GT2 Belt 6mm vs 9mm

Hi there, I'm building a new Ultimaker-style 3d printer with around 300x300mm print bed. To drive the XY carriage I want to use GT2, but I'm not sure about the width of the belt. 6mm should already provide enough force to drive the print head, but 9mm might provide a smoother operation as there is more material "to grip on" the pulleys. What do you think?

I've been wondering the same thing, as the 6mm belts do look pretty skinny. However, once you start tensioning them it's difficult to imagine needing anything thicker for a printer, at least for 180 degrees of belt-pulley contact. One of the pulleys for my Z drive currently only has a little over 90 degrees of contact, so it will be interesting to see if that jumps. If it does I'll add an idler to reroute the belt and increase the amount of contact.

No more experiences? No one?

I use 9mm wide GT2 belt on the X axis in my printer. I haven't used 6mm wide belt so I have nothing to compare its performance to, but in theory, there should be less stretch in a 9mm belt than a 6mm belt. Does a 6mm belt stretch under 3D printing load? Maybe it depends on your acceleration and jerk settings, and the load you're moving. If you'll trying to throw a 400 mm square bed around in the Y axis, the 9mm belt is a better idea than a 6mm belt.

I don't think there's much difference in price between a 6mm and 9mm belt and the pulleys for it, so for a couple extra $ you can err on the side of caution.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Hi guys,

Quote
the_digital_dentist
I use 9mm wide GT2 belt on the X axis in my printer. I haven't used 6mm wide belt so I have nothing to compare its performance to, but in theory, there should be less stretch in a 9mm belt than a 6mm belt. Does a 6mm belt stretch under 3D printing load? Maybe it depends on your acceleration and jerk settings, and the load you're moving. If you'll trying to throw a 400 mm square bed around in the Y axis, the 9mm belt is a better idea than a 6mm belt.

I don't think there's much difference in price between a 6mm and 9mm belt and the pulleys for it, so for a couple extra $ you can err on the side of caution.

I still wonder where do that "stretch factor" came from???
I believe that legends comes from ballscrews comparison, because it can stand a lot of load,
and you should get a huge belt to reach that performance. But even ball screws have pro and cons.
For example, long strokes and speed gives brings a lot of problems to ball screws.

I've read about max stretch of belts, but it's only about max stretch before rupture.
And it's always about a small percentage for 1 meter of belt.
Belts are not meant to stretch during operation. Neither any slip factor.
They're reinforced anyways at the pitch, the cheapest with glass fibre (which is not elastic at all), and only pitch matters.
If you don't overtight and/or overload your belts, you will have no stretch at all, whatever 6 or 9mm system are in use. Definitely.

Belt selection depends of several parameters : load, pulleys diameter, lenght of the belt, teeth in mesh, rotating speed...
I recommend the reading of guides, like the abondant ones from the SDP-SI company, like this one.
Oversize a belt helps in nothing. You just want the right one, calculated within a 20% extra safety factor.
I believe the smaller possible is the better, because of the weight of the overall system and efficiency.
Bigger belts have more friction and inertia. You want the best performance, so...
On my big CNC, I used HTD 5M 15mm belts because it was the right awnser for load/torque/speed requirements.

The bed's weight is more important than its dimensions, even if there's a relationship indeed.
But it depends also of the driving mode. What if the bed is moving only in the Z direction ?
The speed factor will be negligible, and you will just consider the load.

Anyways, let's calculate what torque can holds a GT2 6mm standard 2mm pitch belt, for a 200x200mm bed with 20T pulleys.
It's maybe the most common reprap configuration. So it will be about 0.2 to 0.19 Nm depending of the shaft RPM (which is never high on our machines).
Inertia of a 200g bed should be about 1g.m² so its load torque should be about less than 0.02Nm if friction is moderate and the bed lays flat.
I believe a standard GT2 belt is pretty enough to drive systems ten times bigger than a regular 200x200mm bed.
(It's late here, I hope I haven't done any mistake in my calculations, my appologises if I did wrong).

If a machinist pass by, maybe he can provide more trustable calculations.

