Reading Fast/Soft Rec Diode Graphs

[richie00boy]

I am looking at the 'peak reverse current vs. di/dt' and 'recovery charges vs. di/dt' graphs on a fast recover diode datasheet. How do I tell what is a good one and a bad one? Surely there should be some kind of graph with a time axis?

[jneutron]

Originally posted by richie00boy
I am looking at the 'peak reverse current vs. di/dt' and 'recovery charges vs. di/dt' graphs on a fast recover diode datasheet. How do I tell what is a good one and a bad one? Surely there should be some kind of graph with a time axis?

The peak reverse current, I_rm, shows the maximum reverse current the device will conduct before it starts to release. The faster the di/dt, the higher that number will be for the same device.

For same species devices, the lower the I_rm, the better.

Caveat: if the stored charge, Q_rr is much larger,, it will just take longer for the device to entirely shut off. The two do not relate absolutely directly, you can have lots of stored charge with a low I_rm, and vice versa.

So, lower stored charge is also better.

Some devices will use recombination to lower either thing, but that usually comes with a price, reverse leakage temp sensitivity being the major thing, which can lead to high temp runaway in a switcher.....(blowed up).

Some datasheets show a time based graph, but it is just a typical pic showing what the terms mean, when the device is tested in a standard setup.

Cheers, John...

(did I answer your question?)

[richie00boy]

Thanks for the great reply. I think I understand I was just expecting to see graphs like typical transient graphs of current vs time. It's hard to picture how quickly the device turns off or if the turn off is smooth.

EDIT: I must have been looking at the datasheet late at night. Having gone back to look again following your new info I see a graph of time vs di/dt on the previous page

[jneutron]

Originally posted by richie00boy
Thanks for the great reply. I think I understand I was just expecting to see graphs like typical transient graphs of current vs time. It's hard to picture how quickly the device turns off or if the turn off is smooth.

EDIT: I must have been looking at the datasheet late at night. Having gone back to look again following your new info I see a graph of time vs di/dt on the previous page

The datasheets aren't easy to read, even in the daylight!!

In the old days, it was the old trr test setup, and darn near every diode manufacturer made hugely sloppy test rigs to see it..The databooks drew this rounded waveform thingy, with slow recovery to zero current...hand drawn, of course, because the actual scope photo typically had these huge overshoots and ringing..the manu's were embarrassed to actually print the waveforms..(I worked at two manu's).

Back in '81, I build this rig that measured 2 nSec diodes accurately, showing turn offs as fast as 250 picosec without overshoot, and it allowed the diffusion guru's to develop ways to soften the recovery on their 15 nano product.

Back in '83 or so, they started using that inductive di dt test setup, as the fast power devices couldn't be read accurately, and most designers had inductive switching circuits anyway..but it still doesn't make it easy to figure out..

Cheers, John