Hi All,

I'll need a Z axis with my gantry, thought I'd discuss that separately.

Did a lot of research on the belt driven triple lead screw configuration, and despite it all looking diabolical it seems very popular.

I'm looking for a 450mm build height on a 300x300 platform, and was looking at using 2x 15mm linear rails at the back with one lead screw in the middle, then a lead screw on either corner at the front.

Seems like people are liking a 1mm pitch on the screw or thereabouts, and I've seen mentioned using 2-1 ratio in the belt pulleys, specifically 20t on the motor and 40t on the lead screws.

Is that what's the good thing?

I read something from jetguy where he said to only use single blocks on the rods/rails as they were only supporting XY movement, and more than that would cause the screws to bind.

Thinking 10-12mm screws?

What you have in mind is exactly what I'm proposing to do with a couple of minor differences. I've got (designed) 2 linear rails on the sides but at the front, two lead screws at the same position and a single one in the centre of the back. My bed will be 400x 400 (about 330 x 340 printable area) but a lot taller (about 800mm Z movement). As the bed will be constrained by the linear rails (I'm proposing to use Open Build V slot, Delrin wheels etc), then the screws are only providing lift, and I think they need to kind of "float around" otherwise they'll be fighting with the linear guides and perhaps cause binding issues. So, personally I can't think why 8mm diameter won't work. I've got 3 off 1metre lengths next to me and for the life of me, I can't think why I'd need anything bigger. Of course, if the screws are being use to constrain the bed from any sideways movement, then it would be a different matter.

Originaly, my plan had 4 linear guides - one in each corner more or less but I soon realised that alignment would be a real pain. The I went down to 3 guides, one near each screw but after a lot a research, found that this too would cause more problems than it solved. Two guides seem to be optimum. Take a look on OpenBuilds. From memory, I think there was something called a tripple D bot or some such that someone built. There is some stuff there that made sense (at least to me it did).

I keep getting blasted about thread pitch but I really don't get this "use 4 start / 8mm pitch" thing. With a 1.8 degree stepper, you get 200 (full) steps per mm with a 1mm pitch screw or 20 steps per 0.1mm. Maybe that's a bit too fine a resolution but what the heck, - the odd skipped microstep ain't gonna matter too much. If you have an 8mm pitch, you get 2.5 steps per 0.1mm. So, you have to rely on micro stepping for positional accuracy which ain't what microstepping is good for. AFAIK microstpepping is just a way of transitioning smoothly between full steps. It'll increase resolution, but it doesn't increase accuracy. Typically, positional accuracy of a stepper is +_ 5 % per step. Having 16 small steps between full steps, doesn't improve this. When I started my toolmaker apprencticeship back in the 1969, we had lathes with multi start lead screws. They were invented to give greater linear movement for a given angular movement - i.e designed for one thing only - speed. We don't need speed on the Z.

Now of course, you can use a course thread and gearing to get the resolution back. With an 8mm pitch screw 1:2 gearing will give you 5 full steps per 0.1mm, which is better than a fraction, like 2.5. So, I guess it's a case of what you can get hold of locally, at the best price. 4mm pitch would be the same as geared 8mm. I mm pitch is maybe finer than you need but if you can pick it up at a good price, why not?

I know I'll get blasted for this and I'm just an old guy who cuts and screws bits of wood together, but I've yet to hear anything which can convince me that I need to be using 4 start, 8mm pitch lead screws which is what everyone seem to be selling.

Edited 1 time(s). Last edit at 06/09/2016 03:14PM by deckingman.

I must have got it wrong about the 1mm pitch being the norm....thought I'd done my homework.

I'm looking again at the Hiwin range:

[www.industrialdynamics.com.au]

And if I'm using a Duet which seems likely, I'd probably be looking at 0.9 deg steppers?

The simple/standard internal recirculating screws they have seem to come in a range with all sorts of pitches starting at 1mm, so maybe 4mm with 2:1 pulley ration would be good?

Obviously it must be a good system with many of the pros swearing by it, but I'm amazed it doesn't all just lock up.

Quote
hobbymods
I must have got it wrong about the 1mm pitch being the norm....thought I'd done my homework.

I'm looking again at the Hiwin range:

[www.industrialdynamics.com.au]

And if I'm using a Duet which seems likely, I'd probably be looking at 0.9 deg steppers?

