0

I want to connect diodes in parallel to increase the current carrying capacity. If I connect two diodes in parallel having same rating(same forward voltage drop), does thermal runaway still come in picture? If forward voltage drop is not going to be same for each diode then why so when I'm using same diode?

fahad shaikh
  • 180
  • 1
  • 9
  • 1
    For a hobby project/prototype: sure that can be done, just make sure the diodes are somehow thermally coupled. For the design of something that's going to be produced in large numbers: I would not risk it, I would use a diode with a higher current capability. – Bimpelrekkie Sep 25 '18 at 06:40

1 Answers1

3

If the diodes attempt to share the same current so that neither carries the full load, which could exceed its own thermal-related max current, then they must have an internal resistance that is well-matched rather than >+/-50% worst case tolerance from different batches and also both have excellent thermal conduction to some heat sink. Otherwise if this internal resistance is well matched so that the rise in temperature of each may result in a 60’C rise equally such that the diode Vf drops x.x mV/mW of heat so that they still share current equally.

The requirement here is that the Vf drop with rising temperature) has both a good thermal conduction and good Rs equivalent match in values (Vf@1A-Vf@1mA)/1A = Rs. ( or as I prefer to call it ESR or “internal bulk resistance”). As long the voltage rise due to current is greater than the voltage drop due to temperature rise and mismatched tolerances, it will be thermally stable.

Neil_UK
  • 158,152
  • 3
  • 173
  • 387
Tony Stewart EE75
  • 1
  • 3
  • 54
  • 182
  • I believe that if I am connecting same diode in parallel then their internal resistance must be same and they should share the same current. am I right? – fahad shaikh Sep 25 '18 at 07:30
  • 4
    Their internal resistance is the same nominally, but it is not a closely controlled parameter so you can get big part to part variations. – Vladimir Cravero Sep 25 '18 at 07:41
  • 1
    you can read any datasheet , compute Rs worst case and get what I saying read 50%, then -3mV/‘c rise times C/W thermal rise then thermal couple to same heat spreader and see if you have runaway or read my theorem again – Tony Stewart EE75 Sep 25 '18 at 12:25