Arduino LVC-DL controller

[Brian McGowan]

I am in need of opinions.
I am making an Arduino low voltage disconnect and dump load controller.
I am sensing battery voltage through a resistor divider. The Arduino takes 5 volts max in as a maximum so my divider happens to work out to 17.1 volts battery equates to 5 volts. Since I have Nickel Iron batteries I am going to set it up to turn on dump loads at 16.5 volts and off at 15 volts. I am going to use .1 volts steps for my resolution. I plan on having 3 low voltage disconnects each turning off at successively lower voltages base on importance. I plan on setting them up so that 4 minutes after I go below a given voltage they shut off and when the voltage gets above a given voltage they turn back on. This will stop them from shutting off when I run heavy loads like my one cup coffee maker or when the fridge kicks in as they do now.

My question is what is the lowest voltage I should worry about so this controller might be useful for more than just me. I was thinking 10 volts might be the very bottom. If you get below that for more than 4 minutes you are either really beating up the battery or it is shot. I can worry about 24 and 48 volt batteries later by changing my resistor divider but for now I just want to get this unit going on my setup and try to account for other battery types.

So what should the absolute bottom voltage be?

Brian

[Julie Haugh]

Your absolute bottom voltage should be programmable, one way or another.

You should also make it so your resistor divider can be jumper-selected for voltage range, if you're going to pre-make them. If you're going to publish a schematic, specify the "top" resistor for different voltages.

If you're making a shield so that Vbatt is fed into a voltage regulator, like a 7805, stick a zener diode in front of the regulator, but after wherever you tap Vbatt for your voltage divider.

[Brian McGowan]

At this moment the divider that I have set up will measure 10-17 volts in 510 steps of 0.0137V/step. I figure that should be enough resolution to determine the battery voltage accurately enough for anything in the 12 volt range.

The Arduino has an onboard voltage regulator and will take a supply voltage of up to 20 volts so attaching it to the battery voltage should not be a problem unless something more catastrophic than the Arduino failing is going on in which case the Arduino failing is probably the least of your problems.

I am still debating sensing battery voltage directly from the battery or from my low current fuse block which comes after a couple of current shunts. I am thinking after the fuse block so I can fuse the arrangement shouldn't make much of a difference for these purposes.

[Julie Haugh]

You should look into putting an 18-20 volt Zener diode before the input to the Arduino's voltage regulator. There are a lot of 24 and 48 volt systems out there. Designing for 48 volts is a lot harder than designing for 24 or 12 volts, because so many parts have Vmax limits that are just too low for a 48 volt system.

Do NOT sense anything without a fuse between it and the source. I had some early prototypes of the part I built go "Poof!" because they were unfused. If you're making PCBs, look at polyfuses. Very handy gadgets, and they are self-resetting.

[Brian McGowan]

I have designed telecom equipment which runs on 48 volts so I have some familiarity with that.
I was thinking of having 2 boards. One for 12V systems and the other for 24-48 volts systems. For those systems I would incorporate a switching regulator that would take up to 60 volts and convert it to 12 volts and let the Arduino board take it from there. Having a jumper for the sensing divider for this board to switch between 24 and 48 volt systems is a good idea.
I was probably going to fuse this anyway. I have fuses all over my system because I am a safety nut. I have also used resettable fuses. That is a good idea.
I am a ways from committing to copper. I have to think about an op amp to convert miniscule voltage from a current shunt to 0-5 volts so it can be sensed. I also want to run a couple of the PWM outputs out to the connection board so they can be used to control solid state relays for variable load control whether AC or DC.

[Julie Haugh]

Brian -

I'll have to get a microscope out, but I've got an SOIC-8 part which is a 20:1 amplifier that will amplify that +/- 100mv output from a shunt with a 1000A = 100mv ratio. I just forget the number.

If you'd be comfy using that sort of part, I also have a PCB layout for ExpressPCB I could convert to something you might be able to convert to something else ;)

[Brian McGowan]

I have no problem using any part.

The shunts I have in my system are 75mV=200A except for the one that came with the Victron battery monitor. I will need to adapt to what is more standard. I think the ones I have are oddball but came with the meters I got on ebay a long time ago.

Let's see if I can get some pics in here.

[Brian McGowan]

It has been a while since I posted and I have some time while on the train and news to report.
This project is nearly complete and after a little testing I wi be ready to implement it into my system. I was having a little trouble with the voltage measurerment but I got that resolved and made great progress last week programming. I only have one more relay left to control then it is all timing adjustments. I am using two relays to control the inverter. One will turn the inverter after the batteries are fully charged for 5-10 minutes and the other to connect the load 20 seconds after the inverter has been started to give it a chance to stabilise before t takes the load. Pogramming for that second relay is the last piece.