The simple/standard internal recirculating screws they have seem to come in a range with all sorts of pitches starting at 1mm, so maybe 4mm with 2:1 pulley ration would be good?

Obviously it must be a good system with many of the pros swearing by it, but I'm amazed it doesn't all just lock up.

AFAIK, 0.9 deg steppers are more expensive. Do you need double the resolution if 1.8 deg give a finer resolution than you need? It's easy enough to do the maths. e.g. 360 degrees in a circle, so a 1.8 degree per step motor takes 200 steps to do I revolution. If the pitch is 1mm then 0,1mm will take 20 steps. If the pitch is 8mm then 0.1mm will take 200/80 = 2.5 steps. That's not good 'cos it's not a whole number so you'd be relying on micro stepping for positional accuracy. Gearing an 8mm pitch screw by 1:2 would give 5 whole steps per mm which is better. Equally, using 0.9 degree steppers and no gearing will give you 5 whole steps. Personally I think gearing only comes into it's own if you need to lift a heavy load with a small motor (which we don't). There are various ways of doing this. I don't think there is a best way (but what do I know?). Just bear in mind that we don't need to lift a heavy load nor do we need to do it quickly. What we do need is to do it in very precise increments of (potentially) 0.1mm or less and for that we should really be looking at whole steps, not fractions of a step. Also, as I said before the typical step accuracy of a motor seems to be +_ 5% but it's non accumulative. However, the fewer the number of steps, the more significant this accuracy may become. I guess if 4mm pitch is significantly less costly than (say) 2mm pitch, then use that with 1:2 gearing (the motor shaft being the 1 and the lead screw pulley the 2). If there is hardly any difference in price, then why not use 2mm pitch (or even 1 mm)?. Either method would achieve the same end. Maybe another factor might be belt length. You'll need a closed loop belt. What's the nearset size you can get and does this mean that you have lots of belt length to lose? If so, then big pulleys might help.

Quote
deckingman

Quote
hobbymods
I must have got it wrong about the 1mm pitch being the norm....thought I'd done my homework.

I'm looking again at the Hiwin range:

[www.industrialdynamics.com.au]

And if I'm using a Duet which seems likely, I'd probably be looking at 0.9 deg steppers?

The simple/standard internal recirculating screws they have seem to come in a range with all sorts of pitches starting at 1mm, so maybe 4mm with 2:1 pulley ration would be good?

Obviously it must be a good system with many of the pros swearing by it, but I'm amazed it doesn't all just lock up.

AFAIK, 0.9 deg steppers are more expensive. Do you need double the resolution if 1.8 deg give a finer resolution than you need? It's easy enough to do the maths. e.g. 360 degrees in a circle, so a 1.8 degree per step motor takes 200 steps to do I revolution. If the pitch is 1mm then 0,1mm will take 20 steps. If the pitch is 8mm then 0.1mm will take 200/80 = 2.5 steps. That's not good 'cos it's not a whole number so you'd be relying on micro stepping for positional accuracy. Gearing an 8mm pitch screw by 1:2 would give 5 whole steps per mm which is better. Equally, using 0.9 degree steppers and no gearing will give you 5 whole steps. Personally I think gearing only comes into it's own if you need to lift a heavy load with a small motor (which we don't). There are various ways of doing this. I don't think there is a best way (but what do I know?). Just bear in mind that we don't need to lift a heavy load nor do we need to do it quickly. What we do need is to do it in very precise increments of (potentially) 0.1mm or less and for that we should really be looking at whole steps, not fractions of a step. Also, as I said before the typical step accuracy of a motor seems to be +_ 5% but it's non accumulative. However, the fewer the number of steps, the more significant this accuracy may become. I guess if 4mm pitch is significantly less costly than (say) 2mm pitch, then use that with 1:2 gearing (the motor shaft being the 1 and the lead screw pulley the 2). If there is hardly any difference in price, then why not use 2mm pitch (or even 1 mm)?. Either method would achieve the same end. Maybe another factor might be belt length. You'll need a closed loop belt. What's the nearset size you can get and does this mean that you have lots of belt length to lose? If so, then big pulleys might help.

The screw on my Zortrax machines is 12mm with a pitch of 4mm.

