AllIn1DC Limit Switch Inputs

[DBennett]

I think I'm confusing the manuals for CNC11 and AllIn1DC.

I had to throw a 3 axis CNC controller together in a short amount of time and didn't really have time to sink my teeth into the documentation (which is not uncommon here).

I have an AllIn1DC controller and I am using Mach3 for commands.

Is it correct that the limit switches have to be Normally Closed?

Also, when I have Z and /Z encoder signal wired, I get counts on the other axes in addition to the one I'm controlling.

Any thoughts?

Best regards,

David

[AMDlloydsp]

Limit switches may be configured either way, but for safety's sake, it is best to arrange them as normally-closed, opening upon hitting the limit.

In that way, if a switch fails in its most common way (not to make good contact on closure) it fails "safe", in that it shows you hitting a limit when you have not actually.

When arranged as normally-open, closing upon limit, you risk missing a limit switch's actuation, and hitting the mechanical stops hard. Not good on balls screws or ball nuts.

Lloyd

[DBennett]

Thank you Lloyd.

They will be replaced in the future, right now we are integrating the controller with an existing tool (that isn't a mill, but has 3 axes of motion).

Any thought on the encoder issue?

When I disconnected signals Z and /Z the counting on the the other axes stopped.

Best regards,

David

[AMDlloydsp]

Are you using the original encoders, or ones supplied by Ajax?

If you get transitions on other lines when Z is making excursions, it has to be a ground fault. Just to be clear, Z- is NOT ground, although you probably know that.
However, your encoders might be single-ended, in which case, Z- might be (supposed to be) ground, and if you tie ground of the encoders to an input, you could see such stuff.

I'm probably preaching to the choir, since you seem to know something about this stuff, but the AIODC board (as I received mine) is intended to be used with differential encoders. If your originals are single-ended, there's that to deal with. I don't recall if the AIODC can be configured for single-ended ones or not, but if so, it would be in the manual.
(That manual is pretty good -- if densely packed. It's pretty much all in there, but sometimes you really have to dig it out.)

IF your encoders are single-ended AND the AIODC cannot be configured for single-ended encoder inputs, one 75172 or 75174 chip on a hunk of perf-board, and a 5V supply would fix the problem.

http://www.jameco.com/Jameco/Products/P ... 7907TI.pdf

Lloyd

[diycncscott]

Limit switches may be configured either way, but for safety's sake, it is best to arrange them as normally-closed, opening upon hitting the limit.

In that way, if a switch fails in its most common way (not to make good contact on closure) it fails "safe", in that it shows you hitting a limit when you have not actually.

When arranged as normally-open, closing upon limit, you risk missing a limit switch's actuation, and hitting the mechanical stops hard. Not good on balls screws or ball nuts.

Lloyd

Actually, on the ALLIN1DC, limit switches MUST be NC as the drive electrically disables motion in a direction when a limit switch input is open.

[diycncscott]

I think I'm confusing the manuals for CNC11 and AllIn1DC.

I had to throw a 3 axis CNC controller together in a short amount of time and didn't really have time to sink my teeth into the documentation (which is not uncommon here).

I have an AllIn1DC controller and I am using Mach3 for commands.

Is it correct that the limit switches have to be Normally Closed?

Also, when I have Z and /Z encoder signal wired, I get counts on the other axes in addition to the one I'm controlling.

Any thoughts?

Best regards,

David

Are you using shielded, twisted pair wiring for all your encoders?

[DBennett]

Thanks for the replies guys.

OK, I can get around the NC limit switch issue with some small signal relays (the switches are actually prox switches).

The encoders are differential, Quadrature (RS422) and although they are twisted pair, they are not shielded.

I scoped them and they look clean, but it makes sense that there could be noise coupling.

I will double check the wiring.

Thanks again folks.

David

[diycncscott]

Thanks for the replies guys.

OK, I can get around the NC limit switch issue with some small signal relays (the switches are actually prox switches).

The encoders are differential, Quadrature (RS422) and although they are twisted pair, they are not shielded.

I scoped them and they look clean, but it makes sense that there could be noise coupling.

I will double check the wiring.

Thanks again folks.

David

David,

I should mention that I'd be VERY surprised if you were getting cross talk on an encoder input that was actually wired to an encoder.

Are the other encoders actually connected?

[DBennett]

Scott,

I think I found the problem.

I had a temporary GND wire jumpered from my Resolver to Encoder converter board to the GND pin on the Allin1DC encoder GND pin. Thus, creating a GND loop.

I re-ran my Power supply grounding scheme so that all grounds converged to a single point (earth) and the encoder signals appear to work fine now.

Thanks for all of your help.

David

[AMDlloydsp]

Wow... there must've been a LOT of noise on that ground! Usually, push-pull inputs like that aren't disturbed much unless the common-mode excursions exceed the V-lo or V-hi voltages (or go so high the inputs clamp). I have to admit, though, that I don't know what receivers the AIODC uses, so I have no idea what their common-mode rejection is.

Good fix.

Lloyd