Timing belts and pulleys?

[Jack Aubrey]

Please excuse me if I am being dumb! I was thinking of using timing belts and pulleys vs. leadscrews for my own design CNC router. Can anybody please tell me where I can find out about the pros and cons of belt drives?

Thanks

Jack

[ToyMaker]

You could use the search function for this site. Belt drive has been discussed quite a lot in the home-brew fora and others.

robotic regards,

Thox
= = = = =
"A man can father a baby but a baby fathered does not make a man."
- - Burda Vandeborne

[Jack Aubrey]

Many thanks ToyMaker

[Lotus14]

You might also find "Stock Drive Products" website, and get a copy of their Data Book 757.
It has some very good info about synchronous belts and drives. Also there have been commercial CNC machines that used "timing" belts. I have an Anilam Crusader CNC Bridgeport mill which uses belt drives.

[harryn]

I found that reading the info on the brecoflex web site (a belt supplier) was very enlightening. There are some very detailed articles about backlash, tooth loading, etc and how to manage it.

http://www.brecoflex.com/

[lgalla]

Jack,I am no tec wizard on the subject,but do some searches on the zone.I thought belts would be good but they need huge gearing down which is costly.One member spent $1,000 bucks on belts and pulleys before realizing the gearing down required and scraped the idea and switched to rack and pinion.
Search for Moondog on the zone.He used X belt drive and is happy.I think he had to gear down 27to1
Larry

[nwrepair1]

I have a 4 x 8 Precix router that runs on AT 5 belts. I recently took it apart to upgrade the X axis belt drive to a Brecoflex belt. It looks to me like a 2 to 1 reduction on a Nemi 34 size Magmotor servo.

[Lemo]

I build a 4x6 table this year with a heavy gantry (150lbs) and very solid linear rails. I use .5 inch timing belts with 1200oz steppers without any gearing down and the thing just flies. I am talking about 10 ips in 1/2 chinese crap plywood with a 1/2 inch deep cut using a 1/4 inch carbide bit and 1000/th of an inch repeat measured with a dial indicator utilizing 20 ft of tool path and 100 reps. It works like a charm. I build it after only listening to people on the forum who actually build something and actually use their machines. I see no reason why a 5x8-10 router should not be fast and accurate as well.

Go, build something!
Lemo

[harryn]

Originally Posted by nwrepair1 I have a 4 x 8 Precix router that runs on AT 5 belts. I recently took it apart to upgrade the X axis belt drive to a Brecoflex belt. It looks to me like a 2 to 1 reduction on a Nemi 34 size Magmotor servo.

Hi, thanks for that info. Just a few more questions if you don't mind to much.

a) Do you happen to know if the AT-5 pulleys are standard or zero clearance type (for backlash control )

b) Do you happen to know approx the diameter of the pulley setup ? The reason I ask, is that 1 revolution of the motor will give a different force on the belt and motor depending on how many teeth (diameter) of the timing pulley. I know I said that about as clear as mud, but basically, if 1 motor rotation = 10 teeth moved, it is different than if 1 motor rotation = 100 teeth moved.

c) The AT5 comes in different widths. Do you happen to know if yours is 25mm (1 inch) 50mm (2 inch ) ?

Thanks, I appreciate the inputs.

[mackeym]

I know GT2 pulleys are very good (precision backlash free). The "Stock Drive Products" that lotus14 mentioned has information on them. The HTD (high torque drive i think it stands for) is an earlier version of the GT2 series. I think the GT2 series is the current benchmark for high quality precision belts and pulleys.

[Lemo]

Some ramblings....

The wider the belt (up to 6 inches btw.....) the more friction is in the belt. Important is not what you 'like' important is which load the belt can take. That determines the type of belt (data chart). However, if you choose a 1/2 inch belt, you cannot go wrong. You can suspend your entire table from such a belt and wheel it around.

The at5 timing belt has a fixed amount of teeth per inch. And no, the amount of teeth is irrelevant regarding 'torque'. The diameter of the pulley IS relevant. The diameter determines also how far the belt moves per pulley rotation.

