Adding Encoders to Stepper Motors

I have been doing some research on accurate homing on a cnc machine. I am using some cheap limit/home switches, so no surprise that the repeatability is +/- .015". From what I have read it sounds like what I need to do is add encoders to the stepper motors, so that it homes to a index location, resulting in a really accurate home. This is for a CNC router, so I don't need mori seiki accuracy, but something that would give me close to +/- .001" would be nice. Is adding an encoder a good way to go or should I find a better limit/home switch? I am using nema23 425 oz/in motors. Any suggestions?

adding encoders to steppers , will not really do you much good apart from , your wallet being a bit lighter
i doubt that your switches are at fault , and the problem is more to do with machine rigidity and perhaps backlash .

i presume this is a wood router , so chances are your pushing it a bit hard
has this machine ballscrews or belts to drive , and is their any pully reduction to the stepper motors ?

dont forget that once you switch off a stepper driven system the steppers will move slightly due to which ever pole is nearest
the magnet at the time ..

once the machine is homed and tools are touched off to work , then see what the repeatability is , but dont switch the machne off .
and always check for movement in the same direction , so for example move say 1in , then setup to measure then move 2" and check , and then move another 2" as see if the distances are the same .

theirs a few things you can do , without spending money to get things better . but without knowing more about the build
i can only make general comments .

How fast are you probing your switches? You may find that you need to home at a slower rate to get an accurate measurement. (Your existing set up may work fine if you go slow enough.)

Hey guys,

My machine is quite rigid and I have no detectable back lash. I have 3/4 aluminum plates to support and gantry and the rest of the framing is aluminum extrusion. I have USA ground ballscrews on all axis. The machine itself is very accurate and repeatable once I have everything zeroed out, after a home. I am wanting to make up some fixturing and need a little better then .015". Well X and Y aren't as big as a deal because most of my parts can float by that much inside the stock, I really want a reliable tool table I don't have to keep setting up all the time. My homing sequence does 60 ipm to the switch, backs off, and then goes at 1 ipm. I am making guitars, and I have checked to see if I am loosing steps pre and post operation (1:30 hr run time) by building a super rigid dial indicator, not the typical contraption for hold one. Again, I am not looking for mori seiki accuracey, just something closer to +/- .001". I like the flag/opto idea, I don't need a full encoder, just essentially the index signal from it.

Thanks

then the obvious answer is to find where the backlash is , and get rid of it ,
you cant beat mechanics , get that right first , it will make life so much easier

in fact adding encoders when you know you have a backlash problem , will only make it worse

If you are needing to check for lost steps every hour and a half (or worse every 1.5 min not sure from above), then it sounds like you need to look at stepping issues, and that is a more likely source of your homing inaccuracy. A properly configured machine should be able to run for days without losing any steps.

If his machine is like mine. I can possibly move my gantry slightly while making a tool change resulting in mismatched toolpaths. Also I believe from the above statements that tjones is wanting to use a fixture that will be attached to the worktable. This means that he will want repeatable results every time he fires up the machine. Each time it is started the machine needs to home to the exact same location.

tjones, I agree with your previous statement and a flag-opto setup will be ideal to improve the accuracy of the homing procedure. Now that this discussion has taken place I am going to do the same to my router.

NitroCH3NO2, would be correct. The fixtures will be attached directly to the worktable. Also my machine has no detectable backlash. My first thought when I discovered the problem is loosing steps, which is why I set up the dial indicator in a rigid setup and mounted it to my table. However I am NOT loosing steps. So again, its down to replacing limit switches or adding a flag/opto sensor, at least thats the best idea so far.

Any ideas on where to get those flag opto sensors??