leadscrew vs ballscrew
-
find the high spot with the dial indicator, heat the high spot with a pencil torch, cold quench with a wet cloth, simple, mind you i have a few years of experience of straightening hydraulic rods offshore.
-
@CaLviNx said in leadscrew vs ballscrew:
Ian @deckingman has a nice video covering lead screws if you want a little education.
For the sake of completeness, here is a link to the aforementioned video https://www.youtube.com/watch?v=o_0xdrKUYVU
-
@CaLviNx said in leadscrew vs ballscrew:
find the high spot with the dial indicator, heat the high spot with a pencil torch, cold quench with a wet cloth, simple, mind you i have a few years of experience of straightening hydraulic rods offshore.
I tried that technique too (using proxxon butane torch) ... same as with "pressing it between two blocks" .. I'm just not made for that
.. -
@arhi 10x2 ACME leadscrews I got were very straight and solved my heavy dual z-axis falling problem - but they are a pain to tune ideally, especially with dual linear rails on each side.... everything needs to be aligned within a fraction of a mm or you'll have rail grinding and artifacts in the print.... but if you do it properly it solves most of the problems.
-
@omni said in leadscrew vs ballscrew:
@arhi 10x2 ACME leadscrews I got were very straight and solved my heavy dual z-axis falling problem - but they are a pain to tune ideally, especially with dual linear rails on each side.... everything needs to be aligned within a fraction of a mm or you'll have rail grinding and artifacts in the print.... but if you do it properly it solves most of the problems.
The purpose of a leadscrew & nut (in the context of a 3d printer) is to convert a rotational action from the motor into a linear vertical raising/lowering action, therefore you only need to concern yourself about that, the major problem I see is people forgetting that and trying to constrain the nut and or the leadscrew too much, if you design the system so as the leadscrew nut is able to "float" along the X & Y planes then you do not need to get too anal, the only reall alignment required is to make sure all the rails are vertical, and alignment shouldnt be he job of the leadscrew to worry about, its only concern is regards to the up or down motion.
The weight of the bed/bed frame pushes downwards against the nut in the vertical Z plane keeping everything where it needs to be when going down, when the leadscrew is rotating it is pushing against the weight and that keeps everything in place on the upwards stroke. too many people try to over design and over think these simple machines.
-
One of the biggest differences you are likely to see going from cheap leadscrews to cheap ballscrews is CPE.
Cumulative Pitch Error.
Or the thing that makes your printer produce objects that are the wrong height.
Cheap leadscrews can be off by 3mm in 300mm.
Even cheap ballscrews are normally in the sub-mm range.
Expensive leadscrews can be just as good.
Expensive (ground, as opposed to rolled) ballscrews will be in the <0.01/300mm sort of range. Depending on how much you want to spend. Although if you want to worry about that kind of precision, linear encoders are the way to go. -
i tough about this i him gonna give a try to some belt , here my first idea using pulley, 20<60,20<60 , one set of pulley give me 480 e step i believe with 2 set should get a lot more resolution him just not sure of the exact number , probably 3x more if somebody can confirm this ?
here the first sketch i believe everything would be more compact as well these pulley are slightly thicker than the one him looking at . still need to think about a belt tensionner system probably adding some kind of arms either spring loaded or with a tightening bolt ,
2 belt linked to the shaft would insure the bed to move without much problem would be fixed somewhere on the platform
-
@Dad003 said in leadscrew vs ballscrew:
20<60,20<60 ,
i think it will give you 240 steps. (assuming 1.8 steppers)
you have twice a 3:1.
your actual belt to move the z is 60 tooth.
so 200/2/60* 1/16 microstepping 26.66
3:1 once is 80
3:1 twice is 240
the solution on the ender 3 is
16<80 which gives a 5:1 ratiothen use a 20 tooth to drive the belt.
200/2/20 * 1/16 microstepping = 80
5:1 with 80 is 400.
-
Just to be clear- you are saying that your 'out of straightness' leadscrew is putting enough lateral load on the build platform to give you Z-banding issues ? That does not make sense.
What may be the issue is the elastic buckling of that REALLY tall slender column. Peek into this rabbit hole - https://en.wikipedia.org/wiki/Euler's_critical_load
There are other theories, but this is the easiest to understand.Some quick and dirty estimating to put it in perspective- the normal range for compression members, we proportion them to a ratio of KL/r <200. For Tension members, we normally try to hole to KL/r<300. Since it a machine that we are try to get some manner of precision out of, I would not go over 80 in general.
