GHurd controller with LiFePO4 battery

[DamonHD]

Hi G,

Finally that controller you sent me ... uh ... several birthdays ago is going to be deployed.  The crowbar 'emergency' stop on my MotorWind has been operating repeatedly which isn't good for any of the system.  (OTOH I know that the crowbar does work.)

The LiFePO4 battery has a built-in BMS and the entire turbine-and-crowbar stuff is upstream of it along with a motley collection of 12V-nominal PV panels.  I'm going to reset the crowbar (which only sees the turbine) to a little above the Voc of the panels and the max output voltage of the battery (which is a shade under 16V when full/balancing), so somewhere around 17V.  I'm assuming that the BMS just disconnects from the input when the battery is really full.  I guess that I'm going to find out...

Battery voltage chart: http://www.earth.org.uk/off-grid-stats.html

So, downstream of the battery I now want to set the controller to start dumping at ~15V (the output never gets above about 15.5V).

1) I assume that I only need to tweak the 5k pot a little to push the regulation voltage up there, no other changes?

2) Since I will only need to dump a sustained ~20W at most I imagine then the one power FET you supplied should be happy, with a ~10K/W heatsink?

Of course, I had to buy myself a bench power supply for testing (I've been lusting after one for ~25 years!)...

Rgds

Damon

[Bruce S]

Damond;
I'll jump in here just a little.
I use one of these units for my 14.4Vdc NiCd packs set to dump at 16V so you should be okay. I only use 1 FET, but I have it on a heat-pipe unit from an old Dell desktop.
 Hope this helps.
I worried about the upper limit ( getting close to the 24V mark and the Zener's limit) but have not had any issues yet, though I did take a very long time getting it dialed in.
Cheers
Bruce S

[DamonHD]

Thanks Bruce, that's the sort of info I needed.  Soldering iron will probably be broken out tomorrow, though naturally we're not predicted any significant sunshine until I'm parked behind a client's desk for a week!  B^>

Rgds

Damon

[MagicValleyHPV]

I'm assuming that the BMS just disconnects from the input when the battery is really full.  I guess that I'm going to find out...

Most lithium BMS's do much more than just switch on/off pack voltages. Because of the sensitive nature of lithium chemistry, the BMS monitors and controls individual cell voltages, both during charging and discharging.  Without the BMS, you risk losing the entire pack if just one cell fails.

[DamonHD]

Indeed, the BMS is clearly doing some sort of balancing when the output voltage sails up over 15V but no higher.  It's what is happening to the input at that time which is more of a mystery than it ought to be.

I think I am right because I have PV connected directly to the inputs with no other controller (as the guy who sold it to me off his workbench suggested), but since I got zero documentation I'm playing it by ear still!

Rgds

Damon

[MagicValleyHPV]

Cell monitor:

http://epbuddy.com/index.php?main_page=product_info&products_id=27

Are we talking A123 cells?

http://terrorhurtz.com/a123/charge.aspx

I'm using paralleled Dewalt 28v cordless packs in a small trike. These are 3.2v 26650 cells by A123 (LiFePO4). Currently, I'm using the recommended Dewalt charger, but that may change if I decide to disassemble the Dewalt packs and combine cells - which means, of course, I'd need to buy an after-market BMS and charger.

[DamonHD]

Some sort of bespoke prismatic cells in the pack.  As I say straight off the vendor's test bench: was destined for one of his own new products.  Very compact.

Rgds

Damon

PS. Cannot open your top link safely.

[MagicValleyHPV]

try this one:

http://www.hobbyking.com/hobbyking/store/uh_viewitem.asp?idproduct=10952

[ghurd]

I do not know much about the inner workings of BMSs or LiFePO4s,
but yes, just tweak the pot.

LiFePO4s want a tighter hysteresis than FLA like?  Might want to use a higher resistance than 100K for RX to tighten the hysteresis.  Should be a 150K (0.08V), 180K (0.06V), and 220K (0.06V) in the kit.  The hysteresis will be a little different due to being a slightly higher operating voltage.  I don't know which is best.

"the one power FET you supplied"?  Should have been 2?
Anyway, IRFZ44N, yes the fet is fine for 20W or 80W.  Even with 40W and high frequency (tight hysteresis), should not need a heat sink at all if the fet is not potted or something.
G-

[DamonHD]

Hi,

Ah, yes, 2 FETs when I look a little harder!  Even betterer!

I have nice heatsinks for them so I should be dandy.

I think that the hysteresis will be more-or-less irrelevant since over ~14V they're "full" and all I'm doing is bleeding off some excess.  I'll let you know how it works out!

Rgds

Damon

[DamonHD]

I've just adjusted my MotorWind turbine crowbar trip to 18V (by putting a 10k 22-turn pot in parallel with the voltage divider here: http://www.earth.org.uk/wind-power-pilot-autumn-2007-MotorWind.html#crowbar ), ie above max *output*/balancing voltage from LiFePO4 battery and below Voc of the PV inputs, so presumably 'safe' for the BMS input, to try to reduce spurious and potentially damaging repeated operation a little when the battery is full.

Then I'll get the dump controller up and running, maybe over the weekend.  (I have some software and words to write today for my living today!)

I may yet adapt the new timer gizmo to switch my ADSL modem automatically between mains and off-grid to make best use of my energy inputs in summer...  (Or I may do it in software from my USB I/O board; that's a decision for another day!)

Rgds

Damon 

[DamonHD]

Just assembled the controller, G.  Easy to follow your instructions, thanks.  I've tested it (without load, on my bench power supply) at 15V threshold and the default hysteresis.

Going to use a couple of 22R 10W resistors as an initial dump load.  Should be almost exactly 20W total at the target load.  If that's not enough I have 2x12R @ 25W resistors and 2x33R @ 10W...

Probably missed by chance to test this for real today but may have another opportunity tomorrow if we get some sun!

Thanks again,

Damon

[DamonHD]

Wired in!

Went with 2 x 33R resistors for the initial basic dump load as I was getting slightly nervous about how hot the 22Rs were getting given my ramshackle initial build.  Don't ask.  The 33R dump's surface temperature still gets above 100C, but it's only 15W total that they're dissipating, so I'm not too nervous.

If I add more (I have 50W+'s worth to hand), then I'll heatsink the resistors properly and may get a fan or whatever.

Anyhow, this will do to start with, and I think that I may get enough sun tomorrow to see it in action!  Woo-hoo!

Thanks again G,

Damon

PS. If my house burns down tomorrow you'll know why...  B^>

[DamonHD]

Dump load operating from around solar noon but not able to hold voltage down to 15V with ~70Wp+ of PV working, so may need to add extra load. Current load making system warm but not unreasonably so; however any significant extra dump load should be sited better and heatsinked.

Slight panic caused by neighbour apparently deciding that Sunday afternoon is a good time to burn plastic in their garden...

Rgds

Damon

[ghurd]

Slight panic caused by neighbour apparently deciding that Sunday afternoon is a good time to burn plastic in their garden...

LOL

Sounds like you need a lot more dump load.
More dump load will (in a 'normal' system) will reduce the duty cycle so the resistors run cooler.
G-

[DamonHD]

Yes, probably.  But at the moment the worst that can happen from insufficient dump load is that the BMS disconnects its input and the turbine crowbar triggers more than I'd like.

BTW, the current hysteresis seems fine and when the cloud cam across and at dusk I saw sensible cycling on and off (typically a few seconds' on time).

Rgds

Damon