http://www.thingiverse.com/thing:97377

just sayin.

So....

Why bother going 90 degrees? Why not just use a Gregs Wade direct gear setup? But a smaller ratio, less than 2:1.

Hve the shaft stick out the side and attach a gear on it. Have the other gear on the motor and attach it so the gear is closer to the frame and the motor away from it. Then print out a mount. In otherwords, similar to the direct coupling idea. But now you can gain more torque by adjusting gear ratio (at the cost of speed of course).

I had never heard of the Greg's Wade  direct gear setup, so I had to look it up.  Most of these setups are in reference to the original RepRap Wade extruder, to which Greg made different modifications for either gear reduction or direct drive setups.

Most of these drive setups were designed to provide more torque to the extruder by using some sort of gear reduction, and almost all of them rely on printed gears.  Those are fine for extruders where backlash is not that important, but they wouldn't do very well as power transmission for the motion control because of their general sloppiness.

Anyway, there were 2 reasons for my suggestion:

1.  The OP was looking for alternatives because he's trying to reduce the amount of backlash that exists between the motor and the drive shaft. 

Replacing the belt with a direct drive setup (motor coupled to the shaft directly with an inline coupler), would eliminate that backlash altogether, but then the motor would have to be installed on the outside.  This setup does have a couple of drawbacks:  Torsional and axial deflection which would be minimal but present non the less considering the small diameter of the drive shaft

2.  90° drive.  While most people would probably balk at adding a motor to the sides of the frame, they would probably be more inclined to do so at the back, since the back of the printer already has a bunch of stuff on it.

A helical gear setup lets you drive the shaft at right angles.  Additionally, helical gears have a few added advantages over spur type gears.  Fast, very quiet, capable of supporting higher loads due to the larger tooth surface engagement.  On helical gears, teeth mesh gradually, and at any given time 3 or more teeth are engaged.  This also translates into much tighter gear mesh with less backlash.

Then there's the location of the gear drive, which would be nearly centered.  If you put pillow blocks on each side  of the bearing (or build a gear box for it), you minimize any of the axial or torsional effects I mentioned before.

The only thing you have to account for is the lateral thrust helical gears produce (they will try to push the shaft left or right depending on the direction of rotation).  End caps and/or thrust bearings would take care of that.

So we took care of most of the slop induced by motor connection to the the "Y" drive shaft.  But that still doesn't take care of any of the problems that the end carriage belts have.  You still need to adjust the belts just right, makes sure the 2 sides are aligned, etc.  Maybe some kind of single belt/pulley setup ???

Why bother going 90 degrees? Why not just use a Gregs Wade direct gear setup? But a smaller ratio, less than 2:1.

Hve the shaft stick out the side and attach a gear on it. Have the other gear on the motor and attach it so the gear is closer to the frame and the motor away from it. Then print out a mount. In otherwords, similar to the direct coupling idea. But now you can gain more torque by adjusting gear ratio (at the cost of speed of course).

This would actually work pretty good, if you could find just the right gears. I looked around and I did find a few small (12mm OD) helical gears with 5mm bore.  One fits the motor shaft directly, the other would have to be drilled out to 6mm for the "Y" drive shaft, then cross drilled and tapped to install a set screw.

The problem is that to make this work, the threads must be at a 45° angle, and most of the small gears I'm finding are somewhere around 30°.  I found lots of large (25mm+ OD) with 45° threads but they're simply too big.

Anyway, just thinking out loud.

I've been toying around with ideas of driving the "Y" drive shaft using gears set at right angles to each other.

Originally I was thinking about using a worm gear to drive the "Y" rod.  The idea was to install a slip on worm on the rod, then drive it with a gear at a right angle with the motor installed on the back of the printer from the outside.

Never the less, I got an earful from some of the readers about how slow this would be. 

It's true that worm gears are usually used in gear reduction setups, but in this case it would have been the opposite.  In a standard setup, the worm provides the rotational input, and the main gear is the output.  Depending on the number of teeth on the gear, and the number of threads on the worm (4 in this case), the gear reduction can be as low as 5:1. 

