Why not to use spring tensioners on timing belts?

I was having a discussion with someone about those spring tensioners for belts. I don't use them. But it seems like on paper that they should reduce vibration from backlash correct?

I am pretty sure it was here. But I seen where someone had ran test on the matter and had a good reason for not using those. Anyone have a clue? I can't find the post and can't remember for the life of me.. Or is this all bunk?

Edited 1 time(s). Last edit at 10/18/2018 12:31AM by dasflux.

My feeling is that springs would be better than loose belts but worse than tight belts. Compared to tight belts, they would increase backlash, not decrease it. Think of the spring as mechanism for the belt to stretch and shrink, depending on acceleration of the carriage. Worst case, the motor is attempting to move the carriage to and fro over a short distance and the spring absorbs all the energy and the carriage doesn't move at all.

The belt is a spring that couples the moving mass of the extruder carriage, X or Y axis, to another moving mass, the rotor in the motor that is in turn coupled to another spring, the magnetic field in the motor. I can't see how throwing another undamped spring into the system is going to improve anything, unless the goal is to make everything bounce around even more than it already does.

It is much better to design-in some means of adjusting the belt tension than to add a spring tensioner after you've built the mechanism. Tensioners don't have to be fancy and screw adjustable, etc. You're only going to have to tension the belts once or twice during the life of the machine. If you can pull the belt tight with pliers and screw a clamp down on it, you're good to go.

Edited 2 time(s). Last edit at 10/18/2018 10:35AM by the_digital_dentist.

---

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

Let's not forget the most important "springs" in the majority of these 3D printers, the frame, motor brackets, bearings tied with tiewraps etc... ! Indeed without these, most would simply seize.
As for belts, most can't be tensioned to specs because of this built in springiness.
So why add one more ? OK they are so cheap and look fancy.
Ask this gentleman, What about shock absorber ?

Edited 2 time(s). Last edit at 10/18/2018 10:44AM by MKSA.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

Shock absorber might work if done correctly. Take an automotive timing belt for example, many use a hydraulic tensioner with an internal spring rated for a specific belt tension.
When vibrations try to push tensioner back it resists. It will move back but very slowly and with increased effort. If it could be made to apply the optimal pressure for any given belt,
There would be no need for adjustment and may also dampen vibrations. I like the idea. Anyone know what 6mm and 10mm gt2 rated for psi? Also, anyone know what belts are good, brand?, stiffness etc? Would that effect vibrations?
Rod

Quote
Tierod
Shock absorber might work if done correctly. Take an automotive timing belt for example, many use a hydraulic tensioner with an internal spring rated for a specific belt tension.
When vibrations try to push tensioner back it resists. It will move back but very slowly and with increased effort. If it could be made to apply the optimal pressure for any given belt,
There would be no need for adjustment and may also dampen vibrations. I like the idea. Anyone know what 6mm and 10mm gt2 rated for psi? Also, anyone know what belts are good, brand?, stiffness etc? Would that effect vibrations?
Rod

Engine timing always run in the same direction and the tensionner is installed on the slack side.
"rated for psi" ???? Is it a tire ???

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

what if the springy'ness of the belt was given a little help with the springs, and those springs counter the ac'cell/de'cel of the steps/belt profile?

You can use springs to tension belts and it will have an effect on the vibration damping and many high-end machines (not printers) do. The spring can be used to easily apply a tension to the belt and maintain tension as belts stretch so less servicing is required. There are three key issues with the springs that clip around the belts though which are:

1. The spring rate is FAR too low, they are not nearly stiff enough so act to make print head positioning inaccurate at high accelerations.

2. There is often no preload. Springs should (almost) always be preloaded which the lack of will result in a very small tensioning force which means your belts are not actually anywhere near tensioned enough. So people think you put the spring on loose belts and it's sorted, which is not the case at all.

3. The loading from the torsional springs is kind of in the wrong direction. The inertia forces on the belt have a mechanical advantage over the spring which reduces effectiveness. Its much easer to pull a belt tight than to pinch it together with a lever arm which is effectively how those springs work.

Quote
Ed3D
You can use springs to tension belts and it will have an effect on the vibration damping and many high-end machines (not printers) do. The spring can be used to easily apply a tension to the belt and maintain tension as belts stretch so less servicing is required. There are three key issues with the springs that clip around the belts though which are:

1. The spring rate is FAR too low, they are not nearly stiff enough so act to make print head positioning inaccurate at high accelerations.

2. There is often no preload. Springs should (almost) always be preloaded which the lack of will result in a very small tensioning force which means your belts are not actually anywhere near tensioned enough. So people think you put the spring on loose belts and it's sorted, which is not the case at all.

