Re: 10 amps to drive steppers

Here are some of the problems that I ran into when I tried to hot rod this
circuit:

I started with a layout that Kevin Carroll sent to me. This is similar to
what Dan is selling, only Dan's is much improved. Kevin said that he had
run this design so hot that the mosfets had unsoldered themselves before.
He even had a switching arrangement that would reduce current to the motor
when it wasn't in use-using AND gates.

I built the circuit, milling and drilling the traces with my 2 amp
controller. I installed the thing in a big old XT type case, with a 24 volt
15 amp transformer and a HUGE heatsink for the mosfets with two muffin
fans blowing on everything.
The low current switching sort of worked- but actually created a lot of
noise in the circuit. When it was physically pulled out of the circuit, it
was audibly quieter. Increasing current was real flakey- missed steps and
lots of noise.
Talking to Kevin, he had some of the same problems. So, out came the low
current switching.

As I increased the current above 6 amps (minus the AND gates), the signal
was still getting flakey. The circuit was missing steps all over the
place. My input signals are shielded using single conductor braided audio
cable, grounded on one end only and tied together at a common ground point.
I physically moved the cables around and found that by turning cables 90
degrees to other cables things improved somewhat. So, I tucked cables in
and wiretied them down the best that I could. Are the mosfets acting as
little antennas, emitting lots of noise from any sharp corner on them?
Satisfied that I had the noise down as low as I could get it, I proceeded
to run the current up. Somewhere along the way I blew an axis card. It was
within a day of jacking the current up. I ended up replacing one phase on
that axis (two mosfets) and then adjusted the current down to 6 amps on X
and Y and 5 on Z. My X and Y have triple stack 34 frame steppers, while
the Z has a double stack on it. It's been about a year ago now. This
controller has worked trouble free ever since and it gets used every day.

Things that I could have done better:
Shielding- use better quality shielded cables and then shield them some
more. I think I used Radio Shack brand on the previous controller.
I will use an etched circuit board next time. I suspect the milled board
probably contributed to noise due to ragged edges etc...
Layout- I will make sure that signals are miles away from motor leads and
power supplies.
Cooling- how about three fans on the mosfets? Maybe an air conditioner in
the shop right next to it!

Anyway, that's my mosfet frying story- no flying pieces or burned fingertips.

Dan Falck






At 03:02 AM 6/13/99 +0100, you wrote:

>From: "Mo" <mo@...>
>
>> know how it goes. The only modification to the board needed to allow
>higher
>> current is to parallel the existing .1ohm sense resistors with another
>.1ohms
>> which will double the adjustment scale to 10 amps.
>>
>> Wish Me Luck!
>
>Matt,
>I wish you luck!
>It is far nicer to explore at the ends of the envelope when you have a
>safety chute.
>Getting the board to run at high Amps is one thing, getting it to run
>reliably for years is another.
>My advice is to analyse and improve your protection system to keep the
>experimenting fun and painless.