20A into 1 milliOhm Load

[cupoftea]

Hi,
We have to put 20A into a 1 milliOhm load from 37Vin.

The duty cycle is 0.001 !!!!!
Even at 13.7kHz, the on_time is just 79 ns !!!!!

We must use the UCC28C43 as attached (LTspice and jpeg)

The only reason it works is that the Current sense filter means the on_time isnt actually so ridiculously low.

Every modern controller we looked at wont work due to the Leading Edge Blanking time....(which is longer than the on_time required !!!)

Why no chips for this on the market?

 

[D.A.(Tony)Stewart]

Not everyone wants to drive 20A into the ground from 37V. That starts at 740W of stored power per pulse.

A more sensible design stores energy from a more matched voltage or transforms from a higher voltage at low input current’s.

 

[cupoftea]

Thanks yes, and the power in the 1 milliohm res is just 0.4W.
But essential for us.
The above is the simplest way to do it?

 

[crutschow]

A flyback circuit as shown in the data sheet, with a transformer or tapped inductor having a large primary to secondary turns ratio (e.g. 1000:1) should give a large output current for a 1000 times larger duty-cycle.

 

[Easy peasy]

In this case a properly designed flyback would be better - any output diode will be, for a schottky, 0.6 * 20 = 12W ....

 

[cupoftea]

Thanks yes, with such a large turns ratio, it would be difficult to get good coupling without expensive means. (eg tape wound primary etc) Also, the Buck is small, and it seems to be doing the job in the Sim.
Ayk, Buck inductor also offtheshelf and likely to be smaller than flyback transformer

 

[mtwieg]

Do you have a requirement for efficiency/power consumption? If not, just add a few milliohms of resistance, now your duty cycle is more reasonable...

In practice, the ESR of your output inductor and freewheeling FET will effectively act as an additional resistance.

 

[crutschow]

the Buck is small, and it seems to be doing the job in the Sim.

Then your problem is solved?