Best Setting for low voltage disconnect?

williaty
williaty Solar Expert Posts: 60 ✭✭
I searched in the forum for a good while and I was shocked to not find a clear cut answer to this question. I have an inverter running a critical pump from a backup battery bank during power outages. I've been testing the system (and you guys have already seen a few questions from the results) and I've discovered something I'm not sure how to handle. I'm using AGM 12V batteries in series to produce 24V. When the pump hits them, the current flow is so large (C/3 constant, C surge) that it causes about 3V of voltage sag, measured at the battery terminals. This is enough to cause the inverter to freak out and shut off. The instant it shuts off, the voltage rises back to well above the restart point and the inverter comes back on. So I'm at a loss as to how to set the LVD voltage level. It's at 22.0V right now. I could lower it to allow the pump to run without error because the batteries really aren't that empty, but that risks the inverter's self-consumption killing the batteries over the long run. I could leave it where it is (or even raise it) to protect the batteries from the inverter self-consumption but then the pump is going to turn off with the batteries still containing a lot of energy.

1) What is the correct slow-drain LVD setting for AGMs to not hurt the batteries if they hit it while unattended?
2) Is there any solution to the high-drain voltage sag triggering the LVD other than to buy a bigger battery bank?

Comments

  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: Best Setting for low voltage disconnect?
    williaty wrote: »
    I searched in the forum for a good while and I was shocked to not find a clear cut answer to this question. I have an inverter running a critical pump from a backup battery bank during power outages. I've been testing the system (and you guys have already seen a few questions from the results) and I've discovered something I'm not sure how to handle. I'm using AGM 12V batteries in series to produce 24V. When the pump hits them, the current flow is so large (C/3 constant, C surge) that it causes about 3V of voltage sag, measured at the battery terminals. This is enough to cause the inverter to freak out and shut off. The instant it shuts off, the voltage rises back to well above the restart point and the inverter comes back on. So I'm at a loss as to how to set the LVD voltage level. It's at 22.0V right now. I could lower it to allow the pump to run without error because the batteries really aren't that empty, but that risks the inverter's self-consumption killing the batteries over the long run. I could leave it where it is (or even raise it) to protect the batteries from the inverter self-consumption but then the pump is going to turn off with the batteries still containing a lot of energy.

    1) What is the correct slow-drain LVD setting for AGMs to not hurt the batteries if they hit it while unattended?
    2) Is there any solution to the high-drain voltage sag triggering the LVD other than to buy a bigger battery bank?

    You could build an LVD circuit which used an op amp to take into account both the battery terminal voltage and the current (from a shunt or other current detector) when deciding what voltage to shut off at.
    You could also build a delay into a standard LVD so that it would ride through the motor starting surge before tripping at a compromise voltage.
    One way to do this would be to add an input filter to the battery sense terminal leads to the inverter so that they did not immediately see a sharp drop in voltage. Or you could take a conventional LVD which was NOT integrated into the inverter and run the output from it through a delayed-on relay or circuit and run the battery disconnect from that.
    And just to make things more complicated, if the batteries are starting out fully charged, then the coup de fouet effect (Google it) can result in the voltage at the battery initially dropping more than it would for the exact same current steady state. This seems to be a characteristic of lead acid batteries, both flooded and AGM.

    Next: If the voltage is dropping 3V on a 24 volt battery with a current C/3, then the internal resistance of those AGMs is not what it should be. Larger batteries would help. Or do you mean that the 3 volt drop is only for C, not for C/3?
    SMA SB 3000, old BP panels.
  • williaty
    williaty Solar Expert Posts: 60 ✭✭
    Re: Best Setting for low voltage disconnect?
    inetdog wrote: »
    Next: If the voltage is dropping 3V on a 24 volt battery with a current C/3, then the internal resistance of those AGMs is not what it should be. Larger batteries would help. Or do you mean that the 3 volt drop is only for C, not for C/3?
    Near full, they'll drop 3V only near C. At about 75% SoC, they'll drop 3V for C/3. At 50% SoC, they'll drop just shy of 3V even at C/12. I didn't think the voltage drop got worse for smaller draws as the SoC dropped, but those are the results that I got tonight. Yes, the battery bank is extremely minimal for this purpose. Getting twice the capacity is already in the budget, it's just waiting its turn for money.
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: Best Setting for low voltage disconnect?
    williaty wrote: »
    Near full, they'll drop 3V only near C. At about 75% SoC, they'll drop 3V for C/3. At 50% SoC, they'll drop just shy of 3V even at C/12. I didn't think the voltage drop got worse for smaller draws as the SoC dropped, but those are the results that I got tonight. Yes, the battery bank is extremely minimal for this purpose. Getting twice the capacity is already in the budget, it's just waiting its turn for money.
    In that case, the time delay route will not buy you much, if you need to go below 80% SOC. (Also do not forget to take into account Peukert's law, which says that at C/3 your total amp hours and watt hours available will be a lot less than they were rated at C/20. This may be causing you to overestimate the SOC of your battery bank, since you cannot measure the SG of the electrolyte.)
    What brand/model batteries are you using??
    The only other thing I can suggest is the voltage versus current circuit which adds some multiple of the current number onto the voltage before feeding it to the LVCO decision maker.
    Or, going whole hog, use a digital processor like an arduino to actually take into account the battery voltage while the pump is off as an indication of SOC and then change the correction factor for current based on the relationship you have noticed between SOC and internal battery resistance.

