Alan, are you using LONG cutters or short ones. I mostly use the SHORT PRO-TECH 1/4 cutter.
About 2 weeks ago I was at H/Ware Centre and bought a nice LONG upcut cutter. When doing the same cut with both cutters, the quality on the SHORT cutter is much better than the long one. Maybe the long cutter has some FLEX or something.

I was running the test patterns with my shortest 6mm solid tungsten slot drill. The mirror frame test was run with a 11.2mm straight router with 1/2" shank - both very stiff.

I never checked the Y rack against the X racks over their full length, but when using short pieces to join them the teeth seemed to mesh well. I know the two settings should be the same but my vernier is not agreeing with that. It was out by 0.2mm over 100mm which should equate to 2mm over 1m - I will run more tests today. Anybody have any ideas how to measure accurately over long distances other than a tape measure? My longest vernier is 300mm.

What about those LASER measuring units..don't know if the tolerance will be fine enough.

I have the same issues. When cutting a 300 x 300 square, I was a fraction out on the Y-axis. Changed the steps on Y differently to the X with the aid of the Mach-3 calibration.
When I cut 2 chair halves out, one lying Horizontal and the other Vertical on the same sheet at the same time, they were not 100% the same size. I then went back and made X and Y the same steps.
Did another cut and now both were the same size.
I was wondering if you speed up or slow down the cut if that has anything to do with accuracy.
Will be following this closely because I also want to CALIBRATE my machine 100% on the max of the axis.

You could drill two holes, wide apart, through a batten, into your spoil board. Then try and dowel the batten to the spoilboard in different directions, looking for the error on the dowel holes.

Or, just cut a X-batten and Y-batten to the same length, lay them on top of each other and check for difference.

you couldn't be more wrong, I have had my belt drives in use for more than a year, no broken belts, no missed steps, best of all no backlash,
they have been performing perfectly with out issue, on both machines if you over load the gantry with to much force the pinions will walk out of the rack,

my drives are better for several reasons, they are steel, which makes them stronger, easier to build, and weld, they are a welded assembly which is stronger than bolting and it does not allow flexing or racking as a bolted assembly will, also they allow a 4 to 1 reduction as most designs are of a 3 to 1 reduction, 4 to 1 gives better resolution, they also out perform the O.M. 7.2 hands down, with better cut quality, better jog speeds, and reduced costs, (cheaper than buying O.M. 7.2)
if they are so poor why do I get such great cuts?
quite frankly they are hard to beat, I do not believe it is possible to build a 4 to 1 drive cheaper, or get better performance from them...
//chopper

not again please! :)

Chopper I am so glad it works for you:rolleyes:

I can only laugh.:D :D:D

The only thing I have to add is everyone has an opinion.

I am happy that chopper's solutions works well for him.
But, if I look at my machinist handbook and such from engineering school, I find that some of the solution being used by Chopper does break the Diameter:Diameter and pitch engagement "rules"

Again, happy it works.

I personally feel belt drives are a good and affordable solution for adding performance to an already good CNC machine design. I must add though, before building my first MM, the Multicam 4896 (1000 series) machine I worked with broke belts about 3x a year, thus I have used the 7.2 Gearbox solution on all 3 machine builds.

....back to the couch to watch the kids "race cars" and dance to SpongeBob SquarePants.

Pablo,
I am sorry that you seem to think that this is some kind of a problem,
it is just an update since my drives have been in use for over a year, with great success, I referenced to Gerald's thread since he stated that the lack of tooth contact could be problematic, which it has not, so rest assured that it is not an attack.....
Hennie,Sean, thanks I am also glad that they are working great, and yes they break the rules a bit, that is true, but after a year and many hours of operation there has been no issues,
Kobus, I am glad you find humor in my threads although I do not see what is so funny in this one...but hey whatever works for you... all I merely did was state why the design was better, none of it was untrue, if you do not agree I guess that is your choice, metal head must think they are a good design since his drives are based on mine, I even drew the parts for him, except for the wings on the plates, seems you guys are kinda thin skinned when some one differs from your opinions, I thought that was what this site was about, to share different ideas and views, just because my ideas vary from yours and Gerald's doesn't make them wrong, :D
//chopper

My drives are also made of metal and after 19 months of working every day there are no issues with them, hey for the price I paid for them they even came with the motors attached to them with also no broken belts ,no missed steps and even no backlash that were noticed and they were made in China.Now that is more superior.:D

Sean, did you build go-karts for the little ones?

I cut a scooter for my little one & he distroyed it within the day...
Now a go-cart will surely get him jumping... Any nice design around?

Ken pm me your e-mail I will send you something

Hennie,
I wish they had real go-carts. They were playing races with their little matchbox cars - dozen's of them all around the couch!

Vroom....vroom.

:D

I was going to purchase a 5:1 gearhead for my NEMA 34 servo. After reading this thread, im going to design my own. I'll post solidworks renderings possibly later tonight.

