Corexy did the trick. Still need to work on balancing belt tensions to get rid of the last of the nipples on the circle perimeters when the stepper reverse. But so much better than the Hbot.
Here are some pictures of the last print as an hbot (the 6 cylinders). Compared with the first print as a corexy (one 20mm cylinder and one 40mm).
The corexy is still exhibiting backlash, just looks like it’s in a different direction now. I see improvement, but that may just be because you were playing with belt tensions.
@ThantiK I think that has a bit to do with running the bars over/under because the tip of the hot end whips. Plus I just found the setsrews I use to take the tension out from the linear bearing to the y gantry printed pieces were over torqued. I have a print going right now and it looks even better than the first print (and the latest print is at twice the speed). Still a WIP, but atleast it is going in the right direction.
I’m just saying, it doesn’t sound like Hbot => CoreXY was the fix. It sounds like simply you being able to play with it was the fix…
@ThantiK actually I could physically rack it by hand as an hbot. As corexy I can’t budge it in a racking motion. I think it plays a part… And in my opinion a fairly large one.
Good to know then! Might convert my H-bot.
What exactly is ‘racking’ ? I have 4 nemas coming from Massdrop for a printer I haven’t decided on … I like the look of yours and @Tim_Rastall 's igentis . I now have a Melzi board to play with as I blew out its thermistor input.
@William_Frick “racking” is the process in which the X axis comes out of alignment in regards to square from the build area.
@ThantiK Thank you ! I thought so but wasn’t 100% sure.
@William_Frick this explains it pretty well http://joshuavasquez.com/docs/jVasquez/Projects/coreXY.html
good job, I figured corexy would fix it for you. One thing to watch out for is don’t let the belts rub where they cross over. I get around that by running the pulleys on top of each other so the belts run at different heights and never come into contact and do not need to cross over.
@Wolfmanjm yup. That’s actually what the 6 cylinders pictured are. The last thing the hbot printed was the stand off parts to convert to corexy. There’s a lesson in that somewhere 
@Wolfmanjm how do you confirm equal tension between the two belts. Trial and error with calibration objects… Or is there a trick?
@Eclsnowman I ping them until they are the same tone
Excellent work, I’m glad the core xy got you some improvement. Did the latest fixes get rid of all the backlash?
It is getting better. But long term I think I need a more ridged mount of the E3D (ezstruder stock hotend mount just isn’t cutting it). And I want to convert to side-by-side x rails versus over under. With these changes I think I will be running smooth. nice thing is I can at least print the parts requiring changes now, where as the Hbot tolerances just weren’t up to snuff.
Horizontal X rails vs Vertical will help a lot.
@ThantiK , do you really think there’s enough slop in the linear bearings for that to make a noticeable difference? I was looking to do the same vertical rail arrangement. In theory, the shear forces caused by the twisting action of the core xy are minimized by spacing the bearings apart about the rotational center, in this case the hot end. When looking down on the printer, @Eclsnowman has only 2 constraint points as opposed to to the 3 constraint points in the layout you suggest (in Eric’s setup, two bushings vertically aligned act as one constraint, and in the conventional 4 bushing setup, only 3 bushings are actually constraining at a time… if it’s done correctly). So the benefit of the horizontal rails would be stretching the furthest distance between two of the bushings, that being 2 bushings caddy corner to each other.
So IMO, I’m thinking he can regain any lost rigidity by spreading the 2 bushings further apart, but again, that would be assuming that the rails and bearings are deflecting enough to cause noticeable errors.
Unless I’m totally missing something, which is likely
@Eric_Moy I think the bigger issue is that the belts are well balanced in xy but the extruder cart has bushings stacked constrained in xz. That allows any slop in y to reveal itself on the part when the steppers reverse.
Just my opinion watching it run.