++JM

Edited 1 time(s). Last edit at 12/07/2015 05:43PM by J-Max.

I think stretch is most likely to be an issue only when acceleration is high, so it depends on the moving mass, the acceleration and jerk settings, and the torque that the motor can deliver. If you can pluck the belt and make it vibrate like a guitar string, it can stretch. The only question is whether the dynamic loads in a printer are high enough to make it stretch enough to affect print precision/surface quality. The motor is a spring and the belt is a spring. I suspect that the spring constant of the motor is much lower than the belt, so the effect of bounce in the motor step position probably masks any artifacts caused by belt stretch. Obviously if you use too small a belt you're in danger of breaking it, so there should be some safe maximum static/dynamic load specs on the belts. The problems are calculating/measuring the dynamic loads in a printer, and finding/trusting the belt load spec. It's hard to identify the manufacturer, let alone find specs, for many 3D printer parts including belts. As with so many other things, when you're not sure, it's better to overbuild. Using a wider belt is like using larger cross section frame members. Even if you don't know the numbers you know the relative effect. When overbuilding adds only a couple $ to the cost, why not?

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Hi guys,

Why not ? Hmm, many reasons I'm affraid.
Everything is a spring in quantum mechanics, true.
You say belts stretch, what are your affirmations based on ?
Vibrations and stretching is not exactly the same.
If a belt can vibrate, it just means it's far from it's breaking point, it's a good thing.

Belts are at last reinforced by glass fibers at their pitch (we know mostly pitch matters).
Glass fibers offers excellent tensile and compressive strenght (4900/1600Mpa).

On a recent project, we had to calculate a belt system.
The calculations showed a GT2 2/6mm belts 16T, 78oz.in Nema17 motors was enough.
We did calculations several times and were not that confident, but we decided to trust our calculations and manufacturer's formulas.
Actualy, we're moving a 2000x1600mm plotter machine with a coreXY system, you will estimate easily the lenght of that belts.
We calculated the load won't exceed 2/6mm GT2 belts, even with 16Tooth pulleys.
We calculated the safety factor is even more than 34% with our 3,3kg 1600mm gantry including tools.
Actualy (beta testing @2800mm.s^²) we reach a precision.repeatability within +-0.025mm on max diagonal moves,
which allows us to think belts don't stretch and our calculations were trustable.
I believe this mecanical tolerence comes for a part from the belts, but also from the light gantry not perfectly stiff and our V-groove guides.
But we already reached our goals as it's just a drawing plotter meant to trace lines over 0.7mm.

Sure a belt can break if you exceed max load, anything will break, not just belts.
I never seen any broken belt except on cars. I've seen bent SBR guides on DIY CNCs.
Just because it was not the right guides for the machine expectations.
If you calculate your belt system with a 20% safety margin, and you replace it according to manufacturer's lifetime limits, it will never break.

I know you believe in heavy constructions (Son Of Megamax is a perfect example).
When you oversize, you feel safe, and you believe it will just be better. In a way you're right.
It's a safe way expecialy if you're walking blind and you don't apply calculations to select a component.
But when you oversize a thing at a end, you should oversize another thing at the oposite end. It's a viral method.
The delta between the 6 and 9 belts price tag is not a lot, but it's more important when it's about pulleys.
When you apply the same principles all over the machine, the result is obviously a heavy and expensive machine.
If you built it well, the machine will print good, but will probably be slow because size and weight also brings disavantages.
As your components are more expensive, mainteance cost will be more expensive too.
Obviously that overvalued machine won't perform as well as a machine designed with the right components to match expectations : neither too much nor too little.
Otherhand, not everyone is rich (or passionate) enough to build an overvalued machine.

You often says "people don't want to put more than $300 in a printer", you're definitely right.
First, over the world $300 can means realy a lot. We're lucky enough to live in one of the richest countries,
but even in our countries, there's millions of people who can't buy any printer even if they would.
Seccond in a living, 3d printing is maybe not a major centre of interest.
People are different, some will give all to their childs, others are car enthusiasts, others loves traveling or home improvement...
$300 can represent a reasonable budget to get a 3D printer for ordinary people. We should understand that.
A $300 printer can print pretty well, and probably match ordinary people expectations.
If you overvalue components you will at last double the overall budget.
I don't know about US prices, but here I probably need 3000-3500€ to build a Son Of Megamax.
I've read carfuly your Instructable which I consider as a preaching for an extreme point of view.
I understand and I respect it, but it don't seems to me necessary to put that money into a printer to print detailed and functional parts.

