Choosing belts and pulleys

Hi everyone,
Since i read this wiki page i'm looking for GT2 belts and pulleys, but this doesn't seems so simple (i'm in EU and i can't find good suppliers).
So i was thinking if an HTD belt 3mm pitch could be the same...even if it as a greater pitch than T2.5
Thanks for the help

Edited 1 time(s). Last edit at 04/10/2012 05:29PM by Pacca.

I'm still using and supply T5 belts as here in NZ I simply cannot find any supplier with GT2 and I can't find any supplier who will ship here either and even if I did find one the cost would probably be prohibitive

---

__________________________________________________________________________
Experimenting in 3D in New Zealand

That wiki article is misleading because it implies T5 and T2.5 cause backlash. That isn't true, either with printed pulleys, or the correct metal pulleys, only with cheap metal pulleys on eBay from Hong Kong that don't have the correct tooth profile.

Apparently you can get timing pulleys with some tooth clearance for high speed transmission but the default type available in the UK does not have clearance and doesn't give backlash.

---

[www.hydraraptor.blogspot.com]

Thanks nophead!
So T2.5 should be better than HTD 3M, because of the smaller pitch...

I have used both T2.5 and T5, sometimes on the same machine and haven't really noticed any difference. T2.5 bends round a small radius better but conversely T5 is stiffer so is less springy.

---

[www.hydraraptor.blogspot.com]

What a mess! Because I was thinking about changing belts and pulleys, having also read the post on prusa iteration 2...

I found these people

[www.beltingonline.com]

Look good for UK users

TC

Yes thank, i already know this site...it doesn't sells GT2, my question was born from here

Pacca Wrote:
-------------------------------------------------------
> Yes thank, i already know this site...it doesn't
> sells GT2, my question was born from here

I'm inclined to go with nophead but everyone to their own.

TC

Any opinion and advice is appreciated
Just the last question: assuming that I want to go to T2.5 belt and pulley, if I use a 16 tooth pulley I don't have to change the firmware configuration, this means that steps/mm remain the same...so where is the improvement in quality?

Exactly. It is only differences in the flexing of the belt, which is why I have been unable to notice any. Somebody needs to do a proper test with measurements to back up these claims.

---

[www.hydraraptor.blogspot.com]

now everything is clear, thanks nophead!

Pacca Wrote:
-------------------------------------------------------
> Just the last question: assuming that I want to go
> to T2.5 belt and pulley, if I use a 16 tooth
> pulley I don't have to change the firmware
> configuration, this means that steps/mm remain the
> same...so where is the improvement in quality?

I have prints which show a very clear vertical "ribbing" effect from using a T5 belt (it has 5mm spacing and only shows up along the belt's dimension). That went away when I changed to a T2.5 belt. I have no idea why others aren't seeing any difference.

Another thing is that both the 608-bearing belt idler pulleys and 8-tooth motor pulleys break the minimum limits for T5 belts (at least for Brecoflex belts, other brands may differ).

What sort of pulley were you using with the T5 belt?

Yes 8 teeth is too small for the T5 spec, which is why you can't buy proper metal ones that size. Ten teeth is the smallest, which gives a slightly lower resolution.

I think 608 is too small for idlers for T2.5 as well. Never the less I haven't had one snap (even after years of continuous use) since I switched to the Mendel X axis version which only bends the belt in one direction.

I have never seen the ribbing effect although other people have mentioned it. I have started using a half twist so the back of the belt runs round the idler, which should be smoother.

---

[www.hydraraptor.blogspot.com]

I bought these belts/pulleys:

[www.emakershop.com]

I havn't tried buiding anything with them yet as I am still finishing off the machine but they appear to run smoothly manually. They arrived super quick and so I would recommend.

I had a lot of belt problems in my Orca v0.2. Both the X and Y belt paths were quite convoluted and included things like back bending over 13mm diameter bearings. The convoluted belts paths required a lot of belt tension to get rid of backlash and the T5 running over these small idlers produced the ribbing effect. And a lot of broken belts.

