Diodes in parallel to increase maximum current ?

How long / how many times do you want your device to work? :)

Reply to
onehappymadman
Loading thread data ...

"John Fields"

....... Phil

Reply to
Phil Allison

"Mochuelo"

** A row of four with a row of three snugly fitted on top.

** It is bound to be all three things - especially if the OP has 5408s on hand aplenty. 12 amp, 1000 volt single diodes are not cheap nor that readily available.
** Err - a DMM will help indicate that.

Just pass the same current through each diode and note the forward voltage drop.

Diodes bought together at the same time are likely to be from the same batch & hence closely matched.

....... Phil

Reply to
Phil Allison

"Ancient_Hacker"

** Not if they are matched and in thermal contact.

Paralleling power diodes is COMMON practice in commercial products.

** The 1N5408 diode is rated at 1000 PIV.

Maybe that was beyond you ability to compute too?

....... Phil

Reply to
Phil Allison

"John Fields"

** YOU must try that JF - ASAP.

........... Phil

Reply to
Phil Allison

I have some 600A 1200V FRE diodes lying about :) (they are paralleled up with 600A 1200V IGBTs though)

I snipped 4 1N4007s from bandolier (I happen to have 1000 or so lying about), paralleled them and ran 3A thru the resultant mess for a couple of hours. the bodies were in fairly close thermal contact. First I did it with nice long leads, and measured the forward drops at the diode body. All within +/-1mV (841,842,842,843mV). After a few hours, they were all roughly the same temperature.

I've done it with 9 1N4007s before (a nasty korean TTL video game pcb that locked up with Vcc > 4.3V, psu had no adjustment) but I would never design shit like that into a product, or give it to a customer - my excuse back then was: its either this, or leave it switched off. that particular game ran happily for another 3 years, until the chassis was stripped down and rebuilt - new monitor, controls, loom, psu.

its still an awful trick.

note that most (if not all) PC power supplies use TO-220 dual rectifiers, wired in parallel. Of course they are in extremely close thermal contact. And of course most PC power supplies are crap.

Cheers Terry

Reply to
Terry Given

It says "3.0 ampere operation at Ta = 75C with no thermal runaway". That does not meant there would be no thermal runaway if the current were higher, or the ambient were higher. It also contradicts the graph on page 2 which shows derating the thing to ~2.5 amps at 75C.

I don't think trying it proves anything, other than it worked or didn't work with the particular diodes in the test.

Ed

Reply to
ehsjr

on

available.

voltage

batch

ST has some intersting app note on the subject:

formatting link
formatting link

and another a bit OT but still interesting

formatting link

--
Thanks,
Fred.
Reply to
Fred Bartoli

Ta muchly.

Cheers Terry

Reply to
Terry Given

Well, there you go. I actually felt that Phil was a bit conservative when he advised derating the 4x1N5408 from 12A to 7A-8A, but better be safe than sorry. At least it doesn't seem so dangerous as some like us to believe.

[snip]
--
Thanks, Frank.
(remove \'q\' and \'.invalid\' when replying by email)
Reply to
Frank Bemelman

"Frank Bemelman"

** I only suggested de-rating because of the close thermal coupling.

The final running temp of a bundle of hard working 3 amp diodes is *bound* to be considerably more than that of individual 3 amp diodes.

....... Phil

Reply to
Phil Allison

"Phil Allison" schreef in bericht news: snipped-for-privacy@individual.net...

Good point. I always get a bit nervous when stuff gets above 50-60 degrees, but I'm a chicken. If not for the temperature, then for the mechanical stress it can cause. Learned a lesson once with vertical mounted TO-220's on a stiff pcb mounted heatsink.

Fred was so kind to post a link to some application notes dealing with the subject. There's also some good advice about the pcb layout:

Bad: ################### # # # # A A A A K K K K # # # # ###################

Good: ################### # # # # A A A A K K K K # # # # ###################

--
Thanks, Frank.
(remove \'q\' and \'.invalid\' when replying by email)
Reply to
Frank Bemelman

"Frank Bemelman"

** Your "bad" & "good " diagrams look identical.

No matter the font setting.

BTW

I am used to seeing diodes and zeners running at 150C in commercial products for years on end without failure.

The worst that happens is their solder pads give up !!

....... Phil

Reply to
Phil Allison

"Phil Allison" schreef in bericht news: snipped-for-privacy@individual.net...

Strange, they look fine from here. On the 'bad' the common tracks enter from the same side. On the 'good', one enters from the left, the other exits to the right.

Yes, I can imagine. Money has never been a very important issue for me, so even for very simple boards I use double sided FR4.

--
Thanks, Frank.
(remove \'q\' and \'.invalid\' when replying by email)
Reply to
Frank Bemelman

"Frank Bemelman"

** Someone's wild fantasy idea of an issue.

Do some estimates and you will see why.

........... Phil

Reply to
Phil Allison

if for no other reason than a whole bunch of surface area goes away. its also trivial to show that close placement increases the peak hotspot temperature.

I still wouldnt do this in a production environment, its just asking for trouble. especially with "jellybean" parts, that can come from a variety of different manufacturers. Murphy will ensure several diodes with wildly different forward voltages end up in some customers equipment. The dual diode approach obviously avoids this though.

Plus of course if you are designing something other than incredibly cheap stuff (eg pc smps), then just use the right diode.

Cheers Terry

Reply to
Terry Given

"Phil Allison" a écrit dans le message de news: snipped-for-privacy@individual.net...

The appnote mentions currents in the 500-700A range, IIRC. When there, every fraction of a mOhm counts, obviously much more than for

10A.

Even then, if you're not careful enough, at 0.45mR/sq of PCB track you can easily reach 5/10mR for 6 diodes with the bad wiring solution, if not careful enough. This will give you 50/100mV imbalance for a 10A current. Not really a fantasy. Good wiring practice costs nothing.

--
Thanks,
Fred.
Reply to
Fred Bartoli

"Fred Bartoli"

** Hardly ** bloody likely ** for a few pigtail diodes in parallel on a PCB.
** Bunkum.

The diodes are packed tight and thick solder coats the copper - so the "thermal contact" criterion is met.

Simply placing the plastic cases in contact does NOT do the job - coupling the copper wires leads close to the junctions does.

** Worse bunkum.

** Shame it has all the marks of a complete wank in this case.
** A blatant platitude.

Nice finishing flourish for someone who thinks in clichés.

........ Phil

Reply to
Phil Allison

the app note says nothing about pigtail. you didnt even read it.

but completely true. just like "good PCB layout costs nothing"

hilarious finish for someone who didnt even bother to read the appnote. which is generic, and applies to *any* diodes.

how many switch mode power supplies have *you* designed, Phil? how many units manufactured? betcha Fred's > 40dB ahead of you....

Cheers Terry

Reply to
Terry Given

"Terry Given" = ADHD sufferer & Sheep Shagger

** But the topic is pig tail, ( 3A) diodes in parallel.

Frank's ASCII pic was related to that - or should have been.

Not that staying in context has ever bothered a sheep shagging moron like Terry Given.

........ Phil

Reply to
Phil Allison

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.