Bypass Diodes

[Sandman57]

Has anyone hear incorporated bypass diodes into their DYI panels.

I have heard of a number of people using blocking diodes to prevent backwards current from draining batteries at night.

I have also read that comercial panels include bypass diodes. As I understand it if one cell is shaded or short circuits a bypass diode can prevents hot spots when the circuit tries to compensate.

Given DYI soldering I would suspect that one might expect a cell to short circuit or in some way fail over its lifetime and the inclusion of bypass diodes might minimize the impact of this failure on the whole system

below is a link to an explination

http://www.oksolar.com/technical/diodes_in_pv_systems.htm

http://pvcdrom.pveducation.org/MODULE/Bypass.htm

[Anttistaatti]

Very interesting. I think I`m going to include those bypass diodes to my next DIY panel. If I uderstood correct it would be better to use as many diodes as you can. Only bad would be the cost of diodes and more work soldering them. When I tested my first panel I noticed dramatic amp drop when one cell was covered with hand. So that might help this also. Maybe 4 cells per diode would be reasonable?

[Mike90250]

Just remember, diodes have their own liabilities. (another thing to fail)
I really like the article at http://pvcdrom.pveducation.org/MODULE/Bypass.htm :
"The maximum group size per diode, without causing damage, is about 15 cells/bypass diode, for silicon cells. For a normal 36 cell module, therefore, 2 bypass diodes are used to ensure the module will not be vulnerable to "hot-spot" damage."

Mike

[Anttistaatti]

That article is great. I still think it is better to use more diodes than 1 per 15 cells. If just small part of the panel is shaded the diode bypasses all 15 cells it reduces voltage of the panel to useless level (-15X0,5V). If we have for example 1 diode per 4 cells and one cell is shaded the voltage drop is only -4X0,5V and the panel is still working. But I am not sure did I understood that correct.

[SunnyLight]

Number of diodes

Hello Anttistaatti,

you're right to say that it is better for performance reasons to add more diodes, but there are several problems with that. First of all it is quite expensive. A diode which can handle the AMPs is not cheap. Adding a diode every 4 cells (or so) will require a lot of routing via buswire, which is not cheap either, and requires a lot of work.

A diode which is in bypass because one (or more) cells of the string is blocked/shadowed will result in a voltage drop. That is the reason to choose a schottky diode. This drop is on average 0.3V (for schottky diode). Suppose you put a diode every cell and suppose 4 cells are in the shade then this will result in voltage drop of 1.2V, but if you have only one diode per 4 cells then this would only cost you .3V drop. The same applies ofcourse for 18 cells, though it is perhaps less likely that a large number of the cells of string of 18cells in the shadow. On the other hand if you have a diode every 18 cells and only one cell is shaded then you will not only have a voltage drop of .3V but also loose the power which would normally be produced by the cells which are not in the shade (17 * 0.5V).

With this I only want to indicate that even though it looks to be the best to have a diode every cell (something the industry would prefer if it were cheaply possible and preferably be integrated in the cell) but it is not always the best way. And for cost reasons only 1 diode every 18 cells is used.

regards,
Hante

[Anttistaatti]

I`m not sure that a schottky diode is necessary when used as a bypass diode. Ofcourse schottky is good when used as a blocking diode but i would use regular diode in bypass. Those diodes doesn`t cost much and the voltage drop happens only when shadowed. Is the panel voltage drop calculated like (lost voltage of bypassed cells + voltage drop of bypass diode)? For example with 1diode/4 cells (4X0.5Vcells+0.3Vschottky) = 2.3V drop in panels output?

[SunnyLight]

schottky

In general the diode will not be in bypass, that's true, but when it is, you have to realize that it will consume quite some power. A normal diode is not that much cheaper nor does it have much other advantages, but it will have a forward voltage drop of 1 to 2V. If you have a 2V voltage drop then you are not only loosing 2V (which is not that much perhaps) but the diode will also have to handle 2V*3.5A. That means you're burning 7W which will generate quite a lot of heat. The only advantage I can see of a regular diode is the current leakage, but that is in general in the uA.

