First, in post #4 (attached schematic) the question was asked if two points should be connected and it seems no one answered... yes, those two connections should be made. The SG3525 is a double ended PWM. The 2 outputs are diode "or'ed" to the gate of the MOSFET to turn it on, but because of the diodes, the PWM can't turn the MOSFET off because the diodes are back biased, so R1 acts as a "pull-down" resistor to pull the gate low during the "off time".

R8, R9 and R10 act as the current sense resistor. During the "on-time" the current will ramp up (E=L*di/dt or (E*dt)/L=di), there will be a voltage developed across the resistors proportional to the current (di). When the voltage gets high enough to turn on T1, it will pull the "soft start" pin low and turn the PWM outputs off. When the voltage across the current sense resitor falls to level where T1 turns off, the PWM will go thru a "normal" soft-start cycle. R2 and C4 are a filter to get rid of the spike that will be there at turn-on. The time constant of R2 & C4 should be made equal to the width of spike. The resistor(s) value (R8, R9, R10) will go down because you want to pass more current before T1 is turned on... just figure out what the peak current will be for your load and size the resistor(s) so that you get ~ .65V across them.

 

Is this right??? R8, R9, R10 are 0.022 ohm in paralell peak is 6.5amps
changed to 0.004 ohm I would have peak 32.5amps


I tried a IRF540A at T2 it woul'nt work, I don't think it's Logic Level, would that of made a difference.

cheers


Sorry I got the resisotor values wrong there .22 ohms

 

At some point, increasing the current requires more than simply changing the sense resistor. A 32A current limit needs to have careful PCB layout since 0.004 ohms is such a small value. Also, your 0.004 ohm resistor peak power is 105 watts!

If your supply is at least 12V, You shouldn't need a logic level MOSFET, but without a heat sink IRF540A will get very hot with more than about 4A.

__________________
de KI6RWX

 

The current sense resistor for 32A (peak current) is .02 ohms or 3, ~.061 ohm resistors in parallel. Yes, power rating of the resistors is of concern, about 21 W each for steady state, but you have to figure out the RMS current... also, the Vbe will change over temperature, so as T1's temp increases, the current limit point will go down (~2mV per C° above room temp).

If max Vin is 12V, you don't need a 100V MOSFET... get a lower voltage rating and it will have a lower Rds on.

 

The MOSFET that is in at the moment is the BUK9535-55A TrenchMOS transistor logic level FET. The Id is 34A @ 25*c 24A @ 100*c

R8, R9, R10 are 0.022 ohm 5w and the Overload projection is 6.5A.
I change them to 0.015 ohm 10w and it was running at about 12A. The MOSFET was very hot so I changed the heatsink to a big cpu one with a fan.

I thought I would use this HUF75343P3 Fairchild ULTRA-FET POWER MOSFET. 55v 77A
Thinking it will not be strained so much at higher amperage. Not sure if this will work but It's in the post.

The Vin is variable. I don't think I would be using more than 12v, but I thought it would be a bonus if I could still change input.

For T1 could I raise the value of R2, would that help.

 

R8, R9, R10 are 0.22 ohm not .022
Sorry.

 

Well, with 3, .22 ohm resitors in parallel you'll only get ~9A pk.

 

I changed R8, R9, R10 to .15 ohms 10watt and it was running at about 12.5A
The MOSFET got really hot so I put on a big cpu heatsink with a fan

The MOSFET is
BUK9535-55A TrenchMOS transistor logic level FET
I waiting for
HUF75343P3 Fairchild ULTRA-FET POWER MOSFET 75A 55V
I got this in thinking that if the amperage rating was higher, the MOSFET would not be under so much stress.

The truth is I don't really know what to look for in choosing the MOSFET, I'm looking mainly at Vds Id and Idm and I think I understand what the Rds(on) is.

The Voltage is variable, but I'll just stick to 12v it will make things simpler.
If I lift the resistance of R2 would that help with the temp increase in T1

 

You need a mosfet with the lowest Rds(on) value to minimise the voltage drop across D-S hence lowering the overall device dissipation...

Changing the value of R2 will do nothing....other than I quote, "R2 and C4 are a filter to get rid of the spike that will be there at turn-on. "

The power dissipation in the mosfet is goverened by load current and supply voltage even when the mosfet is turned hard on...

What are you actually trying to achieve?

__________________
Eccentric millionaire financed by 'er indoors

 

I would like to get the PWM to run steady at about 28Amps.

But I would like to learn as much as possible as well, if I'm going to make changes, I would like to have some kind of reasonable idea want I'm doing.

This is my first project.

The HUF75343P3 has a much less Rds on 0.007 to 0.009
BUK9535-55A is 0.024 to 0.032

 

At what frequency are you running the PWM? You also have to be a little careful changing MOSFET’s. As Rds on goes down, you will find that total gate charge goes up, which basically means that you’ll need more current to switch the MOSFET at the same speed. If you don't, switching losses will go up. You may want to consider using a totem-pole driver for the FET.

 

I think he is attempting to drive a motor, PWM frequency, motor inductance and supply voltage all go into the mix as I'm sure you know....I was under the impression that the SG3525 had Totem pole outputs ?

Whether 28 amps is achievable is a bit of a challenge....imho

__________________
Eccentric millionaire financed by 'er indoors

 

Yes, but your going thru diodes on both outputs, so they don't help for turn-off. Also, the SG3525 is only good for ~200mA (500mA max).

28 amps shouldn't be that big of a problem with the right MOSFET's and a proper driver.

 

I would like to keep the frequency variable. It's adjustable from 100hz to 5000hz, will this be a problem.
I'm still waiting for HUF75343P3 I bought them online so have to wait for delivery. I got 3 just incase.
Totem Pole ??
I did a search and it looks like the gate is connected between 2 transistors.
I think I understand, it supplys a larger peak current to help charge and discharge.
Is this part of the chip or will I have to add this.

Cheers

 

Try Googling a tc4427 which is a mosfet driver.......

__________________
Eccentric millionaire financed by 'er indoors