Battery "Bounce"

A shunt regulator has been proposed above. It could look something like this :

It's actually a series regulator because the regulating element, M1, is in series with the load. Using a transistor with very low VDS(VCE) sat, would have minimum impact on the normal operation of the system. Any shunt regulator would require impedance in the circuit as has been discussed, and that would impact the system's operation. A complete solution might include series/shunt/filter elements, but the OP already said he isn't interested in such a solution. Unfortunately for him, this is quite possibly what will be required.

Something else could also be happening is that the water heater has a lower cold resistance and thus the VA rating is too high for the initial turn-on.

You may be able to mitgate the surge with a series resistor and a delay that drops out the series resistor. It doesn't have to be extremely high power because it would only be on for a short time. This I had to do with an amplifier that had 40000 uF of capacitance on a 50 V rail. My circuit based timing on 2/3 the power supply rail. If the amp got a fault, the resistor would vaporize as designed.

Well from my last two posts you will see that an shunt is not going to do it. At least according to my tests. It would appear now that if the caps don't work then what I will need is to briefly disconnect the PV array when a surge occurs. To me now it would appear that the controller is reading a very low battery voltage (12,4V) during the high load so is dumping in lots of power to charge the batteries. However the batteries are not really discharged so the whole system immediately goes high until the controller recognizes the over voltage and corrects it.

I think now that your voltage comparitor switching OFF the PV array briefly might be the way to go. Also with this new info I will call the controller tech support, I know that there are adjustments to the PWM that can be made, perhaps this might help?

What is "The OP" ?

The heater takes 150 amps the battery bank is rated at 3,300 (CCA) amps.

OP is "Original Poster".

I think your best bet is to modulate the heater to start over time.

There is a simple PWM controller at Jameco. e.g https://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_120539_-1 You would just have to make a MOSFET driver able to carry 200 Amps or so. Inherently, the PWM driver contains a soft-start circuit, but you may need to ramp the setpoint.

Would you happen to know the material that the heating element is made out of? Can you find it out? Wire AWG, length and material?

Are you capable of measuring the cold resistance of the element? You would need to be able to measure the current through the device and the voltage across it. As long as you have the resolution in voltage, you may be able to determine it.

So, I see two solutions:

1) A high power resistor. I once bought a 1000 W 0.5 ohm adjustabe resistor, but they are expensive. The downside is the contactor will draw current too.
2) PWM and create a slow turn-on for the heating element
3) An inductor in series with the heating element. An inductor won't allow the current to change instantaneously.
4) Resize your inverter

This really looks like the case that the cold resistance of the element is too low and exceeds the VA capacity of your inverter. Only #2 seems practical.

Dear KeepItSimple; I think you are missing the point here. My heater is fine, if you re-read the threads you will see what I am trying to accomplish and now it would appear that this dummy load is NOT the answer. Additionally the heater is not on the inverter it is direct 12V feed.

Thanks.

Tomorrow at the highest PV input time (about noon and about 40-50 amps if a clear day) I am going to repeat today's test with the water heater load but with 2 of the 3 circuit fuses for the PV circuits removed. This will leave an input of about 15 amps and if the Vmax then does not climb above 14.8V I think I can assume that the problem could be cured by simply disconnecting these 2 out of 3 PV' circuits for a few seconds via a relay. Of course I will then need a voltage comparitor to swicth the relay.

Thanks Guys.

I got the impression that when the Heater turns on, the inverter glitches.

Which could mean that it pulls the battery voltage down where the inverter will turn off briefly or the charger over responds and pushes the voltage up slightly.

So, if there is a cause/effect with the heater turning on, then slowly turn on the heater.

Read all the posts please. Start at the beginning would be good I can't repeat everything written here.

Hi again,

Just to clarify, a shunt regulator for this system needs no extra series impedance to work against because there is already series resistance inherent in the system, and this fact was proven with one of the system test results previously posted. Also, the shunt regulator with extra series resistor, the series resistor is sized small enough to allow the regulator overall to function as a shunt regulator.
To refresh, a shunt regulator is a regulator that shunts current to ground and draws the voltage down in doing so. It doesnt matter what the impedance of the shunt regulator is except that it be small enough to be able to shunt the required current to draw the voltage down to a low enough level during a time of worst case voltage rise or power feed. A resistor in series with the shunt leg would be there mainly to remove heat from the power transistor so a smaller heat sink can be used. Alternately a larger heat sink can be used and possibly no resistance or less resistance.

