What's the best placement of EndStops in a CoreXY and are mechanical ones still used by you guys?

Quote
lkcl

Quote
realthor
Nono, i've kind of given up that idea -I understood about the NO shortcomings and the exposed contacts risks- ... was focusing on switchless homing by reducing the motor current so it hits the wall, skipping steps as it thinks it's still moving, until enough coverage has been traversed to be sure it's there at the end of the rods, then coming back and counting the steps on its way.

.... blerrgh! ingenious idea but one that has me kinda shuddering at the calibration you'd need to perform, and the testing - how many printers are you planning to get through during the testing phase? hey i know - how about, instead of changing the current (which you can do with RepRapFirmware anyway) you change the microstepping rate to something mad like 1/256? my main objection to the "bang it against the wall" idea is that it's not very accurate, you'd get juddering down the belt, you'd be relying on the elastic properties of the belt as well and it really might not like the large steps when subjected to tension. i know they survive wall-banging but it's not a nice sound. buuut if you're using silly-ridiculously-accurate microstepping, not only would the torque be reduced but the accuracy would go up, and, also, you'd be forced to run at a much slower speed.

for the z-axis... the thought of banging the z-bed into things... you'd better have a good bed design! or have just a z-probe.

Very interesting idea, the two can be combined you know. I was going to suggest the following scenario (has already been explained somewhere above but doesn't hurt):
1) read the hardcoded X_travel and Y_travel (say 250mm and 250mm) for a bed that is 200x200
2) assume Y is at the far-est position, namely touching the close Y clamps (it has to go all the way of 250mm before reaching the far-end Y clamps)
3) reduce motor current to a low value but enough to be able to get the carriage there in a decent time without skipping too many steps (+increase microstepping to 32 -steppers must be able to work with those values so if you have 1/256 steppers why not)
4) command a travel of 300mm so you are sure it touches the other side
5) increase motor current to 1/2 of normal (keep microstepping)
6) travel 250mm back to the start
7) increase current to 100% and reduce microstepping to whatever value suits your needs

This would be the homing procedure for the X&Y axis and only for the Z going downward.... Z on the upper side, with bed close to nozzle should be tackled with a probe => less vitamins, less cables, less clutter, less things to go wrong.

What do you say, do we create a pool to petition dc42 to add this to reprap firmware? Maybe more people will buy duets and focus on the Z probe.

Edited 1 time(s). Last edit at 05/16/2016 05:30AM by realthor.

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

Quote
realthor

Very interesting idea, the two can be combined you know. I was going to suggest the following scenario (has already been explained somewhere above but doesn't hurt):
1) read the hardcoded X_travel and Y_travel (say 250mm and 250mm) for a bed that is 200x200
2) assume Y is at the far-est position, namely touching the close Y clamps (it has to go all the way of 250mm before reaching the far-end Y clamps)
3) reduce motor current to a low value but enough to be able to get the carriage there in a decent time without skipping too many steps (+increase microstepping to 32 -steppers must be able to work with those values so if you have 1/256 steppers why not)
4) command a travel of 300mm so you are sure it touches the other side
5) increase motor current to 1/2 of normal (keep microstepping)
6) travel 250mm back to the start
7) increase current to 100% and reduce microstepping to whatever value suits your needs.

Increasing microstepping when driving against the physical endstop wont help. The motor won't just stop when it hits the endstop, it will keep jumping back by an amount corresponding to about 2 full steps whatever microstepping you use. For a typical build with 1.8deg motors and 80 steps/mm @ 16x microstepping, that's 0.4mm of movement.

For a CoreXY machine, I still think that a better solution is to use one IR Z probe to home all three axes. The Ormerod uses it to home both X and Z, but with the bed moving in the Y direction and the gantry in the X direction, there wasn't anywhere you could put a Y homing tab. On a CoreXY machine, I can't see any reason why you can't have a homing tab for Y as well as one for X.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42

Quote
realthor

Very interesting idea, the two can be combined you know. I was going to suggest the following scenario (has already been explained somewhere above but doesn't hurt):
1) read the hardcoded X_travel and Y_travel (say 250mm and 250mm) for a bed that is 200x200
2) assume Y is at the far-est position, namely touching the close Y clamps (it has to go all the way of 250mm before reaching the far-end Y clamps)
3) reduce motor current to a low value but enough to be able to get the carriage there in a decent time without skipping too many steps (+increase microstepping to 32 -steppers must be able to work with those values so if you have 1/256 steppers why not)
4) command a travel of 300mm so you are sure it touches the other side
5) increase motor current to 1/2 of normal (keep microstepping)
6) travel 250mm back to the start
7) increase current to 100% and reduce microstepping to whatever value suits your needs.

Increasing microstepping when driving against the physical endstop wont help. The motor won't just stop when it hits the endstop, it will keep jumping back by an amount corresponding to about 2 full steps whatever microstepping you use. For a typical build with 1.8deg motors and 80 steps/mm @ 16x microstepping, that's 0.4mm of movement.

For a CoreXY machine, I still think that a better solution is to use one IR Z probe to home all three axes. The Ormerod uses it to home both X and Z, but with the bed moving in the Y direction and the gantry in the X direction, there wasn't anywhere you could put a Y homing tab. On a CoreXY machine, I can't see any reason why you can't have a homing tab for Y as well as one for X.

Why make it simple, reliable, smart, cheap when it can be made utterly complex, without any benefits, prone to failure but lead to amusing discussions ? Yet not as funny as the "free energy" crowd

Quote
realthor
Very interesting idea, the two can be combined you know. I was going to suggest the following scenario (has already been explained somewhere above but doesn't hurt):
1) read the hardcoded X_travel and Y_travel (say 250mm and 250mm) for a bed that is 200x200
2) assume Y is at the far-est position, namely touching the close Y clamps (it has to go all the way of 250mm before reaching the far-end Y clamps)
3) reduce motor current to a low value but enough to be able to get the carriage there in a decent time without skipping too many steps (+increase microstepping to 32 -steppers must be able to work with those values so if you have 1/256 steppers why not)
4) command a travel of 300mm so you are sure it touches the other side
5) increase motor current to 1/2 of normal (keep microstepping)
6) travel 250mm back to the start
7) increase current to 100% and reduce microstepping to whatever value suits your needs

What happens when your smooth rods get dirty and require a bit more force than what you have your current set to?

1. You got to where you thought was x-min
2. You don't make it there, but your printer THINKS it is there.
3. You set current higher thinking that you are homed
4. You try to go to x + 200 and bang against something at full force.

Then what?????

-os3dp

@dc42: 1.8deg motors, gt2, 16teeth pulley => 1 full step =0.16mm ...2 full steps would be 0.3mm error (0.15mm for 0.9deg motors) ... what does this mean for X and Y? On the other hand, if the Z probe could be made to test all 3 axes having endstops is pointless be they of any kind or any method. Unless you want redundancy like the industrial machines have.

@os3dp: valid point, I have no experience of how much of a difference can be between a clean and a dirty rod, what means a dirty rod (when is it too dirty? should one let it get that dirty in the first place, as long as the more friction is there the more current it consumes and maybe other consequences too).

---

RepRap Lander concept on Concept Forge
RepRap Lander concept on RepRap Forums
My Things, mostly experimental stuff

I'm suddenly reminded of the old Apple II floppy drive-- Legend has it that Woz used a 40 track controller on a 35 track drive, so the infamous "Chug-Chug-Chug-Chug-Chug" [www.youtube.com] was the head banging against the drive stop 5 times.

... the drive in that link has some serious problems aside from the controller design, though.