Solar Panel Disconnect (controller)

[commanda]

This is my Solar Panel Disconnect unit under construction. Eventually it will contain 2 identical circuits, but only the left one is going at this stage. Been playing with it all afternoon. It contains 4 x IRF540 fets, an MBR4060 schottky diode, a 25 amp circuit breaker, a 25 amp ammeter (haven't printed a meter face for it yet, but it has been calibrated), a 50 amp, 50mV current shunt for the ammeter, and a big double pole, double throw, center off toggle switch. The switch gives me Auto, Off, and always on.

Each circuit is driven by the first 2 outputs from the dump load controller I described back in my diary some time ago.

I'm testing it with a 24 volt, 20 amp, switchmode power supply in place of the solar panels I don't have yet. That's it on the right in the first photo.

The multimeter with the yellow lead is monitoring the temperature of the heatsink. It's showing 29 degrees C. The total voltage drop across the circuit is about 2 volts at 25 amps.

Under the meter is the hand-drawn circuit and a printout of the pcb layout.

The rear panel has 4 sets of 35 amp binding posts for the 2 solar panel arrays, and 2 sets of cabling to the batteries.

I built this.

Amanda

[oztules]

Hi Amanda,

Just a question.

Rds on for 1rf540 is .077r ..... so 4 in parallel should be about .02r.

At 25A this should only lose about E=IR = 25x.02 or about .5v for around 12W loss. 2v@25A gives you a loss of 50w... what am I missing here... or put another way ... why have you chosen to lose 50w of your valuable solar input.

Where did the other 1.5v get lost in the innards?

Would a relay mitigate the loss better than the mosfets, or is there something else I failed to grasp.

PS, feel free to mess your desk up some more, then it will resemble mine (TIC)

........oztules

[commanda]

That 2 volts is total loss. Some is the fets, most is the shottky, some is wiring.

It needs more work, yes. But it is amazing how much is dropped across a few inches of wire at 25 amps. Especially for someone who's used to working with milli and micro amps.

Amanda

[DamonHD]

Amanda: I know what you mean about being used to milli- and micro- (and even nano-) amps.

I still think of anything over about 10mA as ... um ... shocking (well, it might be if it were over 9V)...

Rgds

Damon

[oztules]

Hmmm...

I'm guessing the shottky is used as a blocking diode.

If so, perhaps an array of buk455 50a (rdson .038/4 = .0095r loss in watts  approx 5W. If this array is switched via an op amp set to look at the differential across this array, and set to turn off if vsolar=< v output and above, this should get rid of the shottky.

The irf544's can be replaced with similar fets, and overall losses from the solid state section would be <10w

I haven't thought this through, (I'm off to "town" now), but this may be a more efficient way to go. There are plenty of better fets than the buk455's, but I happen to have a lot of these...

oztules

[commanda]

The schottky is the blocking diode, yes.

In practice, I expect the unit to run closer to 15 amps than 25. This means an array in the 300 and something watt range.

So, I beefed up the main current carrying tracks on the pcb with solder overlay. Changed a couple of component values to get the Vgs to 15 volts rather than 11 point something.

Now, at 15 amps, I have a total drop across the unit of 1.25 volts.

The schottky diode drops 0.42. The fets drop 0.525.  The other 20% is all wiring and contact losses.

Amanda