the resistors/heat sink, maybe on the inverters as well. Those could either trigger the shunt trip relay in my main DC breaker (normally open), or could sit inline with the DC input to the inverter (normally closed).

After I have the full-power version stable, I'll add in another inverter via one of the cheap PWM units, maybe one with filtering and a third without.

 

[fafrd]

I had some time last night to play around with the M250's; here's what I've discovered:

• They will power up when connected directly to a battery, HOWEVER:
  • These models (and iq series) perform a DC resistance check at startup, and if too low they mark themselves offline (solid red LED) which persists across reboots. Per docs this can be rectified using an Envoy; I've ordered one from ebay. Scratch 2 m250's for now.

Interesting - Thanks for the discovery. I don’t know whether my NEP Microinverters perform a DC resistance check or not but I was thinking about connecting 50mOhm on Series and after your discovery, I’ll do at least that and use an old spare unit to boot the first time I try.

• Inserting the 4 Ohm power resistor I use for precharging my inverter permitted it to boot up and begin producing power. MPPT pulls voltage all the way down to the minimum 16V supported by the m250, leaving a ~10V drop through the resistor at 10V/4R=2.5A=25W burned as heat, 2.5A*16V=40W output to AC. Not particularly efficient

Yes, 26VDC from the battery - 16VDC Vmppt_min means a maximum of 10VDC/4 Ohms = 2.5A into the inverter, so 2.5A x 16VDC = 40W of maximum output power.

• Bypassing the resistor after MPPT stabilized didn't trigger the "DC Resistance Low" error. (yes risked 1 more for SCIENCE), but no power was produced - my ammeter suggests MPPT was repeatedly hunting and sleeping.

I wonder whether we can calculate the level of resistance it needs to lock-on.

Ref = Isc / (Voc - Vmp) = 1.093 Ohm for my 335W panels and 1.69 Ohms for my 380W panels, so I’d guess the resistance check can’t be looking for resistance of much more than an Ohm…

• I tracked down a couple of 0.1 Ohm power resistors in my junk drawer.
  • Adding one to the DC input generated full power (9.5A, 248W AC, 10W heat loss on the resistor) when 4W was removed
  • So I tried putting two in parallel (0.05 Ohm) which also worked, and as a bonus only generates 2.5W on each.
  • Still probably need a heat sink for this and for the 4 Ohm bootstrap resistor.

I decided from some crude modeling that I would put 50-100mOhm power resistors between my battery and my Microinverter to make it easier for the Microinverter MPPT to find a pair of suitable operating points and your experimental results perfectly confirm that theory:

My current thinking is that I'll use an 80A SSR to switch DC to the inverter(s), in parallel with a relay and in-line 4 Ohm resistor.
The ardunio/esp32/rpi (haven't decided) would sequence the relay for bootup, then the SSR, then turn off the bypassed relay.

What is the contact resistance of your SSR? It alone might be able to replace your 50mOhm power resistor...

I want to thank you again for sharing this experiment - I've got some old inverters I might be prepared to sacrifice but this gives me an important heads-up for what to plan for. There is little difference between a 1-Ohm resistor and a 4-Ohm resistor for passing DC-resistance check in terms of cost or complexity, but since most solar panels seem to represent an effective resistance of ~1 Ohm past Vmp, I'm thinking that might be a safer resistance to use in case a specific microinverter also build in a 'resistance too high' fault mode.

But some way to pass the resistance check seems like an important safety check to avoid locking out microinverters.

You can of course progran your controller to control both switches, but an easier alternative would be to use a current switch on the AC output of the microinverter so that as soon as 40W is being generated (or 160W if you can pass the resistance check with a 1.0 Ohm resistor ), the 50 mOhm resistor can be switched in. There is also no harm in leaving both resistors connected, so a simpler solution is to have a 1/4 Ohm resistance inline in parallel with a 50 mOhm switched resistance (or just an SSR in parallel if it represents enough resistance alone).

1 Ohm in parallel with 0.05 Ohms = 0.04762 Ohms so over 95% of current / power will be flowing through the SSR + power resistor branch and less than 5% of the current / power will be flowing through the unswitched 1 Ohm resistor (or even less if you stick with 4 Ohms).

So one AC relay/SSR to turn the microinverter ON and OFF and a second SSR controlled by a current switch once it detects that the microinverter has started generating power would be the Luddite's solution to this problem that some of the most active contributors to this thread such as kundip would prefer...

Before leaving this unattended I want to put snap-disc thermostats on at least the resistors/heat sink, maybe on the inverters as well. Those could either trigger the shunt relay in my main DC breaker (normally open), or could sit inline with the DC input to the inverter (normally closed).

After I have the full-power version stable, I'll add in another inverter via one of the cheap PWM units, maybe one with filtering and a third without.

View attachment 146647View attachment 146648

Nice. Are planning to test any cheap DCDC boosters as well?

 

[fafrd]

It is an LC circuit. If the resonant frequency and under-damping allows, it may ring wildly. If a huge bulk capacitance, then my approximation of DC voltage applied to L alone might be OK.

In this case we are talking about an inductance of 0.3-1.0mH and a capacitance of 10-20mF - would a ratio of 10-67 qualify as 'huge'???

If you use a transistor to pull an inductor up to +24VDC, current starts ramping up. When transistor turns off, inductor snaps to negative infinite volts. Or to the point of breaking down transistor. Parasitic capacitance (of inductor windings or circuit) may slow the rise (negative edge) time and limit voltage somewhat.

This is the part I’m still struggling with. Current is flowing through inductor when transistor turns off. That same current continues to flow when transistor turns off, meaning voltage between output and input continues to increase.

If there is no reservoir of charge at the inductor input, I see how that could translate to negative infinite volts, but if there is any capacitance at the inductor input, voltage decrease will be much slower.

And if capacitance at inductor input equals capacitance at inductor output (what we’ve been calling the Input capacitance), the rate of voltage decrease at the inductor input will equal rate of continued voltage increase at the inductor output (both also dropping at a rate driven by average current output.

Of course, that added capacitance at the PWM output will bring back all of the high-harmonic heat that jimbob32 was originally motivated to avoid when he added the inductor in the first place, but since that added capacitor does not need to be the same size as the inout capacitance, it still offers an option to limit negative voltage swing during the OFF phase, right?

