Slant bed lathe, tool changer, pneumatic 5c drawbar, DYNOMOTION control, PICS!

Quote:

---

Originally Posted by kn6za

Thanks Hezz,

Well I could put an encoder in both locations, but it should not be necessary, even with lots of backlash.

The issue with the motor and encoder separated by the backlash came down to servo tuning and the backlash interfering with it. When the servo drive would try to close the position loop it would not know that the motor shaft was moving until it hit the edge of the backlash and started to read movement on the gear box output shaft encoder. By this time there could be significant motor shaft speed achieved. This would create an oscillation that would be difficult to stabilize.

Now that the encoder is mounted directly on the motor shaft no such oscillation is possible. The backlash can now be dealt with in software easily. I am not using the index motion to cut anything, so as long as I can reasonably accurately position the turret, I don't care how much backlash there is. Once the turret is locked, and it will always be locked before cutting starts, the backlash is no longer an issue.

One of the beautiful things about the Hirth coupling is that the servo does not need to have perfect rotational accuracy on index. The thing can take a few deg. of misalignment and correct it in the clamp up movement. Once clamp has occurred, the index servo is disabled. At this point If the servo thinks it is out of position, because of small errors, it wont even try to correct the error until it is re-enabled at the next tool change request.

Hope that cleared up my thinking on this.

---

Thanks for the reply. I forgot about the Hirth coupling. And I can see your reasoning makes perfect sense.

Quote:

---

Originally Posted by kn6za

HaPpY BiRtHdAy!

---

Thanks Andrew, got my 1st present delivered today from Amazon: a nice box with 4 each of 2 sizes of deep throat Kant-Twist clamps. Goody-goody; I always get myself the perfect gift! A few days early but went ahead & ordered 'em on Cyber Monday.

Nice!

Kant-Twist, the perfect gift for the guy who has everything.

I would have had to buy them to make this machine if I didn't already have some!

I am very interested in the pneumatic collet closer. I have thought about trying to design one for my jet lathe after I get some other projects complete. I'd love to see someone make a diy version successfully.

Wow, chips everywhere again!

The last few days went very well. I got the machine work on the closer finish and the unit assembled. There was no part dimensions on the drawing I found, so I faked all the clearance values for the o-ring fits, and the bearing fits.

This had me a little worried that I might see more run out at the back end of the unit than I wanted, but that turned out not to be the case. Run out at the back is only about .002

I ended up leaving .004/ side clearance for all the o-ring fits, then cut the o-ring grooves so that the o-ring stood proud by a minimum of .006/side.

The bearings are a light press fit on the od and id. They slide of with about 50 lbs of force.

I changed the way the draw bar interfaced with the inner rear bearing. In the drawing the draw bar is held in the rear bearing cap with 2 set screws and a shoulder on the end of the draw bar. I ended up threading the draw bar into the rear bearing cap, so that I could use a 1-3/8 dia. x .120 wall 4130 tube for the draw bar.

The front bearing is pressed onto the lathe spindle adapter, and the adapter is a press fit onto the end of the spindle. I gave the spindle adapter .004/side clearance to the draw tube itself, as this interface moved forward and back and I did not think this area should be in contact at all.

In addition to the light press fits, the bearings also have snap rings retaining the bearings on the id and the od.

I have a feeling the press fit I used on the spindle to spindle adapter may be insufficient to keep the units together when opening the collet. The forward move to open the collet will try to push the closer off the end of the spindle and I think I messed up the fit by a .0001 or .0002.

It is a press fit, but I had intended this to be a tight press fit, the tightest fit on the unit, but I blew it and it came out looser than I wanted.

During initial testing the unit operates very well, no air leaks, and it releases the collet from the spindle taper just fine, but my fear is that the heat that is generated from the closer bearings during operation will make the fit between the closer and the spindle loose, and it will not push the collet out, but rather push the closer back off the spindle. If this turns out to be a problem, I will decide on either making a new spindle adapter or putting 4 set screws in the adapter to lock it to the spindle.

The stuff I have read online indicated these units get hot during high spindle speed operation, and the more clamp pressure you run, the more heat that will be developed. This makes total sense to me, as the entire clamp force is transferred through the double row bearings. Bad news for my spindle adapter, so I see myself making a new one, darn!

I am not in a position to test this much farther until I get the machine moving under its own power. At this point I need to start getting the wiring routed. Lots of wires.

Quote:

---

Originally Posted by dbsharp

I am very interested in the pneumatic collet closer. I have thought about trying to design one for my jet lathe after I get some other projects complete. I'd love to see someone make a diy version successfully.

---

dbsharp,

If I were you I would not hesitate to build one if you want one. Its not too hard. There are a few press fits and some o-rings to work with. I make my own o-rings, and always have, so I don't build to fit off the shelf o-rings.

