Tweet
Sounds about right for a DIY panel. Note that the loose-cells you get are NOT grade A, but factory rejects so they may have accellerated heating issues dropping you down to the 14v range. A panel that exhibits 20v when cool, and only gets to about 160F in the sun may not drop as much - maybe down to 17v OCV. But so far ok for DIY.
Cool. That means for these typical lead-calcium ups-type agm's, you don't exceed 14.7v charge, and have a typical maximum inrush current of about 0.25 to 0.3C. That means you should not be supplying it with anything larger than 1.25a to 1.5a. Sometimes that is printed on the case, but reputable manufacturers like PowerSonic, B&B etc have online docs and charts. Your diy panel is supplying about 2 times the recommended max for a 5ah agm.
That means you don't have any sort of absorb voltage limiting - in the case of these agm's, in a cyclic application, no more than about 14.6 to 14.8v should ever be applied. If you do more, you are corroding the plates with high voltage, basically doing an EQ on them which most manufacturers say not to do. Look for venting, hissing, buckling plates. A little handheld IR temp gun can help you keep an eye on it.
You'll also need current limiting, and perhaps the easiest is to just make sure your system is not capable of more than about 1.5a. Since that 5ah battery is now probably nicely toasted from your high-voltage, I'd upgrade to a 10ah agm, and program that arduino to never go above 14.7 volts.
Quite natural. If the leds draw no more than C/20 to C/10, (250 to 500ma total), then you are not heavily exceeding Peukert, and the voltages below will be in the ballpark even under load. Ideally, these voltages should be measured after a period of at least 4 hours of no charge and no load:
12.8 + 100% SOC
12.5 75% SOC
12.2 50% SOC
12.0 25% SOC
Your battery manufacturer will have more specific voltages and may include handy discharge/current charts to help calibrate the actual SOC under a wider variety of loads. Note that most people avoid drawing any larger than 0.25C otherwise your battery will only last minutes instead of hours for example.
Let the battery do the regulating - all you have to do is provide a maximum absorb voltage and limit the maximum inrush current. It is the typical 2 or 3 stage charger.
With your charger set to run at no more than 14.7 volts, the battery will eat up as much current as it can in the bulk stage. Your panel is capable of too much for these batteries, so either cut your panel output in half, or even easier, use a 10ah agm!
Once the battery voltage reaches 14.7 volts, and with your arduino charger limited to not ever go higher than this, the *BATTERY* will naturally self regulate itself by absorbing what it can and current falls naturally as dictated by the battery. Kind of like you eating a desert after a heavy meal. You are already full, but perhaps you can absorb just a little more pumpkin pie.
When to stop absorbing? If you have a way to measure current, then you can do either:
Charge until you reach C/100, and then float at 13.5 volts ....or
Charge until you reach C/1000. Float optional.
Better yet, since all batteries age from natural causes, trying to reach a specific current may not ever get there due to electrolyte dry-out. In this case, program your system to detect no major changes in current for 3 hours, and then drop to float. If no current sensing is available, then perhaps you could just incorporate a timer, say limiting the absorb time from when the Arduino detects 14.7v to no more than 4 hours when a forced change to float 13.5 should occur. Having both timers and current sensing is ideal.
This is basically what most 2 or 3 stage "smart" chargers do. From your description, it sounds like your Arduino is not capable of maintaining a voltage level, but only turning a circuit on and off at preset voltage levels. In this case, the only thing you can do is a 1970's style hysteresis "ping pong", where you turn the panel on when the battery voltage reaches 13.2v at the low setting, and turn off at the 14.7v setting. I don't advise it though for battery longevity.
You could do the old school fake-pwm by putting your low and high voltages very close together, ie 14.2v for low and 14.7 high. Still, read the Morningstar archival docs on why they went to a true pwm duty cycle and not a tight-hysteresis model.
At this level, if you really want to have some fun, and not worry too much about current limiting a 5ah agm, then look into Hawker Cyclon monoblocs. Pure-lead agm. Takes massive current input (within reason), but still DONT exceed 15 volts. They look like baby JCI/Optima's, but are really Enersys/Odyssey/Hawker family. Thing is, these babies like to see about 0.4C minimum under cyclic ops, so you'd want to supply a minimum of 2a charge from your solar/charge-controller system.
I am charging a Lead Acid AGM 12V 5Ah battery with it.
When it is disconnected for some time (hours?), it balances at voltages that make sense, like 13.x after a charge. However, when the panel is connected, the voltage spikes up to 15-16V.
You'll also need current limiting, and perhaps the easiest is to just make sure your system is not capable of more than about 1.5a. Since that 5ah battery is now probably nicely toasted from your high-voltage, I'd upgrade to a 10ah agm, and program that arduino to never go above 14.7 volts.
And on the opposite side, when I put some load on the battery, in the form of power leds that it's supposed to run on a daily basis, the voltage spikes down.
12.8 + 100% SOC
12.5 75% SOC
12.2 50% SOC
12.0 25% SOC
Your battery manufacturer will have more specific voltages and may include handy discharge/current charts to help calibrate the actual SOC under a wider variety of loads. Note that most people avoid drawing any larger than 0.25C otherwise your battery will only last minutes instead of hours for example.
How then do I check the real voltage for the charger, so I can know when to stop charging to avoid over charging, and whether I have to disconnect the power leds because the battery is too empty, to avoid over discharge?
With your charger set to run at no more than 14.7 volts, the battery will eat up as much current as it can in the bulk stage. Your panel is capable of too much for these batteries, so either cut your panel output in half, or even easier, use a 10ah agm!
Once the battery voltage reaches 14.7 volts, and with your arduino charger limited to not ever go higher than this, the *BATTERY* will naturally self regulate itself by absorbing what it can and current falls naturally as dictated by the battery. Kind of like you eating a desert after a heavy meal. You are already full, but perhaps you can absorb just a little more pumpkin pie.
When to stop absorbing? If you have a way to measure current, then you can do either:
Charge until you reach C/100, and then float at 13.5 volts ....or
Charge until you reach C/1000. Float optional.
Better yet, since all batteries age from natural causes, trying to reach a specific current may not ever get there due to electrolyte dry-out. In this case, program your system to detect no major changes in current for 3 hours, and then drop to float. If no current sensing is available, then perhaps you could just incorporate a timer, say limiting the absorb time from when the Arduino detects 14.7v to no more than 4 hours when a forced change to float 13.5 should occur. Having both timers and current sensing is ideal.
This is basically what most 2 or 3 stage "smart" chargers do. From your description, it sounds like your Arduino is not capable of maintaining a voltage level, but only turning a circuit on and off at preset voltage levels. In this case, the only thing you can do is a 1970's style hysteresis "ping pong", where you turn the panel on when the battery voltage reaches 13.2v at the low setting, and turn off at the 14.7v setting. I don't advise it though for battery longevity.
You could do the old school fake-pwm by putting your low and high voltages very close together, ie 14.2v for low and 14.7 high. Still, read the Morningstar archival docs on why they went to a true pwm duty cycle and not a tight-hysteresis model.
At this level, if you really want to have some fun, and not worry too much about current limiting a 5ah agm, then look into Hawker Cyclon monoblocs. Pure-lead agm. Takes massive current input (within reason), but still DONT exceed 15 volts. They look like baby JCI/Optima's, but are really Enersys/Odyssey/Hawker family. Thing is, these babies like to see about 0.4C minimum under cyclic ops, so you'd want to supply a minimum of 2a charge from your solar/charge-controller system.
Comment