How to protect AGM batteries from edge-of-cloud effect using MPPT controller?

RegGuheert

Since I just chose Sun Xtender AGM batteries and Midnite Solar Classic 200 charge controller for my sister's off-grid system I was interested to read Ralph Day's and Dave Spark's comments about overcharging and outgassing that can occur due to edge-of-cloud effects:

Ralph Day wrote: »

I see this happen often, but the surge is not a problem, unless perhaps you have fully charged AGM batteries and the surge causes a little outgassing to occur.

Dave Sparks wrote: »

I just changed out a neighbors AGM's for the second time for this problem of spikey voltage peaks when in absorption. Keep a load going if you must have these batteries.

We have lots of clouds around here and I have seen power levels as high as 3300W from my own 2880Wp array.

So, given that I have this combination, I am wondering how big of an issue this can be and how to best remediate it. Here are some specific questions:

1) How is MPPT related to charging of the batteries in absorb mode since MPPT is ONLY used during bulk charging? (Or perhaps it is not related?)
2) Has Midnite done anything in the design of the Classic controller to improve upon this situation? More generally, I'm wondering how the loop is closed to control the charging in absorb mode and whether it is fast enough to correct for the changing light conditions.
3) Dave suggests to "keep a load going." How does this help? (Again, not knowing the theory-of-operation for the controller in absorb mode makes it hard to understand how it would respond to rapidly-increasing light levels.)
4) I'm willing to set up diversion loads if required to help prevent battery outgassing, but how do you arrange for such a load when the conditions are poorly understood? IOW, what amount of load *specifically* needs to be added and under what exact conditions in order to prevent any outgassing?
5) Any other thoughts on this subject?

Dave, if you could tell more about your neighbor's situation such as battery, controller and array specifications and how long they lasted, etc., that would be very educational to me!

TIA!

Reg

Cariboocoot

Re: How to protect AGM batteries from edge-of-cloud effect using MPPT controller?

RegGuheert wrote: »

So, given that I have this combination, I am wondering how big of an issue this can be and how to best remediate it. Here are some specific questions:

1) How is MPPT related to charging of the batteries in absorb mode since MPPT is ONLY used during bulk charging? (Or perhaps it is not related?)

MPPT functions in all modes: it simply converts incoming power to a predetermined outgoing power as best possible. In Bulk it's trying for maximum charge Amps. In Absorb and Float it's adjusting current to maintained fixed Voltage levels against varying loads.

2) Has Midnite done anything in the design of the Classic controller to improve upon this situation? More generally, I'm wondering how the loop is closed to control the charging in absorb mode and whether it is fast enough to correct for the changing light conditions.

I'd bet the Midnite Classic and the Rogue both handle this situation very well, as they (according to my understanding) adjust output continually rather than doing "sweeps" at intervals. It's that which creates the problem with edge of cloud: charge controller has determined a proper Voltage/Amperage setting when suddenly the input jumps up. This spikes the output Voltage because the microprocessor isn't adjusting at the moment. (I could be wrong about this or just explaining it poorly.)

3) Dave suggests to "keep a load going." How does this help? (Again, not knowing the theory-of-operation for the controller in absorb mode makes it hard to understand how it would respond to rapidly-increasing light levels.)

Again the problem occurs at full charge point. Current under no-load conditions is low. If there is a load on, it can help sink the sudden Voltage increase, preventing it from spiking too high and causing gassing.

4) I'm willing to set up diversion loads if required to help prevent battery outgassing, but how do you arrange for such a load when the conditions are poorly understood? IOW, what amount of load *specifically* needs to be added and under what exact conditions in order to prevent any outgassing?

You could calculate a potential for Voltage spike. The larger the array and the higher its nominal Voltage the greater the potential for dangerous spikes. I've seen 30% (at the panel) over power on mine, but the system sucks it up handily. Array Voltage is only 2X system Voltage, and the VFX3524 chews 20 Watts on its own.

5) Any other thoughts on this subject?

Yes; don't look down your nose at FLA's. They have some decided advantages over AGM's.

Dave, if you could tell more about your neighbor's situation such as battery, controller and array specifications and how long they lasted, etc., that would be very educational to me!

TIA!

Reg

I think we'd all be interested in the dangerous details. Could add valuable info to the Big Book of Solar Power that is this forum!

Dave Angelini

Re: How to protect AGM batteries from edge-of-cloud effect using MPPT controller?

