Information needed for upgrade

[Hack]

I am currently using allthread with a delrin nut for the lead screw of my cnc router (1/2-13) and 276 oz steppers, xylotex board and 24 v power supply and Turbocnc.

In order to get some more speed for my rapids, what would make more impact:

Upgrading to a more efficienct lead screw (probably acme due to budget)
Increasing stepper size (can't go much bigger with the xylotex board)
Upgrading to a bigger power supply (can run up to 35 volt max with xylotex)
Changing computers and software to drive everything (perhaps to Mach 2)

Thanks

Hack

[CJL5585]

This is an educated guess but believe it is a toss up between the leadscrew and the power supply voltage.

A different leadscrew will definately help. However, replacing a leadscrew is a lot of re-engineering and work with the bearings, motor couplings, etc.

The 24 VDC supply is probably the least expensive route to take as it will also give you additional speed in rapids. Also, it is probably the least amount of work to change out.

Steppers with higher inch/oz rating should not make any difference in speed. Just in torque. However, higher voltage should also increase the torque rating at the present speeds.

Hope this helps,
Jerry

[ger21]

I'd try 1/2-8 2 start acme. 6 ft is only ~$25 at http://www.mscdirect.com
You can also get premade anti-backlash nuts from here. http://gonebowlin.com/dumpsterCNC/


Like you said, you can't reall use bigger steppers. And anything over 30V can easily damage a Xylotex. I've heard people say Mach2 runs smoother, but I doubt it will go much faster. Maybe a little.

If changing to the acme doesn't help, you probably need to go with Gecko's and at least 48V. You could build drives from http://www.embeddedtronics that will run @ 48V. Not sure what they would cost, though.

[jgro]

Another option is to change the microstepping of your Xylotex. I ran my machine with 1/2-10 ACME with a 24 volt power supply and 116 oz steppers and was getting around 60 ipm. That was with it set at 1/8 microstepping. I've since changed it to 1/4 stepping and am getting around 115 ipm. If you figure it out:
200 steps per revolution X 10 threads per inch X 4= 8000 steps per inch.
1 inch divided by 8000 = .000125 inches per step.
I can live with that kind of resolution.

jgro

[Hack]

I have thought about changing the microstepping as well. I read an article on the net saying that microstepping is bad for torque in one way or another. http://www.machinedesign.com/ASP/vie...strSite=MDSite (Thanks GER21)

Also think that I can turn the VREF setting of the xylotex board up a bit as well. Currently set a 3.4 and I beleive the max is 3.62. I'm sure that would help some as well.

I was under the impression that with a power supply, if you double the voltage, you will double the performance. (with in reason). So I am thinking that going from 24 v to 30v power supply, I would gain another 25% performance. Is this correct? Any recommendations for a good 30V power supply?

Thanks all?

[anoel]

Originally Posted by ger21 I'd try 1/2-8 2 start acme. 6 ft is only ~$25 at http://www.mscdirect.com You can also get premade anti-backlash nuts from here. http://gonebowlin.com/dumpsterCNC/

I'll second DumpsterCNC...

I just had him make up 3 1/2-10 2 start nuts for me. He was very quick and the nuts are very cleanly made. The quality is seriously as nice as the simliar ones from Nook and other places.

[DieGuy]

Don't turn up the Xylotex Vref. It only limts the amps and a .2 increase to 3.6 volts is on the ragged edge. A 30 volt supply is a better option, you will need to find a 20-22 volt transformer. I run mine at 28.5vdc. It depends on the back emf your setup produces as the total you can drive it to is 35 vdc. Make sure you have adequate active cooling for the heatsinks, I run 2 12 vdc fans in series to blow diredctly on the heatsinks. When my vref was set to 3.6 vdc I lost steps left and right and the machine would not accelerate without dropping steps, Mine is now set to 3.4 vdc and it works the charm.

[Hack]

I emailed Jeff at Xylotex and asked if he could supply jumpers for me to switch to 1/4 microstep instead of 1/8 (he said I could find them at a local computer or electronics store - Man is his support fast - Excellent job Jeff), and he wrote this:

"The only reason you should want to go to 1/4 step is your pulse
generator can not make a fast enough pulse train. You will loose
some of the microstepping smooithness by doing so, and it is not
particularly recommend to run at lower than 1/8th step."

Ok can someone translate this to English for me. What is the pulse train? I was thinking this had more to do with the computer processing speed than anything and not necessarily of importance here. Shows how little I know. I also can't see loosing much smoothness by switching to 1/4 microstep.

[strat]

Originally Posted by jgro Another option is to change the microstepping of your Xylotex. I ran my machine with 1/2-10 ACME with a 24 volt power supply and 116 oz steppers and was getting around 60 ipm. That was with it set at 1/8 microstepping. I've since changed it to 1/4 stepping and am getting around 115 ipm. If you figure it out: 200 steps per revolution X 10 threads per inch X 4= 8000 steps per inch. 1 inch divided by 8000 = .000125 inches per step. I can live with that kind of resolution. jgro

am i glad to see this i just changed to 1/4 instead of 1/8 also and i also nearly doubled my speed ..... glad to see the same effect from someone else

[ger21]

Originally Posted by Hack I emailed Jeff at Xylotex and asked if he could supply jumpers for me to switch to 1/4 microstep instead of 1/8 (he said I could find them at a local computer or electronics store - Man is his support fast - Excellent job Jeff), and he wrote this: "The only reason you should want to go to 1/4 step is your pulse generator can not make a fast enough pulse train. You will loose some of the microstepping smooithness by doing so, and it is not particularly recommend to run at lower than 1/8th step." Ok can someone translate this to English for me. What is the pulse train? I was thinking this had more to do with the computer processing speed than anything and not necessarily of importance here. Shows how little I know. I also can't see loosing much smoothness by switching to 1/4 microstep.

The pulse train is how fast the computer can send steps. I'm guessing you 2 were using TurboCNC, which can have a pretty wide range of step rates from PC to PC. Mach2, otoh, will always output 25,000 steps/second. With 1/2-13 and 1/8 microstepping, you need 20,800 steps per inch.

You didn't say what control software you're using, but go ahead and set it to 1/4 step, and see if it makes a difference. I'm a bit curious, actually. I'd be surprised if your speed doubled, though.

You didn't say how fast you're going, but to go 30ipm, you have to spin the motor at 390rpm. @ 24V, most steppers torque curves start dropping to about 1/2 or less by then. Consider that 1/2-13 is not very efficient, and you're wasting a lot of power there. 2 start acme screws are probably at least twice as efficient, and the motor will only need to spin about 1/3 as fast, where you'll have much more torque available.

I'd start with the cheap fixes first, and work you're way up until you get what you want.

[strat]

i'm using mach 2 but an old p300 and when testing for timing i had to go with the slow clock

[Hack]

I am using an old Pentium 200 MMX for my computer and TurboCNC V4 for controller. Is this computer to slow? How do I know or figure the pulse train?

I will try switching to 1/4 microstepping tomarrow and see if it works, if not, i will probably asdjust the vref a little higher then look for a different power supply. I think this is probably the route to go for the cheapest fixes.

Currently I am only getting about 22 IPM max and thats stretching it. I keep it set at 20. I was realy anticipating getting closer to 30 IPM with the 1/2-13

Thanks all

Hack