A 700mm long TR8x8 leadscrew has an effective solid cross section of 6.5mm. This results in a KL/r of 461! Critical buckling load is about 17 pounds. This is kinda why TR8's dont scale well beyond i3 lengths. Look at the critical stress curve and about where i faked it in and plotted the resulting critical stress.
What might be an easier and faster mod to try and cure that is put the leadscrew in tension instead of compression. Use a stop collar and thrust bearing at the top and HANG the platform from above. Tension is self righting. You are mostly doing the same thing with a hanging belt drive, but adding ain a bunch of complexity with all the jackshafts.
-
@CaLviNx Yes, I am completely aware of that principle and my first thought was exactly what you are talking about, but in practice making a floating nut in a dual z-axis setup like mine (the z axis is holding the x-axis bar) and a height of 500mm that made more problems than solve. When the leadscrews are very long (like mine - almost 600mm) and you do not align them perfectly - the nut and/or screw tend to "dance" a bit which translates into visible z print artifacts. Using a springy motor coupler with a small ball bearing inside) helped a little bit with the constraint problems. When I was talking about aligning - I was thinking more about aligning the linear rails - since I have 4 of them for Z - 2 for each side od the Z axis, and aligning 4 linear rails to do a constrained up/down motion together is a pain. Also it's much more difficult when the lead on the leadscrew is 2mm instead of 8mm which is more common on printers since it's much more sensitive to misalignment.
-
@omni said in leadscrew vs ballscrew:
................ When the leadscrews are very long (like mine - almost 600mm) and you do not align them perfectly - the nut and/or screw tend to "dance" a bit which translates into visible z print artifacts.
If the linear rails are doing their job, and the screws are unconstrained, then any "dancing" of the nut or screw should not translate into any movement of the build platform. If it does, then there must be unacceptable play in the linear guides.
When I was talking about aligning - I was thinking more about aligning the linear rails - since I have 4 of them for Z - 2 for each side od the Z axis, and aligning 4 linear rails to do a constrained up/down motion together is a pain.
Which just means that you are using too many guides. You only need two, preferably located at opposite corners of the build platform. Located and aligned correctly, this is enough to prevent any movement in X, or Y and also any rotation.
Also it's much more difficult when the lead on the leadscrew is 2mm instead of 8mm which is more common on printers since it's much more sensitive to misalignment.
How so? How does the thread form affect the ability to align the screws? An 8mm lead, 4 start screw has the same 2mm pitch as a 2mm lead, single start screw.
For info, I use three unconstrained 8mm diameter 2mm lead screws, 900mm in length driven by a single motor and continuous belt, which raises and lowers a 400mm square build platform constrained by two linear guides arranged as detailed above. No Z artefacts, no need for any sort of software bed levelling or flatness compensation.
-
@deckingman said in leadscrew vs ballscrew:
Which just means that you are using too many guides. You only need two, preferably located at opposite corners of the build platform. Located and aligned correctly, this is enough to prevent any movement in X, or Y and also any rotation.
EXACTLY!
-
@omni said in leadscrew vs ballscrew:
@CaLviNx Yes, I am completely aware of that principle and my first thought was exactly what you are talking about, but in practice making a floating nut in a dual z-axis setup like mine (the z axis is holding the x-axis bar) and a height of 500mm that made more problems than solve. When the leadscrews are very long (like mine - almost 600mm) and you do not align them perfectly - the nut and/or screw tend to "dance" a bit which translates into visible z print artifacts. Using a springy motor coupler with a small ball bearing inside) helped a little bit with the constraint problems. When I was talking about aligning - I was thinking more about aligning the linear rails - since I have 4 of them for Z - 2 for each side od the Z axis, and aligning 4 linear rails to do a constrained up/down motion together is a pain. Also it's much more difficult when the lead on the leadscrew is 2mm instead of 8mm which is more common on printers since it's much more sensitive to misalignment.
All this says to me is: BAD DESIGN....