The International Space Station spends 36 minutes of it's 96 minute orbits in the shade. They change batteries often and protect them with a crowbar circuit independent of the charge controller. When something is available like that for offgrid and when the numbers make sense financially, I will consider a sealed battery and give up the ability to read specific gravity. Until then I always reccomend against AGM/GEL for my offgrid people!

RegGuheert

Re: How to protect AGM batteries from edge-of-cloud effect using MPPT controller?

Thanks for the detailed response, Marc!

Cariboocoot wrote: »

MPPT functions in all modes: it simply converts incoming power to a predetermined outgoing power as best possible. In Bulk it's trying for maximum charge Amps. In Absorb and Float it's adjusting current to maintained fixed Voltage levels against varying loads.

But I would say that in absorb and float the controller is no longer tracking the maximum power point, but rather is operating at some other point.

Cariboocoot wrote: »

I'd bet the Midnite Classic and the Rogue both handle this situation very well, as they (according to my understanding) adjust output continually rather than doing "sweeps" at intervals. It's that which creates the problem with edge of cloud: charge controller has determined a proper Voltage/Amperage setting when suddenly the input jumps up. This spikes the output Voltage because the microprocessor isn't adjusting at the moment. (I could be wrong about this or just explaining it poorly.)

I hope you are right about this! It looks like a nice controller and given the amount of time it has been in the field, hopefully they have improved over previous designs out there.

Cariboocoot wrote: »

Again the problem occurs at full charge point. Current under no-load conditions is low. If there is a load on, it can help sink the sudden Voltage increase, preventing it from spiking too high and causing gassing.

Makes sense. Unfortunately, I do not have a good picture of all the load profiles this system will see. In the wintertime I expect some pretty steady loads during the daylight hours. During the rest of the year, however, there could be long periods with very low load conditions.

Cariboocoot wrote: »

You could calculate a potential for Voltage spike. The larger the array and the higher its nominal Voltage the greater the potential for dangerous spikes. I've seen 30% (at the panel) over power on mine, but the system sucks it up handily. Array Voltage is only 2X system Voltage, and the VFX3524 chews 20 Watts on its own.

If we see such voltage spikes on this system, perhaps just a resistive lump load makes sense. I wonder how quickly the Classic can bring a dump load online...

Cariboocoot wrote: »

Yes; don't look down your nose at FLA's. They have some decided advantages over AGM's.

I don't. In fact, I originally specified FLAs for this system. Unfortunately, as you know, I would have had to spend more than twice as much as anticipated to handle the charging current available. Would the FLAs have lasted more than twice as long? Perhaps, but they just as possibly may have died faster than the AGMs. I think we can agree that AGMs have some decided advantages over FLAs, also.

At the end of the day, even though what is failing here is the batteries, I will say the the *problem* lies with the charge controller, NOT with the batteries. If the charge controller is supposed to supply current in such a way as to maintain a particular battery voltage, then it should do this so that the voltage is maintained and not exceeded. The designers of charge controllers should design their equipment for the conditions which will exist in the field and clouds are a part of the equation which they should properly address. You still could run into issues related to step-function changes in the load current on the battery, but that should also be considered in the design of the controller, IMO.

AntronX

Re: How to protect AGM batteries from edge-of-cloud effect using MPPT controller?

I agree with you that output voltage overshoot is fault of a designer of controller, not battery. Outback FM-60 suffers from this problem, at least one with old firmware. To fix that, you can setup resistive dump load to absorb excess output current. FM-60 has AUX output that can drive solid state relay in PWM mode. I did this in this video. I wanted to capture all available solar energy for data logging purpose, regardless if my batteries are full. I have setup my controller to never go into Absorption mode. Notice how current meter jumps around during diversion? To fix that I have made all analog PWM diversion load controller to take place of FM-60 for driving solid state switch.

Ralph Day

Re: How to protect AGM batteries from edge-of-cloud effect using MPPT controller?

I have an oil filled radiant heater (120vac) plugged into a solid state relayed outlet (relay controlled by MX60). The delay and hysteresis are set to 0 so there's no delay, hold or anything. You can watch the little indicator light on the heater flash faster than the MX indicator screen showing AUX on or off.

Works well for me. It's great on a windy night, full SOC situation. The wind controller is very slow to react and divert to it's resistance heat load, a real voltage surge from it. The oil heater will clip the spikes really well. Not great in the summer, but few windy nights in the summer to worry about.

Ralph