SCR frequency for Itsm and its applications

Whats the limit on how often an SCR can conduct =>75% of its Itsm current?

Are there applications were SCRs are implemented only to conduct some amount of its Itsm or were SCRs not made to be used only for their Itsm capability?

I guess at least once per second without overheating the die. Wear due to thermal stress and long term reliability is a different point.

There are pulse applications of SCR, e.g. capacitor discharge spot welders.

Lots of new brand name puk SCRs available on ebay but its formfactor don't have a compact heatsink.

SCRs are used for line frequency rectification
because one can control the turn on, it can implement a soft start
and by removing the gate, turn things off in less than 1 line cycle

so 50/60 times per second works, and they conduct for 50% of the time in single phase, and less for 3 phase

the data sheet should show thermal information and duty cycle information

From https://www.vishay.com/docs/93711/vs-50ria.pdf

a 20mS 300amp pulse will nominally bare 0.1Rj-Hs and at 300amps the Vf is 2V which yields 600W for that pulse going off



Disregarding any other thermal resistance to heatsink or heatsink to ambient, with these pulses whats the nominal thermal rise for the package at 500mHz?

This is for nominal thermal rise calculation, 20mS is a large duration for such a current.

do you mean 500 milliHz, (0.5 Hz) or 500MegaHz?

no such thing as an SCR at high frequency
they just don't turn on or off that fast

that's why they're used at line frequencies or as non-repetitive crowbars

20 ms is 0.02 s
on bottom graph, that looks like 0.1 K/W so at 600W, looks like a 60 K rise
if ambient air is room temperature, 20 deg C, the die will rise to 80 deg C
well with parameters

Figure 6 page 4 of data sheet
for 20 ms, maximum non-repetitive pulse current is 1000 A
(assuming i read the graph properly)

Figure 2 page 4 of data sheet
square wave conduction and average current
if your pulse is 0.02 sec, what is the period of the signal?
convert that to conduction angle as if the period was 360 degrees
see how much average current vs case temperature
they already applied 0.35K/W from maximum junction temperature

wwfeldman said:

do you mean 500 milliHz, (0.5 Hz) or 500MegaHz?

Click to expand...

Megaherz is typically MHz SCRs don't work at those frequencies, perhaps MCT thyristors would be able to.

wwfeldman said:

20 ms is 0.02 s
on bottom graph, that looks like 0.1 K/W so at 600W, looks like a 60 K rise
if ambient air is room temperature, 20 deg C, the die will rise to 80 deg C
well with parameters

Click to expand...

Period is 2S it averages to 11.882W from the chart 0.1Rj-hs 11.882 * 0.1 = 1.18c of thermal rise. The mass of the stud can take that sort of dissipation without a heatsink. I used LTSPICE to average a 600w load.



Is it 1.18c of thermal rise or is there another object which wasn't computed to obtain correct values?

I don't understand your temperature calculation. If you have 11 W average power dissipation, the related thermal resistance is junction to ambient, I guess something like 15 - 25 K/W. Without a small heat sink, the devices can't handle 11 W.

A programmable unijunction transistor (PUT) acts as a thyristor similar to an SCR.
You can create your own from whichever combination of PNP and NPN as you wish. Parallel multiple devices to handle large Ampere surges. Add heatsinks as needed.

Link to tutorial with sample circuits:

www.instructables.com/id/Circuit-Collection-of-the-Programmable-Unijunction/

BradtheRad said:

A programmable unijunction transistor (PUT) acts as a thyristor similar to an SCR.
You can create your own from whichever combination of PNP and NPN as you wish. Parallel multiple devices to handle large Ampere surges. Add heatsinks as needed.

Link to tutorial with sample circuits:

www.instructables.com/id/Circuit-Collection-of-the-Programmable-Unijunction/

Click to expand...

Id rather have a bidirectional semiconductor current switch rather than a discrete latch.