2 axis to 3 axis conversion

[drembedded]

Hello, I've been lurking in the shadows for a bit. I'm looking to upgrade my 2 axis Bridgeport Series II with a 3rd Z axis control. Currently my Z axis has a linear glass encoder.

I aquired a Z axis quill drive made by Elrod Machine, http://www.elrodmachine.com/CNC_Z_Axis_Quill_Drive.htm. Its an older model which doesn't include the integrated glass scale so I figure I'll attach a rotary encoder to the end of the ball screw ont eh quill drive. In fact this is exactly how the Anilam Z quill drive is built.

So heres my challenge. I need to replace the glass scale with the rotary encoder and the electrical output of the rotary encoder needs to be exactly the same as the glass sclaes as I don't want to have to change anything on the control, an Anilam Crusader M three axis control. The Anilam glass scale on the machine now is a "System D-10". The Z quill drive takes 5 revolution per inch.

This kinda sounds like one of the word problems back in elemetry school. I'd appreciate your thoughts on this.

Thanks,
Michael

[unterhaus]

I assume you mean you need the same resolution. Let's say your encoder is 20 micron per division, 5 micron in quadrature (I'm making this up, do you know the resolution?). The z feed is 5 revolutions per inch, or 25400 microns/5 rev = 5080 microns per rev. Looks bad for the home team. You need an encoder count every 20 microns, so CPR= 5080/20 = 254 line encoder to get exactly the same resolution. If it was a 1 micron scale, you would need a 1270 line encoder.

You can get 254 and 1270 line encoders from BEI http://www.beiied.com/incremental_encoders.html
That's not a standard resolution so it might go for over $500.

I wonder if you can change the resolution on the control? Tons of encoders are available on Ebay for less than $50.

[drembedded]

unterhaus, part of my problem is that I can't find any info on the Anilam D-10 glass scale. I'm hoping that someone out there knows the d-10's resolution. I guess I could use an o-scope to see what the glass scale puts out. I'm confident that my conversion isn't rocket science as a buddy of mine has an Anilam Crusader GXM with a rotary encoder on the z axis.

[drembedded]

unterhaus, I hooked up the oscope to the glass scale and was able to count the transitions on the A and B outputs when I move the Z axis. I did this several times but the results didn't seem to make sense. For instance, I'd move the Z axis appr 0.01102 inches and I'd get 28 transitions. A real head scratcher! Then for kicks and gigles I changed the CNC controls output to metric and... for 28 transitions the Z axis moved .28mm Ahhh that looks right.

So correct me if I'm wrong here. This glass scale looks to be a metric one. Would the resolution be 0.0001 or 0.0002 or 0.00005? or something else?

So that means for every inch of movement, the scales produce 2540 pulses (transitions on both A and B channels in quadrature) (this sounds very close to your original example). So if my ball screw is 5tpi (5 turns per inch) then I need a rotary encoder which will provide 2540/5=508 ppr(pulses per revolution)? I found one at http://www.renco.com/106022.htm.

Does this sound right? Thanks, Michael

[unterhaus]

That sounds like a 40 micron scale or 10 micron in quadrature, which would explain the designation D-10. I think PPR just means the number of pulses pre-quadrature, so actually you would need 127PPR.

To make sure, you are watching both channels and on each channel you get 7 low to high and 7 high to low transitions in the .28mm?

Eric

[drembedded]

Yes, I get 7 low to high and 7 high to low transitions on each channel when the Z axis moves 0.28mm. I also heard back from Anilam, they said that this is a 0.00005 resolution scale. Does that jive with what I am seeing? And if so, can you explain the relationship between resolution and the "40 micron scale"? -Michael

[HuFlungDung]

If the scale resolution is .00005", then 1/.00005 = 20000 pulses per inch. Divide by 5 for the number of turns your screw makes per inch, gives you 4000 pulse/rev. Divide by 4 for quadrature and you are left with a 1000 line encoder, which I believe would be a pretty common one.

[unterhaus]

If Anilam was correct you were off by about a factor of 8 in your measurements. Do you believe that?

1micron is .001 mm is .00003937 inch.

I suppose it's possible the machine is old enough to have inch encoders, but it has been a long long time since anyone made them. We have some bridgeports with very old dro's installed, and they have micron scales.

[drembedded]

HuFlungDung, but I'm not sure that jives with what I see when I connect up my oscope to the A and B lines on the encoder. For every transition(positive or negative) on either channel (A or B) the Z axis moves 0.01 milimeters.

[HuFlungDung]

Best ask Anilam whether their resolution figure was metric or inch. Maybe somebody just took a wild guess

Why don't you simply continue to use the glass scale anyways? Its actually more accurate than an encoder (so far as overcoming any backlash), and the servo loop doesn't care where it gets its feedback from.

[3t3d]

Originally Posted by HuFlungDung Best ask Anilam whether their resolution figure was metric or inch. Maybe somebody just took a wild guess Why don't you simply continue to use the glass scale anyways? Its actually more accurate than an encoder (so far as overcoming any backlash), and the servo loop doesn't care where it gets its feedback from.

Just another "vote" here, but I have never heard of a scale not in metric
units. Even when they list the scale in inch resolution, the basic scale is
actually in metric, and the inch resolution is "close" to that.

Are you really convinced that you can't mount the Z axis scale somehow?

Pete