[rootboy]

Hi guys,

I'm upgrading an old Takamatsu lathe with a Yasnac 2000G controller to CamSoft for a friend.

The drive is a Servopack CPCR-QR756S and in uses a resolver mounted on the motor for feedback. Standard stuff.

The CamSoft board comes with encoder feedback, so it won't read a resolver, but there are things that we could do:

1) Drill and tap a hole in the end of the motor shaft, mount an extension shaft to the hole, run the shaft outside of the resolver housing, and connect an encoder to the outside of the resolver housing. Easy-peasy - and this is how we fitted encoders to our old Demag motors so we could control them with Thor vector drives back at Saturn. Worked great.

2) The Servopack board had a header with what suspiciously looks like encoder signals on them. I would have to put my scope on them to find out for sure (which I am going to do regardless, curiosity and all that). But I'm not keen on modding the drive...

3) A resolver to encoder converter. I've done some research, and I could get one for about $500. This would be the simplest way of doing things.


But how important is the position of the spindle anyway? The original setup couldn't have known what the spindle angle was, it was controlled by an analog speed reference, and reported its speed back to the controller via an analog output.

I'll have to look, but I'm not even sure if the belt driving the spindle is toothed (posi-drive belt). So how would it keep in time in the first place? I can only see an issue with threading, otherwise controlling the drive's speed is the only concern that I can see.


Please be gentle, I'm not a machinist, I'm a controls guy with plenty of time on his hands.


Thanks!

John

[Al_The_Man]

Is it possible to remove the resolver and fit a through hole encoder?
It is too bad that Renco was taken over by Heidenhain, as Renco had a very wide range of through hole versions, they used to dump over-runs on ebay before the T.O. Cheap!
The range is a little limited now and probably Heidenhain prices.
Al.

[Jim Dawson]

Why not hang an encoder directly on the spindle, that takes the motor out of the equation. That's what I have on my lathe. For threading I just electronically gear the Z axis to the spindle encoder, and pick up the index pulse to time the engagement. Since CamSoft is most likely using a Galil controller, this becomes really easy.

Don't worry about spindle speed control, it doesn't have to be exact. Just sending the analog command voltage without any feedback is normally fine when using modern VFDs, especially in sensorless vector mode.

[BBMNet]

Hello.

I agree with rootboy as the change mostly depends on the control equipment you already have. If it accepts encoder signals then you´re very much in the right track. If your equipment does not have that option you might well have to risk the cost of conversion.

As rootboy mentioned there are resolver to encoder converters but to my experience they cost more than encoders at the same resolution. And at least the ones I know provide the absolute angular position of the shaft and not a series of pulses as you might be needing. Therefore you would have to add something that would generate pulses at the change of the position and according to the direction of the change. To my taste to complex and costly to be acceptable.

Again I would do the trick of adapting an encoder to the shaft of the motor or somewhere in the driven train to get the encoder signals.

Under some circumstances that may even requiere the replacement of the original motors. I had to take that decision when retrofitting one MECA chain quilter. Not a too difficult decision as the motors were already in need of major overhaul.

Hope this helps.

Regards.

[rootboy]

Well, I feel like an idiot...

When the owner got home from work today, I explained the situation, and he poked his head into the machine to look for a place to mount an encoder, and asked, "What's this?". I took a look, and lo and behold, there was already an encoder coupled to the spindle. Duh.

In my weak defense, there are no prints, and not all that many wire numbers on this machine, so it's usually something new every day.

BTW, if someone has prints on a Takamatsu / Yasnac 2000G, they would be greatly appreciated.

In any case, I'm golden. I'll find have to out what the wiring on it is, as well as the ratio between the spindle and the encoder, but that shouldn't be a big deal.


Thanks so much for your help, I really appreciate it!

John

[Jim Dawson]

John,

We've all been there

I actually had the prints for my lathe for all the good they did. They were kinda close, but were generic for that model but not for my particular serial number so not exact. I wound up physically tracing out every cable to identify the I/O, then created an accurate I/O map. Felt like there was about a million I/O points by the time I was done.

I lucked out and found an intact data tag on my spindle encoder, led me right to the pinouts with a quick Google search. Not bad for a 30 year old machine. Mine is 1:1 ratio, 1024 line, differential quadrature output.

Sounds like a fun project. Best of luck.

[rootboy]

John,

We've all been there

I actually had the prints for my lathe for all the good they did. They were kinda close, but were generic for that model but not for my particular serial number so not exact. I wound up physically tracing out every cable to identify the I/O, then created an accurate I/O map. Felt like there was about a million I/O points by the time I was done.

I lucked out and found an intact data tag on my spindle encoder, led me right to the pinouts with a quick Google search. Not bad for a 30 year old machine. Mine is 1:1 ratio, 1024 line, differential quadrature output.

Sounds like a fun project. Best of luck.

Thanks Jim

I've been re-creating the prints as I go, and it is already paying off with telling me how I am going to have to connect the CamSoft controller. That, and it will make life easier for the next guy.

But it is taking up a significant amount of time, but since I am hibernating at the moment it gives me something interesting to do. But I would like to get this up and running soon for Thurston to make money with it.

And yeah, it seems like a million wires, mostly gray, mostly unlabeled, and all of them a nuisance. And once past the bulkhead connector, the wires naturally change size and color. But of course, why wouldn't you would change them?

The one consistent up to this point is that the 110 VAC wires (or 100 VAC for the folks back in Japan ;> ) were larger than the control or 24 VDC wires outside of the main controller box. But I noticed that the wire that pulls in the Master Control relay (NRD1) is of the smaller (and not even remotely rated for 110 VAC) type of wire. Can you say "fusible link"? I'll be looking for a suitable replacement for this bad boy before I fire it up. I suspect that there was a repair done at some point, and this was a convenient replacement. But it's held up so far...

Finding the data for the encoders online would be nice. But if not, I can pop the encoder apart and trace the wires to find which ones power it up, and then hook up my scope and figure out the rest.


Thanks again!

John

[lukemart]

I agree with rootboy as the change mostly depends on the control equipment you already have. If it accepts encoder signals then you´re very much in the right track. If your equipment does not have that option you might well have to risk the cost of conversion.

[lukemart]

I agree with rootboy as the change mostly depends on the control equipment you already have. If it accepts encoder signals then you´re very much in the right track. If your equipment does not have that option you might well have to risk the cost of conversion.

Hope it was helpful! mobdro
tubemate