Omron Rotary encoders - stepper motors.

More
25 Aug 2017 08:21 #97983 by Moronicsmurf
So ive been fiddling with this 7i76E board and not really gotten anywhere useful yet.. But i figured it wouldnt be so easy.. Anyway.

I have box of these omron rotary encoders with 2000ppr - and im thinking wouldnt it be neat to connect these to my dual shaft stepper motors and get a position feedback.. then realized fast, no effing clue how to get about that..

www.ia.omron.com/product/item/2463/

Is it even possible to connect these for position feedback, thus closing the loop on the stepper motors?!
My goal is obviously to be able to run faster without loosing steps cutting with a plasma.

Please Log in or Create an account to join the conversation.

More
25 Aug 2017 11:52 #97989 by rodw
I think it is possible but:
For a plasma machine, there are no cutting forces to contend with so the risk of missing steps is unlikely. I'm able to move my X & Y axes at around 20 metres per minute with rack and pinions. I spent some time retuning the Z axis for maximum performance and its running at 6.6 metres per minute with a NEMA 23 on a direct drive 5mm ball screw. I'm using a 7i76e. I've tested the X & Y axes with a dial indicator over 20 metres and it was out by 0.03mm which is way under the accuracy required.

I think you'd have to add some extra encoder inputs but I think its unnecessary. You will have enough to do to get it cutting!

Please Log in or Create an account to join the conversation.

More
25 Aug 2017 13:18 #97993 by Moronicsmurf
Not sure if we were gonna cut the same things.. but i would rather be within the 0.01mm regardless. Especially if im pushing the speed of the machine in thin stuff and a bit of higher amps.

So i would still wanna add the rotary encoders to my setup. Anyone got an idea of how? its a pulse train generator so should be able to connect it to a digital input im guessing.

Please Log in or Create an account to join the conversation.

More
25 Aug 2017 13:21 #97994 by tecno
7i85 or 7i85S most likely, I am using the 7i85S to my mill

Please Log in or Create an account to join the conversation.

More
25 Aug 2017 13:40 #97997 by rodw

Not sure if we were gonna cut the same things.. but i would rather be within the 0.01mm regardless. .


Let me restate what I said earlier in a different way. Error was 0.03mm per 20 metres or 0.0015mm per metre or 0.00015% error.
Plasma is not a precision operation so your target of 0.01mm is beyond what the technology can deliver with or without encoders.
There are so many other sources of error that it is immaterial.

Please Log in or Create an account to join the conversation.

More
25 Aug 2017 14:08 #98002 by Moronicsmurf
Well if u add the precision, the machine can do more things than just plasma. And with the fine cut stuff i have for my plasma i can do 0.15mm kerf. but only in thin stuff. (max 45A)

Thanks Tecno.

Please Log in or Create an account to join the conversation.

More
27 Aug 2017 11:35 #98058 by raglanlittlejohn
I've used a Mesa set up, 5i25, 7i85S, 7i78, to get closed loop control with steppers for a lathe (with a lot of help from this forum).I also used a couple of Omron rotary encoders, the same resolution as yours. I'm still learning about all this, but I think the advantages are, 1. the position you see on Linux is actual position instead of commanded position.
2. any following error will be detected, and if outside the following error set in the INI file, stop machining.
3. a homing position can be set up using the index pulse from your encoders, (you also need a proximity switch for this)
this should be accurate enough to home the machine even if you have to re run a program after a following error/power failure etc.
The PID will correct for missedsteps etc, but this is not much help. This is because this set up will not slow the other axis down while the one with missed steps catches up. I guess this is what happens with proper servo closed loop setups.
I'm still working on this (not yet fitted to a lathe but got the electronics working). It's taken a while to get a working set up, it may be more trouble than it's worth for a plasma table.

Please Log in or Create an account to join the conversation.

More
27 Aug 2017 13:24 #98064 by rodw

it may be more trouble than it's worth for a plasma table.


That was kind of my thought when I posted above. I have substantially increased my Z axis velocity recently so I thought I better test it today so I stuck a dial indicator under the torch and moved the Z up and down 200 times using relative moves at maximum speed. The torch ended up exactly where it started so no missing steps after about 6.6 metres of travel and hundreds of stops and starts. I thought the cheap stepper I'm using were only good for about 750 rpm max but I've pushed them to 1320 RPM by aggressively tuning the step length settings. The 7i76e is pretty amazing gear....

Please Log in or Create an account to join the conversation.

More
28 Aug 2017 18:33 #98103 by Moronicsmurf
Well as i state above, it would be interesting for other application that plasma, the setup for a router and a plasma differs very little. And i can't plasma cut plastics, but i would be able to route it. And when plowing thru material with restrictions i found that closed loop has been useful on servo based machines and well in real cnc (i run siemens 810M on my milling machine). But yeah for plasma only i guess its not usefull for more than the learning exercise.

Please Log in or Create an account to join the conversation.

More
30 Aug 2017 22:50 #98238 by andypugh
You should be able to use the encoders, but they might not give you all that you hope for.

The problem is that steppers lose torque at higher step rates. If the system uses PID to recover lost steps then all it can do is step faster, and that will have exactly the wrong effect, typically.

It _might_ be possible to connect the pid error value to the adaptive feed pin so that step-rate reduces while still making up the lost steps, but I don't know if anyone has ever tried it.

Please Log in or Create an account to join the conversation.

Moderators: PCWjmelson
Time to create page: 0.154 seconds
Powered by Kunena Forum