How to tell if Charge Controller is functional?

[kjmclark]

I have an off-grid solar power system for my barn. 12V, with a 20V 250W Canadian Solar panel, a Tracer 20A MPPT controller, six deep cycle batteries, and an inverter with a transfer pump as the primary load. The inverter says the load from the pump is 220W when the pump is running.

This summer, I realized that my 6 year-old battery bank had boiled, and needed to be replaced. (One of them went bad, and took out the rest too fast for me to catch it.) I put in all new batteries, after I'd charged them at home. I put the new bank back into service about a month ago.

The pump runs to fill a 1000 gallon water tank each day (it's arduino-controlled). Usually the pump runs continuously at about 9am for 45 minutes, then the rest of the day (til about 7pm) the pump kicks on and off as needed to keep the tank full. I'd estimate the pump actually runs about 4hours total a day, but it's lots of bursts, so off for a few minutes, on for 20 seconds, off for a few minutes, etc.

What I've seen in the past is that running like this, the panel manages to keep the batteries fully charged, even with the pump running during the day. It is a little late in the season, however, and many years we would have shut the pump down for the season, but this year we're leaving it running. I also need to adjust my panel tilt, as it's still set for late May. Now, we're getting the batteries bumping down to 12v before my arduino decides the power is too low and shutting off the pump for low power.

I'm wondering if having the earlier battery bank very close to completely dead may have cooked the controller in some important way. And that brings up the question of how in the world one can test a charge controller? The voltage from the panel is fine, the voltage from the batteries is not what I'd expect (but they were fully charged when I hooked things back up, but how would I know if it's a problem with the charge controller besides buying a new one and swapping it out?

[ButchDeal]

Did you connect the charge controller up to the new batteries before reconnecting it to solar?

your pv module has degraded a bit and with your long use of the pump into fall, you might consider adding a second pv module...

[sensij]

A clamp on DC ammeter would tell you how much current is going into the battery, if your charge controller doesn't report it (or if you don't believe it).

That you are still able to run most of the day suggests nothing catastrophic... I would guess it is mostly related to running later in the season. Adjusting tilt might help, but days are getting shorter and there isn't much you can do about that except adding panels.

[Mike90250]

that panel seems too small for all the batteries, might be OK in summer, but days are shorter now

[NEOH]

Does the Charge Controller indicate the Charging Mode?
a) Bulk Mode until Solar Noon
b) Absorb Mode in the afternoon
c) Float mode before sunset

Whats is the voltage of the Battery bank during each mode?

WARNINGS:
Connect battery terminals to the charge controller BEFORE connecting the solar panel to the charge controller.
NEVER connect solar panel to charge controller, until the battery is connected.
NEVER disconnect battery without disconnecting the solar panels first.

[kjmclark]

Butch, yes, I was very careful to connect at least one battery to the charge controller before connecting the panel. IIRC, I had two connected when I attached the panel. Made that mistake on a completely separate system for our hoophouse, and had to replace that controller.

Sensij - that is one piece of equipment I don't have, and thought I ought to pick up at some point. Excellent idea, I'll do that and check.

Mike - yeah, we're not used to having to run the pump this late. Somebody is welcome to have their heat and drought back - we'd be happy to get rid of them. But because we're running the pump, the crops are doing quite well.

Neoh - that would be a not-bad reason to get a new controller - this one tells you very little. It uses two LEDs to tell you all you're going to get. I'm reading voltages off of the inverter and my DVM. The lights I'm getting indicate that the PV is providing power (during daylight - this goes out as expected at night), the battery light is on solid, indicating it's charging the batteries, they're not full, and not bad. My wife tells me (I'm color blind!) that the battery light is green not red, so not a fault. When the bank is fully charged, it changes to a slow flashing green instead of solid.

So, I'm going out tonight, and I'll check things, tilt the panel back up (it's about time anyway), reprogram the Arduino to shut off the pump for below 12.2V, and assume things are working until I can get a clamp-on ammeter. I was concerned about the voltage this past weekend, so I unplugged the pump Sunday. We've had rain and clouds for the first two days, then ample sun today, so I should see a little better voltage when I get there.

Seems kind of funny that the most intelligent part of the system is the only thing that there aren't good tests for.

[NEOH]

Does your battery ever get to 14.8 Volts before sunset?
If not then what is the highest battery voltage?

Could an Internal Setting have changed when the Battery Bank was disconnected?
Maybe, the Charge Controller reset to Gel type or AGM type vs Flooded type battery?

[kjmclark]

Neoh, If there's an internal setting on this controller, I'm not sure how to change it. Another reason to upgrade the controller. I'll try to dig up the manual.

Looking at things last night, I think you're all right that I'm just pushing the amount of energy I have coming in. So, I checked the voltage with a voltmeter, and the voltage on the inverter and voltmeter were both 12.5V. When I run the pump though, it immediately drops to 12.3, then slowly drops from there (this was after sunset, so it's more of a battery test). I shut it off quickly, and the voltage rose back to 12.5.

One of the things I have the arduino doing is logging the voltage, and that suggests the controller was at least working fine when I last changed out the SD card. That graph is attached. Starts on 9/2. 9/4-9/6, and the last three days, 9/18-9/20, were no-load. I unplugged the pump because we weren't going to need the water, and I wanted to see if the voltage rose. Looking at it now, that looks completely normal. 9/4-9/6 were all sunny, and things rose nicely. 9/18 and 9/19 were cloudy, and not much changed, but the last day rises nicely, and that was a sunny day.

I only get out there every three days or so, and can only see the system running when I'm there on a weekend. The datalogging is pretty helpful to see that things are working. I just hadn't checked it for three weeks.

So, I raised the angle of the panel, and raised the minimum the arduino will allow the pump to run at. That should protect the batteries better, and slightly increase the light getting to the panel. Clamp ammeter is on its way. If someone's looking at this later, make sure you get a clamp ammeter that reads DC current. You have to get a more expensive one to get DC current readings.

I'll try to check back in a week or two with a last update.