Regardless what is causing the increase, a shunt regulator would in fact work. However, since you are saying that there may be a way to adjust the array controller to help the original situation then that would be the first thing to try.

MrAl. I really appreciate you staying with me on this. I guess I am out of my depth now because I don't understand the difference between a shunt regulator and the dummy load I placed on the system today which actually ended up re-introducing the problem.

However I bow to your better knowledge in these matters. Tomorrow (high sun levels permitting) I will repeat today's experiment (after the batteries are fully charged) but remove 66% (2 fuses out of 3) from the solar inputs. If the inverters then do not trip I think we are on to something.

Thanks.

A shunt regulator can be thought of as a variable "dummy load" that is controlled by some means such as voltage.

A series regulator can be thought of a variable resistor which is controlled by some means such as voltage. Example: LM317

The Setpoint(SP) is the desired value. The (MV) is the Measured Value, e.g. voltage.

See: https://en.wikipedia.org/wiki/Linear_regulator

Interesting problem. Sounds like the controller has a slow response to battery over-voltage. Probably designed that way. So let me throw out my thinking and see where the holes are. Your batteries have a higher internal resistance than what is expected by the charge controller. The controller is counting on the batteries holding the voltage below 16 volts at 40 amps for at least 2 seconds. Yours don't.
So you either need to drain off some current with the shunt regulator thus dropping the voltage from the PV. Or, as you mentioned you might be able to disconnect 2 during the over-voltage to increase the internal resistance of the PV system. Either might work. The other thought is to add a very small resistance (.015 ohms) in series with the PV to drop a few 10ths of a volt. Might be enough - hard to say.

SirSparks said:

MrAl. I really appreciate you staying with me on this. I guess I am out of my depth now because I don't understand the difference between a shunt regulator and the dummy load I placed on the system today which actually ended up re-introducing the problem.

However I bow to your better knowledge in these matters. Tomorrow (high sun levels permitting) I will repeat today's experiment (after the batteries are fully charged) but remove 66% (2 fuses out of 3) from the solar inputs. If the inverters then do not trip I think we are on to something.

Thanks.

Click to expand...

Hi again,

The shunt regulator i was talking about would be like your dummy load except that it would know exactly when to kick in and kick out, and do it smoothly so everything always looks normal. A series regulator would work too, but it would have to be connected all the time which would reduce efficiency. The shunt regulator would be effectively non existent (draw very little current) while it is not needed.
Are you perhaps saying that the voltage goes up even with the dummy load? If so, what was the value of the dummy load?

Certainly if you can get your tech support to help you adjust the controller so it works better without any extra circuitry, that would be the way to go.

Hi ronv. Well I did that last check and to my amazement dropping out 2 of three PV circuits for 30 seconds and then reconnecting them led to an immediate shut down of the inverters each and every time ! In fact at 16.2V every time which is higher than I have ever seen. I think now that there is very little choice but that it IS a battery problem. I have decided to purchase better batteries but not until Jan 3rd (for tax reasons). In the mean time when the capacitors and choke arrive I'll try those too but in the last couple of days I am seeing a serious reduction in reserve power from my batteries. Voltages etc. are fine but they are depleting more quickly now.

Thanks for your input.

MrAl.Tech support only repeats that it is a battery problem! Coupling this with the fact that I am now seeing reduced reserve capacity (despite them being only 3 months old) I have decided to bite the financial bullet.

I will however still try out the caps and chokes on order.

Realistically I need about 2Kw of reserve after sundown the new batteries should supply 2.8Kw. Of late the old ones are below 1Kw.

Thanks.

I sometimes have the same problems with a old set of deep-cycle batteries (secondary power) when they are fully charged, the charge controller (C40 12vdc) is in float and the PV voltage is high (20+ volts) with a light load. The primary set never has the problem. (Costco golf-cart batteries)
I have my own DIY battery monitor that just moves the charging to another set of cells just as old set goes into float mode and monitors the AC inverter output for glitches. If everything is fully charged with little loading the anti-bounce logic disconnects the charge controller until a battery needs the power.

That seems like a perfect solution. I fairly obviously need a new battery bank but it would be a shame to discard the old one.

Thanks for the input.

I have a physical space placement problem; I can only place the new batteries within 7ft ( 0.1 voltage drop at 50 amps) the old set would then need to be moved a total one way wire length of 20ft ( 0.3V drop at 50 amps) away. I am considering a mechanical (B1, B2 or both) switch instead of your automated switching. I could choose a "absorb/float" volts a little high for one set and a little low for the other. In other words; Recommended V plus 0.2V