So my idea was that if we shift the input to the gear instead of the worm, then the worm (in our case the "Y" drive shaft) would rotate 5 times faster than the input, but it could be slowed down as needed.  The one thing I didn't take in consideration was that the torque required to drive this setup would be more than the standard motor is capable of.

So it's not very practical to use a worm gear.  However, that brings us to helical gears.

Helical gears can engage at right angles as well, and 2 identical gears would provide a 1:1 gear ratio.  The additional advantage of the helical gears is that the have a lot of surface engaged in the mesh, which tends to reduce backlash dramatically.  The disadvantage is the additional friction speeds up wear and tear on the gears.  This can be compensated for (to an extent) with good lubrication.

So what about something like this?  Let the pundits rip me to shreds...

http://www.soliforum.com/post/55362/#p55362

wardjr wrote:

switching to spectra fishing line helps a lot but not for the Y-motor drive belt.  I would guess that if you have that kind of backlash you haven't put the pillow block bearings on that rear Y-drive rod.
http://www.soliforum.com/topic/4824/how … -for-ever/

Yeah, I've got the pillow block, first mod I installed.

So is the backlash caused by the Y-belts on the side?  Spectra line will help in that case as you can tighten it up to the point of breaking things.  Same for the X-belt.  It takes a few runs to get the line tight but once that's done you really don't have to mess with it again.  It's a very cheap upgrade and if you don't like it just switch back to belts.

switching to spectra fishing line helps a lot but not for the Y-motor drive belt.  I would guess that if you have that kind of backlash you haven't put the pillow block bearings on that rear Y-drive rod.
http://www.soliforum.com/topic/4824/how … -for-ever/

Yeah, I've got the pillow block, first mod I installed.

With the x motor, where exactly is the wobble?

I recently did the belt calibrations, multiple times, the past few weeks. (3 more major problems and I'll be an expert ) All to discover I tweaked the pulley on my X motor when I was moving some things around and it's got a wobble in it now. I'm hopping I can fix it without having to buy a whole new motor. It's only a couple months old. I added a fan and that's made it possible to get usable prints again so I'm kinda staling out on pulling the motor off and inspecting it further.  Also finally getting some use out of my clear PLA

Yup, I've been experimenting with direct drive on my SD3 recently. The motor does stick out, but it's not too much (I just had to take a chunk out of my side panel to clear it. The printer is running well, but unfortunately I'm having trouble getting the alignment right. In summary, I have less backlash but now even more wobble. I'm hoping an experiment with some flexible couplers (instead of my average lathework) will yield worthwhile results.

Your idea of a gear drive is great, but note how close the y shaft is to the back wall - the motor wouldn't fit where the shaft has to line up, e.g. for a straight bevel gear. Hypoid bevel drives exist, which allows for a (specific) offset, but this may not be worth the cost / difficulty of procurement compared to having the motor stick out!

I have yet to see a gear set that doesn't have some play in it.  I think your (grob) on the right track with the direct drive.  I also think a good spider coupler will solve most of your issues.  I just wish you would hurry up and perfect it so I can upgrade mine

Your idea of a gear drive is great, but note how close the y shaft is to the back wall - the motor wouldn't fit where the shaft has to line up, e.g. for a straight bevel gear. Hypoid bevel drives exist, which allows for a (specific) offset, but this may not be worth the cost / difficulty of procurement compared to having the motor stick out!

Has anyone considered, or is aware of, a gear drive solution to replace the closed-loop belt on the Y-axis stepper?  I have the motor set as tight as I can against the belt, but I have to command about .15mm of Y movement before the Y shaft (the horizontal one across the back) starts to rotate.  The stepper motor itself rotates instantaneously as commanded.  I can put a dial indicator on the loop belt, and watch the slack come and go as the Y motor changes directions.

Such an arrangement would require some bevel gears and a short vertical shaft, but surely it would be more precise than the current belt configuration.

My total lash is about .2mm before the X-carriage moves, so I probably have the long Y belts about as tight and evenly balanced as one could expect.

An alternative might be some sort of idler arrangement on the Y stepper belt.

p.s. crazy thought:  is there somewhere in EEPROM Config or Printer Config that I can 'pre-load' a value to compensate for the lash?

p.p.s I saw a direct-drive mod someone did on their SD2, but that requires mounting outside the enclosure; I don't want to go there.