Rectifiers...

[richp0169]

Good deal or no?  Chinese made...most stuff is.



http://www.ebay.com/itm/10-50A-1000V-Metal-Case-Bridge-Rectifier-SEP-KBPC5010/290629696057?_trksid=p2047675.c100010.m2109&_trkparms=aid%3D555012%26algo%3DPW.MBE%26ao%3D1%26asc%3D23772%26meid%3D7872499607278815606%26pid%3D100010%26prg%3D10164%26rk%3D1%26rkt%3D24%26sd%3D261371447516

[Mary B]

You will find a high failure rate from the chinese junk.

[richp0169]

Technically I would not need something this hefty for my test based alternator, but thanks for the input. 

[OperaHouse]

Run very fast away from these.  I bought these same ones at a real discount from a seller who admitted in the listing that they didn't seem to handle that much current and that there had been failures.  I needed some quick for the repair of a buddy's 20A 36V battery charger.  I even put two in parallel with leads long enough to do appropriate load sharing.   At a 15A charge they ran too hot to touch with the factory heat sink.   Later I tested them on a supply running  just 3A.  Forward voltage was twice as high as the 25A bridges I has. They also weighed half as much.

[Mary B]

That style bridge requires an external heatsink. Depending on how hard you push it screwed to a metal chassis to a large heatsink with fins may be required.

[AzSun]

I am building a battery charger using the 70 amp version of this style. In short, even when mounted to a heat sink with  conductive paste and a fan blowing on the heat sink, the diode failed at 35 amps, I blamed it on overheating the connections when soldering. I used "Faston"  terminals on the second test. (no solder, new diode)  The terminals became very hot at only 10 amps.  It seems like the diode junction is not transferring heat to the case. The vendor is sending me an 80 amp version. Not sure if I had 2 defective ones or this style is way over rated.

[joestue]

try these
http://www.ebay.com/itm/Semikron-SKB30-04A1-Bridge-Rectifier-30A-400V-/251542744932

yes its possible bad contact between the terminals contributed to overheating, but the reality is their "rating" is simply lies.

[Flux]

What they don't tell you is that the claimed rating is only for resistive or inductive load. With a capacitive load ( which a battery charger is as far as a diode is concerned) there is a big de rating factor. This added to the fact that the claimed rating is at 20deg C base temperature means that even a genuine manufacturers device will not get near the nominal rating for this application.

In reality the 35A devices are good for about half rating. Add in the limitations of the faston connectors the bigger ones are even more pushed. Many of the un branded things are rejects so you can easily get something very limited in capacity.

I only use them for convenience in applications where their life is easy. For anything else I use the stud mounted diodes or the press fit ones for less demanding things, they are much better and although you have to make special heat sinks, they actually need less heat sink than a potted bridge for the same current.

Semikron are respected manufacturers and their devices will meet the spec. but the spec is for resistive or inductive load. They are expensive but sometimes you can get a good deal. As far as I know these things are not common enough to be concerned about fakes but that can be a big concern for audio transistors and some common ics.

Flux

[DanG]

On top of the circuit flavor (resistive/inductive/capacitive) the diodes get inserted into - there is duty factor of the waveform the diode is being fed, think waveform graphed and conduction per time period. I have a box of 'supposed' 50A bridges, that out of all available spec sheets, I found only ONE one explaining rating test procedures -- rated using a sawtooth |\_|\_ waveform versus Sine ~ or square pulses ||_||_, and after all things considered are good for 20-25% of their rating for higher RMS or DC work. Anyhow, rated using sawtooth wave, 50 Amps for 15% of the time... I can see it helping with inrush surges but that's about it.

[joestue]

In theory the diode does not have to be de-rated.
The reason why is because its forward voltage drop is constant.
In reality, there is an additional resistance component, which is also temperature dependent.

For example, the 100 amp reverse blocking diode of a transistor i have dozens of:
1 amp 0.56v at 125C, 0.7 volts at 25C
10 amps .75v at 125C, 0.9 volts at 25C
70 amps 1.15 volts at both 125 C and 25C
100 amps 1.3 volts at 125C and 1.25v at 25C

for that semikron diode block, at 30 amps each diode is going to drop 1.1 volts at 150C and 1.15 volts at 25C
this is a pretty typical diode "rating"

so assuming a resistive load, 30 amps rms ac going into the diode block is going to dump 66 watts into the diodes, which means the diodes are going to be 46C warmer than the case is. so if the case is held at 75C, that is totally reasonable and the diodes are going to be running at 120C.

