Z axis: ball screws vs belts
-
@fma Ball screws or belts will let the bed drop if you don't do anything to prevent it.
Gear/belt reduction will increase resolution, increase lifting power of the motor, and can prevent the bed from dropping when power to the Z motor is lost. I used a worm gear drive in my printer's belt lifted Z axis, which ensures that the bed doesn't move when power is cut.
Someone at the makerspace built a printer using one of these to lift the bed and it produces excellent quality prints. I don't know what gear ratio he used.
Planetary gear reducers can be used but they have odd gear ratios that lead to weird steps/mm and so weird whole-step-multiple layer thicknesses.
You can screw around with clutches, brakes, etc. but they usually cost as much or more than gear reduction and require messing with the firmware to get the drive signals for them. You won't get increased resolution or lifting power.
-
@mrehorstdmd worm gear drive is a good idea. But the backlash may be an issue, as in 5 axis printing mode, Z axis will be driven back and forth...
Frédéric
-
@fma unless your planning on running with a z accel of over 9.81m/s (so ~10000mm/s +/- a little friction), gravity will still win, so you'll only be loading up one flank of the worm gear teeth, whether going +ve or -ve, and backlash will never come in to play
-
@fma Gravity keeps the gear teeth meshed in the same direction no matter which way the bed moves (assuming the weight of the lifted part is sufficient to overcome friction in the guide bearings, and that the guides are properly aligned) , so there's zero backlash. A worm gear is essentially a lead screw in this regard.
-
@fma for ball screws, back driving isn't guaranteed - but for the price (for quality parts at least from real manufacturers) you better be doing some true design and analysis work. Ball screws have been used (to great success) here, but in many cases elsewhere the results have been quite poor.
As a note, I'm happy with my ball screw implementation (on my current printer), but I am not going to claim that I did real solid engineering before purchasing & building. If I ever close out my current build, I know I won't be flying by the seat of my pants on the analysis for the next one ($$$).
A good check on ball screws here - be prepared to get into the weeds of manufacturer catalogs. https://www.linearmotiontips.com/how-to-determine-if-a-screw-will-back-drive/
Quality reading material - https://www.boschrexroth.com/documents/12605/25316270/R999001185_2021_04_EN_Gewindetriebe.pdf/b455f68f-90f4-03dc-758a-6360c9a7dcaa
Quick math for my setup (two ball screws belted together, 10kg bed assembly) the ball screw assembly frictional torque is 7x that of the back drive torque. I'd be hard pressed to put an additional 60kg on my build platform.
-
Thanks for your suggestions.
But in order to keep the price low, I think I won't use any reduction system: if the bed falls too hard when motors are off, I will just add shock absorbers at the bottom...
-
@fma said in Z axis: ball screws vs belts:
Thanks for your suggestions.
But in order to keep the price low, I think I won't use any reduction system: if the bed falls too hard when motors are off, I will just add shock absorbers at the bottom...
I believe your next problem will be potentially destroying a driver on your board from the back EMF.
-
@sebkritikel, Maybe I have been lucky all this time but having moved my bed manually many a time (with driver being inactive), I have never blown a driver chip
-
@jens55 said in Z axis: ball screws vs belts:
@sebkritikel, Maybe I have been lucky all this time but having moved my bed manually many a time (with driver being inactive), I have never blown a driver chip
I've carefully moved all of my gantries by hand and haven't blown any either - but its always good to try and mitigate risks, especially with some sort of uncontrolled movement.
-
@mrehorstdmd said in Z axis: ball screws vs belts:
Planetary gear reducers can be used but they have odd gear ratios that lead to weird steps/mm and so weird whole-step-multiple layer thicknesses.
Not always true. My first test used low priced units with 5.18 to 1. Once I was satisfied that belt drive was feasible I upgraded to more expensive units with 5 to 1.
The bed does not drop when power is removed.
-
Another option, although adding more complexity is to run a counterbalance system with ballscrews. I am implementing this on a new build but it's not meant to eliminate back driving but to make the system more responsive and repeatable fast Z hopping a 7kg bed assembly and allow the steppers and NSK C3 ballscrews to be the best they can. Purely theoretical but who has time to test and evaluate everything haha.
From an accuracy standpoint you might be better off with belts if the plan was to use cheap C7 (probably worse with high fluctuation), rolled ballscrews. How about belts and a counterbalance? With a little thought a very clean design may be possible.
High quality, low error lead screws are available from companies like Helix for a fraction of the cost of a THK, NSK, etc ballscrew.
And lastly it's good practice to consider using whole steps when setting resolution as micro stepping once was shown to be far from accurate. Maybe that's changed with the new drivers in the last few years but I doubt the motors have. -
@3dpmicro said in Z axis: ball screws vs belts:
And lastly it's good practice to consider using whole steps when setting resolution as micro stepping once was shown to be far from accurate. Maybe that's changed with the new drivers in the last few years but I doubt the motors have.
The counterweight idea is interesting.
I'm using a mix of 3x cheap ballscrews and a belt reduction with a single motor. The bed never drops, but I have 2500steps/mm at 16:1 microstepping. I'm sure, with three motors, it will not drop either with belt reduction or without.
I reduced u-stepping to 4:1 once but than the mesh-leveling moves of the Z-motor turned from the common sizzling/hissing sound to farting. -
@o_lampe what is the lead on your ballscrews?
-
@3dpmicro
1204 from robotdigg -
@mrehorstdmd I am using ball screws with a 300x300x10 bed and they do not drop when the motors are disengaged, but, the bed does move easily when pressed with I would say about a pounds worth of pressure.
-
@cdl1701yahoo-com how many ballscrews and what is the lead?
-
@sebkritikel said in Z axis: ball screws vs belts:
I believe your next problem will be potentially destroying a driver on your board from the back EMF.
I never destroyed a driver, even by moving hard the gantry.
Have a look at this vidéo (from quebec): https://www.youtube.com/watch?v=PCpfXJV89WA
-
@fma said in Z axis: ball screws vs belts:
I never destroyed a driver, even by moving hard the gantry.
It depends on your driver... Twice I have killed a TMC2100 by moving an axis while unpowered (and not fast either). Granted those were the old Waterott boards, you can buy an extra board to protect them now.
I have not blown a driver otherwise, and not for lack of trying. But it can happen (if rarely) depending on your hardware.
On an unrelated note: I think I remember you from the ORDBot days, do I?
VORON V2 CoreXY + Duet3 Mini5+ Ethernet v1.0 with Mini2+ expansion, VORON V0 CoreXY + Duet2 Maestro
-
@fulg said in Z axis: ball screws vs belts:
Granted those were the old Waterott boards, you can buy an extra board to protect them now.
Some boards such as SKR use small driver modules that can be easily replaced with no tools. Not as elegant and dense but functional.
-
@zapta I believe Waterott was actually the first provider of Trinamic drivers in Pololu Stepstick form-factor (with the chip on the back side so you could put the heatsink on the correct side of the PCB!), but the lack of physical space means no protection diodes. They are easy to replace but they are also quite fragile, at least the TMC2100 model.