I have a question! how do i calculate how close i can place 2 pulleys? I want to place them as close as possible to save space. specifically a 11 tooth 1.313 pitch diameter pulley AND a 34 tooth 4.058 pitch diameter pulley using a 3/4" kevlar "L" (not XL or MXL) belt. How can i calculate the correct distance for maximum tooth to belt contact. Thank you! :)

If you are like me, (return my maths to my school teacher the moment after my exam) There is an easy way out, you can do it graphically by drawing to scale on any cad program, then measure.

Hope that helps:

http://www.sdp-si.com/Cd/default.htm

YRD! That is perfect!
KENC- I was going to do that, just haven't got that far yet haha. time will tell! Thanks!

Essentially it looks like you at least want roughly 40-45% of the teeth on each pulley engaged.

Attachments:
"Belt diagram" show a normal minimum center to center distance, way to close because i do not have enough teeth engaging in the small pulley (in my opinion).
The picture next to it is the minimum distance that elicits the maxmimum tooth meshing in the smaller pully.

"belt diagram 2" Utilizes the minimum distance from center to center, and modifies the belts path with 2 ball bearing aluminum rollers that are adjustable. They also act as belt tensioners. The original short belt cannot be "forced" and used. A longer belt must be measured for the new modified path. This allows minimum distance used, as well as engaging maximum amount of teeth with the pulleys. Like i mentioned in diagram 2, i will have to find the extra distance it takes to allow the belt to be slipped onto the pulley, say 1/2" or so.

Excuse the typos. I hate Microsoft paint haha.

I should of made the servo motor about 5 times the size of what it was, but you all get the picture.

I'm shooting for 1056 IPM (which just so happens to be exactly 1 mile per hour)

to calculate what gear ratio i need i did the following.

1" face diameter output x 3200RPM / 1MPH x 366 (constant)
= 9.52:1

I could easily find a gear 1" gear and mate it to a 9.52" pulley, but that would be a very expensive pulley most likely and take up a whole lot of room.

I'm going to try and use a multi gear reduction system to save space, it will use 4 pulleys and 2 belts. essentially a 3.09:1 X 3.09:1 giving me a 9.54:1
See attached picture below for details.

Toothed pulleys from "quality transmission components"
http://www.qtcgears.com/RFQ/default2.asp

QPELS75011F16
11 Teeth
Double flanged
.5000 Bore
1.313 Pitch Diameter
1.283 Outside Diameter
.866 Hub Diameter
1.457 Flange Diameter
1/4-20 Tap
(2) pieces needed @ $5.91 each

QPELS75034F16
34 Teeth
Double Flanged
.5000 Bore
4.058 Pitch diameter
4.028 Outside diameter
4.409 Flange diameter
(2) needed @ $15.55 each

McMaster:

RACK
20 degree pressure angle
6' Length
12 PITCH
1" HEIGHT
1" WIDTH
5174T23 PART NUMBER
(1) piece needed @ $82.74

SPUR GEAR (1 needed)
20 degree pressure angle
12 Teeth
1" face width
0.75" Hub diameter
1.17" Outside diameter
1.62" Overall length
1/2" BORE
5172T31 Part number
(1) piece needed @ $17.21 Price

Tonight ill model it in solid works. The simple 2D drawing does not show the spring retention system that holds the spur gear against the rack, bearings, a number of stand off that would have gotten in the way of the overall picture, and the belt tensioners. Hopefully tonight ill have a modeled concept that i can work with.

These are NEMA 34 gearboxes.....for the right price just in case people are looking.
About the same cost after you consider all those pulleys and bearings.

http://planetarygearbox.net/products...ID=381&cID=151

http://www.sdp-si.com/gearhead/toc.htm

Hey thanks! The reason i started looking into belt drive gear reductions was because of those gear drives. The backlash expressed in arc minutes is horrible unless you really want to spend a good chunk of change. The SDP-SI nema 34 frames have 6 minutes of backlash, if you use there double it goes up to 10minutes, triple goes to 14 minutes.

These units below are about $240 dollars and possess 5-6 arc minutes of backlash. For some wood workers that might be totally reasonable and fine. I mill a lot of aluminum. The price goes up quickly to remove the backlash completely with a gear head. If you use mach 3 or another controlling program you might be able to lesson the backlash with software, but i would rather spend a fraction of the $, save my time, and do it correctly for the beginning =] I'm deff not trying to spend $240 on each axis either with that much slop.

I suspect ill have $140 bucks into each one. Simplicity, reliability, and serviceability are key factors. (1/2" plate isn't cheap, you could probably get away with 1/4" but eh..)

Keling gearheads: Nema 23 and 34
http://www.kelinginc.net/CNCGEARBOX.html

Noted and reasonable.

Hi diyengineer and others

I have been recently drawing up some belt reduction drives and the same thoughts came to me about putting two set of gears in the one reduction drive.