A reverse-biased diode across inductor while applying +24V will clip that negative infinity down to about -0.7V (not above +24V). Inductor will drive current through it for a while. That's the basic operation of a Buck regulator. A MOSFET in place of (or in parallel with) diode reduces dissipation and wasted power.

I understand that but feel like there should be a way to limit negative voltage swing so that it is at most done multiplier of positive voltage swing. From a starting voltage of 25V, the same computation you performed earlier means a ~100uH inductor will have a maximum of 8A flowing within the 20uS ON period. That should mean the maximum current that can be driving negative voltage swing at the beginning of OFF period is 8A + 10A = 18A.

18A for a full 20uS is 360uCoulomb which would drive a negative swing of only -36mV to 23.964V on a 10,000uF capacitor at inductor input.

If that added capacitance were dropped to only 100uF, wouldn’t that translate to worst-case negative voltage swing by the end of the OFF period to less than -3.6V to inductor input voltage that can’t be lower than 20.4V?

If what you have is a solenoid or relay you want to turn off fast, the freewheeling diode may keep current conducting too long. A resistor and diode in series is a snubber, and voltage is limited to I x R, can be sized to remain within transistor's voltage handling ability.

I don’t know protections a PWM may have to prevent negative voltage on the output, but I’d rather find a solution that keeps all voltages between 0-57.6VDC and does not need to rely on diodes or reverse-voltage protection.

The key epiphany I learned from you is that inductance translates to a maximum current change within an interval of time.

That means an inductor can limit current increase or decrease from a square wave and can maintain a known current swing around an average DC current.

I’m not going to bother with Spice yet, but I might try working op a piece wise-linear spreadsheet-based model such as the one I used to convince myself that 50-100mOhm between a battery and a Microinverter was needed to make the MPPT function more easily.

I’m going to try:

-Switched idealized current source (floating output when OFF)

-into smaller output capacitance

-into ~100uH inductor into 20,000uF input capacitance

-into idealized voltage source (MPPT).

I’ll report back when I have results to share.

The orthogonal L R impedances might be useful for calculating current through transformer with secondary open, especially in saturation where inductive reactance is reduced. Conjugate impedance matching involves adding a "C" which resonates with "L" at operating frequency, so series LC presents zero ohm impedance. That leaves just "R", with lower impedance magnitude so higher current and power than what SQRT(R^2 + XL^2) let through.

At elevated frequencies, "skin effect" means resistive component of inductor is not same as DC resistance.

Bunch of EE stuff, some of which I occasionally apply. Mostly, I used Ohm's Law and Geometry.

I appreciate all if this EE stuff but am hoping I can skate around it.

One question I do have for you, though:

If we’ve determined that a ~100uH inductor will protect a capacitor from seeing more than 8A of ripple current over a period of 20uS, doesn’t that mean that the same inductor would protect the same capacitance from the same >8A ripple current from a switched current source such as one of the cheap DCDC boosters (at least for current ripple at 25kHz or higher?

The inductor needs to be sized to handle the full DC current (8.3-10A) but since there is no ‘OFF’ period with a current source, when the Microinverter moves from Vmp to deliver target power @ programmed current (8.3A for 250W @ 30VDC) to much lower power and current at some voltage beyond prigramned maximum voltage of 30V, the inductor will act to limit both the rate f current increase as well as the rate of current decrease.

But since the timescale or inverters switching is 16.7ms (60Hz) versus the 40uS (25kHz) switching rate we’ve been talking about for PWM controllers, it should be a non-issue, shouldn’t it.

Can you see any reason an inductor should not be added inline between a DCDC booster and a Microinverter to protect the input caps of the Microinverter from high-frequency ripple current that might emit from the switched DCDC booster?

"If we assume inductor current can drop to 0A before the end of OFF time"
That is "discontinuous conduction mode". If inductor is big enough for a given load, you get a DC current with sawtooth ripple riding on it. There are various SMPS modes of operation, producing various noise that may matter in sensitive applications.

Noise does not matter in the case of a Microinverter being powered by a PWM controller, merely ripple current, so I think it is precisely that ‘DC current with sawtooth ripple riding on it’ that we are looking for.

Hopefully with your help I’m getting closer…

I think you've got the first-order idea, which is about what I've got. I did go through the different switching regulator architectures in a college class. Someone who specializes in it would be more on top of the analysis. I've bounced around from microprocessor design to missiles to multi-chip modules to atomic mass spectrometers to vacuum applications involving electrons, and EMI/EMC. It keeps me confused and I have to dig up examples of how to do things.

Your understanding goes leagues beyond mine and I appreciate your efforts to educate me. Between the evolving theory and the incoming experimental results, I’m increasingly confident this nut is on it’s way to being cracked…

 

[fafrd]

Check this out! https://amzn.eu/d/3dHCsBG

These are what I'm using, 3 in series. Resistance is under 0.1 ohm but not zero so I think it can be considered as a LCR filter along with the input capacitors. They will take 15a but no idea what the saturation current is..

This is a 100uH inductor supposedly rated for 20A: https://www.amazon.com/gp/aw/d/B07F...&pd_rd_r=d1d32954-df3f-47f6-9cc4-f55a83238570

It’s using 0.8mm diameter / 20AWG wire which from everything I can find means maximum current is 11A (for chassis wiring).

So are inductors under DC current subject to different ampacity limits than what would apply to those same size wires carrying DC current in free air???

 

[fafrd]

This is a 100uH inductor supposedly rated for 20A: https://www.amazon.com/gp/aw/d/B07FK5N39T/?_encoding=UTF8&pd_rd_plhdr=t&aaxitk=2a305a3117d8ef49464dd7f47fc94077&hsa_cr_id=9983892060501&qid=1682550876&sr=1-2-3c6b3b04-89d4-46ee-857c-1e2f0de6a70e&ref_=sbx_be_s_sparkle_scm_asin_1_title&pd_rd_w=eyNsI&content-id=amzn1.sym.75872814-4917-4a48-8af3-2760678d170e:amzn1.sym.75872814-4917-4a48-8af3-2760678d170e&pf_rd_p=75872814-4917-4a48-8af3-2760678d170e&pf_rd_r=75VMH2R496Y7NCMVZ4G9&pd_rd_wg=285RR&pd_rd_r=d1d32954-df3f-47f6-9cc4-f55a83238570

It’s using 0.8mm diameter / 20AWG wire which from everything I can find means maximum current is 11A (for chassis wiring).

