4

As an example: https://www.onsemi.com/pdf/datasheet/mbrb20100ct-d.pdf

I've seen discussion where it says diodes in parallel may not increase the current capacity. BUT the datasheet clearly says it's 20A per device (or 10A per leg)

Does that mean Onsemi has extra "trickery" in their device to double the current capacity?

I had a look at the graphs provided but unsure which one applies to the doubling of current capacity.

I have a PCB here which uses 4 of these 3-legged diodes to make a bridge.

Math Keeps Me Busy
  • 23,260
  • 4
  • 21
  • 75
mrjayviper
  • 399
  • 1
  • 7
  • From the datasheet: SCHOTTKY BARRIER RECTIFIER – MiNiMe Nov 12 '23 at 05:28
  • Another datasheet, which gives examples of applications: https://www.mouser.com/datasheet/2/196/bas40_bas140series-60208.pdf – MiNiMe Nov 12 '23 at 06:07
  • 2
    There's nothing in the data sheet that says the individual diodes can be connected in parallel doubling the permissible current. What it does say is both can simultaneously carry their full rated current. – stretch Nov 12 '23 at 12:50
  • 2
    @MiNiMe What are you trying to emphasize about it being a rectifier? Nothing here suggests that the asker doesn't know that. – Hearth Nov 12 '23 at 15:51
  • @Hearth What I had in mind is the other half of Sphero's answer. I'm not sure why a dual Schottky would be used in parallel instead of one diode rated 20 A (to match the example). But I'm sure there are other applications where the load can go up to 2 x 10, so I back 1/2 step. – MiNiMe Nov 13 '23 at 22:03

3 Answers3

6

There is no guarantee that the current will split evenly if you parallel the diodes. It’s up to the application to ensure that the per-device limits are respected. Nothing is lost in paralleling the diodes- it can’t make things worse- however the datasheet does not address the question that I can see, so you shouldn't count on any particular improvement.

Also, note that with a bridge configuration only half the diodes conduct at any given time, so you get a doubling of rating from that if you can assume the input is AC at around mains frequency or higher.

Spehro Pefhany
  • 397,265
  • 22
  • 337
  • 893
  • 6
    You actually answered this 7 years ago: https://electronics.stackexchange.com/questions/213604/why-do-schottky-diodes-have-two-anodes – MiNiMe Nov 12 '23 at 05:54
  • 2
    @MiNiMe Thanks, lol. A slightly better answer since I mentioned that leakage current would double so it's not all gain in paralleling them. – Spehro Pefhany Nov 12 '23 at 05:56
6

If you take two single diodes, they are unlikely to share current as well as two diodes in the same package, which as well as sharing the temperature, might have shared the manufacturing environment as well.

If the manufacturer is implying that they share well, then they may well have shared all their manufacturing variations.

Bear in mind that the 10/20 A figures are Absolute Maximum ratings, you will normally run the package well below this .

Neil_UK
  • 166,079
  • 3
  • 185
  • 408
6

Does that mean Onsemi has extra "trickery" in their device to double the current capacity?

No there is no trickery other than to keep the individual diode junctions thermally connected as much as possible. And, if the data sheet doesn't give assurances about this then, you can't assume that they'll share current at all evenly. What can happen is the following...

As soon as one diode starts taking a little more current than the other, it will heat a little more and, when this happens, it'll take a little more current, warm a little bit more and take a bit more current. Hopefully, you can see where this is going so, maybe look at this modified graph in the data sheet to see what I mean: -

enter image description here

I've drawn a trajectory on the graph that shows a slight tendency for thermal runaway (increasing power dissipation as you move from a 5 amp current on the right to 7 amps on the left).

But, it's tricky to pin-point what trajectory the thermal effects take (as temperature rises from ambient). There may be a net effect that keeps the currents the same or, there may be a net effect that causes a big disparity. To demonstrate this I've chosen the same trajectory in the graph above. That trajectory suggests that as one diode warms, the currents become significantly different and, the final junction temperatures are 150 °C and maybe 105 °C for the two diodes respectively: -

enter image description here

Both diodes begin conducting 5 amps but, over a short-time, one is conducting 7 amps (at a junction temperature of 150 °C) and, the other is conducting 3 amps (at a junction temperature of about 105 °C). The combined current is still 10 amps but, individually they are different.

I can also easily imagine this less problematic scenario: -

enter image description here

How closely the diode junctions are thermally connected is unknown from the data sheet.

Andy aka
  • 456,226
  • 28
  • 367
  • 807
  • The better the matching the better the sharing. Because there is slope to the VI curve both devices will carry current. If you are approaching the current rating of one device then using two identical thermally matched units will gain some. As Spehro mentioned you change leakage. You will also change capacitance and do very little to package power dissipation limits. If I had two diodes in the wee SMD device I would use both unless there was good reason NOT to but would use the benefit of the second to improve reliability and not to try and exceed rating of single device. – KalleMP Nov 12 '23 at 16:36
  • 2
    I’ve seen one regular Si dual diode once where the datasheet guaranteed low enough thermal coupling to allow them to be used in parallel without thermal runaway. Most don’t. – winny Nov 12 '23 at 17:30
  • "Hopefully, you can see where this is going so" For poiwer amplifiers with multiple transistors in parallel, it's called "Firecracker Mode Failure". – WhatRoughBeast Nov 12 '23 at 22:43
  • @WhatRoughBeast I've never heard of that definition but, I can see why it might be called that. – Andy aka Nov 14 '23 at 11:57
  • @Andyaka - It goes back a long ways, and it's a mark of my age that I know it. The thing is, the potential is so well-known, and the fixes so well-known, that almost nobody does it any more. But the institutional knowledge was hard-earned. As the saying goes, "Good judgement comes from experience. Experience comes from bad judgement." But even now, people get in trouble with LEDs. – WhatRoughBeast Nov 14 '23 at 18:36
  • 1
    It horrifies me that so many folk get LEDs so badly wrong @WhatRoughBeast – Andy aka Nov 14 '23 at 19:25