Bypass Diodes

j2mc

I have acquired some solar panels that have some blown and shorted bypass diodes. The diodes I pulled out are 20a and 40v, but I'm struggling to find any diodes that match. Does anyone know a good place to find bypass diodes? Do I need 20a(on ebay I'm able to find some up to 15a), my panel's short circuit current is 8.71a and open circuit voltage is 44.7 and has 3 diodes installed. I'm planning on using these with micro inverters is it still important to get replacement diodes, or are hotspots only a real problem on strings?

Thanks!

mike95490

Re: Bypass Diodes

I'd skip ebay and go to a legit electronics vendor, digikey, mouser, newark. Get first run parts, and install them with proper heatsinking, and they should last longer than originals.

j2mc

Re: Bypass Diodes

Mouser and Newark were the first places I checked but I am having trouble finding any matching diodes, 20a 40v axial style, this is the closest I can find:
http://www.mouser.com/ProductDetail/Vishay/VSB20L45-M3-54/?qs=cwrkoD4LB%252bOFqnVF3xTraw%3d%3d

Problem is lead time is 21 weeks and minimum quantity is 1050. Similar for the 15a, which is what brought me to ebay, so far that is the only place I can find them in smaller quantities up to 15a. Besides the one above I haven't found 20a axial style anywhere. The junction box on the panel is pretty tight so I don't think I can make any other style work easily.

mike95490

Re: Bypass Diodes

Wow, there used to be more options. you can use larger spec diodes. 20A is pretty high for axial leads, no way to dump the heat. You could look at the stud mount styles, use a ring terminal for the bolt connection. And try for Schottky diodes, less loss. http://en.wikipedia.org/wiki/Schottky_diode

here's 3 @ digikey
VT4045BP-M3/4W
STPS30M60SG-TR
SK1545-TP 88 devices found in stock (likely at least 20 in small lots) You likely don't need the full 20A, but you do need some heat dissipation, or the diodes will fry

Cariboocoot

Re: Bypass Diodes

j2mc wrote: »

I'm planning on using these with micro inverters ...

Well that's a "STOP!"
Micro inverters means grid-tie.
Grid-tie means panels need to be certified.
Repaired panels are not certified, especially if repaired with unoriginal parts.
This application would technically be illegal.

Such damaged panels are best left to the realm of experimentation.

There's the warning. What you do next is up to you.

j2mc

Re: Bypass Diodes

Cariboocoot wrote: »

Well that's a "STOP!"
Micro inverters means grid-tie.
Grid-tie means panels need to be certified.
Repaired panels are not certified, especially if repaired with unoriginal parts.
This application would technically be illegal.

Such damaged panels are best left to the realm of experimentation.

There's the warning. What you do next is up to you.

Using that logic I would need a new engine every time a spark plug was worn out! The panel's manual even lists the diode as a field replaceable part, and this panel is designed for grid-tie systems. I can understand that repairs that require breaking the weather sealing would be a big no with grid tie systems, but swapping out the diodes is as easy(or easier) than replacing a pv lead on these panels.

You are right though, I should be using the same as the original part, which I have yet to find, hence my post here. I have also contacted the manufacturer, but have not heard back yet.

Cariboocoot

Re: Bypass Diodes

j2mc wrote: »

Using that logic I would need a new engine every time a spark plug was worn out!

Uh, no; that would be "authorized service".
Cars don't require UL certification either.

Really the panels should be covered under warranty because it's pretty difficult to abuse them in such a manner that would blow those diodes (although possible). Your contacting the maker is the best course of action. Hope they are still in business (something that happens too much in the industry these days).

jcheil

Re: Bypass Diodes

I thought most modern (non 12v/24v panels) didn't even have diodes in them any more? That was all handled by the CC.
Just curious why they would still put diodes into newer panels (if they do). I never opened the box on mine to see.

BB.

Re: Bypass Diodes

Blocking diodes are there (generally) to prevent back feed from the battery bank at night. Since almost all modern systems have battery chargers/controllers which prevent back feed, there is no reason to have blocking diodes anymore.