Again my trusty linear guy from Hiwin (who has shown a keen interest and unlimited patience) has told me that he can do 12-16mm screws for the same size as 8-10 (he keeps saying that lol) in various pitches starting from 1mm.

I see these on ebay at 5mm pitch (they only do them in 16mm):

[www.ebay.com.au]

I can only go on what I've been reading here and elsewhere, but I'll probably go with 4mm pitch, 1:2 pulley ratio with 20T on the motor and bit of a heavier duty 1.8 deg stepper to spin it all.

Another member mentioned the 0.9 deg steppers for the corexy gantry, which is where I thought they might have a positive effect. You're probably right about it being overkill on the Z axis I'd say.

I don't think there is a definitve right and wrong way. 4mm pitch with 2:1 gearing will do exactly the same as 2mm pitch and no gearing. 4mm pitch with no gearing would probably work too. It would give you 5 full steps per 0.1mm but my gut instinct tells me that may be a bit too course. I still think 16mm is much larger than is necessary but I guess it won't have any negative impact. Can you get pulleys to suit though? You'd certainly have to use big pulleys and gear them because the outside diameter of a 20 tooth gt2 pulley is 16mm. Or you'd need to get the ends of the screws turned down. GT2 pulleys are quite common in 5 and 8mm ID here in the UK but larger ID pulleys are hard to find. One other thing that may affect your decision is how you intend to fix the pulleys to the screws. I was just going to use "normal" pulleys with grub screws to clamp them to the threaded shaft but I don't know if it'll work. Some people go with lead screws that have been turned down at one end. I guess ideally, they ought to be turned down and then a milled flat.

No they have an 8mm section at the bottom that the pulley will slip onto.

Id never clamp a pulley onto a threaded part

I'd really like to go with 12mm screws, as I agree that 16mm is just too big and heavy.

1204 is a fairly common size in the cheapy chinese's, but I can get any pitch I want from Hiwin.

It's one area where I might use the no name brand to save money, as I've got to get 3 of them. I'll see what the prices for Hiwin are like first.

Smallest layer height I'd EVER bother with would be 0.05, and to be honest I'd use that once in a blue moon but that's where I'll set the bar. I always just use 0.2/0.3 depending on the size of the print and am pretty happy with those. I occasionally use 0.15 for finer detailed prints....to be honest I'd set the bar at 0.1.

I'd like the bottom bearing mount to mount flush on the floor, and through the floor with the pulley underneath.

Is there an issue with these pillowblock bearings supporting a vertical load by being bolted straight onto the floor?

[www.ebay.com.au]

If not I could clamp them on the 10mm machined section and have my pulleys straight onto the 8mm section, and that bit of self aligning might make my Z resolution nicer? I don't see it done often, so I'm guessing those big block mount things are the only way.

Ended up going with 3x (supposedly) genuine TBI Motion 1204 ballscrews off alibaba.

[www.aliexpress.com]

I went with a 600mm length overall with the 4mm pitch as it was a standard off the shelf size...sort of, and would suit my need for 450-550mm Z travel.

Both this guy and the Australian distributor said exactly the same thing, it came in plain lengths and had to be ground to suit.

Difference was that the Aussie distributor was a rude, arrogant, grunting Western Australian pig who had absolutely bugger all interest in helping at all, and the Chinese guy just couldn't do enough to help and spoke good english. He steered me towards a cheaper mount that was the same accuracy class as the screw as well.

And the price.

Some good articles on his page about how to spot fakes, and according to the pics there my Zortrax has a genuine TBI motion screw, and I spotted a definite fake in another sellers page.

Stung a bit to fork out this cash so early in the piece, but while I can print/mock up imitation rails, I really need the screws with their support flange/mount from the start.

We'll see how it goes.

Edited 1 time(s). Last edit at 06/13/2016 04:58PM by hobbymods.

Yes I think 0.05mm is a good place to set the bar with a view to potentially printing at 0.1mm (although personally, I do most things at 0.3mm and rarely bother with less). But you never know what you might want to do at some point in the future...........

When you think about it, if the lead screws are only lifting the bed and not being used in any way to provide any sort of lateral stability, all that's needed is something that'll take the weight of the bed plus whatever the printed object itself weighs, without stripping the thread or wearing out too much. Logically, something like a single 5mm threaded rod would do that, so logically something like 3 off 5mm rods would be overkill. Personally, I couldn't bring myself to go that small but I can't think of a technical reason why not - it just doesn't feel right. My Mendel variant has two 5mm threaded rods for lifting the Z and whilst this printer has many, many faults, Z movement has never been a problem. I guess that's why I'm happy to go with 3 off 8mm lead screws for my new build.