So are inductors under DC current subject to different ampacity limits than what would apply to those same size wires carrying DC current in free air???

These 1.5mH inductors are rated for 20A but is using 1.8mm diameter / 13AWG wire with a chassis wiring ampacity of 35A: https://m.aliexpress.us/item/325680...irect=y&gatewayAdapt=glo2usa&_randl_shipto=US

Wouldn’t this be a better-suited inductor for the type of 20A PWM (+/-10A ripple on top of 10A DC) smoothing we are trying to achieve?

 

[Hedges]

In this case we are talking about an inductance of 0.3-1.0mH and a capacitance of 10-20mF - would a ratio of 10-67 qualify as 'huge'???

1/(2 x pi x sqrt(0.001H x 0.01F)) = 50 Hz resonant frequency.

Way below switching speed of your speed control. But very close to ripple imposed by microinverter, so likely to build resonance unless damped. Have to evaluate impact of parasitic loss and source/load impedance.

This is the part I’m still struggling with. Current is flowing through inductor when transistor turns off. That same current continues to flow when transistor turns off, meaning voltage between output and input continues to increase.

Output voltage continues to rise.
Input was being pulled up to +24V (while it was pulling down, opposing the applied voltage).
When you stop pulling up, it snaps negative, gets clamped by diode if one is included.
Positive voltage increased current, negative voltage decreases it.

If there is no reservoir of charge at the inductor input, I see how that could translate to negative infinite volts, but if there is any capacitance at the inductor input, voltage decrease will be much slower.

Yes, interwinding capacitance and transistor drain capacitance reduce the voltage from negative infinity. But commonly enough inductive kick to kill devices.
We've blown a few bench supplies driving our magnetic components with DC.

And if capacitance at inductor input equals capacitance at inductor output (what we’ve been calling the Input capacitance), the rate of voltage decrease at the inductor input will equal rate of continued voltage increase at the inductor output (both also dropping at a rate driven by average current output.

Two identical capacitors in series is half the capacitance.

Of course, that added capacitance at the PWM output will bring back all of the high-harmonic heat that jimbob32 was originally motivated to avoid when he added the inductor in the first place, but since that added capacitor does not need to be the same size as the inout capacitance, it still offers an option to limit negative voltage swing during the OFF phase, right?

Would avoid adding capacitance on PWM output. In fact, often circuits (op-amp) driving large capacitive load do better with a series resistance. That helps their internal and external feedback achieve stability so they don't oscillate.

Snubbers are likely to be R, C, Diode in some combination. SMPS use FET for higher efficiency.

I understand that but feel like there should be a way to limit negative voltage swing so that it is at most done multiplier of positive voltage swing. From a starting voltage of 25V, the same computation you performed earlier means a ~100uH inductor will have a maximum of 8A flowing within the 20uS ON period. That should mean the maximum current that can be driving negative voltage swing at the beginning of OFF period is 8A + 10A = 18A.

Inductor can be pumped up to a much higher steady-state current than magnitude of ripple (depending on duty ratio and load). Smoother output but slower transient response.

18A for a full 20uS is 360uCoulomb which would drive a negative swing of only -36mV to 23.964V on a 10,000uF capacitor at inductor input.

If that added capacitance were dropped to only 100uF, wouldn’t that translate to worst-case negative voltage swing by the end of the OFF period to less than -3.6V to inductor input voltage that can’t be lower than 20.4V?

Probably. Capacitors reduce the voltage and the rate of rise. Ignition systems used capacitor across points so points could fully open without arcing, but still high enough di/dt to make high voltage on secondary.

I don’t know protections a PWM may have to prevent negative voltage on the output, but I’d rather find a solution that keeps all voltages between 0-57.6VDC and does not need to rely on diodes or reverse-voltage protection.

I would add external protection diode, sized according to inductor and driving conditions. I'm not sure a motor it was meant to drive would make as big a spike, so not sure what protection built in.

The key epiphany I learned from you is that inductance translates to a maximum current change within an interval of time.

For an applied voltage.
That's pretty much the definition of inductance.
(Start thinking about cores, saturation, hysteresis and things get messier but more real.)

That means an inductor can limit current increase or decrease from a square wave and can maintain a known current swing around an average DC current.

I’m not going to bother with Spice yet, but I might try working op a piece wise-linear spreadsheet-based model such as the one I used to convince myself that 50-100mOhm between a battery and a Microinverter was needed to make the MPPT function more easily.

Some math in Excel or Matlab is useful.
Programs that understand complex numbers are useful.

At half a dozen components, coding a passive filter circuit in Matlab got too messy for me.

Install LTSpice. Try an example filter circuit and an example SMPS IC circuit from their library.
Much easier to build off a copy of a running simulation.

I’m going to try:

-Switched idealized current source (floating output when OFF)

-into smaller output capacitance

-into ~100uH inductor into 20,000uF input capacitance

Look at its frequency response with and without resistance added.
Then try resistance over/under/equal to critically damped, look at time domain response to step function.

-into idealized voltage source (MPPT).

I’ll report back when I have results to share.

I appreciate all if this EE stuff but am hoping I can skate around it.

One question I do have for you, though:

If we’ve determined that a ~100uH inductor will protect a capacitor from seeing more than 8A of ripple current over a period of 20uS, doesn’t that mean that the same inductor would protect the same capacitance from the same >8A ripple current from a switched current source such as one of the cheap DCDC boosters (at least for current ripple at 25kHz or higher?

I think so. Within operating current range of inductor.
Unless you end up with a high Q tank circuit that amplifies voltage (that's how compact fluorescent starters work.)

The inductor needs to be sized to handle the full DC current (8.3-10A) but since there is no ‘OFF’ period with a current source, when the Microinverter moves from Vmp to deliver target power @ programmed current (8.3A for 250W @ 30VDC) to much lower power and current at some voltage beyond prigramned maximum voltage of 30V, the inductor will act to limit both the rate f current increase as well as the rate of current decrease.

But since the timescale or inverters switching is 16.7ms (60Hz) versus the 40uS (25kHz) switching rate we’ve been talking about for PWM controllers, it should be a non-issue, shouldn’t it.

