120V AC power path for solar charge controller

[davegravy]

I have a 65W MPPT solar charge controller module from Genasun:

https://sunforgellc.com/product/gv-5-mod/

I want to be able to use this charger from a 120V AC source when it's present, solar panel otherwise. I'm planning on using an off-the-shelf AC-to-DC converter followed by a buck converter to provide 3.5A constant current.

I'm looking at the LTC4412 as a means to provide the OR logic between solar panel and the AC source.

I found this circuit which I figure is mostly what I want:

https://circuitdigest.com/fullimage?i=circuitdiagram/LTC4412-Power-Path-Controller-Circuit-Diagram.png

Can this work?

The "load" in this circuit would be the "panel" terminals of the Genasun. Primary would be the AC source and Aux would be the solar panels. I'm not sure that any capacitors are necessary since it's not a problem if the voltage drops when switching between sources (the Genasun always has a battery connected).

I guess I'll want back-to-back MOSFETS rather than the single configuration or else I could have AC source power flowing back into the solar panel.

Anything else I should consider?

[Peabody]

I haven't used the LTC4412, but it would need to switch to the AC source if it is present, regardless of the panel voltage.  If there's any light, the panel voltage will rise to its maximum open-circuit voltage at the current illumination level.  If that's higher than the AC source voltage,  a switchover based on relative voltage levels wouldn't work because the panels may not be able to provide any current.  So you need the 4412 to give absolute priority to the AC source if it is present, even if it is below panel voltage.

As for the mosfets, your example circuit has mosfets in both lines, both oriented to prevent backflow when they are off.  I don't think you need more than that.

[davegravy]

I haven't used the LTC4412, but it would need to switch to the AC source if it is present, regardless of the panel voltage.  If there's any light, the panel voltage will rise to its maximum open-circuit voltage at the current illumination level.  If that's higher than the AC source voltage,  a switchover based on relative voltage levels wouldn't work because the panels may not be able to provide any current.  So you need the 4412 to give absolute priority to the AC source if it is present, even if it is below panel voltage.

You're right, I hadn't considered that.

I thought a while about it and don't see a way to make the 4412 (or a pair of them) work to assign priority (see attached).

I found the 4418 which may be a better fit, about to study it.

https://www.analog.com/media/en/technical-documentation/data-sheets/ltc4418.pdf

[Peabody]

These power mux chips can be very complicated, and often come in very hobbyist-unfriendly packages.  But I wonder if we can back up and rethink the problem.  Does it really matter if the solar panels take over providing charge current even if the AC source is plugged in?  In other words, what would be wrong with a two-diode setup?  The source with the higher voltage would power the load.  The diode voltage drop presumably wouldn't matter much in the AC line, but if it's a problem in the panel line, then that diode could be replaced by a mosfet.

In the circuit below, the panels could still take over via the body diode if the sun is strong enough, but I'm not sure that would be a problem.

[davegravy]

I like the simplicity but doesn't this have the same issue due to the fact the panel can be higher voltage than than the AC source even when it's not actually sunny, (ie can't provide meaningful current)?

I measured my panel Voc face down on the floor (almost no light condition) and it's higher than I expect to get from the AC source. I expect voltage drops sharply as soon as any current is drawn in this condition, but does that then lead to oscillation where the MOSFET turns on, voltage drops, MOSFET turns off, etc?

[Peabody]

The panel voltage will only go above the main source voltage if the sun is strong enough for the panel to provide *ALL* of the load current. Up to that point, the panel will provide as much current as it can at the voltage set by the AC source, which functions as a low-impedance "sink" at its voltage.

Solar panels behave strangely, and you have to keep in mind that they are basically current sources, not voltage sources. 

Below is a modified circuit which adds a schottky diode across the mosfet.  In effect, this diode takes the place of the body diode, and should be more able to handle the charge current.  When the AC source is not present, the mosfet turns on, which bypasses the extra diode as well as the body diode, and there is no voltage drop.  But even when AC is present, some current will flow through the extra diode depending on how bright the sun is.

Both schottky diodes should be 3A or whatever your maximum charge current is.  The SB220 shown in the previous reply is a 2A diode, so it wouldn't be enough.

[davegravy]

Thanks, ok I'm going to try this.

I thought of another possible issue, since it looks like in this scheme both sources can be pushing current simultaneously.  The panels are rated to 50W, the AC brick is 60W, and the Genasun is 65W. Should I implement some form of limiting to prevent overloading the Genasun?

[Peabody]

I don't think either source will be "pushing" current.  I think the Genasun will draw what it needs.  But the Gensun should be able to handle the maximum voltage the panel can reach.

[davegravy]

The panel voltage will only go above the main source voltage if the sun is strong enough for the panel to provide *ALL* of the load current. Up to that point, the panel will provide as much current as it can at the voltage set by the AC source, which functions as a low-impedance "sink" at its voltage.

Thanks for this, I've been testing it today in cloudy weather. It works very well. The only (very minor) issue is that while it will charge from AC (ie fast) when AC is connected first, if AC is connected second or if AC is briefly interrupted then it seems to charge from the panel (slow).

It's really only an issue if it's both a day where AC is unreliable AND it's cloudy AND there's been been a string of cloudy days with no AC  (ie the battery is low)... so not a very common scenario.

It seems that even though the panel can't provide high charge current desired by the charger it's voltage DOES rise above the ac brick (I measured the panel at 17.0 and the ac brick at 14.2)

I should test if night is dark enough for AC to take over.

I'm going to think on if there's any simple modifications to address that - if you have any ideas I'm all ears.