    The simplest thing would probably be to just bump the threshold voltage by a fixed amount (using a resistor and a relay) whenever more than a certain amount of current is flowing. That would give you reasonable protection over a fairly wide range of SOC.
    SMA SB 3000, old BP panels.
  • BB.
    BB. Super Moderators, Administrators Posts: 33,590 admin
    Re: Best Setting for low voltage disconnect?

    Are the battery cables heavy enough? I.e., is the 3 volt drop at the battery terminals or at the inverter DC input?

    I would have suggested that, if possible, you play with the starting capacitor (reduce size a bit and see if pump still starts with lower surge)--But that may not be possible if the pump is sealed (or if it even has a starting cap).

    You could try something else--Use a cheap/long AC extension cord and see if the added resistance will reduce surge current (i.e., a 50-100' 16 AWG extension cord).

    But the difference between C Starting Current and C/3 Running Current is not that big of spread (a 5:1 ratio for pumps can be pretty typical).

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • williaty
    williaty Solar Expert Posts: 60 ✭✭
    Re: Best Setting for low voltage disconnect?

    Battery cables are only about a 6' round-trip run from bank to inverter. They're 2AWG. I'm carrying around 70-90A surge (it's quick enough it's hard to catch) and 33A sustained when the pump is on.

    The voltage drop is measured at the battery terminals. The inverter doesn't seem to see anything more than a few hundredths of a volt difference, if that.

    I think it sounds like the best option is just going to be to double the size of the bank and buy higher quality AGMs like Sun Xtender or Lifeline (what's the difference, really?) when I do get to upgrade it.


    If I do go the simple route of just getting more bank to make the sag less, what really should the inverter LVD be set to?
  • BB.
    BB. Super Moderators, Administrators Posts: 33,590 admin
    Re: Best Setting for low voltage disconnect?

    To protect the batteries from over discharge--Probably around 23.0 volts--But if your surge is heavy and the cutout does not have a time delay, then it probably won't work.

    Is there anyway you can justify using a smaller AC (or DC) pump+Motor? I would guess that your pump is not on very long when it cycles (or the bank would be discharged in a few hours).

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • williaty
    williaty Solar Expert Posts: 60 ✭✭
    Re: Best Setting for low voltage disconnect?

    Unfortunately, no. The house it's trying to keep dry receives up to 1400GPH into the basement sump during prolonged rain. The pump will run 30 seconds on, 20 seconds off. At peak flow, the battery system effectively just buys us 10-20 minutes to get the emergency generator out of mothballs and running. The real benefit to it is that on the day after the rain, when the pump still needs to run for about 20 seconds every 15 minutes, we don't have to have the generator on except to re-bulk the batteries as necessary. Then by day 2 or so, the hope is that we'll be running the generator no more than once a day. The goal is a BIG savings in generator fuel and the ability to actually sleep normally rather than have to try to keep shifts 24 hours a day. Long term (hopefully sometime late this year), we're going to add a new Midnite Classic and the 870W of panels we already own to it so the generator will be what gets us through the rain, then the generator bulks the batteries and the solar absorbs them the next day, then after that the solar can handle everything on its own.
  • waynefromnscanada
    waynefromnscanada Solar Expert Posts: 3,009 ✭✭✭✭
    Re: Best Setting for low voltage disconnect?

    30 seconds on / 20 seconds off is one heck of a lot of starts and stops. Hope you have a second pump for backup.
    Second thought that may work - - Two pumps. One about half the size/power consumption of the one you have now, that will ALMOST handle peek flow, and a second smaller one that would kick in when and if needed to handle what the first one may from time to time not be able to handle. It seems reasonable that you wouldn't always have that massive peek flow, especially at the start of a heavy rain, therefore this has the potential to lessen the loading on the batteries, especially that more than once a minute monster start surge they're now having to endure, and still give you lots of time to hope your batteries don't die before you can figure out why you can't get your generator running.
    Just something to think about from a different angle.
  • Ralph Day
    Ralph Day Solar Expert Posts: 1,020 ✭✭✭✭
    Re: Best Setting for low voltage disconnect?

    Just set the float switches (as per Wayne's suggestion) with the weaker pump carrying most of the load (first to switch on) and the "peak" pump set to come on at a water level above the "on" setting for pump 1. If pump 1 is running and can't get ahead, or is overwhelmed, pump 2 will kick on. Ask me how I know...now I have a drain from the sump pit to the ditch. Thousands of $$ for peace of mind. Sump pump only runs when it's checked for operation. 15 yrs and counting.

    Ralph
  • williaty
    williaty Solar Expert Posts: 60 ✭✭
    Re: Best Setting for low voltage disconnect?

    Two smaller pumps is not something I had considered. That's definitely worth exploring.