You are not logged in. Please login or register.
SoliForum - 3D Printing Community → Solidoodle Discussion → Odd circle test results
E is backlash on the extruder and should probably always be 0....
Might be relevant to geared extruders though... Certainly 0 on a direct drive like a solidoodle.
I tend to agree, I won't be using lash compensation myself. I'd rather just spend the money on hardware.
Your choice. it's not an either/or. Having this option in the slicer or firmware doesn't force anyone to use it, so whats the harm.
Do let me know when you get the backlash gone with hardware upgrade. So far I can tighten the Y motor belt as much as it can be tightened and still get backlash, and no binding.
So either the belt type is bad, pulley type is bad, frame is bad, or you need to find a sweet spot between too tight and too loose and I've spend too many days on that.
Writing slic3r is much easier when you make the assumption that machines will accurately reproduce the GCode.
As well as writing firmware is?
I don't understand why it being easier matters.
it would be easier to not write any code for handling situations where something goes wrong, but would that be good code? Less time to write, but less useful.
I understand where you're coming from but think of it this way instead. Machines which won't accurately reproduce it, won't reproduce it the exact same way. So backlash compensation will not be dependent on the machine. The values which need to be entered will be, but not the feature itself.
If lash correction belongs anywhere, it is in the firmware, in my opinion
That would work too. I don't personally care where it goes, I just believe it should be there, somewhere.
Backlash code for Marlin was never 100% complete and never added to the official branch.
With all the topics popping up on the forums (not just solidoodle, makerbot, printrbot, ultimaker, prusa) with people having the exact same problem, lacking skills to fix it in hardware, either the slicer or firmware developers and 3d pritner sellers should take this a bit more seriously than telling everyone to "tighten the belts".
Otherwise forget about consumer 3d printer market.
I mean seriously, it's been almost 4 freaking months of me owning a 3d printer and not using it for anything useful and rather fixing it. I and I believe most of the of the folks posting the same issue have more important things to do than turn fixing a printer into a hobby.
I've personally wasted a lot of my free time of 1/3th of a year of my life on this. Not to mention the money I spent. i could probably buy another Solidoodle 4 by now. Not everyone can afford this.
In short: ideally it should be fixed in hardware. In reality, there should be a software fix for people like us.
Interesting discussion. My $.02: while firmware seems preferable, I only see a slight difference between backlash compensation for an axis and the setting in slicer to extrude an extra length after retraction. Both compensate for some hysteresis in the system.
Extrude after retract is filament-specific. Axis backlash is not.
Axis backlash belongs in the machine's firnware, as it is machine specific.
Well, on the subject of circles, I've just printed my first test cylinder after getting the Y direct drive modification installed, and it is looking pretty much entirely backlash free. I've got a link to a video on my cylinder test page:
http://home.comcast.net/~tomhorsley/har … nders.html
The light reflection off the cylinder no longer shows any flattened or pointed spots, other than the blobbing at layer change, the cylinder seems perfectly uniform. So I am definitely liking this mod.
I think the direct drive mod gets too much flack because it sticks out of the frame. Big deal. Printers are tools!
Nice work as usual 
If the all-bearing conversion on the back rod doesn't do it for me, I'll probably follow in your footsteps.
dont know what slicer you used but with Slic3r, the spiral is result of option 'randomize layer start point' being checked and in either case its to prevent having a seam in 1 place that can crack/separate easy causing a weak spot on model.
Well, on the subject of circles, I've just printed my first test cylinder after getting the Y direct drive modification installed, and it is looking pretty much entirely backlash free. I've got a link to a video on my cylinder test page:
http://home.comcast.net/~tomhorsley/har … nders.html
The light reflection off the cylinder no longer shows any flattened or pointed spots, other than the blobbing at layer change, the cylinder seems perfectly uniform. So I am definitely liking this mod.
Good job and nice photos and videos
However, you're making side shots and videos where backlash is pretty hard to spot when you have some <2mm backlash left like me (no pointy dents).
I would try this mod out, but it appears I need a longer rod and I'll need to order one and wait it to be shipped.
But I do have access to brass rods. You think i could use those? Since there isn't a belt bending it int he middle anymore and I'm going to keep the pillow block, maybe it would work for this mod even though its weaker than steel?
Extrude after retract is filament-specific. Axis backlash is not.
Axis backlash belongs in the machine's firnware, as it is machine specific.
I agree the firmware is a better place but imho these are just tools and sometimes already rather far from perfect ones at that so I dont think it's terribly horrible to give people extra practical options even if it's wrong from a dogmatic point of view. While extrude on retract has a filament dependency preventing it from being practical to fix in firmware, it is also partially machine dependent (whether you have Bowden etc) so I dont see it as a terrible crime if some might find it useful and more convenient.
However, you're making side shots and videos where backlash is pretty hard to spot when you have some <2mm backlash left like me (no pointy dents).
I've added a photo of the print of the circle.stl file that appears way back up near the start of this thread, and it looks pretty darn good:
I would try this mod out, but it appears I need a longer rod and I'll need to order one and wait it to be shipped.