As for how to retain the end of the lead screw, I'm not convinced that those pillar block thingies are the way to go. You have the weight of the shaft bearing down so you'd be relying on the ball contact with the side grooves in the bearing housing to take the weight. Maybe deep groove bearings are OK but again, it doesn't "feel right". I've designed my bottom bearing blocks as a kind of blind hole. A dedicated thrust bearing, slightly smaller than the shaft diameter sits in the bottom to take the downward force. Then above that, I'll have a "normal" roller bearing to take the lateral forces on the belts, with the pulley above that. So basicaly, two seperate bearings doing two seperate jobs.

As an aside, DC42 who's firmware I use (Duet electronics) has it on his list of things to do, to implement automatic bed levelling by using 3 independently driven lead screws and his IR height sensor. I don't know if that is anything that would interest you but I'm designing my printer in such a way that I could easily add 2 extra Z motors and take out the single looped belt to be replaced by 3 very short ones (one for each screw/motor). I might give it a try. If it doesn't work, it'll be easy enough to put it back to a single motor.

Edit. Looks like we were both typing at the same time. Glad you got sorted anyway.

Edited 1 time(s). Last edit at 06/13/2016 05:19PM by deckingman.

Just checked the link you posted - those screws look like the dogs' danglies.

Quote
deckingman
What you have in mind is exactly what I'm proposing to do with a couple of minor differences. I've got (designed) 2 linear rails on the sides but at the front, two lead screws at the same position and a single one in the centre of the back. My bed will be 400x 400 (about 330 x 340 printable area) but a lot taller (about 800mm Z movement). As the bed will be constrained by the linear rails (I'm proposing to use Open Build V slot, Delrin wheels etc), then the screws are only providing lift, and I think they need to kind of "float around" otherwise they'll be fighting with the linear guides and perhaps cause binding issues. So, personally I can't think why 8mm diameter won't work. I've got 3 off 1metre lengths next to me and for the life of me, I can't think why I'd need anything bigger. Of course, if the screws are being use to constrain the bed from any sideways movement, then it would be a different matter.

Originaly, my plan had 4 linear guides - one in each corner more or less but I soon realised that alignment would be a real pain. The I went down to 3 guides, one near each screw but after a lot a research, found that this too would cause more problems than it solved. Two guides seem to be optimum. Take a look on OpenBuilds. From memory, I think there was something called a tripple D bot or some such that someone built. There is some stuff there that made sense (at least to me it did).

I keep getting blasted about thread pitch but I really don't get this "use 4 start / 8mm pitch" thing. With a 1.8 degree stepper, you get 200 (full) steps per mm with a 1mm pitch screw or 20 steps per 0.1mm. Maybe that's a bit too fine a resolution but what the heck, - the odd skipped microstep ain't gonna matter too much. If you have an 8mm pitch, you get 2.5 steps per 0.1mm. So, you have to rely on micro stepping for positional accuracy which ain't what microstepping is good for. AFAIK microstpepping is just a way of transitioning smoothly between full steps. It'll increase resolution, but it doesn't increase accuracy. Typically, positional accuracy of a stepper is +_ 5 % per step. Having 16 small steps between full steps, doesn't improve this. When I started my toolmaker apprencticeship back in the 1969, we had lathes with multi start lead screws. They were invented to give greater linear movement for a given angular movement - i.e designed for one thing only - speed. We don't need speed on the Z.

Now of course, you can use a course thread and gearing to get the resolution back. With an 8mm pitch screw 1:2 gearing will give you 5 full steps per 0.1mm, which is better than a fraction, like 2.5. So, I guess it's a case of what you can get hold of locally, at the best price. 4mm pitch would be the same as geared 8mm. I mm pitch is maybe finer than you need but if you can pick it up at a good price, why not?

I know I'll get blasted for this and I'm just an old guy who cuts and screws bits of wood together, but I've yet to hear anything which can convince me that I need to be using 4 start, 8mm pitch lead screws which is what everyone seem to be selling.

awesome info!
Thank for sharing