Choosing a spindle encoder

andypugh wrote:

jmelson wrote:

andypugh wrote:

I was suggesting quadrature on the motor and index on the spindle.
it isn't ideal, but should be OK up to a point.

Well, it might work for one-pass tapping, but depending on the amount of belt slip or the accuracy of the belt ratio to encoder scale, it could be a real problem for peck tapping or long threads. I really don't like this idea as Vee belts wear and the ration changes as the belts slip lower in the Vee groove.

If you are going to go to the trouble of putting a one-pulse/rev sensor on the spindle, you might as well just do it right.

Jon

"I was suggesting quadrature on the motor and index on the spindle.
it isn't ideal, but should be OK up to a point."

And this option is very doable. I can make an index on the handwheel at the tail of the spindle and mount the sensing circuit to the headstock, no problem. And I can attach the an encoder to the motor shaft with out too much hassle, so long as the motor won't cause noise issues. Shielding the encoder would likely help, though, right?

Sound possible? If that's the case, it may be much easier to do it this way, rather than getting quadrature with index encoder on the spindle. There's room for an index on the tail of the spindle at the collet closer, but the movement of the collet closer and limited space is making me think I'd probably have to get a large bore encoder wheel, which I'm tracking down and waiting on emails for, and machine an adapter for the spindle nose for that wheel. That option is kind of sketchy due to the movement of the spindle and it's proximity to the enclosure, coolant and flying parts. Making a shield is easy enough, but no guarantee that oil wont' permeate the enclosure and cause issues due to it's size at over 4" in diameter.

The index doesn't have to be a pulse of light, right? If it can be a magnetic type, that's an easy mount and wire situation, while I run a timing belt to the encoder and motor in the motor compartment.

I may be able to make a custom mount ecoder for the tail of the spindle at the collet closer, but I'm not 100% sure of the placement being possible, and if it is, it still leaves the wheel somewhat out in the open and easily damaged.

I guess I could even turn the handwheel down and attach a custom encoder wheel or even a timing belt pulley to it, but the wheel or pulley will likely be about 7 or 8 inches in diameter. That would prove an issue for the gear ratio between the encoder and that wheel It would be easier to machine the slits for a custom encoder, giving it much more space being so large in diameter, but what a can of worms.

Either way, thanks everyone for the input. This problem is not insurmountable, but I guess the options are limited.

Scot

I'd rather do it correctly than make-shift. But the options are really limited on the placement of the devices necessary to monitor the movement. I'm not in front of the machine today. But I'll be there soon. As soon as the parts I ordered are in, I'll be in there with mics and an imagination to see what I can do to fit either a wheel or a timing belt pulley to the tail of the spindle. If there is a half inch of non, axial-moving shaft at the collet closer region, I can machine an adapter for either a custom wheel or a timing pulley. But no room and we're at that index being put on the spindle with the quadrature at the motor.

I can be daring, too, and see about putting something on the spindle's nose, as sketchy as that is. It's an easier adaptation, but the likelyhood of damage is higher. I'd like to avoid an estop situation, if I can.

Worse case scenario, I have another machine which has the right configuration to do a proper placement of the encoder. And I was planning to retrofit that one or repair the fagor 8055 controller on it. It's just a pricier option right now.

No matter what, thank you for that input, again.

Scot

andypugh wrote:

I agree. I wasn't suggesting this is a best option, just a possible fallback.

As the belts wear the thread pitch would change, but by having the index on the spindle you should be able to single-point thread OK. Possibly tweaking the pitch to suit.
Let me point out, once more, that I don't consider this a good solution.

So, again, let me put this link up :
pico-systems.com/bridge_spindle.html

This is how I put three sensors inside the head of a Bridgeport mill to pick up the gear teeth on the bull gear.
These sensors are quite small, and sense the passing of a gear tooth, not the PRESENSE of the tooth, and so work really well over a wide range of speeds. You could also make it work on a timing belt sprocket.

I've been doing rigid tapping with this since 2009, it has worked very well.

Jon

If I can help it, I'd rather not crack open the head stock to get into it. I've read the bearings can be a hassle and these were solid when they turned last The lathe is a dovetail bed, if that makes a difference. But I don't think it is.

As for the photo, it's clear enough. The problem isn't the clarity, the heads are different and like I said before, I don't want to open it if I can help it. The collet closer was fitted with a pneumatic cylinder on it's 2nd retrofit. I'd use a f pneumatic collet closer if I had the budget for it, but I'm pretty sure it' cuts off access to the spinning parts of the drawbar, so I'm probably just going to stick with it.

I've also seen the elevator encoders you suggest from your past suggestions. they're the right ID, but the width is an issue. I did, however, find some companies which make that kind of thing in the states which may have loose wheels that I can adapt. I'm waiting on hearing back from them. And of course, there is always the potential to do a custom encoder wheel. But price matters. Those things are spendy for one-off's.

So, I'm kind of at square one again. I can also just abandon the idea of single point or precision motion for threading for a while, too. So long as the machine can turn parts and keep tolerances without threading, I can make it happen for a while.

Thanks again. Your ideas are great and very much noted. The machine will turn, it may not turn to the degree I hoped for a while, is all.

Scot

This option seems doable. But it's not in my skillset, yet. I don't really know how to wire it, but I can learn.

I have both a budget and time limit, which makes it even harder. The contract to make this machine go is likely going to be extended, but it's not in stone or even on paper yet. So time matters, too. I'm sure it looks like I bit off more than I can chew. But I'm pretty confident I can come up with a solution, even if it's to abandon the notion of controlling the spindle precisely for a bit.

Regardless to my path, thanks a billion. Because a million can't buy much any more.

Scot

Well, if there is a collet closer, there must be some sort of extension out the back of the spindle. What you need is to put some kind of slotted steel disc on the part that turns.
Some of these have 4 socket head screws which would be ideal for mounting a home-made disc. Or, you could get a large XL or MXL pulley and bore it to fit over the spindle. Put one pulley on the encoder and one over the spindle, and you've got your encoder system.

Jon