Home switch addition to rotary table

I have a 6" Parker rotary table I would like to use for light milling and indexing work with a NEMA 23 on my mill.

Unfortunately, it didn't come with the home switch option and I'm not real inclined to drop $500+ to Parker for the add-on.

So, I would like to fab up one.

The only info I could find about the stock Parker switch was that it was "magnetic". I'm really unclear as to how a magnetic field type switch would provide repeatability commensurate with the accuracy of the device (it's supposed to be repeatable to 12 arc-seconds!), but apparently that's what they use.

After thinking on it a bit, I'm leaning towards a photointeruptor type optical switch where a beam must be broken to trigger the switch. I'm starting to think the key to repeatability with rotary type motion is that when homing, the switch must always be approached from the same direction (i.e. CW vs CCW).

What do commercial units use?
Am I better off with a different type of setup?

Any info would be appreciated!

I adapted a 10" manual Troyke rotary table to be my 4th axis, and there was no switch. Since it has a scale with a vernier, I'd just pulse the table till the vernier read ZERO and home it that way.

Worked just fine. When the program contained a "home" or 0.0 command, it simply went to where I'd zero'ed it at start up.

Sounds like that would work well.
Unfortunately there is no vernier nor is there a good place to adapt one.

That thing looks heavy!

... It is heavy. But it gets the job done!

Without some physical reference, you've got no absolute to go to, and installing a switch would be a random location anyway... so....

Why not put some physical feature into your 4th axis chuck or faceplate that you can use as an absolute reference? It could be something as simple as a hole for a removable dowel pin or tooling ball, which you could use an edgefinder or indicator on.

There's a dozen ways to do it.

Quote:

---

Originally Posted by aubreyj

I'm really unclear as to how a magnetic field type switch would provide repeatability commensurate with the accuracy of the device (it's supposed to be repeatable to 12 arc-seconds!), but apparently that's what they use.

What do commercial units use?
Any info would be appreciated!

---

You do not mention what controller you have?
If the original intention by Parker is for systems that can read the encoder marker then the nature of the switch is not important, it is just used to arrive at the slow down point and from then on looks for the encoder marker.
Al.

Thanks all for your replies and info!

The table was designed for use with an AC servo & encoder. I have adapted it to a stepper with no encoder.

Maybe adding an encoder will be the best resolution -no pun intended <sic>- to the problem?

Seems the trouble with adding an encoder to the motor would be that it would encounter the home position 90 times for every revolution of the table. The software could only keep track of which home position was the true one if one never turned it off.

I suppose an encoder could be somehow adapted to the table itself -which is not fundamentally different than my photointerrupter idea (blocking/unblocking a light beam to trigger).

The difference is that the photointerrupter beam is a lot wider than the slits in the encoder. So, it would seem that approaching the beam from one direction would trigger the switch at a slightly different position than approaching it from the opposite position. If you only homed traveling in one direction, it should be repeatable.

Any reason why this wouldn't work?

Al:

I just re-read your post.

Having a switch to trigger a slowdown for the encoder marker makes a lot more sense if always approached from the same direction.

I don't know for sure, but I imagine Mach3 does not come stock with that kind of functionality built in (?), so I probably need to explore other methods.



If I can get away with it, I would prefer to have an automatic repeatable switch controlled by the software and not need to use an indicator every time.

Well you can pick up a switch from mouser.com, digikey, etc. You could drill and tap a small hole in the table (on the side) and insert a threaded insert to serve as a point. The use either an optical or mechanical [limit] switch that uses the insert as the home point. You need to machine a sturdy bracket to mount the switch to the non-rotating part of the rotary table. I woudl recommend a very sturdy bracket so that it does deflect if its bumped or bent causing an error. I assume you prefer good precision. If you use solid alum or steel you could machine an L shape where you mount the switch to the non rotating part of the table. As the table rotates the insert would swing by the switch and either bumping into the switches lever (mechanical) or blocking the optical source (optical). Although a mechanical switch may be a better option since chips and coolant might false trip an optical switch. If you're running coolant, be sure to choose a water tight switch.

I would imagine you could have something useable in a few hours (once you recieve switch). I doubt machining and tapping the table, to mount the bracket and for the insert would take very long. The only problem may be if the table top is hardened causing you grief to drill and tap the hole.

I would also recommend setting up the switch with a connector so you can remove the cable when you setup\clean up the table. Otherwise you risk damaging the cable. Mouser\digikey also sell connectors. Maybe an BNC Bayonet connector might be appropriate (old ethernet RG-58 cable connector), since they are fairly rugged and can easly be connected and require a twist to insert\remove (less chance of getting accidently disconnected).

Hope this helps!

There are many ways to home out an axis that is driven by a servo with encoder feedback. The important point to realize is that repeatability does not come from the home switch but rather the Z pulse on the encoder. As an example here is one homing sequence commonly used with servos.

1. Send the axis in the direction of the home switch at a fixed speed.
2. When the home switch is encountered stop the servo.
3. Reverse direction and move slowly setup to latch the Z pulse.
4. When the Z pulse is encountered zero the position counter.
5. Stop motion

Quote:

---

Originally Posted by aubreyj

Al:

I just re-read your post.

Having a switch to trigger a slowdown for the encoder marker makes a lot more sense if always approached from the same direction.

I don't know for sure, but I imagine Mach3 does not come stock with that kind of functionality built in (?), so I probably need to explore other methods.

---

There are a number of CNC systems you could use, for example LinuxCNC. I don't know the specifics about Mach so maybe asking on a Mach forum would get you there faster.

Quote:

---

If I can get away with it, I would prefer to have an automatic repeatable switch controlled by the software and not need to use an indicator every time.

---

That would most likely require encoder feedback to get good repeatability and reliability. You can get all sorts of precision switches, but getting reliable results over time is a pain.