Why 4 Wheels on the Guides?
-
Hi All,
I'm curious to know why the guides for Delrin wheels that run in the open build v-slot often use 4 wheels rather than 3? With 4 you either have some very tightly loaded or you always have 1 wheel lose. Wouldn't three be easier to set up and less likely to introduce a backlash like effect as the carriage rapidly traverses?
-
There are a number of designs that actually do use three wheels for exactly the reasons you mentioned. Such as:
https://www.thingiverse.com/thing:2155251
I have never talked to anyone that used three wheels personally so I don't know how well it works. I use 4 on my CoreXY and manage to have all 4 tight without putting too much load on them so I have no real motivation to switch.
-
Thought it may have just been and 'easier to CAD' thing! Cheers for the link I'll have a look at that. I was just a little concerned there may have been a more important reason.
Yes, 4 on my CoreXY print carriage, on the guides at the end of the beams, and on the z-axis guides. I've recently discovered a flat-spotted delrin wheel that had been locked onto the v-slot too tight which set me thinking.
-
@doctrucker I have heard of stuff like that, which is one of the reasons I use the Xtreme(Polycarb) wheels rather than delrin. I have had no problems with those whatsoever.
-
I saw them when I was building up a parts list for a second set of z-axis guides. I find it so tempting to 'developer gold plate' the system. Trying very hard to hold myself back and only change/upgrade things when I've spotted a problem on the parts. Now I've flat spotted them I guess there is justification to move to the more expensive options!
The idea of a second set of guides being a contradiction to the thought process of dropping from 4 to 3 wheels on the guides but want to constrain the elevator a bit better to control the resonances a bit better.
-
My CR10 has 3 wheeled carriages and works extremely well. When I used to have a kossel mini with wheeled carriages that had three wheels too. I haven't seen a four wheeled carriage.
-
3 wheels works great on both of my printers as well.
-
Thanks for your comments. Looks like the 4 wheels used on the build I followed was an exception rather than the norm. When I've got time I'll CAD up some 3 wheel equivalents.
I built the 300x300 version of the following and have converted to direct extruder to play nice with some soft TPU. Now running three screws to, but need to move the original two screws forward in the frame to the other side of the elevators c of g. Clear improvement in side wall in the three screw vs two screw setup.
-
I'm using this 3 wheel z axis mod on my DBot.
https://www.thingiverse.com/thing:1802733
And this one for Y axis
https://www.thingiverse.com/thing:1930200
I didn't find it too difficult to get the 4 wheel versions tight but they didn't seem to want to stay even over time. Haven't had a problem with the 3 wheel versions.
I also switched to a 3 screw single motor setup for the z axis with wheel guides in front and back. It's much more stable. In my opinion it is the only same way to go.
You can see my arrangement here
-
I think when it comes to 3 versus 4 wheels its analogous to bed levelling points 3 defines a plane 4 can deform the bed. A three-wheeled carriage forms a triangular contact with the extrusion, and should discourage any unwanted movement.
-
Yeah, changing to three point on the bed correction is on the radar. It is shelved at the moment because I'd like to look at getting a back plate for the elevator machined out of something insulating with a higher combustion temp than the poorly fitting ply sheet I am currently using, but that may as well wait until I've looked into a silicon heater with a bit more punch than the 300W unit on the system at the moment. With a list of mods as long as that it may remain shelved until I am sure I don't need to increase the build area at all!
Thanks again for the links. I'll be looking through them when I've got a few moments!
-
It's a lot like a 3 legged stool vs a 4 legged stool, have you ever sat on a wobbly 3 legged stool?
I use 3 wheeled carriages on my delta, the triangular load distribution makes it super simple to get the clamping load on the wheels balanced and stable.
4 wheels will definitely still work though, it may be more work to get the carriages stable, and they may be more stable in the long run once everything is balanced.
-
@doctrucker said in Why 4 Wheels on the Guides?:
Hi All,
I'm curious to know why the guides for Delrin wheels that run in the open build v-slot often use 4 wheels rather than 3? With 4 you either have some very tightly loaded or you always have 1 wheel lose. Wouldn't three be easier to set up and less likely to introduce a backlash like effect as the carriage rapidly traverses?
I think maybe it just stems from a few OpenBuilds YouTube videos and also if you were to buy (say) a gantry kit, it comes with 4 wheel sets. I started using 4 when I didn't know any better but now I use 3 where possible. My X carriage runs on twin parallel 2020 and I use 6 wheel sets but that's 3 per rail if you understand.
As you say, it's easier to get the tension right (and it saves a wheel set worth of cost). Ideally, you need the third wheel to be centrally located between the opposite two, otherwise you can end up with uneven loading, and sometimes it's not possible to get that position.
When I started out, I used OpenBuilds metal gantry plates but I found that sometimes, the tension was just too tight, even with the eccentric spacers at maximum adjustment. I fixed it by opening up the holes in the gantry plates to give the bolts a bit more clearance. Nowadays, I print all the plates but I always have a front and back plate so that the wheel bolts are axels, supported at both ends. It does mean double the number of spacers and longer bolts but I find that once set, the tension never changes, unlike having them fixed at one end to a single plate. The other thing I've come across is a bit of variation in rails. Sometimes you can fit a gantry to a length of 2020 and it's really tight, even with the eccentrics backed right off, but you can fix it simply by turning the 2020 90 degrees so it uses the other two slots. Obviously not possible with 2040 etc. I tend to design my plates with an extra 0.5 to 1mm distance over the OpenBuld spec between opposite wheel holes. It's cured the "too tight" thing and I've never had a problem of them being too loose.
-
Just wish I'd considered it before redesigning the front and rear carriage plates to take the titan and under slung stepper.
As I've been thinking about the design a little more I realised the guide wheels on the x-bar are loaded axially, rather than direct onto their contact patch, and then realised that the cross bar really should be rotated 90 degrees about the x-axis so that it's most rigid in comparison to the loadings applied to it from the y-axis motions of the head and loading the wheels better. I'd like to run the calcs at some point to work out if the difference it would make would be worth working on. There are negatives to the re-orientation; belt orientation wouldn't be as neat though and you'd probably end up relocating the cross bar to clear the belts. Wouldn't mind betting the accelerations that would warrant the work would flex the rest of the chassis far more than the bend in the cross bar.
-
I've got to figure out the assemblies on FreeCAD first. Too many rookie mistakes recently. Most frustrating of which was taking 5mm as the flat to flat measure on an M3 nut rather than 5.4. Luckly I'm building with PETG so the flex in the material seems to have coped!
-
@doctrucker said in Why 4 Wheels on the Guides?:
Just wish I'd considered it before redesigning the front and rear carriage plates to take the titan and under slung stepper.
As I've been thinking about the design a little more I realised the guide wheels on the x-bar are loaded axially, rather than direct onto their contact patch, and then realised that the cross bar really should be rotated 90 degrees about the x-axis so that it's most rigid in comparison to the loadings applied to it from the y-axis motions of the head and loading the wheels better. I'd like to run the calcs at some point to work out if the difference it would make would be worth working on. There are negatives to the re-orientation; belt orientation wouldn't be as neat though and you'd probably end up relocating the cross bar to clear the belts. Wouldn't mind betting the accelerations that would warrant the work would flex the rest of the chassis far more than the bend in the cross bar.
If it's any help, I've got the wheels loaded both ways on my printer. That is to say, on the Y carriages the wheels run in the upper and lower slots, but on the X carriage, they run in the side slots. The X carriage is "kind of wrong" but it was the only way I could do it. I've been throwing a few kgs of mass around for getting on 2 years, and have had no issues in that respect.
-
This post is deleted!