If you use stepper motors, simply use a 2 inch diameter pulley directly connected and you will be good with pretty much anything you can throw at it.
I use that together with a 1200oz stepper to move a large gantry with a 2hp spindle. The stepper has two shafts and each side is connected to a shaft ahich drives one pulley. I use a belt on each side of the long axis. There cannot be any inaccuracy because of that. No slack, no following, pure precise movements. The theoretical resolution in this case is the result of pulley diameter and the amount of steps per revolution. My concept is based on microstepping drives. Or it will be terribly coarse. But with microstepping the motor resolution is better than the possible resolution considering al the other parts involved (bearings etc.). Also both sides have to be tensioned the same. Or the gantry would lag on one side. Which is easily corrected in case....

If you use Servo's the low rpm high torque characteristic of steppers is not available any more and you have to use a gearbox like setup. Servos have to be driven as fast as possible to yield the benefits. You have to calculate from the top down (3000 rpm or even faster) and divide it down to the table speed you anticipate/wish for. It can yield anything from a factor of 1:3 to 1:60 depending of the servo drive. Servo drives make a belt driven table more complicated. IMHO for no reason. The Stepper setup with good quality drives runs just fine. Servo's add cabling issues, more complexity you have to be able to trouble shoot, and additional requirements regarding gear ratios.

I use a servo setup.... which I (after my buddies version of my design using steppers runs perfect...) regret in the meantime a bit. My system will be able to jog much much faster than his stepper version, but will only be able to cut as fast as the stepper version because the stepper version already maxes out the theoretical feed speed ratio for all the materials we throw at it. Thus... the servo would not bring any additional benefit while cutting. The maximum we can safely jog with the stepper based table is 20-30 ips right now. WHich.... honestly.... is scary fast and makes us have a hand hover over the e-stop. The servo can top that easily, in case I probably start praying and wearing splatter proof underwear.

However, from a geek perspective, and a engineering marvel / conversation piece the servo setup (three power supplies for drives, motors, logic, e-stops, limit switch network, complex drives with USB ports each for tuning, firmware updates, encoders and so on...) is way more shiny, slick, mystic, and thus in a bewitching way more desirable than a plain and simply working stepper setup consisting out of a single power supply, three geckos, and a breakout box.

Sorry.... where was I... ahhh... belts...

Regardless what you build out of belts, keep a rule in mind regarding the amount of teeth which have to be meshed all the time. What I mean with that is that you have to have a certain amount of teeth in the grooves of the pulleys to provide the necessary grip to be able to take load. With my 2 Inch pulleys I pretty much have to wrap the belt about 180 degrees around the pulley. To achieve my 1:3 ratio I needed to introduce an idler pulley as the large pulley was creating a steeper angle on the smaller pulley. I had to take a longer belt, and had to use an idler to push everything into place. More complexity, another bearing, and a tick less motor power for the gantry due to friction. Not a lot... but again, something the stepper setup did not need.

Another thing to consider is 'prospective mayhem'. My stepper design can crash with all the force of 50-60 lbs into the hardware limits. Considering the rigidity of everything... not traumatic. I had the occasional bend 1/4 inch mount for a limit switch while getting used to things. Nothing 'permanent'. My Servo table (I am so scared already...) has 2 HP plus a 1:3 gear down (yielding a whopping 6 HP in theory) which can dig into the environment. Which in case.... will most certainly suffer greatly.

Again, servo's are cool and fun to play with, make small things real expensive and complex, and IMHO should only be applied when things are big and bad enough to scare a stepper's torqued ego. And after building two large tables.... I think I only need a servo if I want to move components fast which weigh more than 500 lbs. Till then, steppers and belts using micro stepping drives are sufficient for 99% of the applications.

Good luck, and read up on things and most importantly, build your own opinion
Cheers
Lemo

PS: My Servo's bigger than yours! HA!

PPS: If interested in the belt components I used please PM me and I can forward exactly which material I used and what I paid for it. I also have some spares left.

[nwrepair1]

Sorry! I didn't keep track of this post. Pulley for belt is 25 tooth (minimum recommended size). Back idlers are 60mm. Belt is 25mm wide due to space constraint. A 32mm would be better, Also upgrade the AT5 belt to a ATL5 belt which is same tooth design but thicker steel cord. Use adjusters (tensioner clamps) at each end of each belt. That way you can square the gantry and correct any racking of gantry. The pulley that is on the belt shaft will be 50 tooth and the motor pulley will be 25 tooth. That is, the shaft that has the 25 tooth pulley for the ATL5 belt will have a 50 tooth pulley on the other end. Last but not least important, the center to center distance between the back pulleys and the tooth pulley has to be at least 100mm apart. Hope you get this.