3. The loading from the torsional springs is kind of in the wrong direction. The inertia forces on the belt have a mechanical advantage over the spring which reduces effectiveness. Its much easer to pull a belt tight than to pinch it together with a lever arm which is effectively how those springs work.

So, to use or not to use ?

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

use 2

Quote
MechaBits
use 2

Confused too ?

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

Quote
MKSA

Quote
Ed3D
You can use springs to tension belts and it will have an effect on the vibration damping and many high-end machines (not printers) do. The spring can be used to easily apply a tension to the belt and maintain tension as belts stretch so less servicing is required. There are three key issues with the springs that clip around the belts though which are:

1. The spring rate is FAR too low, they are not nearly stiff enough so act to make print head positioning inaccurate at high accelerations.

2. There is often no preload. Springs should (almost) always be preloaded which the lack of will result in a very small tensioning force which means your belts are not actually anywhere near tensioned enough. So people think you put the spring on loose belts and it's sorted, which is not the case at all.

3. The loading from the torsional springs is kind of in the wrong direction. The inertia forces on the belt have a mechanical advantage over the spring which reduces effectiveness. Its much easer to pull a belt tight than to pinch it together with a lever arm which is effectively how those springs work.

So, to use or not to use ?

This thing:



Absolutely not.

Something like this:



Potentially with some careful design.

None of them in fact.
Explained in previous posts.
"careful design" ? Please show.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

Anything which acts as a point mass on a belt is a probably a bad idea. I discovered this accidentally when I tried joining the belt on a Delta printer by clampint the lengths teeth to teeth in an aluminium clamp - the belt vibrated erratically about the node formed by the clamp and this was an additional source of artifacts on the print surface.

Having a spring mounting on the idler to maintain tension may well be a good idea and having having a damper on the tensioner as well probably even better. A damper may be something as simple as sliding parts lubricated with a super thick grease like Rocol Kilopoise 0860.

Mike

Mksa, ok,ok, was late, bit tired,lol. Foot pounds it is. Ok, back to timing belt thing. I was using that example because of the type of damper used. It’s internally spring loaded and oil dampened which also acts as a strong resistance to backward movement only. Moves quickly the other way. Unfortunately weight would be an issue because the best placement would be the carriage. But some form of self adjusting self locking arrangement may help with belt vibrations. Anyone tried mounting adjustable weights on motors, etc to deal with harmonics/vibrations that way? I’ll try and find some time to do that with an accelerometer hooked up and see if it has any effect.
Cheers

My random thoughts, in no particular order:
Springs are great on clothes pins, not so great on bidirectional drive belts that shift a weight. They allow the working portion of the belt to grow and retract...the opposite of precision.

Want to reduce ghosting? Switch to a wider belt. Just as with a string instrument, it lowers the vibrating frequency to where it can be tuned out with firmware settings easily. Reducing the weight being moved is another path, but not always so easy.

Tension a belt once, fine. Tension it twice, order replacement. The tooth pitch is changing slightly along with that stretch, and wear will occur on the toothed pulleys, as well as introducing even more print artifacts as the now 2.1 mm belt teeth are "snapping" into 2mm pulley teeth repeatedly.

Tension a belt just enough to make a very low note, not more. They'll last and last.

Want to shift a heavier weight with belts (bed plates)? Use two. Still ghosting? Use two wider belts, they're available.

I have no idea why 6mm belts are the standard. I use one to hold my bed wires up off of the limit switch, and that is it.

Edited 1 time(s). Last edit at 12/09/2018 05:19AM by Diggrr.

Quote
MKSA
None of them in fact.
Explained in previous posts.
"careful design" ? Please show.

As I said in my previous post which it looks like you didn't actually read yes, springs are a viable option for tensioning a belt. In fact I can name 3 industrial machines right off the top of my head that use a spring to tension a GT2 belt and they are all capable of sub micron positioning at up to 1G acceleration with a head weighing close to 1 kg - so far more rigorous than what a 3D printer will ever experience. I'm not talking about the cheap spring you would find in a bic but something with a much higher spring rate that has been properly specced and thought through.

To be clear the way people use a spring currently is not the same and I agree that it will ruin your tensioning but properly designed and specced spring has several benefits over simply tensioning a screw. Those being:

1. Ease of assembly - assembly lines will struggle to over tension or under tension drives.
2. Reduced servicing - belt stretch is compensated for by the spring, yes it's not ideal but better to have a belt slack off a little bit than a huge amount.
3. Reduced influence of thermal effects - as aluminium members expand / contract you limit the amount of undue stress applied to the belts.
4. Vibration damping - the spring will help isolate some of the vibration from the drive transmitting into the machine which can often cause resonance.
5. Belt life - due to the reduction in sharp stresses belt life can be greatly improved.

