Home build linear bipolar driver card!

[essele]

Hi,

I've started working on a home-grown driver card using many of the principles from the linistepper, to actually driver bipolar steppers using linear outputs rather than PWM/current chopping.

I've built a couple of other controllers now using more traditional modules or TB6560's, but I really want to look at the linear option.

My goal is a low cost (£40ish) 3-Axis controller, probably with PWM control for the spindle - that anyone can build at home.

I've just posted a blog covering (in a long winded way) my though process and some of the preliminary work, it's very much still in the formative stage, so I'd love feedback either there or here.

Random Thinker Blog

Many thanks,

Lee.

[doorknob]

No comments yet, just wanted to let you know that you're not talking to yourself...

I'm only vaguely familiar with PSoCs and Atmels, though, and being already a bit overwhelmed by trying to force-feed my brain with the likes of ColdFire and Cortex M3, I'm a bit gun-shy about the prospect of trying to make sense out of other architectures at this time.

So I may end up just sitting on the sidelines watching but without a whole lot to contribute...

[bharbour]

Hi,
I've started working on a home-grown driver card using many of the principles from the linistepper, to actually drive bipolar steppers using linear outputs rather than PWM/current chopping.

What kind of driver are you thinking about? Voltage Source? Current Source? Either will dissipate a LOT of power as heat. The current source could be configured to give higher step rates than a voltage source, but it will dissipate more power.

I am curious why linear as opposed to the current chopper architecture that is commonly used now?

Regards,
BobH

[H500]

It looks like a fun learning project. However, I see problems. I'm not trying to rain on your parade, but if you're looking for tb6560 level performance, I think you will be disappointed. If you use a low voltage (ie: 5v) supply, you won't get more than about 60-150 rpm out of it if you're lucky. The motor inductance will limit the current at higher speeds.

If you use a high voltage, your transistors will dissipate enough energy to heat a small room. You will need big heat sinks and a big power supply to handle the wasted power. 2 amps is really pushing the linear design beyond where it works well.

Why did you abandon the tb6560 board? The missing steps problem is easily solved by bypassing the opto-couplers.

[essele]

Thanks for the comments so far, especially H500 ... I did suspect there was a flaw in my plan, it all seemed a little too easy.

I had intended using 5v to avoid the heat problems, I did wonder whether there would be performance issues as a result.

I will probably still build it, I've learnt a huge amount doing this stuff and it's much more tangible than my usual software stuff.

On the TB6560 ... I spent quite a bit of time experimenting and debugging a chinese card, I did manage to get it working fine with a combination of changing the timing caps and removing the opto-couplers, but I guess I lost confidence in it. Also, from the spec you need to use very long pulses, something that seems to be difficult with Mach3 (although I now use EMC2 anyway), so I'm convinced most people are using the driver outside the published specs.

(Relevant stuff starts about half way down the second page: Zen Toolworks LLC • View topic - Cutting polycarbonate?)

I then built myself a card around three Pololu A4938 modules (details are on my blog) and included a slow-start PWM controller for the spindle motor. There's something really nice about using your own stuff ... it's now my default controller card, and I used it everytime I mill.

But ... I really don't like the hiss/whine from the motors when they're stationary, I know it's probably fine, and with the current set low it's not bad at all. Also the in-motion noise difference between the A4938 and the TB6560 was quite startling (at the same settings) so I know there is definitely scope for improvement.

So I guess, I don't really *need* to do this, but I'm keen on understanding more ... and whether it works or not, I acheive that!

Cheers,

Lee.

[H500]

I'm quite sure 5v will give you horrible performance compared to the tb6560 with 30v. You shouldn't need long pulses if you bypassed the opto and make sure that the step and dir signals are configured with the proper phasing...at least from what I see.

If noise is an issue, the 60v 6amp drive I posted in the open source controller board forum has no hiss when the motor is turning. When stopped, one of the positions hisses softly.

You can even use the power section with your PSOC instead of the CPLD. I can get 1800rpm with my motors.