A switching power supply control IC would adjust its PWM to regulate voltage as needed.
Your speed controller won't, might end up current limited (hope it has that feature) if constant duty ratio switching causes inductor circuit to run away. But no guarantees on output voltage. Except, I think for buck architecture it will stop at input voltage. Boost architecture goes higher than input (of course) and there may be no limit except something failing.

Can you see any reason an inductor should not be added inline between a DCDC booster and a Microinverter to protect the input caps of the Microinverter from high-frequency ripple current that might emit from the switched DCDC booster?

Noise does not matter in the case of a Microinverter being powered by a PWM controller, merely ripple current, so I think it is precisely that ‘DC current with sawtooth ripple riding on it’ that we are looking for.

Hopefully with your help I’m getting closer…

Your understanding goes leagues beyond mine and I appreciate your efforts to educate me. Between the evolving theory and the incoming experimental results, I’m increasingly confident this nut is on it’s way to being cracked…

Decades will do that to you. But mostly what I've learned being around it. I've dabbled in SMPS and seen a few things blow up, but haven't been full time on them. One thing I've seen is quality brands perform better. I tried an off-brand SMPS chip and while reference design worked, modified parameters for a different output never did. Better to stick with TI, LTC, Analog etc. and use design tools plus LTSpice. It was still an effort to model one unstable op-amp circuit. Output impedance varied with frequency according to the data sheet but not in model, so I had to add it externally so I could then simulate the fix.

Then I'm off to something different, so never an expert in the area.

 

[Hedges]

So are inductors under DC current subject to different ampacity limits than what would apply to those same size wires carrying DC current in free air???

Typically rated at the current where inductance drop to 80% of what it was at low current.
The magnetic core gets toward saturation and becomes less effective.
Heat is an issue too, but likely heating of the core affecting inductance is limiting factor.

 

[Hedges]

I used capacitor in series, inductor in parallel for impedance matching.
At elevated temperature performance of the circuit was poor.
When I touched the inductor and burned my finger I understood why.
After that I only did tuning (impedance matching) with capacitors, no inductors. I modifed the RF transformer I was driving so I could get away with that.
11 MHz, 700 Vpp into a capacitive load (part of an instrument.)

 

[fafrd]

1/(2 x pi x sqrt(0.001H x 0.01F)) = 50 Hz resonant frequency.

Way below switching speed of your speed control. But very close to ripple imposed by microinverter, so likely to build resonance unless damped. Have to evaluate impact of parasitic loss and source/load impedance.

Too much more below for me to comment on now, but I did want to ask a follow-on question about this:

Does this mean that with a 100uH inductor rather than 1mH inductor we’d get a resonant frequency of 160Hz?

And if so, is that far enough away from 60Hz to avoid concerns associated with Microinverter ripple?

The 25kHz frequency is so fast I don’t believe it will be a factor other than the amount of 25kHz current ripple reaching the 10,000uF input capacitance of the Microinverter.

So any thoughts on what minimum resonant frequency avoids potential issues with 60Hz Microinverter ripple would be appreciated.

And also, in terms of damping the resonance, if I’m ready to add one or 2 50mOhm power resistors into the circuit to increase damping, where do you think they would be most effective?

Between the inductor output and the input capacitance?

Between the PWM output and the inductor input?

In parallel with the inductor?

Some combination of the above?

 

[twrlp]

I decided from some crude modeling that I would put 50-100mOhm power resistors between my battery and my Microinverter to make it easier for the Microinverter MPPT to find a pair of suitable operating points and your experimental results perfectly confirm that theory:

Thanks for sharing your calculations - it's great to see model and reality line up so closely.!

What is the contact resistance of your SSR? It alone might be able to replace your 50mOhm power resistor...

Far too low, alas - I'm using the Thornwave 24v 80A ssr which consists of a driver and 6 parallel NTMFS5C604NL MOSFETs (60 V, 1.2 mW, 287 A). Measured voltage drop of 0.2mv at 10A == 0.2mOhm aligns pretty well with those chips' individual Rds(on) of 1.2mOhm.

So one AC relay/SSR to turn the microinverter ON and OFF and a second SSR controlled by a current switch once it detects that the microinverter has started generating power would be the Luddite's solution to this problem that some of the most active contributors to this thread such as kundip would prefer...

I'm a programmer by nature so I think in terms of state machines and such, but you're right hardware should be considered as well.

My next steps will be to clean up the installation, add a heat sink and thermal cutoff, and then let the inverter run for a week to confirm my assumption that DC resistance check only happens at startup. I'll also test whether the inverters are capable of extended operation at peak amps in normal solar panel usage that only happens for a few hours each day.

 

[Hedges]

160 Hz sounds correct.
Away from the fundamental, but close to 180 Hz 3rd harmonic (which should be much lower amplitude than 60 Hz, much less to excite resonance.

To decide how much series or parallel resonance, you have to try an equation for Q (Q is voltage peak above stimulus signal, if I remember correctly.) Or SPICE. The driving and load impedances will also be a factor.

 

[fafrd]

Thanks for sharing your calculations - it's great to see model and reality line up so closely.!

Far too low, alas - I'm using the Thornwave 24v 80A ssr which consists of a driver and 6 parallel NTMFS5C604NL MOSFETs (60 V, 1.2 mW, 287 A). Measured voltage drop of 0.2mv at 10A == 0.2mOhm aligns pretty well with those chips' individual Rds(on) of 1.2mOhm.

That’s a nice SSR! And yeah, you’ll need another ~50mOhms in series with that SSR.

But you don’t need a second switch on the 4 Ohm resistor - if you leave it permanently connected and only turn on the Microinverter begins outputting power, you’ll have ~0.07 Ohms in parallel with 4 ohms, meaning 68.8 mOhm parallel resistance. So 98.3% of the current will be running through the SSR + 0.05 Ohm resistor and only 4.3W will be flowing through the 4 Ohm resistor.

There is a tiny bit of added heat loss but it is not worth worrying about - assuming 10A total we’re talking 172mA through the 4 Ohm resistor for less than 120mW of wasted power…

By the way, I looked more closely at passing the resistance check with a 1 ohm resistor and you are lucky it passes with 4 Ohms for your Microinverters. 100W power resistors are easy to find but 200W power resistors, not so much and you’re right at the limit of 100W with 10A through a 1 Ohm resistor (and over that limit if your battery charges up past 26.0VDC)…

My Microinverters have a minimum MPPT voltage of 22V, so I only have 4VDC to worry about at 26V on the battery but as much as 6.8V if the battery is fully-charged to 28.8V. Even with a worst-case voltage of 6.8W, I’d only be looking at 46W through a 1 Ohm resistor…

I'm a programmer by nature so I think in terms of state machines and such, but you're right hardware should be considered as well.

