Imformation overload, Homebrew CNC
(Post #1 )

Ok, now that I've been throughly (much by my own doing) drown in imformation.

Ok, This is where I'm at right now. (Things are working, or at least moving in a linear fashion).

I'm using this controller,
http://www.stepper3.com/31.htm

I'm using these three motors,
http://cgi.ebay.com/ws/eBayISAPI.dl...item=2582180840

I've hooked into the +lines 1 ohm resistors(no idea what the watts are).

PSU, is standard older ATX 300 watt.

I'm running KCAM4.0. (Have noted, I should use TurboCNC).
on a Win98 system p166 -w- 32 mg ram standard laptop.

I was noticing some flatness on circles, say larger than 3" dia.
This flatness occurred on the x-axis.
(I was told this could be caused from a machine/mechanical problem).
Check system, Same type x-axis configuration as y-axis.
(would think flatness would occur in both directions if mechanical).
Check motors, x-axis motor seemed to be really light in the torgue area.
(this was while running KCAM4.0).
Tried TurboCNC, the torgue output was there for the x-axis but seemed to stutter somewhat.)

??Idea's??
I downloaded the resistor calculator. If my controller has "FETS"
my motors require 5.00 ohm /7.2 watt resistors.

If my controller has resistors. my motors require 4.25 ohm/6.12 watt resistors.

FETS and RESISTORS are alien to me. as are the watts for the motor power resistors.

Can I find the required power riesistors given the info above?
And do I have a decent setup for a homebrew system?

Thank you much in advance for any input.

I think that board has transistors for outputs. You need 2.55 ohm resistors, at least 10 watts. Probably you can find some 10 watt power resistors at www.mouser.com

Don't use the 1 ohm resistors unless you are using 1 per motor, in which case it needs to be 20 watts. If you are using one per motor, then it's probably ok.

If you are using Turbocnc you need to run it in DOS mode, not a dos box. Click shutdown and "restart in dos mode". If you are running it in windows it will be just as rough as Kcam.

Slow the cut speed down a bunch and see if you are making better circles. It sounds like you are loosing steps.

Eric

Those are some good suggestions from Eric (BalsaMan).

If you're not losing steps, there are a couple of other things that you could check, too.

First is to verify that your configuration is, in fact, correct...I know you said it was correct, but it never hurts to go back and verify it, again. Over a few hundred, or a few thousand steps, errors will be cumulative. Check to make sure your config is as perfect as you can make it.

Second is backlash. This is an error that depends on the fit of *each* screw/nut pair, individually. It's possible that there is greater backlash in the X axis than in the Y axis...this would account for a slightly flattened circle, too. It rears its ugly head, whenever you change directions.

-- Chuck Knight

Slack in a system is often overcome in one direction only by the rotation of your cutter. This will present itself as a flat on one side only.

When doing test cuts, if possible, use a plotter pen or ball point pen chucked in your tool holder. ( remember not to power the router on).

Easier to try, You might try to also cut your test circle ( or square in two passes on opposite directions. This will mildly exagerate the slack because of cutter rotation and exibit itself on the horizontal parting line between the two cuts. Keep the overall size of your cuts down to less than caliper span. The error on a 1" cut will be the same as on a 16" cut if it mechanical.

You might also send your g-code to a willing volunteer with a known tight machine and re-assure yourself the code is good.

If none of this defines your problem, then you have control problems.


Best of luck,

Phill Pittman
digicarve@verizon.net
www.masterwerkes.com

I think I've gotten the bugs worked out, I switched to TurboCNC.
And tried making spirals, I think it turned out pretty nice.