[Peabody]

Sorry, it's been a while and I've kinda lost track of which circuit you're testing.

But remember that the upper schottky diode has to be forward biased for the AC brick to take over.  I'm not sure what's going on, and particularly don't know what your Genasun is doing.  It's MPPT, so maybe it's cutting back on current draw when the panel is providing power, and doesn't know to draw more from the AC brick when it could.  That keeps the voltage feeding into the Genasun too high, so the schottky is reverse biased.  I know you have a lot invested in this charger, but kinda wish you had maybe used a cheap linear charger, although I don't know of one that would handle these currents.  Anyway, when it's not shifting over to the AC as it should, you need to measure the voltage at various points and see what's going on.  And maybe I just gave you a bad idea for the circuit.

[davegravy]

Mainly I went with the Genasun because I care a lot about efficiency and I expect they have a better performing charger than what I can develop with my limited resources (from cursory review of MPPT IC datasheets). It also has a bunch of useful features like low voltage disconnect, battery fault detection, etc.

I was testing the attached circuit FWIW

[Peabody]

Well, as I said, I suspect the MPPT charger is adjusting its current draw based on what produces the most power from the panel even though that is less than it would like to draw.  That makes perfect sense.  But the resulting voltage apparently is higher than the AC brick source, so the brick's diode won't pass its current to the charger - it can't flow uphill.  And the MPPT panel voltage will be even higher on a sunny day.  I should have considered this before suggesting that circuit.  I think it works properly only if the AC side is always a higher voltage than the panel (higher than the maximum open circuit voltage of the panel in full sun), or if the panel is capable of providing all the current the charger can use, which apparently it can't.   So I don't know what would make this circuit work properly.  Can you tell me what batteries are being charged?

[davegravy]

The batteries 12V LiFePO4 (4 cell in series). Capacity varies depending on physical location, typically between 10 and 100AH.

[Peabody]

Ok.  Well, I'm sorry, but I don't know how to fix it.

[davegravy]

Ok.  Well, I'm sorry, but I don't know how to fix it.

That's ok I appreciate you trying. I might circle back on the LTC4418 although it's a pricier solution.

[NiHaoMike]

I want to be able to use this charger from a 120V AC source when it's present, solar panel otherwise. I'm planning on using an off-the-shelf AC-to-DC converter followed by a buck converter to provide 3.5A constant current.

Just use the buck converter as a second charger? Set the voltage limit at 13.2-14V depending on how much you want the batteries charged from AC.

[Peabody]

In your testing in reply #8, what mosfet did you use?  I'm wondering if you picked one that is NOT logic level, but had as high a threshold voltage as possible, it might make the mosfet turn off whenever there's AC power.  But you might need to take away the second schottky, and possibly have back-to-back mosfets.  Well, I guess I'm making some assumptions here - that the mosfet is turning on when you have 17.0V and 14.2V.  But that may not be true.  Does it work better if you take away the second schottky that's in parallel with the mosfet?

[davegravy]

In your testing in reply #8, what mosfet did you use?  I'm wondering if you picked one that is NOT logic level, but had as high a threshold voltage as possible, it might make the mosfet turn off whenever there's AC power.  But you might need to take away the second schottky, and possibly have back-to-back mosfets.  Well, I guess I'm making some assumptions here - that the mosfet is turning on when you have 17.0V and 14.2V.  But that may not be true.  Does it work better if you take away the second schottky that's in parallel with the mosfet?

I'll try removing the schottky but in the meantime the MOSFET used is FDS4465 (datasheet)

[Peabody]

That mosfet has a very low threshold voltage, 0.6V typical.  Maybe if you found one that was 2.5V or so, it would be less likely to turn on even when AC is present.  But I'm not sure it makes any difference.  If the panel voltage is high enough, it will just go through the body diode even if the mosfet is off.

But just to be sure about what's happening, when you're stuck in this situation where the weak panel is supplying the load even when AC is present, what are the voltages on the three mosfet pins?

[davegravy]

That mosfet has a very low threshold voltage, 0.6V typical.  Maybe if you found one that was 2.5V or so, it would be less likely to turn on even when AC is present.  But I'm not sure it makes any difference.  If the panel voltage is high enough, it will just go through the body diode even if the mosfet is off.

But just to be sure about what's happening, when you're stuck in this situation where the weak panel is supplying the load even when AC is present, what are the voltages on the three mosfet pins?

When in the scenario you described:

• Gate is steady at 15.1V
  
• Drain and source both hovering between 19.2 and 19.5V.
• 0.3A going to battery

When I unplug the solar panel (leaving AC connected) the voltages change to

• Gate is steady at 14.6V
  
• Drain and source both 14.3V
• >3A going to battery

[Peabody]

When in the scenario you described:

• Gate is steady at 15.1V
  
• Drain and source both hovering between 19.2 and 19.5V.
• 0.3A going to battery

19.2V?  What is the panel's voltage?  I mean nominally.  Is it a 20V panel?  24V?

And I assume there is no voltage drop across the mosfet.  If the gate is 15.1V and the source is over 19V, then the mosfet should be fully on.

When I unplug the solar panel (leaving AC connected) the voltages change to

• Gate is steady at 14.6V
  
• Drain and source both 14.3V
• >3A going to battery

Well, if the gate is 14.6V and the source is 14.3V, then the mosfet must be off.  And if it's off, I don't see how there can be 14.3V on the drain.

Anyway, the panel voltage is so much higher than the AC brick that I doubt this will ever work the way you want without a more capable Or'ing chip.  At least I don't see how it could.  I should have thought about the MPPT implications, but I didn't.