A 1-4 Ohm permanent resistor with a typical SSR in parallel which gets turned on a current switch sensing AC output current is a nice simple solution to getting past the Resistance Check…

My next steps will be to clean up the installation, add a heat sink and thermal cutoff, and then let the inverter run for a week to confirm my assumption that DC resistance check only happens at startup.

Heatsink with thermal cutoff being added where, on the 0.5 Ohm power resistor?

The current switch solution has the added benefit that if there is a grid outage shutting down the system, it will automatically come back online once he grid is back up and the Resistance Check has passed…

I'll also test whether the inverters are capable of extended operation at peak amps in normal solar panel usage that only happens for a few hours each day.

Do you plan to use a fan or any other active cooling technique or passive only?

 

[jimbob232]

I had some time last night to play around with the M250's; here's what I've discovered:

• They will power up when connected directly to a battery, HOWEVER:
  • These models (and iq series) perform a DC resistance check at startup, and if too low they mark themselves offline (solid red LED) which persists across reboots. Per docs this can be rectified using an Envoy; I've ordered one from ebay. Scratch 2 m250's for now.
• Inserting the 4 Ohm power resistor I use for precharging my off-grid AIO permitted the m250 to boot up and begin producing power. m250’s MPPT settled down to the minimum 16V supported by the m250, leaving a ~10.6V drop through the resistor at 10V/4R=2.5A=25W burned as heat, 2.5A*16V=40W output to AC. Not particularly efficient
• Bypassing the resistor after MPPT stabilized didn't trigger the "DC Resistance Low" error. (yes risked 1 more for SCIENCE), but no power was produced - my ammeter suggests m250’s MPPT was repeatedly hunting and sleeping.
• I tracked down a couple of 0.1 Ohm power resistors in my junk drawer.
  • Adding one to the DC input generated full power (9.5A, 248W AC, 10W heat loss on the 0.1 resistor) when 4W was removed
  • So I tried putting two in parallel (0.05 Ohm) which also worked, and as a bonus only generated 2.5W each.
  • Still probably need a heat sink for these and for the 4 Ohm bootstrap resistor.

My current thinking is that I'll use an 80A SSR to switch DC to the inverter(s), in parallel with a relay and in-line 4 Ohm resistor. The ardunio/esp32/rpi (haven't decided) would sequence the relay for bootup, then the SSR, then turn off the now-bypassed relay.

Thanks AGT, this is really useful information.
I'm now not sure why i haven't bricked my inverter (yet) on the resistance check - the first test i did was direct powered (like Kundip's setup) although i was using a desktop psu not battery. I don't have the AC connected when i initially apply the DC but i don't think that's the reason.

With the PWM unit i always start at zero duty and wind it up a bit until the inverter goes through it's boot and test cycle then wind it up after that (although only to 4.5a so far due to PSU limits) and that's been fine as well. You do see the inverter pulsing the current several times before it goes into production mode, i guess this is the resistance check?

Is it maybe because the PSU and / or PWM setup have a higher internal resistance than a battery would? I am now concerned that if i use a battery with the PWM unit it may well fail the resistance test....

As per farfd's graphs it seems like a series resistance would be useful to settle the MPPT down, i don't have any to hand to try but i may well get hold of some.. i'll be very interested to see what you find.

 

[jimbob232]

I can't recall whether dynamic control of battery-driven power is a priority to you or not, but as you add inductors it occurs to that you might want to get ahold of one of these: https://www.amazon.com/AITRIP-Conve...8TWKK5Z9/ref=psdc_10967761_t3_B0756HQTRM?th=1

I picked up 3 of these from AliExpress for less than half the price of each these 2 from Amazon, so once you start adding up the cost of inductors, PWM, etc... this quickly becomes a cost-effective fully-integrated solution.

8A max without added cooling = 200W max input power from a 25VDC battery translating to 160-180W max output assuming efficiency of 80-90%. By adding a fan, you should be able deliver 250W (or you can use 2 units in parallel).

You will need to boost by at least 2 volts so if your battery varies between 28.8V when fully-charged to 25V when fully discharged, you'll need to configure output voltage for 29.0V and then set the output current to 5.5-6.2A (depending on measured efficiency).

Current ripple and stability of output are the two criticisms that have been levelled against these cheapo DCDC boosters but since these are precisely the areas you are looking at and characterizing, could make sense to include one of these in the mix as a reference...

I did look at this sort of thing but i need to be able to control the output current - i want to be able to control the power output of the inverter so that i can supply my base load but not feed any back to grid (i will use an arduino for the control so 0-5v input on the pwm board is ideal, hence trying this route) - unless you know of any of these that are externally adjustable?
What i want to end up with is inverter output controlled by 0-5v arduino input, whatever the easiest and most robust method is!

 

[kevin93]

"It’s using 0.8mm diameter / 20AWG wire which from everything I can find means maximum current is 11A (for chassis wiring)."

It has two (maybe 3) conductors in parallel as can be seen form the image on the Amazon web site.

kw

 

[fafrd]

Thanks AGT, this is really useful information.
I'm now not sure why i haven't bricked my inverter (yet) on the resistance check - the first test i did was direct powered (like Kundip's setup) although i was using a desktop psu not battery.

Your desktop PSU has a current limit. Do you know what it is? 30?

So both you and kundip successfully powered up Enphase inverters with a PSU (DCDC-booster in kundip’s case) which strongly suggest that a PSU / DCDC converter provides a sufficient current limit to pass the Resistor Test.

ADT bricked without 4 Ohms inline when powered-up directly from battery while kundip successfully powered up from battery through a relay (which is a resistance of possibly as much as 0.25Ohms).

Especially since you want to control output, there is no need for you to power direct from a battery and unless you want to risk bricking an inverter in the interest of science, I’d suggest you stick to your PSU.

I don't have the AC connected when i initially apply the DC but i don't think that's the reason.

