bmachining

Ive come up with .001 ohms for R2.
R2 sets the off time of the mosfet driver , so i guess this isnt as critical? too large an off time would mean more ripple??

Ubergeek63

No, that would be the bottom of R2-the low side of the big cap is circuit ground. Remember, it is a current regulator running in peak current mode. All it needs is the current sense voltage.

I get 0.007 for R2 for 35Apk. What did you use for your "nominal frequency"? Lower Toff makes the max frequency higher (with the associated troubles) and the inductor smaller.

Just remember that the duty cycle is from equation 2: Toff/Tperiod=1-Vout/Vin and when you set the off time (Toff) you set the minimum duty cycle since the blanking period is around 0.45uS. That sets your low end drive current. If you get the other app note I suspect it will work fine for lower output voltages just switching the R1 to ground.

Constant frequency peak current regulation does not like running over 50% duty cycle (if you read up on in you see the term "subharmonic oscillation") which prevents it from regulating well when the output voltage is greater the 50% of the input.

Constant off time peak current regulation has the off time and blanking period limiting the duty cycle which limits how low a voltage you can generate from a given input.

On the bright side, outside of those limitations, they are easy to stabilize.

bmachining

My Bad!! R2 is currents sense,
R1 is frequency.

0.007 for R2! i got 0.008 @30A, and rounded it up hehe.

R1..... Would it be a good idea to use a trimpot to adjust this?

Ubergeek63

is the input 40VDC or 50VDC? 40VAC would give you 48-65VDC peak (the valley depends on the load). Don't use a pot, pots are horrendous. 178K would give you 8uS off time and 5% ripple for 10V@30A.

In equation (4) the 0.3 is "Assuming a 30% peak to peak ripple in the output current"

The easy way is to plug the formulas into a spread sheet and play with the numbers, unless you actually have a full blown math package that you understand

bmachining

I will bump the input up to 50DC. Maths isnt one of my favourites, so yeah some of those formulas are a little daunting hehe.

Does the rest of the circuit look like it will be ok?

Ubergeek63

Seems workable other than what i mentioned above

bmachining

Thankyou very very much!!! i will do an order and build! Much appreciated! I hope u can make use of it also!

Will let u know how i go with it

bmachining

http://www.irf.com/product-info/data...rfp4110pbf.pdf

How would this Moset go? the Rds on resistance is different...

Ubergeek63

Quite good actually. I would estimate less than 37W dissipation and one of the best junction to case thermals I have seen. (as long as the case to sink is there...heat sink grease or silicone pad)

a quick worst case is:

40A*40A*0.004R=6.4W + (conduction losses, actually multiplied by duty cycle)
(88ns*2+67ns*2)*100KHz*50V*40A*0.5=31W + (switching losses)
43nC*2*100KHz*50V =0.2W + (drain to gate charge)
210nC*100KHz*12=0.25W (gate charge actually in the driver)

bmachining

Kewl! there available locally so if i smoke 1 or 2 during testing i can easily get more!

How critical is the inductor size? in uH?

Ubergeek63

no very, higher lowers the ripple current and lower changes the frequency. Either way you run into a limit that you can tweak. At least with peak current control you do not have to worry about instability

bmachining

I picked up a couple of toroidal cores, 40mm OD 24.1mm ID 14.5mm L, and with an inductor calculator i found, using 8 gauge wire they should be acceptable. will have 2 of them.