Constant Current power supply, help please

[oliverthepig]

Hello everyone,

Ive bought the k142 driver boards from oatley electronics here in Australia.
I decided not to purchase the linear constant current supply as I thought that I could make a pwm constant current supply that would run more effeciently.

I made the circuit described in the following website,

http://www.eio.com/crntctrl.htm

Note that the base of the transistor is not grounded looks like a mistake..

I'm using a p type fet a 9540 in place of the darlington transistor.

I've found that the fet gets very hot, and does not seem to be switching cleanly, I've found that placing a electro cap at the output of the opamp tends to dampen the response of the system and works a little better as the fet runs cool and a constant current can be achieved at standstill, however this idea appears unsuitable as it stops the supply from sustaining a constant current while stepping the motor.

Any ideas or thoughts would be greatly appreciated

thanks David

[niclatrique]

On the schematics these are NPN transistors. Why don't you use N channel fet instead of P ones? they are easier to control and cheaper to buy.

[NC Cams]

Try building the circuit "as is"/as designed before you start swapping in different parts.

You might also want to do some investigations into the LM3524 or SG3524 application notes as this is essentially an industry standard IC for PWM switching power supplies.

THe neat part about logic fets is that you can use logic level devices and switch them directly from only 5 volts but they switch faster and harder with 10 VDC Vgs.

If you use regular fets, they work best at somewhere in the 15-18 volt Vgs range. Hot fets are either oscillating or are not being fully enhanced when being turned on.

The faster you try to switch them, the more you need a true fet driver - you can't drive them fast or hard enough with a direct connection to the LM3524, especially at higher switching speeds.

[Gootee]

I'm not an expert, and am not familiar with the devices you are using, but that circuit seems to be not quite suitable for using an FET directly in place of the Darlington. The best thing to do might be to try it first with a proper Darlington, or a couple of BJTs in a Darlington configuration.

FETs must be switched fully, and quickly, to run cool. Also, many of them are typically used with a small resistance in series with the gate, which might help to prevent some ringing/oscillation that might be contributing to your overheating problem.

Depending on the driving voltage and the FET specs, you might have to add a small BJTransistor (or other) switching circuit, to use part of the power supply rail to drive the gate, just to drive the gate hard-enough for the FET to work well-enough, if you insist on using an FET.

It can be made to work. Maybe you should google for FET OR MOSFET switch OR switching circuit OR circuits.

[Torchhead]

You need one of these "Constant Current" controllers per motor. You have to limit current based on the current to each motor. It's a brute force way to do what modern chooper drives do anyway.

You can't simply replace a bipolar design with a FET. Bipolars turn on with current and on the PNP shown the base only need have enough current flow from the base to turn on the device. The FET uses voltage and will put the FET in the linear range (high power dissipation) unless full turn-on voltage is acheived. Drive signal for on/off need to have crisp rise and fall times.

Takes a lot of electronics to fix a bad design

[oliverthepig]

Thanks for the responses everyone, its giving me some very good idea's to try....
I've just got some lmd18245's in the mail so I may admit defeat yet, and build a picstep.

Thanks again,

Dave

[apollonono]

Give me an answer if I agree to see postscript next.
http://www.cnczone.com/forums/showth...225#post602225