But I do have access to brass rods. You think i could use those? Since there isn't a belt bending it int he middle anymore and I'm going to keep the pillow block, maybe it would work for this mod even though its weaker than steel?
I don't have any idea if brass would work or not, but harder to get than a new rod is a coupler - the oldham coupler I got from Misumi took a week to ship, they said they could have the rod in one day. You certainly don't want to use one of the helix couplers, since they are basically designed to add backlash.
Claghorn, could you post the product page for that coupler? You may have posted it on your site but I'm on my phone so searching is difficult.
Claghorn, could you post the product page for that coupler? You may have posted it on your site but I'm on my phone so searching is difficult.
The part number is MCOCG17-5-6, which hopefully you can type in a search field, because the URL is insane:
How about a 3d printed coupler?
http://www.thingiverse.com/thing:299663
How about a 3d printed coupler?
http://www.thingiverse.com/thing:299663
You could certainly give it a try and switch to different one if it doesn't work, but the oldham coupler doesn't require totally precise shaft alignment (via some ingenious mechanics)
Okay, so I've ordered a 5 to 6 coupler to give this direct drive mod a try. Before it gets shipped in few weeks I'll try a 3d printed coupler instead and post my results here.
Somebody on google groups commented on the trapezoidal MXL teeth with this
Hmm. Trapezoidal belt teeth? Could be the wrong type of belt for precision motion control. There are some good trapezoid-tooth belt types, but generally 3DP builders use GT2 (round teeth).
MXL:
GT2:
Anyone think this might be a reason and I should try replacing the closed loop belt and the two pulleys its on with GT2 and see how it goes?
I can test this out, just need to know which closed loop GT2 belt I should get to replace the MXL one.
Did you get the GT2 pulleys?
You can check the length of your existing MXL closed belt, see a mark on the belt or
measure it, or calculate it by the distance of two centers of pulleysx2+diameter /2 sq,x3.14 of pulley.
go to "Misumi USA" website to find the right GT2 belts, you can find the right GT2 pulleys
there.
GT2 pulley: 2mm pitch, for 6mm wide belt, 22 or 24 or 26 tooth, e.g. 22*2mm=44mm (pitch diameter) with 6mm hole for rear rod, you need 2pcs with hub and set screw, and 2 pcs without hub and Set screw.
GT2 belt: 2mm pitch, 6mm wide, for the length of closed belt you can check my lat post.
If you like use GT2 on X carriage, buy more GT2 pulleys and cut a closed GT2 belt, or
Buy a short opened belt.
All parts you can get from "Misumi USA" online.
This is what Solidoodle uses: http://sdp-si.com/ss/pdf/81002015.pdf
the terminology is still confusing to me. Ian said the length is 6.75". That's all I managed to grasp.
I checked it for you, if you like to use GT2 closed loop belt on Y drive(SD4) to
Run the Y carriages.
You can check the mark on your existing MXL closed belts,
It is Z334, it is groove code from SDP-SI catalog (your last post),
It shows the length of MXL closed loop belt on your Y drive is 678.69 mm or
26.72 inches, you can go to "Misumi USA" to find GT2 closed loop belt with
6 mm wide, 2mm pitch, and 675mm long(or other close #) etc.
you can buy 3 pcs, tow for Y drive, cut third one for X carriage belt.
I don't know why you talked about 6.75", maybe it is the short belt
On Y motor, but if you try to make a direct Y drive, you will eliminate
This short belt. the Y motor will drive the 6mm dia. Rear rod directly.
So you only need two GT2 closed belts for Y drive(left and right carriages),
And one open belt for X carriage.
http://us.misumi-ec.com/vona2/detail/11 … h=Gt2+belt
You check this link, there is 676mm long 2GT closed belt, you have make a
double check to ensure the length of belts is right, not too long or short.
I don't know why you talked about 6.75", maybe it is the short belt
On Y motor, but if you try to make a direct Y drive, you will eliminate
This short belt.
The point of this test I want to do is to find out if we can reduce backlash by replacing the trapezoidal belt on the Y motor, without modding the printer and making it direct drive.
For the side belts I already have tried fishing line vs MXL belt test and I see no difference.
The all info. I have provided are only for some people who like to use GT2 timing belts and pulleys, and a direct drive on Y drive, because these mods can reduce the backlash on
Y drive, many users have tested and verified it.
Okay. But any idea what GT2 belt to use instead of the MXL closed loop belt for the Y motor? That's what I asked.
You can try fishing line on the motor as well so you don't have to worry about tooth pattern at all. Google for a good knot to use and allow for a couple of wraps on each pulley. Getting the length right can be tough, but I did mine so it's possible. It's also a lot cheaper than getting all the pulleys you would need.
You can try fishing line on the motor as well so you don't have to worry about tooth pattern at all. Google for a good knot to use and allow for a couple of wraps on each pulley. Getting the length right can be tough, but I did mine so it's possible. It's also a lot cheaper than getting all the pulleys you would need.
+1 but what do we know 
SoliForum - 3D Printing Community → Solidoodle Discussion → Odd circle test results
Powered by PunBB, supported by Informer Technologies, Inc.