By-pass diodes are there to prevent "dark/shaded" cells from getting more than ~12 volts or so voltage across them. Solar cells are just very big diodes and when dark go high resistance and look like a reverse biased diode. When you have other panels in series, you can get >>12 volts across these dark/shaded cells and over voltage them, and ruin them.

You have by-pass diodes across every ~36 series cells to allow the current to bypass the dark cells.

Diodes generate a lot of of heat (voltage drop * current = ~1 volts * 7 amps = 7 watts per operating diode)--Solar panels are already very hot under the sun, and have very poor thermal transfer (glass top, plastic backing, plastic j-boxes)--And I always wondered how long by-pass diodes would last if used as designed.

I am very surprised we don't hear about more failures (diodes can fail shorted or opened).

-Bill

jcheil

Re: Bypass Diodes

BB. wrote: »

I am very surprised we don't hear about more failures (diodes can fail shorted or opened).
-Bill

Now, the interesting question, how would you know if one blew on (for example) a GT panel like mine that has 108 cells?
I am guessing there would be 2 (or 3) diodes in there on each panel?

Cariboocoot

Re: Bypass Diodes

jcheil wrote: »

Now, the interesting question, how would you know if one blew on (for example) a GT panel like mine that has 108 cells?
I am guessing there would be 2 (or 3) diodes in there on each panel?

As Bill said they can fail either way. If it shorts you'll see a drop in power production. If it opens you won't notice a thing because most of the time the bypass diodes do nothing. Guess what? They can short, get subject to the current from the rest of the cells in their weakened state, and then open.

How many diodes are included depends on how they laid out the cells. Having seen many variations for the same number of cells I wouldn't hazard a guess how they did it with 108 of them.

BB.

Re: Bypass Diodes

Open by-pass diode-No easy way in an operating string... You could try "shading" parts of each panel. An open diode string would go high resistance" and the string current should drop dramatically and/or Vmp-string would drop dramatically when the small shadow covers cells protected by an open by-pass diode.

Shorted by-pass diode-String voltage would be down by ~12 volts or so, and/or drop in string current.

You may see browning/delamination over the panel J-box if you look closely (excessive heat from shorted diode and/or excessive shading on panel???).

-Bill

jcheil

Re: Bypass Diodes

BB. wrote: »

Open by-pass diode-No easy way in an operating string... You could try "shading" parts of each panel. An open diode string would go high resistance" and the string current should drop dramatically and/or Vmp-string would drop dramatically when the small shadow covers cells protected by an open by-pass diode.

Shorted by-pass diode-String voltage would be down by ~12 volts or so, and/or drop in string current.

You may see browning/delamination over the panel J-box if you look closely (excessive heat from shorted diode and/or excessive shading on panel???).

-Bill

[trying to understand] So if, for example, 4 parallel strings of 3 panels in series with a VOC of 33 each (for example), and a diode failed in a way that dropped 12v or so off "one" of the panels in "one" of the series strings, would that then cause the "entire" array to start operating at the reduced voltage of 87 instead of 99v (again, just throwing numbers out there)?

BB.

Re: Bypass Diodes

Think of a bunch of series D cell flash light batteries--Each which has a by-pass diode. Short one of the batteries, the string will loose 1.5 volts for every shorted battery.

Short a ~18 cell bypass diode, then you short out ~9 volts on that string.

That will reduce Vmp-string by ~9 volts... So that string will probably have lower output current when compared to identical parallel strings.

If this was a Vmp=18 volt array--The string with the failed diode would see zero amps.

If this was two Vmp=400 volt strings (such as a Grid Tied inverter), then 9 volt drop will probably only result in a small amount of current loss relative to the other string (Vmp 400 volts in parallel with Vmp 390 volts is almost a "no difference" situation). Vmp-array might drop a bit and be at Vmp-array 395 volts... Have not thought out the details.

And this gets to one issue I have between a 48 volt battery bank and a 12 volt battery bank.

