Question about Microstepping and axis resolution

[Tom Kay]

Hi All;

I am working on a router design and hope to start soon. I wanted to cnc a mill, but I've opted for a router instead. Either way, this hobby is too fascinating to ignore for much longer.

So, I will use chains and #40, 12 tooth sprockets to drive the x axis direct from the stepper motor. Assuming the pitch diameter of the sprockets is just under 2 inches, this gives me an x axis movement of about 30 thous per step of the motor. I'd like better resolution than this.

So, I have a few choices. I could use smaller sprockets (I won't) or I could use timing belts to drive the sprockets at a geared-down ratio (more likely).

Another choice is to rely on micro stepping. Years ago I bought bipolar stepper motors and drivers from Dan Mauch's Camtronics. The drivers are half-stepping, not microstepping. I assume the best resolution I could get is about 15 thous per half-step. Is this concept correct?

Or let's say I buy Marriss' Geckos and use the 10 microsteps. I assume that I'd get a resolution of about .003" per step (actually 1/10th of a "normal" step)?

Is the 10 microstepper really 2000 steps per rev? Is each step as reliable as the basic 200 steps per rev of the motor without microstepping? Do people typically and successfully rely on microstepping to achieve the tight resolution that they want? Or is microstepping not that reliable?

Any thoughts would be appreciated. If I need to build in more of a timing belt ratio, now would be the best time to know.

Thanks, and I hope this wasn't annoyingly wordy. Tom.

[jcc3inc]

Sir,

#1) I would respectfully submit that you should consider rack and pinion drive instead of sprockets and chain. Sprockets with few teeth produce a non-linear output distance vs input angle change. Also, you can achieve lower distance per revolution if you use, say 24 DP gearing.

#2) Recommended rack and pionion would be 24 DP, 20 degree pressure angle components with the pinion being a MINIMUM of 18 teeth to avoid undercutting. Gear rack can be had from www.stdsteel.com/gear.htm . Although they claim to stock 20 DP and coarser, they do make 24 DP as we used to buy a lot of it. Pinions can be had from Boston Gear, Browning Gear and Sprocket, and others. You should spring load the pinion into the rack with a force which = Max drive force x (Tan 20 Degrees) which is Max drive force x .364. Thus if your max driving force is 50#, you load the pinion into the rack with 18# of force.

#3) As to microstepping with Geckos, I use a set of these on my test bed machine and have attached encoders to the steppers so that positioning could be confirmed. My results are that running the machine without the load imposed by the router's cutting, the microstep increment is exactly on 99 % of the time. (my smallest increment of motion is .001") Thus although some have said that you cannot count on the accuracy of microstepping, I have not found their comments to be true.

Best wishes on your enterprise.

Regards,
Jack C.

[Tom Kay]

Hi Jack;

Thanks for the detailed reply. Much appreciated.

I had a bit of a notion that the longitudinal movement of chain-driven parts might not be smooth, or completely linear as the 12 tooth sprocket turns and pays out chain. Now it becomes a question of how non-linear it is. I would assume that there aren't wild back and forth fluctuations as the sprocket feeds movement to the chain. But this is an assumption.

Part of my desire on this first build is to make a low cost demonstrator. I'd like to get some repeatable use out of it, and possibly carve intricate household items, such as trim around doors, or fireplace mantlepieces, but not super-accurate machine items. Eventually I'd like to see if it can carve full scale Spitfire replica propeller blades out of laminated wood, and do 5 of them as identical as reasonably possible (without going nuts to get it dead-on). So I'd be looking for a resolution of .003" to .005" in this first attempt at homebuilt CNC equipment. Considering lots of homebuilt aircraft have hand-carved, or "follower"-carved props, if I can get this process repeatable, the accuracy should be acceptable.

The reason I have ruled out ballscrews for this first model is cost, given the 7 or 8 foot desired length in the x axis. That's a lot of ballscrew, especially if I do get picky about accuracy. Whip effect is also a concern at higher revs. So this limits me to other drive means.

Jack, I did consider timing belts, and very long belts can be obtained. Do you know if this method suffers from the same non-linear jaggies as small sprockets? And while we're here, do you have any idea of the amount of plus or minus fluctuation in sprocket-driven chains?

I'll have to investigate the rack and pinion approach, but my gut feeling tells me that this takes me out of my cheap demonstrator range. I'll follow your links though and see where that takes me in terms of cost/do-ability.

Lastly, in terms of microstepping, I am happy to hear that you have had success. This is a very handy way of increasing your resolution by an order of magnitude. I don't have Geckos but I'm certainly aware of them.

Cheers, Tom.

[jcc3inc]

Tom,

You can email me
jccinc-at-owc-dot-net

and I can send yo some pix etx.

Regards,
Jack C.