DsPIC - Battery Monitor

[GarethG]

2 of our batteries had to be scrapped because of improper care. Our battery management was OK but it could have been better. The current battery management powers up and holds a large 24 volt relay closed (not good for power consumption). It then monitors the batteries, when the general output voltage drops to low it turns the relay off.

What we really need is more comprehensive monitoring of the battery state. We need

• Monitor voltage of the entire battery
• Monitor the current with something like this http://onecall.farnell.com/texas-instruments/ina138na-250/ic-current-shunt-monitor/dp/1564888

Communicate this over RS485
This module will advise the pilot (high level mission planner node) that the batteries are getting low. If the pilot decides to shut the power off then we can do so.

Any ideas on how to shut the power off without using too much power to keep it on in the first place?

[MrFoamy]

Probably something of worth here: http://www.swarmrobot.org/PowerManagementBoard.html

Also here although i've not read it all yet: http://www.societyofrobots.com/member_tutorials/node/302

And this looks like it has detailed lipo safety information: http://www.rchelisite.com/lipo_battery_charging_and_safety_guide.php

[GarethG]

These all look like good sources

Before we were just using a potential divider to measure the voltage output of the battery, obviously the big advantage there is it cost barely anything to make. However if you think an actual IC meant for this sort of thing would be more useful then by all means put a bill of materials together and we'll get something ordered

[alexsleat]

Might be something useful in here, it's for USB at 3.5-6v but yeah figured it was worth a link

http://hackaday.com/2012/01/28/lipo-charging-circuit-tutorial/

[MrFoamy]

How much current does the relay replacement need to handle?

[GarethG]

The Logic is somewhere in the region of 1.5Amps.

The Motor channel is about 4Amps when all the motors are going at it

[MrFoamy]

A possible relay replacement? http://onecall.farnell.com/fairchild-semiconductor/fqp7p06/mosfet-p-to-220/dp/SC08203

That's just one randomly picked one, there are loads more if it is not suitable.

[GarethG]

Looks good

Got any ideas coming together? The quicker we get this done the sooner you can move on to some more ROS stuff if you fancy

[MrFoamy]

Yeah i've got a load of ideas. Just got a couple questions before i start working on a design:

1) When you say "monitor voltage of each cell" do you mean each cell of each lipo pack or each lipo pack as a whole?

2) I guess that same question for the whole battery if you mean a collection of lipo packs and then also how many of them will there be?

3) You mentioned temperature monitoring earlier, was that a requirement for this?

4) Do you know what the 'MC34063A' is being used for on the schematic you sent me?

5) Encase the pilot does not listen to the low battery warning, should there be a certain point where the battery manager shuts off power by itself?

6) Is any logging or sending of battery status data required?

Thanks

[GarethG]

Right I think that we need to tone this down a bit, I wrote it just after we had a battery fire in the lab so I was a little paranoid. And I would rather this done as simple as possible and then you can get one with some coding.

There is one battery per supply, they are in separate boxes on the sub. 2 Batteries in use at any one time on the sub.

• Measure the main output voltage of the battery, don't bother with individual cells, you only need to check this when charging and we have a tester for that now.
• Don't bother with temperature.
• MC34063A if memory serves that is there because it was a 24 volt relay with a 12 volt nominal coil voltage so they had to boost the voltage from 5 to 12 in order to trigger the relay
• Send the battery voltage over the RS485 bus. Dan is working on the PIC to ROS interface so you will have to work with him to know what you need to send.
• Send the voltages to the pilot, which will determine if it needs to surface or not. The PIC will decide when to turn off the power. We're reducing the messages and complexity here, who cares if the sub surfaces but still drawing a little bit while its standing still.
• No need for logging on the PIC, we can do that easily as a function of ROS called bagging

[MrFoamy]

Ah right, so each battery is on a separate circuit and so the monitor will need to handle them individually? What voltage are each of them running?

Also if the sub is just going to surface and not cut off the batteries then is the relay or replacement necessary?

[GarethG]

Each battery is in its own little yellow box on the sub.

Hmm that's a good point, come to think of it we never got to a point where the batteries had to be shut off and not having to worry about that would save quite a bit of power.

Use this reg http://onecall.farnell.com/jsp/search/productdetail.jsp?SKU=1696320

its a dc/dc converter so really efficient, noisy as hell though, but it doesn't matter in the battery compartment. use plenty of decoupling caps and consult the datasheet for any design considerations

[MrFoamy]

I've got a prototype comparator battery monitor circuit now. just got to program and test.

If that goes well then i've selected a nice little 8 pin PIC12 with two comparators on it so i'll order some samples and get to work on the final board.

[GarethG]

Cool, I'm not sure where the RS485 bus is but if you work the premise that each monitor will receive a request for the voltage over serial and then send back the current voltage. That will be awesome.

I'm trying to collate some work that's been done on the sub in the last few weeks, if you have any schematics or just a photo of the breadboard that would be great.

Good job