I eventually rebuilt the thing as a Prusa and with the simpler belt paths and the bigger idlers the effect is not so visible. I ran a T5 belt on one axis for a while (with a 8-tooth motor pulley) and I believe I could still see a minor "wobble" in the carriage movement. This is what was causing the ribbing, the carriage slows down and speeds up (almost imperceptibly) as the belt teeth pass over the idler and the slower spots get a bit more plastic than the fast spots.

ttsalo Wrote:
-------------------------------------------------------
> I eventually rebuilt the thing as a Prusa and with
> the simpler belt paths and the bigger idlers the
> effect is not so visible. I ran a T5 belt on one
> axis for a while (with a 8-tooth motor pulley) and
> I believe I could still see a minor "wobble" in
> the carriage movement. This is what was causing
> the ribbing, the carriage slows down and speeds up
> (almost imperceptibly) as the belt teeth pass over
> the idler and the slower spots get a bit more
> plastic than the fast spots.


Do as nophead suggests and run the smooth side of the belt over the idler.

---

Help improve the RepRap wiki! Just click "Edit" in the top-right corner of the page and start typing. Anyone can edit the wiki!

The odd thing is that I don't do that on my Mendel and the belt runs over 13mm bearings and I still don't get any ribbing in my prints. Also I worked out the error caused by the teeth on a smooth pulley and it was minute. Maybe some belts have softer teeth and behave differently.

---

[www.hydraraptor.blogspot.com]

If people are worries about running the teeth of the belt over the idler, why not use a pulley instead of a smooth idler ? That way the idler's profile would be a perfect match.

Because you need a bearing for the pulley. I don't think you can buy timing pulleys with bearing, so you have to print a pulley that fits round a bearing. If you use a metal motor pulley for accuracy it doesn't make sense to have a printed idler and, even with a printed pulley, a printed idler is a second source of error.

Twisting the belt solves all these problems.

Edited 1 time(s). Last edit at 04/14/2012 02:14PM by nophead.

---

[www.hydraraptor.blogspot.com]

To get a nice toothed idler, you could cut a bearing sized hole into something like this : [blog.makezine.com]

---

Most of my technical comments should be correct, but is THIS one ?
Anyway, as a rule of thumb, always double check what people write.

@Deuxvis, that's not going to work as the gaps between the teeth are opening up when the belt it bent backwards, so the belt won't mesh with the idler that uses the same belt. To mesh with the belt running over it, the gap between the teeth of the idler needs to stay the same to accommodate the teeth of the belt. The teeth on the idler need to become narrower.

I've used 10-tooth T5 pulleys for a long time, now using my modelled scad file here: [www.thingiverse.com]
Compared to the original Mendel 8-tooth pulley, I find that these cause less camming of the belt as it moves around the pulley, and less lash as the tooth fit is accurately modelled and there are more teeth in contact at any one time. Sure, positional accuracy is slightly less using 10 teeth rather than 8, but *repeatable* accuracy and consistency is improved. I can see the point of changing to metal pulleys, especially if you are having trouble printing an accurate, round pulley with a central motor shaft hole. I either print a toothed idler that fits tightly on two 608 bearings at the other end (which could be a source of error, if the idler width was inconsistent) or just run the toothed-side of the belt on a double 608 bearing idler, which I just can't see introducing error. I'm not a huge fan of bending the belt backward; the diameter of 608 bearings is 22mm, which is smaller than the minimum diameter of 30mm for an idler, and you're supposed to use a bigger 15-tooth pulley if bending the belt backwards. See [www.beltingonline.com]
The idler size is supposed to be 30mm for T2.5, too, see [www.beltingonline.com]