No offence but the benefits of a properly designed spring tension unit are fairly obvious but I know that no matter what I say you wont change your mind so that renders any kind of productive discussion here pointless as you "know best". Unfortunately in this case you're wrong and I think I'll chose to follow the lead of the professionals, who have engineered products far out performing the needs of a printer, and conducted rigorous testing on them on this occasion instead of someone on the internet with no credentials and nothing to back up anything he says.

Edited 5 time(s). Last edit at 12/09/2018 05:59AM by Ed3D.

Quote
Diggrr
........................................

I have no idea why 6mm belts are the standard. I use one to hold my bed wires up off of the limit switch, and that is it.

Just a quick reply from memory - I will look it up the numbers again when I have the time:-

The correct loading for 6mm GT2 belt is already at or beyond the radial load limit for most NEMA17 stepper motors. Using 10mm would mean that the belts are under tensioned as well as moving the pully out another 2mm and increasing the leverage and hence the radial load on the motor bearings. The correct answer is of course to use two external bearings with associated coupling for the pulley and 10mm belt. Another problem is that belting has a small but significant mass itself so becomes another part of the mass/compliance/elasticity/reliability compromise.

Mike

Yep, sorted. No ringing allowed here.

Quote
Ed3D

Quote
MKSA
None of them in fact.
Explained in previous posts.
"careful design" ? Please show.

As I said in my previous post which it looks like you didn't actually read yes, springs are a viable option for tensioning a belt. In fact I can name 3 industrial machines right off the top of my head that use a spring to tension a GT2 belt and they are all capable of sub micron positioning at up to 1G acceleration with a head weighing close to 1 kg - so far more rigorous than what a 3D printer will ever experience. I'm not talking about the cheap spring you would find in a bic but something with a much higher spring rate that has been properly specced and thought through.

To be clear the way people use a spring currently is not the same and I agree that it will ruin your tensioning but properly designed and specced spring has several benefits over simply tensioning a screw. Those being:

1. Ease of assembly - assembly lines will struggle to over tension or under tension drives.
2. Reduced servicing - belt stretch is compensated for by the spring, yes it's not ideal but better to have a belt slack off a little bit than a huge amount.
3. Reduced influence of thermal effects - as aluminium members expand / contract you limit the amount of undue stress applied to the belts.
4. Vibration damping - the spring will help isolate some of the vibration from the drive transmitting into the machine which can often cause resonance.
5. Belt life - due to the reduction in sharp stresses belt life can be greatly improved.

No offence but the benefits of a properly designed spring tension unit are fairly obvious but I know that no matter what I say you wont change your mind so that renders any kind of productive discussion here pointless as you "know best". Unfortunately in this case you're wrong and I think I'll chose to follow the lead of the professionals, who have engineered products far out performing the needs of a printer, and conducted rigorous testing on them on this occasion instead of someone on the internet with no credentials and nothing to back up anything he says.

I thought we were speaking about hobby 3D printers ? I was !

Fine, so you consider it would be a benefit for our 3D printers. Show us your design then.
Planning to include it in your 3D printer project ? Would be great, you could show the results with and without.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

Quote
Tierod
Mksa, ok,ok, was late, bit tired,lol. Foot pounds it is. Ok, back to timing belt thing. I was using that example because of the type of damper used. It’s internally spring loaded and oil dampened which also acts as a strong resistance to backward movement only. Moves quickly the other way. Unfortunately weight would be an issue because the best placement would be the carriage. But some form of self adjusting self locking arrangement may help with belt vibrations. Anyone tried mounting adjustable weights on motors, etc to deal with harmonics/vibrations that way? I’ll try and find some time to do that with an accelerometer hooked up and see if it has any effect.
Cheers

Indeed, the key here is this damper that dissipates the vibration energy. A spring doesn't dissipate energy !
And again, easier for an engine as it is installed on the belt slack side.
In fact, as noted by others here too, the vibration is more likely due for the most part to the stepper rotor, not the belt.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

I decided to add a spring tensioner to my printer's belt just to see if it's as horrible as people say. Long story short: it's not.

I doubt anyone can guess which one of those two small enclosures for a 40x40mm board was made with a belt spring and which one was made without.

I'll post more pictures if just this one isn't enough for a blind test (heh).

Hint: they're both silky smooth in real life

This one should make it a bit easier to see the extra ringing that the spring may or may not add (I won't tell which one of the two was printed without, you have to use your calibrated eye to tell)