UCC28780: UCC28780

What  is the Vout? 

Are you saying there is a 30mF cap on the input? If so, it can hold the Vin above 110V to provide 400A over 0.5mS. Vdrop = I*t/C = 400 * 0.5ms/30mF = 6.6V. So your input is close to a DC input. 

However, 10us response time means at least 100kHz bandwidth. Flyback typically cannot reach to that bandwidth. You will need a large output capacitor to provide 400A instantly. 

I don't under stand this comment "What battery charger can operate in a bosst mode?"

Please share more details.

Hello Ning:

Let us chat about the issues, shall we<

The main problem to solve here is to design a " constant current " pumping of a tube operating from  a 100V  energy source! Such that it can draw constant current of 400 Amp over a Pulse duration of 500 usec only.

Repetation rate is 40 Hz.

Imagine any dc dc converter - such as PSFB-- realize that from the local capacitor across the bridge, the diagonal current can be 30 amp even in medium power levels..this is flowing in a short loop formed by the switches, Lpri(reflected), leakage, Caps in the loop ...etc...that has little to do with DC current drwan by the power stage from the source, right? Albeit, one can calculate DC current from the pulsating currents in the brridge.

Our case is not different: but the way we need to provide 400amp pule current is a bit different: it has to be constant. During the pulse duration.

So we are planning to do a boost/flyback dc dc converter( 36V to 108 V in reality)  such that it can charge up a capacitor bank. We sense the current in the  load then feed it back in the converter to modify its output voltage. Such that the current during the pulse remains constant..

Our load is different, but you can take a look at this typical behaviour.

Device_charact.pptx

As you can see, if we maintain feedback of the sensed current , we can maintain it by controlling the voltage applied across the device.

But the device is ON only for 400 usec every 25 ms.

A DC current source wont help: because it is not going to keep the current level at 400 amp during the pulse.

Because the charge is delivered by a input capacitor( bank, if you will)- and it is losing it,  thus voltage across it. The power source is not fast enough to make up for the loss over this short a time.

So we need to make a dynamic current source such that there is no current droop during the Pulse- it means that the output voltage has to increase to keep the current from drooping. During  400 usec every cycle.

Does it make sense to you now? That the loop BW has to be fast enough to ramp up the voltage during 400 usec every cycle.

Let me know

-r

If I understand it correctly, the Vin is 36V, Vout is 108V, either by boost of flyback. Sounds like isolated topology is not a must.

To get a faster response time, non-iso topology is typically better. You can just use a regular boost controller and build a current feedback loop to respond. I'd recommend a COT (constant-on-time) boost controller for your application. But based on my calculation, the Vout won't drop much with the output caps you are using. Did I miss anything here?

BTW, I understand the application need. I used to do laser design in the past. 

Ning:

very exciting to know you are familiar with such drivers!
I am nubie & using exp in similar domains to cut my teeth here but must be a unique competitive approach.

Your suggestion of COTT is one such.

in order to handle the current sense( 400 amp!), I have a very uniq solution.

Furthermore, the feedback point to use current sense is also unique. If you send me your TI email. we can further discuss including which COT...if you have time & inclination to do so.

thnx again

r

Robin,

Please leave your email here. 

But from what I understand, you just need a boost controller with a large output cap. 

I am closing this thread.

robin@ganmartechnologies.com

Opting for COT : but can it be effective with a flyback configuration? Also, must have an external switch + an external Vref.

Hope to hear from you.

r