Diodes and bypass diodes

[john p]

This advice I gve you have to be careful with ok? it is far better if you can mount the panels where its IMPOSSIBLE under any circumstance to ever have part of the panel in shade.. if you do this there really is no need for any diodes.., if also you very careful in wiring them up there is no need for any output diode either, as all chargers have the ability to stop back flow to the panels after sun down..
diodes are only really to protect back flow of current. but if panels designed and placed well there simply is no need, they just reduce output by .7v. for standard diodes..

its like many power supplies have a output diode to protect device in case you connect the input wires the wrong way.. they serve no other purpose

[martinjsto]

thanks John,
no wories with the safety, thanks for the thought though. I have a restricted electrical licence and although im of the thought that a little knowledge is more dangerouse than none, I am not ready to fry myself for a solar panel. I am 47 now and worked on 240v Appliances so i will take precautions, thanks anyway for the warning. I have a cover for the panel so when im wiring, it will be fully covered and should not be producing any power.
my shed on my block is nessled amongst trees, beautifull but unfortunatly there is no place where i get direct sunlight all day and the nearest padock in the open means major cable runs and size. shame there not A/C panesl.
so diodes looks like my best choice, I have 4 80w panels set up already at the moment but now i have learned about mismatch and hotspots from shadows i will be re-wiring them to include bypass diodes in the array.
should get my cells today, 6x6 4w 8amp monos, after 2 weeks researching i am itching to get started. i will open a new post with pics etc of my method soon.

[john p]

ok it seems you just going to have to use diodes ,, you will have too masny problems if you dont . shade and solar panels are a very bad mix

[RifRaf]

i am at a loss to find any info on how i will lose 50% from using a matched diode pair in to-220 package, from that assumption it would be just as useful to not even connect up the second leg. john if you could please show some real examples or results from experiments, or calculations i can make using values from the datasheet linked to in a previous post that show this 50% loss, am about to waste a few hours to prove to myself that it will be ok using an ammeter connected to each lead of the diode directly and sending some large currents though it. I understand what you are saying for an unmatched diode pair, but not for 2 identical diodes in the same package that are already sharing one lead.

this will be the test jig, http://rifraf.dyndns.org/solar/36-diode_test_jig_1.jpg http://rifraf.dyndns.org/solar/37-diode_test_jig_2.jpg http://rifraf.dyndns.org/solar/38-diode_test_jig_3.jpg the wiring is 25A+ the meters are as close as i can get em, the meters are 0.003R max so hopefully an acceptable test unit

[john p]

Rif raf I dont have figures at hand to show you nor are there calculations you can do..I think this is good oppertunity for you to do this test yourself and see what happens..

[martinjsto]

i have wired up some diodes in the junction box, the diodes i got are MBR 3060PT in the TO-220 . I think its right, but i am not sure about where to put the resistors.
pics attached, i hope.
pos and neg from array or controler goes in on far left and right, tied in is the diode circuit.
hope you can make sence of the setup and please offer changes as you see fit.
its in an old BP junction box from an old failed 2nd hand panel.
thanks.

[RifRaf]

martinjsto, the diode connections seem ok however from here it appears your 2 outside diodes are bypassing 16 cells each and the center diode just 4 cells. if you were to connect the center diode across the next closest place then it should be more even, each diode will bypass 12 cells. visualise or draw your entire panel as 1 long string and the diodes appear to make more sense, then arrange the whole string into a panel layout keeping the diodes connected to the same places.

with the resistors i beleive you would want one on each leg of each bypass diode, they would also need to be about 5w resistors. will get some to do testing with as well, however if the resistors are required large single diodes seem the easier idea

john p, initial diode testing shows the current being distributed evenly testing with a 3A current and voltage controllable desktop power supply, the current stays the same, or very close, through both legs from 0A though to 1.5A. the diode is still rather warm with just 1.5A per leg after 30 minutes. testing a single leg at 3A seems to produce the same amount of heat. one good thing about the to-220 package is being able to heatsink it well, will be adding some horizontal strips of bus wire for this. will test with batteries and light globe array after work to get more current to test the higher end

[martinjsto]

Hi, I have tryed a different layout, how does this look John,
thanks mate. two other pics follows.

[martinjsto]

the other two

[RifRaf]

if you want to use resistors to balance the load across the diodes then one needs to be on each anode, the 2 external leads of the to-220, before joining the resistor leads together. the cathode, located at the center pin or heat sink tab, does not need a resistor. from the images is hard to see the rewiring though. i am still under the impression that if one leg of your diode can already carry the current then using both for a safety margin should be ok without resistors. will offer an alternate way to hook them up where the leads will not join for several inches, and hopefully the bus wire resistance may be enough, need to draw it though

[martinjsto]

i also found this. is this better?

[martinjsto]

so place resistor here?

[RifRaf]

nope in series with each leg, hopefully the image makes it clear

[martinjsto]

ok, got it, thanks for the pic. do you have to use two resistors per diode, (1 for each side).
the cell layout is doing my head in to have every 12 cells protected. i can see how to split the panel in half so that 18 cells are protected but for 12 or 9 cells. hmmm need to think on it some more.
any pics or commentrs on rhe pice ive posted would be apreciated.
thanks

[RifRaf]

using resistors depends on what diodes you use, the MBR20100CT i been testin seem just fine without them, have made a basic video that could have better lighting i guess but watch the needles, the increase was stable and shared consistently as the load increased. so why the resistors? martin i was just assisting you to install them as balancing resistors for the diodes, thats all, which cannot hurt btw and would add an extra layer of safety though i do not think it would be used if heatsinking correctly and staying at less than 90% maximum load of a single diode.

link to video http://www.youtube.com/watch?v=cAbdnA4t0tg

also getting diodes rated for more than 6A is very hard at local electronics places and even on ebay, digikey, mouser etc, i picked these up at jaycar, and when buying a few at a time adds about $10 to a large panel (seven to-220) so seems worth it to have the low voltage drop, 20A capacity and matched diodes, testing with and without bypass diodes in shady conditions made me want to use them and make them safe. so these should actually be far better than the standard inline 6-8a power diode sold on ebay and also preinstalled on many manufactured panels.

as to wiring up the panels for 12 per diode, you were close the first time and mentioned a way to fix it, hopefully the penny has dropped by now, if not will try explain again

martin, i'd assume your mbr3060 is a 30A 60V dual diode package in 15A max branches. same manufacturer as mine and i cannot even see an issue if you choose to not connect the second diode with less than 10A cells. so forget the resistors for these, and test for yourself but i reckon you are safe to double them as well as you originally did

back to the test, at the 9A or so the globes were drawing the diode was very hot, too hot to touch but did not fail. the heat using one leg at the 9A was just as hot, need to find thermocouple to test heat better than fingers, the inline diodes must get very hot at times in commercial panels under shading, with the sideways bus wire heatsinking i want to add should make things even better, will run them out to back of panel so they can be in air but seal with an epoxy for weatherproofing