My CNC PCB Mill

It's finally coming together. I finished the mechanical part of the xy plane (just need to mount the table) and I tested my x-axis with a breadboarded L297, 4 Microchip TC4420s (rated @ 6A, if my memory serves me correctly), and a 555 pulse gen. It works great, amazing stepping resolution even in full step. However, the TC4420s like to get quite hot if I step the motor at a higher speed. I'm not sure if I've done something wrong or just need to heatsink the TC4420s. I have the outs of the L297 directly into the TC4420s (each is a half bridge) and then a pair going to one coil.

I'm not really as good with mosfets, would I need a resistor between the source and gate?

(I don't have time to draw up a schematic right now, but I'll post one soon. After I breadboard the entire controller I'm going to have it cut a PCB for the controller, then I'll switch it over and solder it all.)

4 Microchip TC4420s (rated @ 6A, if my memory serves me correctly),

Click to expand...

That's 6Amps of peak current I might add.

However, the TC4420s like to get quite hot if I step the motor at a higher speed.

Click to expand...

You are driving MOSFETs with the TC4420s and not the steppers directly I hope!

kchriste said:

That's 6Amps of peak current I might add.

Click to expand...

Thanks...

kchriste said:

You are driving MOSFETs with the TC4420s and not the steppers directly I hope!

Click to expand...

Why would I use the 4420s then? My steppers are relatively tiny (from old printers and scanners), but I haven't done any math yet either I guess.

how hot is "hot"?

if you can touch them without getting a burn, then they're not hot. a lot of industrial electronics are rated to run at 80 to 100 C, some hard core parts at 150 C... check your datasheets!

The TC4420 is intended as a buffer/driver between your microcontroller and a high power mosfet.
Although it is rated at 6A it has a high on resistance of about 2.5 ohms and so would get hotter at high load compared to say an IRF730 which is rated at 5.5 amps and only has 0.75 ohms resistance.

Or, for that matter, an IRF540 which is rated at 27A, 0.085 on resistance, and only costs a bit more than the IRF730 (about $1.50 each)...

of course, you need to drive the gate with more like 8-10 volts (not 5v logic levels) for switching high currents, which is exactly the point of something like the TC4420...

The fact that the TC4420 is referred to as a "MOSFET driver" should clue you in to the fact that it's NOT for driving a load, it's for driving a MOSFET which drives a load. switching a mosfet requires charging/discharging the gate capacitance, and for power mosfets the gate is large and has an appreciable amount of capacitance, therefore to switch it FAST you have to drive the gate HARD, and this results in very short spikes of high current when it is switched, which is why the TC4420 is rated for 6A peak but you don't even see a rating in the datasheet for continuous current, because there is effectively no sustained current when driving a MOSFET. So you should NOT be using it to drive a load with a continuous current, you're just going to destroy them. If you use it to drive mosfets like you should, I would expect that neither the chips nor the mosfets will even get more than a little bit warm driving the small motors you are using.

If you want a single-chip solution to driving stepper motors, look for the all-in-one stepper driver chips, I don't know any part numbers off the top of my head but I know I've come across a few in the past (in fact, they're often salvageable out of old printers and stuff), and many of them not only include enough outputs to drive one or two steppers per chip, but they usually include logic decoding (so you just feed it two logic lines, one for direction and pulse the other to step it), and sometimes include chopper drive circuitry so you can run your motors with constant current and get way better performance without completely abusing your motors. They're not cheap, but that's the price you pay for not having to use a bunch of MOSFETs and drivers and other parts.

Thanks for all the help...I think I'll buy some L298s to compliment the 297s...

And I did a quick calculation...I was feeding 4.5 amps of continuous current through the 4420s. Oops

Did you take the on-resistance of the motor coils AND the 4420 into account there? If you were actually feeding 4.5A of current through them, with an on-resistance of 3 ohms as it claims to have, that would be like 60 watts of power dissipated, those little 8-pin DIP packages would incinerate immediately... So chances are you weren't putting nearly that much power through them if they're still in one piece. On the plus side, when you switch to proper power transistors or a dedicated driver, the motors should run better

I didn't account for the 4420, and I measured 2hm: across one of the coils with my crappy analog meter (analogue for anyone who speaks 'cool' English ). So V=IR, R = 2hm: V=9 volts...

Revised- 9=5I

I = 1.8A

They're the five pin TO-220s...the DIPs probably would've been better though.

I'm looking to the best (read: cheapest) way to get the motors up and running. Thanks everyone and I hope to have the mill running in the next few weeks...I'll post some pics!

Ok...update. I think I'm going to get some FQP13N06L s...here's the datasheet:

https://www.electro-tech-online.com/custompdfs/2006/12/FQP13N06L.pdf

Do they look ok to everyone?

100 mOhms still seems kinda high... couple questions

what is your gate voltage going to be?
is low Qg something you're specifically looking for?

the FDP20AN06A0 offers 17 mOhm Rds at 10v (probably closer to 30 at 5v), the Qg is a bit higher

https://www.fairchildsemi.com/ds/FD/FDP20AN06A0.pdf

I'm looking for something that'll do at least 12 volts. The Rds doesn't matter too much, although faster switching times would be better for stepping the motor faster (though I don't think I'll go that high with my step speed). The only reason why I wouldn't buy the FDP20AN06A0 is that it's a bit more expensive...

If you want realy low heating you can use the IRF3205 whith only 8mOhms.but they cost about 2 Bucks a pice