I've found lots of factors that effect quality: having parallel z-rod/z-leadscrew (REALLY important), frame stiffness, bed levelling, getting slicing settings right, good fan cooling, and most recently, really good temperature control - I've just started using PID temp control in Marlin, which, after tuning with M303, keeps the temperature +/- 0.1 degree at the target temperature. Layer consistency has dramatically improved, as it used to swing +/- 10C, mainly because one thermistor wire was a bit loose! I admit I have never used anything else other than T5 belt. I would change the M8 rod on the z-leadscrew for proper leadscrew before changing the belt, but that's not to say I wouldn't like to try T2.5, GT2 2mm or 3mm, or HTD 3mm belt, though I think printing pulleys for this size belt is difficult - the sharpness of the tooth-shape is compromised unless you use a very small nozzle. I just don't think the difference would be significant for the reasons nophead outlines.

Here's a recent print at 0.15mm layer height (83mm overall height) to show the quality I get, which I'm pretty proud of! [dl.dropbox.com]

NewPerfection Wrote:
-------------------------------------------------------
> Do as nophead suggests and run the smooth side of
> the belt over the idler.

Firstly, there is no need since T2.5 belts and printed pulleys are performing pretty much perfectly, and secondly, back bending the belts over a 608 bearing idler would break the manufacturer's limits even for T2.5 belts, not to mention the stiffer T5 belts. I have a too big pile of broken belts to do that even though some are getting away with it.

And for the record, this shows the artifacts I was getting from using T5 belts:



This was on Orca v0.2, in which the belt paths are pretty awful. On a printer with sane belt path design, such as Prusa, this isn't anywhere near as this big a problem.

I don't think back bending is a problem if you don't bend it forwards as well. I think the problem occurs when it bends both ways.

It doesn't say so on the datasheet, but in normal circumstances you would never get a belt that only bends backwards as the teeth would all face outwards. I have seen drive systems where the belt twists, so it is a legitimate use.

A belt used to last ~200 hours on my Mendel when it bent both ways. Since the new X axis and on three later machines I have built I haven't had any more belts snap.

Also bear in mind that the belt spec covers high tensions and very high speeds. I suspect we get away with violating the minimum diameter because we don't push the belt on speed and tension. A more detailed spec would have graphs that show the effect of tension and speed.

I have never seen a cogging on any of my machines.

---

[www.hydraraptor.blogspot.com]

I have the exact same "cogging" on my machine. I thought it was normal.... The funny thing is that when I changed the x-axis belt to a GT2 belt and pulley, the "cogging" got proportionally smaller. If there is a way to eliminate it, I would be interested...

As discussed above, using a single idler (needs an alternative Y motor bracket on a Prusa) and a half twist in the belt will get rid of any cogging due to pulley teeth running on a smooth idler. If there is still cogging it means the pulley tooth profile does not match the belt.

---

[www.hydraraptor.blogspot.com]

I had a very similar problem. The problem: belt was too loose, and the belt teeth were climbing a bit on the gear then sliding back into the groove as it turned. Tightened the belt and the problem disappeared.

CHINA POST Air Mail to Europe from RobotDigg, China Source of GT2 Pulley and Belt.

fwiw, gt2 belts are also called mxl from a different (japanese) manufacturer.

brought my gt2 pulley in a belt shop. salesperson tested it on one of theirs and told me its called mxl###. last 3 nos are the no of teeth on the belt.

hope this helps.

f3rDz Wrote:
-------------------------------------------------------
> fwiw, gt2 belts are also called mxl from a
> different (japanese) manufacturer.
>
> brought my gt2 pulley in a belt shop. salesperson
> tested it on one of theirs and told me its called
> mxl###. last 3 nos are the no of teeth on the
> belt.
>
> hope this helps.

GT2 = metric (2mm pitch)
MXL = imperial (2.03mm pitch)

It just looks "almost" same but this are two different pulleys and belts, using GT2 belt with MXL pulley and vice versa, create huge backlash and slip.

Best Regards
Martin

www.gadgets3d.com - parts for your 3D printer