[Al_The_Man]

You obviously have the skills to do the mechanical work, so fitting the encoders should not present too much trouble.
First remove the resolvers, then once you have the encoders in hand you can get a better idea of what it entails, as long as the physical size of the encoder body will enable any back cover to go on and it is just a question of working out the mounting method and come up with a suitable helical coupling.
If Barry has encoders, then you can get the dimensions etc from him.
Al.

[jshumer1]

How do I pick the right encoders for my motors?

[Al_The_Man]

Electonic Specification wise preferably you need 5v A,B,Z Differential output, 1000p/rev. This give you complementary 2 channel with marker pulse output ( 6 terminations + power).
The mechanical specs are what ever you can get in physically.
When the motor has a rear shaft, through-hole type (no shaft) can be used and is often cheaper and preferable.
If you go to the Renco site, you can learn more about encoders in general, Koyo make nice compact fairly low cost encoders also.
Al.

[jshumer1]

I took off one of the resolvers. Does the encoder mount up the same way? Shaft to flexible coupling, and run the leads to the connection plug…Also should I look into new cables? I think I have two due to not enough wires in the cables, plus the fact they are kind of hacked up. Any suggestions on where and what to get?

[jshumer1]

The resolver mounts the same way as this type encoder. Would something like this work?

[Al_The_Man]

That should, you want output type 7 by the look of it.
Al.

[jshumer1]

It's users adjustable, so any value is acceptable. But even a low cost 100 ppr encoder will be quadrupled by the system to 4 times to become 4000 pulses per rev which equates to .0002 accuracy =1/4000



This is what a tech from CamSoft had to say about what encoder to buy. What do you guys think about this? Sound right?

[Al_The_Man]

How do you mean 'User adjustable"?
The resolution depends on your Linear distance/pulses per motor rev.
What will this be?
Also you are using 4000 instead of 400.
Al.

[jshumer1]

That was word for word in the e-mail he sent me. As far as the numbers, I don’t know if he forgot a zero or added one. (One of the reasons I posted if here.) I send these guys at CamSoft an e-mail with a question, and they send me something back referring to the “frequently asked questions” section in the manual. I have the manual and I looked it up three times already!! GRRRR!!!! It takes three to four e-mails on each question before they take the time to answer it in some detail. I don’t know.

[Al_The_Man]

I don't know how they can give you resolution without knowing the motor rev/distance moved.
If you can post this, I can give you the resolution for a given encoder count.
I think what the CS guy meant was that whatever the count was you can scale it accordingly, but this does not determine the actual final resolution that you will get.
Al.

[skmetal7]

It's users adjustable, so any value is acceptable. But even a low cost 100 ppr encoder will be quadrupled by the system to 4 times to become 4000 pulses per rev which equates to .0002 accuracy =1/4000



This is what a tech from CamSoft had to say about what encoder to buy. What do you guys think about this? Sound right?

no it doesn't sound right...

Isn't 100ppr x 4 (quadrature) = 400? if so and you have a .200" lead leadscrew, then 400 ppr / .200"per rev = .0005" per pulse. Assuming no lead or other positioning errors/backlash.

That is exactly what the CamSoft guy said? I'd question his knowledge...

[jshumer1]

Can anyone tell me if the shield wire the encoders are wrapped in gets soldered to a pin on the connector plug, or do I leave it unattached at the plug and just connect it to ground by the input device?

[123CNC]

You should run the shield as much the length of the signal wire as reasonably possible. You should only ground it at one end, typically at the controller or drive end, whichever your signal is running to.

[jshumer1]

I’m soldering the cable of the encoder to the connection plug. Do I solder the shield wire to the connector plug also? (with the encoder and power wires.) Or cut it off there and just attach it to ground at the other end?

[jshumer1]

Can someone recommend some limit switches for me to use on my boss 5? I also bought a Centroid M39 retrofit package for my Shiz. I think I need to put new switches on that also. Barry, did you put new switches on your shiz when you retrofitted?

[Barry_ward]

J,

I did not replace them, only cleaned them and changed the wiring. My setup is likely different since you had a different controller.

The moving parts (XY) have taper blocks that intersect 3 switches at a fixed location. Upper sw is left, lower is right, and middle is either limit. I might have the upper/lower reversed because I'm not at the machine and its from memory.

B

[ratrace2]

The Galil 1832 only accept sinusoidal commutation for AC servos, for brushed or BLDC you need quadrature encoders, but it will accept both differential or single ended 5v or 12v.
Also dual encoder, if need be.
The problem with resolvers, although more rugged than encoders are much lower resolution capable, they were often geared up in the early systems that used them.
I use the Galil line quite extensively and find them an excellent product to work with.
Al.

Al,
you guys are doing some great work. Hey, could you help me set up a Galil 5-axis system? I'm thinking about using the Galil DMC-4080 (5-axis).....Then I need to decide on servo drivers and amps......If you could give me some input I would appreciate it....Thanks Al, and ALL.

[jshumer1]

I decided to stop screwing with The Boss 5 for a bit due to frustration, and I bought a Centroid retrofit for my Shiz. The Centroid M-39 was a more complete system, and therefore much easier to install. The total retrofit took me about two weeks from start to finish. I couldn't be happier with the Centroid control, and Doug Laursen at Machines in Motion. It only took two weeks for the crate to show up at my shop. Doug really knows his stuff. Any questions I had, he was able to answer on the spot. Not only answer the question, but fix the problem. If I had it all to do over again, I would buy two of the M-39 retrofits from Doug. I'm hoping I didn't buy a piece of @$%& from CamSoft, and can get my Bridgeport up and running once I pay a tech to come out. However, I don’t think the supply has enough power to run my motors. I'm not sure about that, but I think that’s why it throws an error light every time I try to tune the motors. I have checked the wiring and pot settings a billion times over with no success. Of course that still doesn't mean I’m right. But I have wasted more then enough time trying to tune the motors myself with no results.
I’m posting some pictures of my Shizuoka retrofit. I may still reverse the motors to shrink this monster and gain some space.

[jmelson]

A typical resolver such as in the pic has a primary pair and two secondary pair, non of them usually have DC on them.
The 1 pair on their own are probably the primary.
If you measure across any of the other two pair and get a low resistance it could confirm it, but according to your pic it definitely looks to be a resolver.
Al.

Just to let everyone know, I now have a resolver to quadrature converter board available, see
http://pico-systems.com/resolver.html
for more info. It runs on 5 V only, and produces single-ended or differential quadrature outputs with index pulse. The resolution is fixed at 4096 counts/rev.

Jon