Your calculation is correct. You will loose the voltage of the cells over which the bypass diode is connected and the forward voltage of the bypass diode.

regards,
Hante

[Anttistaatti]

My new panel with bypass diodes is finished and it works great. I used 8 diodes for 36 cells. You can see it here http://sites.google.com/site/diysola...i/solarpanel-2

[solarrules]

bypass diodes in full light

Does a bypass diode subtract from panel performance during times of full light?

[Mike90250]

Does a bypass diode subtract from panel performance during times of full light?

Generally, no.

[solarrules]

Thanks Mike. That's what I thought but I needed confirmation.

Other than a small cost increase and a small complexity increase, I don't see any reason why a DIYer shouldn't be fairly generous with bypass diodes.

[Mike90250]

My thought too, I'd put them in every 3 volts, that way, if shaded, losses are lower.

Many DIY have no idea how diodes work, how to orient them, and how to conservatively rate them to work with out heatsinks.

[Solmark]

specify diodes

My thought too, I'd put them in every 3 volts, that way, if shaded, losses are lower.

Many DIY have no idea how diodes work, how to orient them, and how to conservatively rate them to work with out heatsinks.

I'm in the process of designing my first panel.

The panel is to be somewhat experimental in design and have decided to include bypass diodes principally due to winter shading of the lower 15% of the panel. Unavoidable thanks to location.

Each panel designed to give a notional 150 watts
using 6"x6" polycrystalline cells 3.5w 7.5a 0.5v (0.6v full sun) each
40 cells total - 5 across, 8 high .

I want to provide a bypass for each of the 8 horizontal strings of 5 panels - running bus bars to a box on an end of the panel where diodes can carry a heatsink if needed, be acessed, either for maintenance in the future or/and to remove them from the circuit for evaluation purposes.

Can someone explain how to assess the rating needed for these diodes, or better still recommend some to me, I do however need to understand the how and why.

I do bricks better than electronics.

Schematic below

[Mike90250]

1) I don't open PDF's from new posters, sorry, too easy to have a dragon lurking in there. I'd open it a week from now, when new AV updates are out.

2) design for shade, will fail. If you over-voltage the panel to allow for loosing 3,6 or 9 volts in shade conditions, then when in full sun, your charge controller will waste the extra voltage, and watts are lost. If you DON't comnpensate by over-voltageing your panel, then when the shade occurs, the 3 volts bypasses, your 17V panel is now 14V and too low to charge a 12V battery (12V deep cycle battery needs 16V to fully recharge, it's not a car battery with brief discharges from starting)

Best is to mount smaller panels, in parallel, so when shade happens, you only loose wattage off your array. PanelDiodes.gif PanelFire.jpg
Click on the thumbnails, and observe the results of plastic & plywood DIY panels.

Panel diodes need to pass the full array amps, so if you plan a 3A array, you need 6 amp diodes. They should be Schotkky Diodes, which have lower Vf and waste less power. You need to research this in an electronics class, as to why, and how, that's beyond the scope of free advice, just enough to get you in real trouble.

Read the earlier posts in this thread, some of which have links to background info on diode applications.

[Solmark]

bypass diodes

Mike, Thanks for such a rapid response, (non dragon gif of said pdf here by the way) pv schematic.GIFnot to mention the heads up regarding ply and plastic, not that I was going down that route.

If I may expand a little on my thinking... (charge controllers were next on my list to investigate and may well deserve some searching on here before I put my foot in it there too)

proposed output of panel was to be, in an ideal sunny cool situation would be 20volts @7.5 amps.
each of the 8 strings of 5 cells contributing 2.5 volts at said current

During the 3 darkest months of the year, when the sun is shining, (rare but appreciated in my part of the world) the lowest 5 cells would be in shadow - not a good thing)

If however they were bypassed I would, or at least could hope for 17.5 volts from the remainder of the panel, enough for a deep cycle recharge.

However, on cloudy days where all cells are getting a diffuse light, the extra cells may help the overall supply to the charger.

Now I appreciate that in high summer I'll be loosing some valuable voltage, but to be frank the power demands on my outbuilding will also be lower, ie less lighting and longer daylight hours to compensate as well.

is this thinking totally silly?
Mark