So you switched AC when powered by your PSU while kundip switched DC power with AC already connected. I don’t know whether kundip switched DC from his DCDC booster to his M250 or switched DC from his battery to his DCDC booster, so perhaps he could clarify.

Either way, it sounds as though AGT might have been switching DC through his ultra-low impedance SCCs, meaning he was not switching AC like you were.

So it may be that having DC switched off when AC is switched on is another way to pass the resistance test. Perhaps AGT could disconnect DC, cycle AC, after a nice long interval like 5 or even 15 minutes, try connecting DC through an 0.1 ohm resistor to see if that passes.

With the PWM unit i always start at zero duty and wind it up a bit until the inverter goes through it's boot and test cycle then wind it up after that (although only to 4.5a so far due to PSU limits) and that's been fine as well. You do see the inverter pulsing the current several times before it goes into production mode, i guess this is the resistance check?

We’re calling it a Resistance Check because AGT was able to successfully avoid the fault condition by switching in DC power through a 4 ohm resistor but not through direct connection through a ultra-low impedance SSR.

Reading his post again, it does not seem he tried powering through an 0.1 ohm resistor - only 0 ohm or 4 ohms.

So it may be that an inline resistance of as little as 0.1 ohms or even 0.05 ohms is all that is needed to pass the Resistance Check.

I was concerned that the infinite slope of a battery would make it difficult for an MPPT to lick in to the two operating points as easily as it does on the less steep slope of a solar panel past Vmp, and that’s the reason I looked into what would happen if a small power resistor was added.

Once it has identified Voc, The MPPT steps to a lower voltage and expects to see current has decreased by a bit. If current changes by too much, it tries the midpoint. If current has still changed by too much, it tries the quarter point, etc…

If it can’t find a point where current is below what it is looking for, it will give up.

So AGT proved 0.0002 Ohms (plus whatever wiring resistance he had) is too low and kundip proved whatever the resistance of his DC relay (probably 0.1-0.2 ohms) plus whatever his wiring resistance is was enough.

So this Resistance Check could just be the MPPT confirming that it can lock on to the necessary control points and as little as 0.1 or 0.05 ohms is enough to do that…

Is it maybe because the PSU and / or PWM setup have a higher internal resistance than a battery would? I am now concerned that if i use a battery with the PWM unit it may well fail the resistance test....

I doubt it. Any current-limited supply will almost certainly allow the Microinverter MPPT to lick on. As I’m going back and reviewing all of the details, I’m suspecting the ‘Resistance Check’ is just the MPPT confirming that it can find the two operating points it needs. On the end it amounts to the same thing: no ability to lock-on = resistance from an infinite current source = too low.

But since we know these inverters can lock on and invert with as little as 0.05 ohms, that may be all that is needed to pass the Resistance / Lock-on test…

From current-limited sources such as DCDC boosters or PWM controllers, there should be no issue in any case (which is why I was interested to go in that direction).

As per farfd's graphs it seems like a series resistance would be useful to settle the MPPT down, i don't have any to hand to try but i may well get hold of some.. i'll be very interested to see what you find.

I’d suggest you get ahold of some 0.05 ohm 100W power resistors: https://www.amazon.com/gp/aw/d/B07H35JR86/ref=ox_sc_saved_image_2?smid=A1THAZDOWP300U&psc=1

You can always start with 2 in series to test 0.1 ohm and then drop down to a single one if that works.

 

[fafrd]

I did look at this sort of thing but i need to be able to control the output current - i want to be able to control the power output of the inverter so that i can supply my base load but not feed any back to grid

Welcome to the club!

I’ve been doing battery-powered load offset for 2 years now using a pair of 1kW GTIL inverters. For various reasons, I want to make a DIY solution using several spare Microinverters controlled by a PLC.

(i will use an arduino for the control so 0-5v input on the pwm board is ideal, hence trying this route) - unless you know of any of these that are externally adjustable?
What i want to end up with is inverter output controlled by 0-5v arduino input, whatever the easiest and most robust method is!

I’m looking at a hybrid DAC approach which will include:

250W Microinverter always direct-powered and switched at AC (at most twice per day)

~250W, ~125W, ~62.5W, ~31.25W switched through DC SSRs to the 2 inputs of an always-on 500W ‘DAC’ dual-Microinverters (dynamically every ~25ms as needed).

I’ve got all the bits and pieces for that 750W 4-1/2 bit solution and if the DCDC boosters can be reliably preset and addition of a simple filter circuit can prevent / limit high-frequency ripple reaching the Microinverter, that’s my brute-force plan.

The PWMs are appealing because they make it so easy to control through external resistors.

I would not attempt to control through an analog signal as you are hoping to do but would use a bank of relays to switch in different control resistors to get my power DAC.

The same thing might be possible with the cheapo pot-controlled DCDC boosters but the wired potentiometer of the PWMs makes it so much easier and PWM may be more stable than the DCDC boosters.

If you are not too far down your analog control path yet, I will tell you that even 32W-level control is overkill, so analog control to offset load is really not necessary.

Depending on the peak power you are trying to offset, a few of you m250s under AC relay control plus another one under 4-bit digital control is more than you need…

 

[twrlp]

Heatsink with thermal cutoff being added where, on the 0.5 Ohm power resistor?

Common heat sink for the 3 power resistors, then just natural convection I hope. Cutoff is relevant should the SSR die or the switch/controller fail to enable it- either way the 4W resistor might be on at 25W for an extended period. (After 3 hours @ 9.5A, the two 100mOhm resistors hit 75C/170F in free air inside a cool garage; the m250, flush against plywood, was at 35C/95F. )

Reading his post again, it does not seem he tried powering through an 0.1 ohm resistor - only 0 ohm or 4 ohms.

So it may be that an inline resistance of as little as 0.1 ohms or even 0.05 ohms is all that is needed to pass the Resistance Check.

That's correct, I haven't yet tried 0.1 at startup - will wait until I can unbrick them.

A bit more info: the North American m250 manual states the following regarding the solid red LED I encountered::

 Solid Red: For all M250 models and for M215-60-2LL-S22-IG, this indicates a “DC
Resistance Low – Power Off” fault. Troubleshoot as described in the following sections

DC Resistance Low – Power Off Condition
For all M250 models and for M215-60-2LL-S22-IG, a solid red status LED when DC power has been
cycled indicates the microinverter has detected a DC Resistance Low – Power Off event. The LED will
remain red and the fault will continue to be reported by the Envoy until the error has been cleared.