My engineering skills are limited as some recent experiences trying to drill pinions concentric for a stepper shaft have reminded me.
None the less the idea of using the commonly available 3x reduction pulleys twice, that is a 15T x 45T twice, really appeals.

1. Can any of you belt drive gurus see any issues with the extra width the stacked pulleys might cause?
2. Adjustment of the two belts means that the simple slotted sliding adjustment of the stepper that is typically used won't work, any thoughts on this?
3. Am I being greedy even thinking about a 9x reduction when 4x would do for all practical purposes?

Interested about what others might have to say.

Regards
Ross

I recon as long as you use the same belt length & same ratio of each sets, the adjusting slot will work just fine. Lets take diyengineer's design as the case, assume using 2 3:1 sets, if you fix the final pulley position coaxial to the stepper motor shaft, & the other 2 intermediate pulleys on a same shaft (lets call it intermediate shaft), you only need to adjust the intermediate shaft, the 2 pulleys belts should have close enough properties to get almost equal tension on them.

Hi Ken

Two belts of equal length is a given... if you look at the diyengineers drawing, the right hand shaft is captive (or a through shaft) between the two plates.
The first small pulley adjusts as the motor is moved in something like the existing elongated adjustment slots on most belt drives.
That leaves the final large pulley on the side opposite the stepper to be adjusted to tension the 2nd belt.
I'm wondering what this mechanism might look like OR how the captive shaft on the RHS could be moved to tension both belts.

Regards
Ross

Which type of timing pulley is better than ?
 

:)Dear Experters,
I am buidling CNC plasma machine with size about 4M X 6M. It's a big size.
I think I will use belt reduction at 4:1 ratio. I am not sure which type of timing pulley is better than. between type XL or AT5 ? I will use 18T and 72T.
What is the advantage and the disadvantage of each type ? Pls help advise.
Here's my information
1) Stepping motor 860 oz-in 2 eaches for y-axis.
2) Stepping motor 860 oz-in 1 eaches for X-axis.
3) Micro-stepping driver with 1600 pulse per rev.
4) Rack with module1,PA= 20 degree
5) pinion gear = 24 T,module =1

Best Regards,
Mongkol

diyengineer,
I tried to build high resolution drives when I built mine,
the problem is that the returns are so minimal after 4 to 1 that it wasn't worth the extra cost, as your reduction gets larger the returns are diminishing the gains are minimal at best for the extra investment, and I really doubt that you would or could get noticeably better cuts, I too cut aluminum, stainless steel,
steel, plastic, etc.. with my machine, I Just did a job the other day for a customer that was done in a hass milling station, the one I did for him was way smoother than the hass, I am sure that it took me a little longer than it did in the hass but none the less the quality was better, as far as the tooth contact goes you can try to change it, add guide wheels etc. I thought of all this stuff before, but in the end practicality and price win the day, I have one less tooth in contact then what is recommended, I know for some this is a big deal....and if I was running a big iron cnc it probably would be a big deal, but for the MM it is not an issue, when you drive the gantry too hard, or hit something hard with the tool head the pinion gears come out of the rack, the belts will not stretch, they will not break, they will not jump cogs, will not break teeth off etc etc etc ... I know that I am kinda a black sheep here and that is fine, I in no way trying to discourage you in building what you want in fact i encourage it, before I would go for higher resolution I would change some other parts of the machine like the spider, and the Y car, and use a spindle, ( a good one like a colombo, or HSD) to eliminate run out at the tool you will gain more there than in the reductions, good luck with your drives I will be waiting to see how they perform...
//chopper

Ton,
I am not familiar with the T5 pulleys I run the XL with the Kevlar reinforced belts, and my bro is going to use the same setup on his plasma cutter, I would use the xl's, you do not need to worry about much with the plasma cutter since the plasma torch will not be pushing a bit through material so all you have to worry about is moving the gantry, y car, and z but there shouldn't be force applied against the drive systems, unless you crash the torch, but the xl's are plenty strong enough to take it if and when that happens...Ton, I see you are going to use 860 ozin motors, if you use 4 to 1 drives that would be 3,440 oz in of force from each motor, I am running 450 oz in motors, which puts me at 1800 oz in of force, but it gives me less rotating mass in the motor (inertia) so it is easier for the motors to change direction faster, which is not a big deal with the cutting I do but it may be helpful when plasma cutting to maintain your cutting speed for smoother cuts so the motors do not have to spend as much time ramping up and down in the corners, it is just a thought, and maybe there is little or no difference in the inertia of your motors vs. smaller ones but it wouldn't hurt to check it out before you use them, since you don't need all that torque to move a plasma cutter maybe you could save a few bucks, and use smaller ones, smaller power supply etc...
good luck..
//chopper

Chopper,
Good note on the "cornering" speed. Noted well in the gecko design criteria for those really interested.
Good luck.
Sean