[Brian McGowan]

I believe I have completed programming for this device. I got the contactor relay logic squared away and added a second dump load control also.
Operation is as follows:

3 low voltage disconnects 1-3 from most important to least important. Must be below the setpoint voltage for 5 minutes before shutting off to insure it doesn’t shut off when the fridge kicks on or I make a cup of coffee for 3.5 minutes.

Off after 5 minutes below On at
#1 11.6v 12v
#2 11.7v 12.1v
#3 11.8v 12.2v

2 dump loads

#1 will drive the main inverter. On after 5 minutes at or above 16.5 volts with a secondary relay contacting the load to the inverter after 20 seconds to insure inverter is up and stable and ready to take the load. Off after 5 minutes below 15 volts to insure momentary voltage dips or clouds on the panels or me making a cup of coffee don’t shut it off. These relays will drive larger 12v automotive 40Amp relays.

#2 will go on after 5 minutes at or above 16.5 volts if the inverter is already on. This is to make sure there is a place for power to go if inverter loads are not enough to consume the power being made and the batteries are full. Off at 15.1 volts with no delay to insure load is removed before getting to the inverter low off threshold. I have not decided what this load will be yet but I have several options.

My interface board still has 2 relays left which I imagine will go for additional dump loads if I find the two I have are being satisfied and I still have additional power being made. I can’t imagine needing more than 3 low voltage disconnects.

I am waiting for delivery of solid state relays for the inverter load contact and 7809 9volt regulators for power to the Arduino from my battery. These should be here Saturday.

I can’t wait to come home and not see 2 hours of absorb time and 3+ hours of float time and find I have harvested and used 4+kWh and still have full batteries.

[Brian McGowan]

My parts arrived on Friday. I put the 9 volt regulator down last night and tested. Next step is the switching outlet box then install to the system. Testing continues but all appears to be good so far.

[Brian McGowan]

This weekend I got my Arduino Low Voltage Disconnect/Dump Load (LVD/DL) controller installed and today was the first day of totally automated operation and I am happy so far.

Yesterday I managed to weld the contacts of one of the miniature relays closed with the inverter turn on port. I am not sure why. Relay is supposed to be good for 50V@500mA and the inverter port supplies battery voltage at 85ma was what I measured. Anyway I put a bigger relay in the line and that seemed to work today.

I think I uploaded some pictures. One of the charge controller stats, one of the battery state of charge and one of a clock showing how long the inverter was on for today.

I think I need more load. The fridge is next up for a relay box. I case the pics don't come out:

Absorb for 2:01
Float for 2:47
State of charge 99.8%
inverter was on for 6:47:58

I will try to post a picture of the arrangement and a picture of what was there before.

Theory of operation is as follows.

5 minutes after batteries reach 16.5V the inverter is turned on. 20 seconds after that a solid state relay is turned on to connect the inverter to the house.
5 minutes after the batteries reach 15V or less the inverter and the relay are both turned off. The 5 minute delay on the low end is to insure that if clouds or a sudden sag from the fridge starting or me making a cup of coffee happen the inverter won't shut off and then start up shortly after due to the panels still making power.
There is a second dump load programmed in just in case the 1st load is satisfied. It goes on 5 minutes after the inverter is turned on and the 20 relay is turned on and goes off at 15.1 volts without delay.

On the low end the low voltage disconnect relays turn off from least important to most important with a 5 minute delay after falling below the low setpoint which is 11.8V, 11.7V and 11.6V for the same reasons as above and go back on at the high setpoint which is 12V, 12.1V and 12.2V immediately with no delay.

I still have two additional relays to program as I see fit. All parameters can be changed through programming.

I hope this helps me use more power.
Brian

[Brian McGowan]

So it's been two weeks and I thought I would give an update on my system.
I completed my refrigerator relay box shortly after my last update meaning that is now part of the dump load. Subsequent days have yielded 5 hours of inverter run time and 1/2 hour of absorb time. I have another dump load programmed but need to decide what load that should be.
I have had no low voltage disconnects yes as I haven't been below the thresholds for more than 5 minutes yet.
So far the controller seems to be doing exactly what I made it to do. So far though I have not been home on a day where the sun was shining enough to dump power. It has been absolutely dismal here for a long time. I would really like to see this work while I am watching it.

[Brian McGowan]

For your amusement. The system controller is working nicely. In my "pile of stuff" I found an old but brand new old timey wheel type power meter. I decided to install it in my system to see how much power I am sending into the house once the batteries are charged.
And of course it is rotated to the left so you will have to turn your head to see it.