-
@deckingman said in leadscrew vs ballscrew:
How so? How does the thread form affect the ability to align the screws?
you must constraint the screw in one point - motor ... so lets assume the other side of the screw is free for the screw tip to "fly around" on the top ( like most of mine design have) with "small 8mm screw and 4x2 pitch" the nut can handle a misalignment of e.g. 1 degree without binding. identical setup with bigger 10mm screw and 1x2 pitch nut "grabs" more threads and at 1 degree binds up, can handle max .2 degree without binding. Not sure why I'd have to draw them both to see why the bigger one binds so much at same angle... the ideal solution for this is with those arresters that ppl used to design where you allow nut to freely move in XY direction but you lock it in Z direction... I never tbh seen a design that I liked and used more than few
the angle I'm talking about if unclear is the angle between the center axis of the nut and the center axis of the screw
-
@CaLviNx said in leadscrew vs ballscrew:
All this says to me is: BAD DESIGN....
as I'm familiar with that build ... it's of course way better design to use single high quality linear rail on each side of the X bar that moves in Z direction, but due to well known situation right now getting a high quality linear rail here is next to impossible and cost more than that whole printer so not really doable. That's why 2 smaller PRC linear rails are used on each side. Doubling them increases their precision / reduces the "play" but increases complexity of the alignment significantly. So the design is driven by what's available and not what's best
... took a while for omni to align those 4 rails and 2 nuts but that machine now works freaking awesome 
... -
@arhi said in leadscrew vs ballscrew:
the angle I'm talking about if unclear is the angle between the center axis of the nut and the center axis of the screw
Good design and proper assembly cures all.....
-
@CaLviNx said in leadscrew vs ballscrew:
@arhi said in leadscrew vs ballscrew:
the angle I'm talking about if unclear is the angle between the center axis of the nut and the center axis of the screw
Good design and proper assembly cures all.....
fairly possible when you do it from scratch, not something you can force when you are retrofitting something like wanhao d9
-
@arhi said in leadscrew vs ballscrew:
@CaLviNx said in leadscrew vs ballscrew:
All this says to me is: BAD DESIGN....
as I'm familiar with that build ... it's of course way better design to use single high quality linear rail on each side of the X bar that moves in Z direction, but due to well known situation right now getting a high quality linear rail here is next to impossible and cost more than that whole printer so not really doable. That's why 2 smaller PRC linear rails are used on each side. Doubling them increases their precision / reduces the "play" but increases complexity of the alignment significantly. So the design is driven by what's available and not what's best
... took a while for omni to align those 4 rails and 2 nuts but that machine now works freaking awesome ...All my rails come from an Ebay seller in South Korea Dy Global and are either good used or new old stock being Genuine from various manufacturers Hi-Win, THK etc, i usually contact him and just ask for what i want instead of trawling his site as he usually has stuff in stock that is not listed on his page on ebay. I have had 7 orders from him now and every order has been perfect.
He supplies with 72hr shipping via fedex to europe included in the rail price.
So the "excuse" of good rails not being available is horsesh**
The best rails I have had so far are Bosch-Rexroth, they just ooze quality you can see it and feel it from just holding them in your hands
-
@arhi said in leadscrew vs ballscrew:
@CaLviNx said in leadscrew vs ballscrew:
@arhi said in leadscrew vs ballscrew:
fairly possible when you do it from scratch, not something you can force when you are retrofitting something like wanhao d9
For some games that you know you cant win, dont play......
-
@CaLviNx said in leadscrew vs ballscrew:
So the "excuse" of good rails not being available is horsesh**
ah ... don't talk about stuff you can't even imagine
- 99% of amazon sellers will not ship to Serbia at all, no matter what shipping method
- fedex/dhl if/when arrive here cost us another 100eur for "customs processing"
- on top of price+shipping+customs processing we need to pay another 10% customs + another 20% VAT on top of all that
- in order to do "customs processing" we need to show proof that it will not be used in food production, weapons manufacturing and provide bunch other "import" paperwork + if customs decide on we have to pay for testing "quality" of the goods... laws didn't change much from Tito era and are made for big corporations to do all the import work and for private persons to be able to have access to only local stuff .. we use tricks to go around it but due corona most of those channels are dead now
For some games that you know you cant win, dont play......
when you know up front, yes, but when you are already waste deep .. you get it done any way you know how
anyhow, we are hijacking thread with irrelevant discussion
