Paralleling Diodes

[ElecDesigner]

I am working on an application (reverse voltage swing on HV capacitors due to extremely rapid discharge) where diodes need to be able to handle currents of about 10kA for 100us or so.

It might turn out to be space (and cost) efficient to use multiple devices in parallel to achieve the requirement. I am a bit skeptical about paralleling but it appears that it can be done, see:

https://www.st.com/resource/en/application_note/an4381-current-sharing-in-parallel-diodes-stmicroelectronics.pdf

Are there any additional pitfalls due to my application being a very short high current pulse. It is common to see inductors used to force leading edge sharing in Thyristors but I assume these are not needed for diodes as I assume the same dI/dt limits do not apply.

 

[danadakk]

If you take the effort to do some matching .....if not then ballast them (AC and DC).....

Diodes In Parallel

If the load current is greater than the current rating of a single diode, then two or more diodes can be connected in parallel

www.daenotes.com

https://idc-online.com/technical_references/pdfs/electrical_engineering/Series_and_parallel_connected_diodes.pdf

Regards, Dana.

 

[Easy peasy]

high di/dt can lead to very high Vf-fwd ( forward recovery voltage ) - worse for faster reverse recovery diodes, the combined Vf, If product could cause die melting in extreme cases.

[ paradoxically, or ironically, slower diodes ( rev rec ) are faster to turn on ]

 

[ElecDesigner]

Forced sharing using resistors seems a bit troublesome due to the very high currents.

Currently thinking of using 6 diodes in parallel to make one 3kA rated (ish) part - about 500A per device.
Above about 200A the parts in question cross over to the positive temp coefficient so that will help sharing.

Not convinced I need any sort of ballasting in this situation but possibly thinking of using a previously designed 100nH air core inductor. Anyone see any floors in that?

 

[danadakk]

Do a test, randomly pull some parts, data log them at some nominal current,
then look at them thermally in parallel and back into what current they are
taking ?

Contact the vendor, get characterization spread data, as some guidance.]
Also get guidance on electromigration, to make sure thats not going to be
a factor for the part thats worst hog. I would even ask about die attach
uniformity, as well as general guidance from the vendor.

Obviously making sure these parts torqued down properly, and heat sinks
flatness appropriate.


Regards, Dana.

--- Updated Nov 14, 2022 ---

Section 5.8, see attached.


Regards, Dana.