If you have a shorted cell in a 12 volt bank, you get 10 volts--A vary obvious failure and most DC devices will not operate.

A shorted cell in a 48 volt bank means 46 volt nominal. The system will continue to operate just fine.

That is one reason I suggest weekly (or at least monthly) carefully checking the voltage of each battery (or cell) with a DMM. On a 48 VDC battery bank, it is pretty "easy" to miss a weak/failed cell in long strings.

And another reason I have not pushed 100-380 volt nominal battery banks--I would think a per cell (or, at least, per battery) voltage monitor would be needed to ensure that there are no shorted or open cells in a string. You would have a huge "nest" of extra wires (more dangers) or a bunch of vampire powered voltage monitor circuits.

And issues like this are why solar arrays are difficult to "fuse/breaker/arc fault/ground fault" protect/self diagnose. When you have huge numbers of low voltage (0.5 volt) cells and electrical junctions/wiring in series--Almost any failure that can happen will happen--And will punch holes in your safety systems.

-Bill

Cariboocoot

Re: Bypass Diodes

jcheil wrote: »

[trying to understand] So if, for example, 4 parallel strings of 3 panels in series with a VOC of 33 each (for example), and a diode failed in a way that dropped 12v or so off "one" of the panels in "one" of the series strings, would that then cause the "entire" array to start operating at the reduced voltage of 87 instead of 99v (again, just throwing numbers out there)?

No, what would happen if you lose a segment of one string is that it's Voltage would be brought up by the parallel string(s) causing its current to nose-dive because it would be forced to operate above the 'new' (failed) I*V point.

Like paralleling severely mismatched panels: put a 17.5 Vmp panel in parallel with a 35 Vmp panel and the output will be around 35 but the lower Voltage panel ends up contributing nothing because the Voltage is held above its Vmp point (it 'thinks' it is on an open circuit). Variations on this theme according to how many panels and strings and what the I*V characteristics of each are.

So if one segment fails (shorts) on one panel in a string that is parallel to other strings you may not notice the drop in power at all as it would look much like the loss from a passing shadow - except it would be consistent and persistent.

j2mc

Re: Bypass Diodes

Cariboocoot wrote: »

Uh, no; that would be "authorized service".
Cars don't require UL certification either.

Sorry, wasn't really trying to argue the certification with that statement, just the fact that a 70 cent part would effectively require the purchase of an entirely new panel.

j2mc

Re: Bypass Diodes

BB. wrote: »

I am very surprised we don't hear about more failures (diodes can fail shorted or opened).

I'm curious about this as well, how many panels out there are running with failed open diodes, and the owner has no way of knowing. Since I know I have at least two panels with shorted diodes and one with an opened diode I was planning on replacing all the diodes just to be safe. Seems safer than unknowingly running a panel with an opened diode.

Getting back to the more technical aspect of this thread, how is the minimum diode rating for a panel determined? I assume they have to be oversized quite a bit for the excessive heat they encounter. Does all possible string configurations play a role in that calculation or do they only need to be sized for that string of cells?

RCinFLA

Re: Bypass Diodes

Bypass diode are really for safety to prevent the possibility of excessive heat build up on a cell that becomes reverse biased during shading.

In normal forward biased (full sunlight) a cell will have about 0.7vdc max. A spot parallel shunt defect of less then an ohm will create a little leakage loss but relative little loss to it rated output current in normal operation. If the cell becomes shaded in a series string of cells then the reverse bias across the cell can be a reverse voltage almost as high as the total panel voltage (30+ vdc). The defect now has a high voltage across it fed by the current of the other illuminated cells in the panel. This causes a hot spot that can melt solder or worse. There are examples on this site of homebrew panels that catch fire.

Commercial panel manufacturers require good cells from supplier or grade them themselves weeding out the really bad ones for shunt defects. These rejects shows good voltage and current under normal illumination but under reverse bias can get very hot. The rejects usually end up on Ebay which is one of several reasons not to try to build your own panel for more then a science experiment.