If the current however is not a sine wave and you're dumping the current through the diodes in 50 amp pulses, then the diode voltage drop is going to increase, to, i'm guessing 1.4 volts. so keeping the average current at 30 amps that means 84 watts instead.

[AzSun]

try these
http://www.ebay.com/itm/Semikron-SKB30-04A1-Bridge-Rectifier-30A-400V-/251542744932

yes its possible bad contact between the terminals contributed to overheating, but the reality is their "rating" is simply lies.

First of all, you started an interesting topic Rich*  Thanks for the link joestue and all follow up comments. I was looking for screw terminal units originally but could not find them at the time. I didn't know a charger saw a battery as a capacitive rather than a resistive load. The testing was done using a resistive load. I am using this to test some old Edison Nickel Iron cells. They have a 75 amp charge rate. 

I could get 4 of those but would need to parallel them. Can you do that as long as the wire lengths are all equal ?

[joestue]

First of all, you started an interesting topic Rich*  Thanks for the link joestue and all follow up comments. I was looking for screw terminal units originally but could not find them at the time. I didn't know a charger saw a battery as a capacitive rather than a resistive load. The testing was done using a resistive load. I am using this to test some old Edison Nickel Iron cells. They have a 75 amp charge rate. 

I could get 4 of those but would need to parallel them. Can you do that as long as the wire lengths are all equal ?

If you can split the transformer secondary into multiple parallel coils you can use the transformer resistance as the current sharing resistors, but I would look for a single diode block that is large enough.
The 100 amp single and three phase diode blacks sold on ebay for 20$ seem a bit too good to be true, you can compare them to a legit 100 amp stud mount diode with a 4 awg copper strap. note however the single diode isn't rated the same as the stud mount types, and four 100 amp stud mount diodes properly heat sinked are equivalent to a 200 amp diode bridge for a resistive load.
also, the high voltage 1200v and 1600 volt diodes have about double the forward voltage drop compared to a 200 volt rated diode, so keep that in mind.

[dnix71]

The older style car battery chargers had diodes embedded in large metal plates to keep the diodes cool. The engine start feature would allow 50 amps for maybe 15 or 30 seconds before a long required cooldown. Even professional shop bulk battery charrgers and engine spinners don't have the kind of rating some of you seem to be asking for.

This discussion reminds me of the old days when guys rewired Pentiums to bypass the software bus multiplier limits. Yeah I saw a P4 clocked at 6GHz, but it was immersed in liquid nitrogen and would only run long enough to boot and take a bios screen cap before all the N2 boiled off.

Maybe your local power company has some out of spec IGCT's they would sell you. All you need is a driver to make then oscillate to make pulsed dc. The smaller ones on industrial drive units can handle 5MW's.

[Mary B]

Name brand 50 amp bridges on a large heatsink. I ran 45 amps continuous through these for an hour with no signs of over heat. Sure they were hot but not beyond my being able to hold a finger on them. Good silver bearing heatsink paste under them for better heat transfer. If memory is right that heatsink is 12x6x2 1/2 inches with over a dozen fins.

[dnix71]

http://www.mouser.com/ProductDetail/Rectron/BR256W/?qs=G5AQjGfRJcKTbQFpq9fxnA==

Only sold in bulk, but $2 at that price. 1.1v forward drop. 600v rated. 25 amps. Probably real numbers.

http://www.goldmine-elec-products.com/prodinfo.asp?number=G19879B $5 for 1

[AzSun]

I was thinking of changing to these, one on each side  to make a bridge rectifier. I don't want to alter the heat sinks and these fit in slot. I think they have a 1200V rating which has a higher forwarding voltage. I would be charging a single cell up to 2.0V or max continuous rating of 75 amps.
Do you think these would work?

Thanks.

 

[joestue]

AzSun,
yes those are overkill and although they are high voltage, they are also very large, 1.5 volt drop at 400 amps. (see figure 15 in the datasheet) so the forward voltage should still be under a volt for the currents you're dealing with.