An insulation resistance (IR) sensor in the microinverter measures the resistance between the positive
and negative PV inputs to ground. If either resistance drops below a threshold, the microinverter stops
power production and raises this condition. This may indicate defective module insulation, defective
wiring or connectors, moisture ingress, or similar problem. Although the cause may be temporary, this
microinverter condition persists until the sensor is manually reset.

An Envoy is required to clear this condition. The condition usually clears with operator intervention
unless conditions causing the event have not been remedied or if the failure is permanent.

If a microinverter registers a “DC Resistance Low - Power Off” condition, you can attempt to clear this
condition. If the condition does not clear after you perform the following procedure, contact Enphase
Energy customer support customer support at enphase.com/en-us/support/contact.

This language is completely absent from the 50Hz version - perhaps the checks are absent in the 50Hz versions.... which would explain why @jimbob232's experience differs from mine

 

[fafrd]

Common heat sink for the 3 power resistors, then just natural convection I hope. Cutoff is relevant should the SSR die or the switch/controller fail to enable it- either way the 4W resistor might be on at 25W for an extended period. (After 3 hours @ 9.5A, the two 100mOhm resistors hit 75C/170F in free air inside a cool garage; the m250, flush against plywood, was at 35C/95F. )

Good plan.

That's correct, I haven't yet tried 0.1 at startup - will wait until I can unbrick them.

Makes total sense and will be interested in the result once you get a chance to test it. I guess once you have the ability to unbrick, the test sequence is irrelevant, but the resistor values I would be interested to see are 1 Ohm, 0.1 Ohm and 0.05 Ohm…

A bit more info: the North American m250 manual states the following regarding the solid red LED I encountered::

 Solid Red: For all M250 models and for M215-60-2LL-S22-IG, this indicates a “DC
Resistance Low – Power Off” fault. Troubleshoot as described in the following sections

DC Resistance Low – Power Off Condition
For all M250 models and for M215-60-2LL-S22-IG, a solid red status LED when DC power has been
cycled indicates the microinverter has detected a DC Resistance Low – Power Off event. The LED will
remain red and the fault will continue to be reported by the Envoy until the error has been cleared.

An insulation resistance (IR) sensor in the microinverter measures the resistance between the positive
and negative PV inputs to ground. If either resistance drops below a threshold, the microinverter stops
power production and raises this condition. This may indicate defective module insulation, defective
wiring or connectors, moisture ingress, or similar problem. Although the cause may be temporary, this
microinverter condition persists until the sensor is manually reset.

An Envoy is required to clear this condition. The condition usually clears with operator intervention
unless conditions causing the event have not been remedied or if the failure is permanent.

If a microinverter registers a “DC Resistance Low - Power Off” condition, you can attempt to clear this
condition. If the condition does not clear after you perform the following procedure, contact Enphase
Energy customer support customer support at enphase.com/en-us/support/contact.

This language is completely absent from the 50Hz version - perhaps the checks are absent in the 50Hz versions.... which would explain why @jimbob232's experience differs from mine

‘An insulation resistance (IR) sensor in the microinverter measures the resistance between the positive and negative PV inputs to ground.’

This is a resistance check but not between battery + and battery -.

Battery + has ~infinite resistance to ground, so that test is obviously not the problem.

What resistance do you measure between - input and ground of your M250s?

I just measured my unpowered NEP Microinverters and they have 775 kOhm between - input and AC ground, so I doubt your 4Ohm resistor is making and difference confirming the negative DC input is not shorted to ground…

The fact that this ‘Resistance’ (Short) check to ground is only performed on some M250s and not others complicates the story, but it is possible that in addition to checking for shorts between positive and negative to ground as documented, the M250 also checks for resistance between positive and negative and/or seeks the two operating points it needs to begin generating power when it first receives DC current (which amount to largely unknown he same thing) and that a failure to lock into an operating point when there is clearly DC power available triggers the same fault condition… (despite the fact that it is not documented as such).

We’ll know as soon as you are n a position to boot-up with only an 0.1 Ohm resistor…

 

[fafrd]

I want to repeat the significance of what this simulated data means since I believe it is central to this discussion about a fault condition caused by insufficient resistance between battery positive and negative:


Focus on the upper left graph with 0.01Ohms inline on positive DC input. That slope is very steep but with 0 Ohms of wiring resistance between battery positive and battery negative, it would be an infinite slope / step function.


An MPPT will quickly lock on to the maximum DC power point at Vbat - delta (32V in this case). That delta is ~0.1V to 31.9VDC with 0.01 Ohms inline and would be 32.0V with 0 Ohms between positive and negative.

So with 0 Ohms of resistance, the MPPT is only going to be able to drive input power and current to 0W / 0A with whatever smallest voltage step it can apply above Vmp of 32.0V. If, for example, the MPPT wnts to find a point where input power drops by only 50% instead of 100%, it will be unable to do so no matter how fine-grained it's voltage control is.

Now look at the upper right graph with 0.1 Ohms of resistance inline. That 0.1 Ohm inline resistor will pull maximum power point voltage back to 31.4V but it makes it easy for the MPPT to find the 50% power point at 29.6-29.7V.

If we look again at how microinverters work when powered by solar panels:



That slope above Vmp is closer to 1 Ohm than 0.1 Ohm or 4 Ohm, which makes it very easy for the MPPT to incrementally reduce power / current input with moderate increases of voltage above Vmp.

(Note the that this graph shows an MPPT locking on to Vmp and then locking on to a delta_V delivering a lower power level at Vmp - Delta_V with increased current and an even lower power level at Vmp + Delta_V with decreased current.

Microinverters need a lower but greater-than-zero power point at voltages above Vmp which is not possible when powering off of a battery with 0 Ohm of resistance (and all-but-impossible with only a few milliohms of wiring resistance).

I was concerned (and still am) that any 'resistance check' between positive and negative might require solar-panel-like effective resistance or ~1 Ohm but the fact that we know the microinvetrer will produce power with only 0.05 Ohms means that a resistance as low as that my be all that is required for the MPPT to confirm it can operate with the delta-V / delta-I slope it sees anove the maximum power point and there is nothing more to it than that.

The confirmation one way or the other will be from ABT once he is in a position to reboot bricked M250s and to try booting with 0.05 and 0.1 Ohm resistors...

