Exactly what IS a bypass diode?

[Jeff7]

I'm wondering about bypass diodes for my prospective solar panels - they're to be 36-cell panels, so 18V max output, at 3 amps. These are of course with the infamous eBay Evergreen cells.

I see bypass diodes mentioned everywhere, but I can't seem to find any for sale.

Is "bypass diode" just a term for any diode used in this fashion - for example, could a 6A05G rectifier be used (50V, 6 amp) as a bypass diode? Or is there a class of diode (zener, rectifier, bypass) that I've not heard of before?

[ghurd]

Just a diode.

They won't do much until a system gets to 36 or 48v.  Most factory types I see are higher V, 600 or 1000V, each bypasses a 6v (18 cell) section of the panel.

G-

[Jeff7]

Won't do much...but would they do something? I just don't like the idea of an entire panel being shut down by a single shaded cell, whether it be from a leaf, a bit of snow, a large blob of something from a passing goose, whatever.

[ghurd]

It really doesn't work like that.  The other cells will 'push' it through.

Shading isn't as bad as it is often made out to be. Usually, if one is shaded they all will be soon. A leaf or 2, or something doesn't have much effect. A bypass diode won't help with that anyway.  'Shading' means shading like a leaf or an overhead tree limb, not duct tape covering a whole cell or 2.

A bit of overnight snow at the bottom melts off pretty fast.  A 1/2" of snow shuts them off like a switch!  But it will again melt off pretty fast, especially if the sun shows up that day.  I try to put panels where a broom can easily get the snow off (not on top of a silo or the peak of a 3 story house).

The standard bypass diode configuration would bypass 6v of a 12v panel. So 2 diodes per 12v panel.  Think about that.  12v batteries need say 14v to charge. The panel is 20v open. The diode bypassed 10v of panel. The diode used 0.7v of the 10v remaining.  That leaves 9.3v (and thats OPEN V) where 14v minimum is needed.  The diode didn't help.

It could help for higher voltage systems like 48V.  Maybe even 36v.  Not often or much, if ever, for 24v (on paper it looks like it could if the SOC is less than 0 and the controller has a low forward V drop).

I have seen one fail to a shorted condition, bypassing 1/2 the panel. Shut the whole thing down. There was nothing wrong with the panel.  The diode caused the problem.  If it was in a 12v, multi panel system, no one would have known this one wasn't working for many years.

G-

[Jeff7]

I'm hoping to hook the panels directly to a grid-tie inverter. Batteries or an inverter is going to be expensive; figure I'll go with the inverter first, and worry about batteries later.

As for the bypass diode - I did a test on a small array of broken cells. It pushes 2.2V @ 1.5 amps. With one cell shaded, it barely manages 200mA.

I just did some tests on my little array, with 2 2N4001 diodes wired in parallel (to get a 2 amp rating):

- In sunlight (not facing directly, and through glass):

2.1V, .5 amps.

- With the diode in place, weakest cell bypassed:

2.0V, .75 amps

- Without the diode in place, without the weakest cell bypassed, with the weakest cell shaded:

1.9V, .015 amps

- With the diode in place, with the weakest cell bypassed, and with the weakest cell shaded:

2.0V, .7 amps.

So just bypassing the weak cell boosted the amperage; no idea what that means. I guess its just considerably lower than the others, so the diodes began to conduct. (?)







But the one result there, without the bypass diode, and with just one cell shaded, the VOM hovered between .02 and .01 amps. With the bypass diode, output was back up.

As for an easy location to get snow off of...unfortunately, the south side of the house is also the side farthest off the ground. Putting the cells on some kind of stand in the yard would be pointless - we have woods right behind the house that keep the yard shaded almost all the time during the winter; the roof however is high enough that it gets decent sunlight.

[Jeff7]

Well there's the numbers; comments, questions, answers, musings?

Also, I was originally planning to just hook a bunch of 12V panels in parallel, but I'm realizing that that would put pretty high current demands on the hookup wire. That's something else I was never entirely clear on; for example, would 12V @ 4 amps need a thicker cable than 24V @ 2 amps, even though they're the same power?

[pyrocasto]

Yes Jeff, if you have 12v @ 10 amps, you would loose twice as much power(I think) as a 24v @ 5 amps system. Or you could run the wire twice as far for the same efficiency.

[ghurd]

Worry more about amps than watts. So yes, 4a should have a thicker cable than 2a.

For some fairly complete explainations, Google 'solar wire loss chart'.

G-

[Jeff7]

Well heck, the panels are going on the roof, and even if it's at the opposite end of the house from where the mains tie into the breaker - the house is probably only 50 feet long anyway, so there's really no "long distance" runs here.

But still, it looks like using a 24V setup would definitely be better. I still am unsure of the bypass diode situation though, due to the numbers I recorded with my VOM.

Also, grid-tie inverters - do they necessarilly need to be phyiscally located at the breaker box? Could I have the inverter mounted under the roof right near the panels, and then run the 120V output from there to the breaker? Very minimal losses there, even if I do just use 12V.

[nothing to lose]

"Also, grid-tie inverters - do they necessarilly need to be phyiscally located at the breaker box? Could I have the inverter mounted under the roof right near the panels, and then run the 120V output from there to the breaker? Very minimal losses there, even if I do just use 12V."

That will probably depend what you have for the inverter. I had read about those solar gorrilas as they call them and checked it out. Those guys were setting up solar pannels and a 100watt grid tied inverter. Basically just mount the panel and plug inverter into a wall. I checked into that as I was kinda thinking about it then decided it was nothing but a joke paying out nearly $1,000 for a 100watt grid tied inverter and the panel to power it. 100 watts, big deal, durring the entire day it might make enough power to run someones bathroom light while they poop

 Better uses for $1,000

Maybe prices came down since then I hope!

Anyway a grid tied inverter like those types you just plug into a wall outlet anywhere. I am sure there are others about the same way. Most decent ones you probably would want  closer to the mains box.

[ghurd]

For a bypass diode you could get fancy and use a Schottky. They have a lower forward voltage drop. Either way, use a diode double the amps and a high reverse voltage.  I would use a 6a 600v with 3a PVs.

Another thing to consider for the wire.  

The peak output will be for a relativaly short time.  If you go for the 2.5% loss charts wire, you only loose 2.5% for a short time, the rest of the time the amps are lower- lowering the losses.

With a 5% wire loss, the loss is only 5% for a short time too. Most of the time it is probably less than 2.5% anyway ().

To me, it comes down to cost.  Say 4 PVs at $350 ea.  It the wire cost is extreme to get a step bigger, that money could be used for a good chunk of another PV.  The extra PV with a 6% loss is better than big dollar wire with a 2% loss.

It isn't that big of a deal for what I do. 12ga is usually good. The biggest I ever needed was 6ga anyway.

G-