Bypass diodes cap the reverse voltage possibility. Usually the panels are arranged in down and back row pair series connected cells, maybe two or three pairs of rows. Each of these down and back rows are bypassed at the junction box at one end of panel making connections convenient. A bypass diode should be placed across no more then twenty series connected cells to limit the maximum reverse bias any single cell might get.

Now for diodes. They should be PIV rated higher then total Voc of panel with current rating greater then max current capable of panel cells with extra margin. The current handling rating is not the whole story. A partially shaded panel with 8 amps of Isc can drive a bypass diode into conduction with the full current capable of the panel Isc. The bypass diode forward conduction voltage with current at 8 amp might be something around 1 volt. That means the diode has 8 watts of power dissipation it must be handling. An axial leaded epoxy packaged diode will get blistering hot. This is the primary reason for bypass diode failures. I believe TO220 packaged diodes with some heat sinking capability is a better solution over axial diodes, if the junction box has the room.

If you get shading regularly, at a particular time of day, the bypass diodes can be severely heat stressed every day without some heat sinking.

PNjunction

Re: Bypass Diodes

I accidentally blew out a bypass diode from a custom-matched encapsulated pair in my Sharp 80W panel terminal box. I was making some temporary connections, heard a quick snap, but everything seemed to work ok.

Voltages, etc were all fine as before, but the current output of the panel was half of what it should be when I was in bulk charge under the same sun conditions.

Inspecting the diodes, sure enough - one was gone. If I hadn't noticed the drop in current, I could have easily overlooked it if I relied solely upon voltage monitoring. 10A schottky diodes are in there now as a temporary measure until I can get a Sharp replacement.

inetdog

Re: Bypass Diodes

PNjunction wrote: »

Voltages, etc were all fine as before, but the current output of the panel was half of what it should be when I was in bulk charge under the same sun conditions.

That does not make a lot of sense unless you have one or more bad cells in the area of the panel covered by that bypass diode or else there was partial shading of that area during your test.

A shorted bypass diode will cost you voltage, both Vmp and Voc, but an open bypass diode should have no effect at all as long as all of the cells are able to produce close to the same Isc. Or to put it more precisely, as long as no cell's individual Isc is below the overall panel Imp.
With one bad cell with Isc right at panel Imp you would lose the .5 volt contribution of that cell from the total voltage at Vmp.
And with one cell with Isc below panel Imp, you would see a drop in the panel current if no bypass diode available or a drop in Vmp at the same Imp, losing the entire Vmp contribution of the section being bypassed.
As long as the bypass diode is OK, both the Voc and Isc values will be exactly the same as for a normal panel, but you may find that the Vmp is a lot lower than it should be when current is drawn at the nominal Imp.
Voc and Isc testing alone is not a guarantee of the health of a panel.

j2mc

Re: Bypass Diodes

Did a little test with two of my panels(both reacted exactly the same):
No shading on panel with 1 shorted diode, Voc was down ~12v.
No shading on panel with opened diode Voc was at spec, Isc was at spec, however when 1-2 cells were shaded(in section with opened diode) Voc dropped 1-3v, but Isc dropped dramatically ~80%.
Same test with good diodes, same 1-2 cells shaded, Voc dropped again 1-3v, but Isc only dropped ~15%.

Cariboocoot

Re: Bypass Diodes

j2mc wrote: »

Did a little test with two of my panels(both reacted exactly the same):
No shading on panel with 1 shorted diode, Voc was down ~12v.
No shading on panel with opened diode Voc was at spec, Isc was at spec, however when 1-2 cells were shaded(in section with opened diode) Voc dropped 1-3v, but Isc dropped dramatically ~80%.
Same test with good diodes, same 1-2 cells shaded, Voc dropped again 1-3v, but Isc only dropped ~15%.