To use 9mm belt and pulleys when any 6mm stuff is definitely enough is just pointless.

On an Ultimaker architecture, you definitely don't need 9mm belts on XY asis.
You won't get more grip or precision (how could you ?), but you will get more friction and more weight and it will cost you more.

++JM

Of course, belts aren't meant to stretch. And frame members and guide rails aren't meant to flex, either, but they do. And we're not talking about quantum mechanics. Things expand with temperature and they flex with mechanical loading, all predictable and measureable. There's no escaping it. I don't know how to calculate the dynamic mechanical loads on a printer's frame or the flex those loads will cause. I do know that using larger cross section frame members will reult in less flex. When I built my printer, I was able to purchase 1.5" square 8020 from a scrap yard for less than the cost of new 20 mm square extrusions. I'd have to be nuts not to use it.

Yes, every kit I've ever seen uses 6mm belts. But most kits use flexible 8mm guide rails, zip ties to hold bearings, etc. When selecting what to copy from those kits, how can you tell what is and isn't "adequate" among the varius "features" of those machines, and whose definition of "adequate" are the component selections based on? My assumption is that the parts in those machies are selected primarily to minimize cost and/or effort to locate something better. I think it's a reasonable assumption.

You were able to do your calculations because you had specs to work with. That means you were buying industrial grade, fully spec'ed goods. Show me any spec besides the width and pitch of GT2 belt from any of the Chinese manufacturers that are supplying belts to 3D printing hobbyists. If you don't have the specs (or can't trust the specs you have) you can't do the calculations or the results are not meaningful.

Duplicating my machine with new components would indeed be very expensive. That's why I recommend that people find their local makerspaces and search their local scrap yards and ebay for used or NOS components. It allows you to keep the cost down and the performance up. If you have access to software that will simulate the machine and calculate all the static and dynamic loads, and you know how to use that software, and you have a lab full of instruments that can measure everything to make sure it agrees with the simulation, by all means, do it. Most people, including me, don't have all that. I think my printer demonstrates what can be done by applying basic mechanical principles and by focusing on print quality instead of printer cost.

Edited 1 time(s). Last edit at 12/08/2015 01:16PM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Hi guys,

Yes, I guess everybody here knows you forcus first on print quality. It's a good point.
I'm looking in the same direction with another method.

I don't have any simulation software, I just do calculations based on formulas anyone can find on the web.
If you look at mechanical lessons and studdies, you'll find easily plenty of them. Then you just need a pencil.
Moment of inertia of any material can be found too, as belts/pulleys specs, rotor inertia of your motors, and so on.
Of course, you won't find the moment of inertia of a threaded rod, because it's just not meant to make a structure.
Instead, you will be able to find its traction torque, because it is meant to hold something.
You will be able to find any information you want about GT2 belts, from low to high quality.
To take the average value is wise and it's pretty enough to validate any system.

Don't get me wrong, I just use zip ties to hold wires, not bearings.
I agree if Jo Prussa made calculations, he would probably not use 8mm smooth rods, but it's not his worst mechanical decision about his i3.
He said his primary goal was to build a printer with cheap materials anyone can find localy. That was his goal.
To me better choices could have been made : a screw is not much expensive than a zip tie.
Then kit resellers optimized anything for profit, you're absolutely right.
Cheap materials or components can't make a good printer, I agree. It's excessive.
Otherhand you don't want to be conversely excessive too.
You'll get a better printer but at high cost.

Calculations helps you to get the right component at the right place.
I'm not a genius and I did not found it difficult to do or to source informations.
Simulteaneously you get a more performant machine.
Even production machines don't use overvalued components.

++JM

Edited 1 time(s). Last edit at 12/08/2015 10:26AM by J-Max.