But be it from the resistance of a DC relay or from an explicit power resistor added inline, I think we’ve seen enough to know that anyone powering a Microinverter from a battery needs to assure there is at least ~0.05 Ohms of resistance round trip between positive and negative leads to the battery…

 

[kundip]

I had some time last night to play around with the M250's; here's what I've discovered:

• They will power up when connected directly to a battery, HOWEVER:
  • These models (and iq series) perform a DC resistance check at startup, and if too low they mark themselves offline (solid red LED) which persists across reboots. Per docs this can be rectified using an Envoy; I've ordered one from ebay. Scratch 2 m250's for now.

FYI All my setups go through the Envoy.
Possibly an important point I failed to detail clearly.
Originally I thought the MicroInverters would not work without it.
Mainly I wanted to monitor production.
I had no idea re the above.

• Inserting the 4 Ohm power resistor I use for precharging my off-grid AIO permitted the m250 to boot up and begin producing power. m250’s MPPT settled down to the minimum 16V supported by the m250, leaving a ~10.6V drop through the resistor at 10V/4R=2.5A=25W burned as heat, 2.5A*16V=40W output to AC. Not particularly efficient
• Bypassing the resistor after MPPT stabilized didn't trigger the "DC Resistance Low" error. (yes risked 1 more for SCIENCE), but no power was produced - my ammeter suggests m250’s MPPT was repeatedly hunting and sleeping.
• I tracked down a couple of 0.1 Ohm power resistors in my junk drawer.
  • Adding one to the DC input generated full power (9.5A, 248W AC, 10W heat loss on the 0.1 resistor) when 4W was removed
  • So I tried putting two in parallel (0.05 Ohm) which also worked, and as a bonus only generated 2.5W each.
  • Still probably need a heat sink for these and for the 4 Ohm bootstrap resistor.

My current thinking is that I'll use an 80A SSR to switch DC to the inverter(s), in parallel with a relay and in-line 4 Ohm resistor. The ardunio/esp32/rpi (haven't decided) would sequence the relay for bootup, then the SSR, then turn off the now-bypassed relay.

Before leaving this unattended I want to put snap-disc thermostats on at least the resistors/heat sink, maybe on the inverters as well. Those could either trigger the shunt trip relay in my main DC breaker (normally open), or could sit inline with the DC input to the inverter (normally closed).

After I have the full-power version stable, I'll add in another inverter via one of the cheap PWM units, maybe one with filtering and a third without.

View attachment 146647View attachment 146648

 

[kundip]

Your desktop PSU has a current limit. Do you know what it is? 30?

So both you and kundip successfully powered up Enphase inverters with a PSU (DCDC-booster in kundip’s case) which strongly suggest that a PSU / DCDC converter provides a sufficient current limit to pass the Resistor Test.

ADT bricked without 4 Ohms inline when powered-up directly from battery while kundip successfully powered up from battery through a relay (which is a resistance of possibly as much as 0.25Ohms).

Especially since you want to control output, there is no need for you to power direct from a battery and unless you want to risk bricking an inverter in the interest of science, I’d suggest you stick to your PSU.

So you switched AC when powered by your PSU while kundip switched DC power with AC already connected. I don’t know whether kundip switched DC from his DCDC booster to his M250 or switched DC from his battery to his DCDC booster, so perhaps he could clarify.

I switch from my battery via the timer.
The booster only comes on after the times switches it on.
So it is always Battery -> Timer -> then all other devices after that.
Originally I powered my timer from the battery but on the new setup I use a 240V AC power supply to power timer. (24VDC)
Just to tidy up my wiring.
I am not sure if the Timer has any effect on inrush current as the battery is always connected to the timer relay
I have bought a UNI-T UT216C Multi-meter to measure inrush current which I have never done.

​

Either way, it sounds as though AGT might have been switching DC through his ultra-low impedance SCCs, meaning he was not switching AC like you were.

So it may be that having DC switched off when AC is switched on is another way to pass the resistance test. Perhaps AGT could disconnect DC, cycle AC, after a nice long interval like 5 or even 15 minutes, try connecting DC through an 0.1 ohm resistor to see if that passes.

We’re calling it a Resistance Check because AGT was able to successfully avoid the fault condition by switching in DC power through a 4 ohm resistor but not through direct connection through a ultra-low impedance SSR.

Reading his post again, it does not seem he tried powering through an 0.1 ohm resistor - only 0 ohm or 4 ohms.

So it may be that an inline resistance of as little as 0.1 ohms or even 0.05 ohms is all that is needed to pass the Resistance Check.

I was concerned that the infinite slope of a battery would make it difficult for an MPPT to lick in to the two operating points as easily as it does on the less steep slope of a solar panel past Vmp, and that’s the reason I looked into what would happen if a small power resistor was added.

Once it has identified Voc, The MPPT steps to a lower voltage and expects to see current has decreased by a bit. If current changes by too much, it tries the midpoint. If current has still changed by too much, it tries the quarter point, etc…

If it can’t find a point where current is below what it is looking for, it will give up.

So AGT proved 0.0002 Ohms (plus whatever wiring resistance he had) is too low and kundip proved whatever the resistance of his DC relay (probably 0.1-0.2 ohms) plus whatever his wiring resistance is was enough.

So this Resistance Check could just be the MPPT confirming that it can lock on to the necessary control points and as little as 0.1 or 0.05 ohms is enough to do that…

I doubt it. Any current-limited supply will almost certainly allow the Microinverter MPPT to lick on. As I’m going back and reviewing all of the details, I’m suspecting the ‘Resistance Check’ is just the MPPT confirming that it can find the two operating points it needs. On the end it amounts to the same thing: no ability to lock-on = resistance from an infinite current source = too low.

But since we know these inverters can lock on and invert with as little as 0.05 ohms, that may be all that is needed to pass the Resistance / Lock-on test…

From current-limited sources such as DCDC boosters or PWM controllers, there should be no issue in any case (which is why I was interested to go in that direction).

I’d suggest you get ahold of some 0.05 ohm 100W power resistors: https://www.amazon.com/gp/aw/d/B07H35JR86/ref=ox_sc_saved_image_2?smid=A1THAZDOWP300U&psc=1

You can always start with 2 in series to test 0.1 ohm and then drop down to a single one if that works.