This is just as should be expected.

mattyp

I have a panel that has blown its blocking diodes twice.
The history is.
1 installed 200w 12v panel connected to a 30a contoller and 2 x 12v 92ah batteries.
2 the installer made a mistake and reversed the charge, instantly blowing the blocking diodes, the charge contiller was in the middle but did not prevent this.
3 replace the diodes with the originals 10a from a new panel
4 ran with one panel for one month, no issues
5 installed the new second panel, but with 8a diodes (only ones i could buy)
5 on day one noticed the 14.4v going to batteries despite cc bring set to limit to 13.8v
Tested 23v at out of panels.  I partially covered each panel to reduce the output, this worked and output went to 13.7v.  
6 on day 2, before it got too sunny i replaced the contoller with a 40a mtt.  System ran well and had upto 15a output
7 on day 3 noticed output was low 4-8a despite sunny.  Tested panels and original panel was down to 12v. 
Presumably the diodes have blown again.

Should i replace the diodes or the panel, or the new contoller!!  I thought the contoller would prevent this, perhaps the new one is faulty?  

mike95490

First - they are bypass diodes , NOT blocking

The controller is supposed to provide the blocking function for nighttime discharge, not reversed polarity

The replacement diodes you obtained were most likely junk diodes and failed quickly

jonr

Consider this one:    https://www.ti.com/lit/ds/symlink/sm74611.pdf

BB.

When a solar panel is connected "backwards" to a source like a large battery bank (and, in your case, the PWM charge controller let battery current through)--The solar cells, just "giant" diodes were "forward biased" and let >> Imp/Isc through the panel. This should not only fry the bypass diodes, but the solar cells too. Any series fuse/breaker is probably not fast enough to protect the reverse connected panel.

With a solar panel connected "correctly", the bypass diodes should not conduct any current except when one or more cells are shaded (bypass diodes are there to allow current "around" the "dark cells", rather than exceeding their reverse voltage rating of ~12-20 volt has or so).

If you have random shading on the solar panel during mid day (lots of sun on panel) from leaves, vent pipes, etc., and you have multiple panels in series, then the bypass diodes can allow current around the dark cell(s). And, at that point, if the bypass diodes are too small, they could overheat and fail (generally fail open, but can fail shorted too).

Remember that the maximum current/power from a solar panel could have been "throttled" by the charge controller doing its job... A charge controller only takes all avialable power/pushing maximum current to the battery bank if the bank needs charging (Bulk mode charging). Once the bank is over ~80% state of charge or so, the bank should be at the controller's voltage set point, (i.e., 14.8 volts for a 12 volt lead acid battery bank) and the battery+controller naturally limit the charging current to the battery bank (Absorb charge mode).

Of course, once the battery bank is full, the charge controller will drop charging voltage to "Float mode" of ~13.6 volts--And the battery will take very little charging current (usually much less than 1% rate of charge). At this point, the controller/solar array will only supply a little current to the battery bank (float charge) and supply current to any DC loads.

So--Looking at the charging current (4 amps, 10 amps, etc.), you need to know what the battery state is (is it charging and under ~14.8 volts--or whatever your set point is), and the charge controller is in Bulk, not Absorb or Float, mode charging. Before you decide to start replacing components.

You need to look at Vpanel voltage, Vbatt voltage, and Ibatt (battery current) at the same time to figure out if everything is really OK, or you have a bad panel or charge controller... You have many "confounding" issues--Possibly damaged solar panels... Not knowing enough about the operating state of the charge controller, and the battery bus voltage/state of charge/any DC loads...

From your history--I would suggest if you are having problems (full sun, low charging current, battery bank/charge controller in Bulk charge), that a new set of panels may be needed--They have been damaged, and "diodes" (many types of electronic/silicon devices) are notorious for failing down the road after being stressed.

If your bypass diodes failed shorted--That would have protected the solar cells and immediate reduced the panel's output voltage and/or current. If the diodes failed open, then the solar cells are not protected, and the solar cells could have been easily damaged. Open bypass diodes and "good" solar cells--The output of the panel is not affected.

I am not sure that your MPPT charge controller needs replacing at this point. But this is all guess work.

-Bill