3D printer bed correction

[shipto]

Happy Christmas all.
I wasn't sure if I should put this in the Design section but here goes. I have had a little trouble with Y axis skipping teeth on occasions (not often but its really annoying when it happens) so I have an idea to add a optical switch to the center of the bed and each time the bed passes it while also going toward the limit switch the firmware could check that the light was shining through a thin slot. If the light is there then all is fine and carry on but if not it could briefly interrupt the print to rehome the axis and thus avoid ruining a print. I was also thinking that if you could get accurate enough it might even be able to just update the position without even the need to rehome. I am thinking of suggesting this as a feature request for the marlin firmware but decided to get some thoughts of the possibility before I embarrass myself 
Thoughts please?

[efrench]

Wouldn't it be easier to fix the skipping tooth problem?

[shipto]

Wouldn't it be easier to fix the skipping tooth problem?

Yes but this to my mind wouldn't be a permanent fix it would be a failsafe. Its not too much fun if the problem ruins a really long print.

[WeldingRod]

I've run prints where I re-zeroed x and y each layer when I was having bad problems with the hot end catching on bumps.  Added a lot of.print time and junk on the zero side.
Yeah, you want to fix your lost strps problem.  Have you checked motor current?  Do you have top surface bumps catching your hot end? 

Sent from my SM-G892A using Tapatalk

[chipenter]

Ir it is skipped teeth it's belt adjustment or sticking rails  , but if it's dropped steps you need to lower the acceleration in advanced setup .

[shipto]

Sorry but I don't have any issues right at this moment with my printer, this isn't to fix an immediate problem. As I said its more of a failsafe in the event of an issue arising say 5 hours into a 6 hour print.

[WeldingRod]

It's easy to add an "each layer" script to recheck zero.  If you go that way, you should think about two prox sensors rather than the usual mechanical switch.

I did it, and then un-did it when I solved the blobs sticking up problem.

Sent from my SM-G892A using Tapatalk

[shipto]

It's easy to add an "each layer" script to recheck zero.  If you go that way, you should think about two prox sensors rather than the usual mechanical switch.

I did it, and then un-did it when I solved the blobs sticking up problem.

Sent from my SM-G892A using Tapatalk

Thats going to add loads of time to each print when it probably wont be needed 9 times out of 10.

[ddmckee54]

I'm not saying this can't be done, because if you try hard enough the software can get around a LOT of hardware limitations. However, as a controls type engineer doing this stuff for 30+ years I can see a lot of gotchyas that you'll need to work around.
1) How accurate do you want this thing to be?  You need to realize that most 3D printers are claiming that they can position to within limits that were considered state of the art for integrated circuit dimensions 40 years ago.
2) Your light source, you'll probably want to use a laser.  A normal light will spread out in more or less a cone shape once it leaves the source.
3) Are you planning on stopping the bed at what is thought to be the correct position, or at least traveling at the same speed?  The hardware will react at a fixed rate and if the bed is traveling at different speeds the "Correct" spot could occur at what seem to be different locations.  Been there, done that - got the Tee shirt.
You can try compensating for this, but I wouldn't recommend it.  I knew a guy once that was trying to keep a large weight suspended 40' below a traveling car from oscillating when the car slowed down and came to a stop.  He worked on that for months and never did get it to work correctly.  The calculations can be done, but NOT by an 8 bit Arduino in any reasonable time.
4)  It's gonna be a BEAR to align this, and keep it aligned.  Think about how much of a shaking the Y axis gets when doing a solid infill on a narrow part.

Again, I'm not saying this can't be done, because my gut feeling is that it can be done.  I'm just playing the Devil's Advocate and giving you some things to think about, so you can avoid some dead-ends.

Don

[shipto]

I'm not saying this can't be done, because if you try hard enough the software can get around a LOT of hardware limitations. However, as a controls type engineer doing this stuff for 30+ years I can see a lot of gotchyas that you'll need to work around.
1) How accurate do you want this thing to be?  You need to realize that most 3D printers are claiming that they can position to within limits that were considered state of the art for integrated circuit dimensions 40 years ago.
2) Your light source, you'll probably want to use a laser.  A normal light will spread out in more or less a cone shape once it leaves the source.
3) Are you planning on stopping the bed at what is thought to be the correct position, or at least traveling at the same speed?  The hardware will react at a fixed rate and if the bed is traveling at different speeds the "Correct" spot could occur at what seem to be different locations.  Been there, done that - got the Tee shirt.
You can try compensating for this, but I wouldn't recommend it.  I knew a guy once that was trying to keep a large weight suspended 40' below a traveling car from oscillating when the car slowed down and came to a stop.  He worked on that for months and never did get it to work correctly.  The calculations can be done, but NOT by an 8 bit Arduino in any reasonable time.
4)  It's gonna be a BEAR to align this, and keep it aligned.  Think about how much of a shaking the Y axis gets when doing a solid infill on a narrow part.

Again, I'm not saying this can't be done, because my gut feeling is that it can be done.  I'm just playing the Devil's Advocate and giving you some things to think about, so you can avoid some dead-ends.

Don

Thank you for the input, what I had in mind was something like the picture below. If for example you had 200mm bed travel every time it got to the point where it should be at 100mm it would check the state of this sensor, if sensor is in correct state then carry on but if not then home that axis. sounds quite simple in my head but as you point out there is much to think about.

[ddmckee54]

shipto:

I wouldn't try doing that every time that the printer got to that location.  Think about what that would make the processor do if you happened to have a short solid infill that crossed that location.  It would probably take a large amount  of computing time just for the processor to say "Yup, I'm here."  Might even drive it off into La-La land.

If you really want to try this most slicers give you the option of adding a g-code script between layers, I'd try it there.  That would catch any major boo-boos in position before they had a chance to screw up more than part of one layer.  Plus it would allow you to approach the position at a fixed speed, no trying to compensate for the effect of latency on different speeds.  I think latency is the term I'm looking for, it's the time it takes for a solid state device to switch from off to on or vice versa.  I don't know, you might find that a printer travels slow enough that latency can just be ignored.

One other recommendation, put your test point outside of where you normally print.  For instance if your print bed is 200x200mm I would use 180-190mm as the test location.  This practically guarantees that you will always be approaching the test location from the same direction, and you will also always using the same side of your slot to activate the sensor.  That way it really doesn't matter how wide your slot is, wouldn't even have to be a slot - could just be a flag with you looking for the sensor to go off.  When the sensor goes off, Marlin (or the g-code) loads a fixed value into the current position and Bob's your uncle.

Don

[shipto]

Well it seems this is probably more trouble than its worth. Oh well!