You can see from the pics, my manual lathe is nothing special, its a second hand Chinese junker from ENCO I think, but i have done all the lathe work for this project on it.

Its not a beginner DIY project, but if you have a lathe and see a need for a closure, don't think you need to spend nearly $3000 on a commercial one!

Good luck!

The PDF file for the design of the pneumatic closer as I found it on the Royal site.

Its alive!

Attachment 342368

I was able to get some of the wiring of the front panel and the dynomotion boards done.


Attachment 342370

Attachment 342372

Attachment 342378

I have got the servos moving and tuned. The spindle motor is up and running as well.

For now I have a 1hp 2500rpm 90v dc motor running in servo mode on a 150v dc bus. This gives me the ability to reach 3000rpm with 1:1 belt drive ratio. I decided to run a servo driven main spindle for 2 reasons. The first being I had the drive electronics already and second, the servo spindle gives me the ability to rigid tap and do so to a blind depth. I think this can be done with a 3 phase motor and sensor vector, but that is out of my price range at this point.

2hp 3 phase is on the to do list in the future as Its the only way I will ever get any real torque and decent rpm without changing gears on the spindle drive. For now I will make due with this setup. I think for the parts this machine has been designed to run, the current setup will be more than adequate.

Aligning the spindle center line with carriage travel, and turret.

Attachment 342374

This took more time and effort than I thought. There is an adapter plate between the steel lathe bed and the cast iron spindle housing, and I ended up using this to bring the final alignment of the spindle height and turret center height into alignment. This required several dis-assembly and re-assembly cycles to get to what I consider close enough. This coupled with the use of shims allowed me to get the spindle center line and turret center line to agree with each other to within .002 tir over 12" of z axis travel. And I am able to cut a shaft diameter to within .0001 in 5" of z axis travel.

The adapter plate is within .0001 in thickness in all 4 corners, but it appeared that i had either one or both of the following problems to contend with. #1 the center line of the spindle was not parallel with the cast iron mounting surface, or #2 the surface I machined for the spindle to sit on was not co-planar with the bed ways. This was what necessitated the use of shims. It took .004" shim on the front end of the spindle casting to get it into alignment.

I guess I could have cut a taper into the adapter plate, but by the time I fully understood the problems I was fighting I was too low on the spindle height anyway. I came really close to dismantling the cross slide and taking a few thou off of it to make it all line up. I got lucky that .004" of shim got the spindle pointed parallel with the z axis ways!

I have to tidy the wiring up, its a mess, but its nice to see it moving under its own power finally!

Backlash compensation on the X axis is set at .0002 and on the Z axis its set at .00005. I think the .0002 in the X is coming from the angular contact thrust bearing setup not having enough preload, as I can put a test indicator on the end of the X axis ball screw and see the lost motion there. Half of me wants to remake the spacer and add some preload, the other half of me wants to let it go. :p

Tool changer and draw bar are working great so far.

Once I get the wiring tidied up, I will start cutting the bores for the ER16 collet chucks to locate into. I figure using the lathe to cut these bores will be the best method. I will try to get a video of some of that process.

I took a small detour the last few days.

I noticed the X-axis drifting when in an E-stop condition.

I had designed this thing with no weight compensation for the slant :eek:, so this is what I came up with.

Attachment 342934

The weight of the X-axis slide components is 55 lbs, @ 45 deg incline. 55*.707= ~40 lbs I need to compensate for.

So looking from behind the machine I installed a cable and gas spring system to compensate the weight. I needed 8" of axis travel, but no room for a 20" oal spring. So I ran the cable around the end of the spring. The cable is anchored to the z-axis slide on one end, and the X-axis slide on the other. This allowed me to use a 16" oal spring with 5" of travel @ 80 lbs of spring force. This arrangement allowed the spring to move 4" and apply 40lbs of force over the 8 inches of X-axis travel.

An easy fix to a design oversight.

Attachment 342936

Attachment 342938

Attachment 342940

The X-axis was falling .010 - .020 when power was removed from the axis motor, and now it only moves .0002

Now back to finish the wiring, unless I find another oversight:p

why not just add a brake to the x axis servo motor?

This completes the turret.

Attachment 343586

Some of the tooling to bore the holes for the collet chucks.

Attachment 343588

Attachment 343590


Rough drilling the holes.


https://youtu.be/XbTcNP_1HZY

I'm so close to having this thing ready to cut some delrin!

Quote:

---

Originally Posted by Runner4404spd

why not just add a brake to the x axis servo motor?

---

Good question. I thought about the brake idea, but did not want to buy one, and I have never made one before. I have built pneumatic counterbalances before for my milling machine, so I knew I could quickly implement a counterbalance solution.

Ok, this is it! machine is done!

Check out the bar puller!


https://youtu.be/Gz3LFrtLHvo


I dig the cur-chunk of the tool changer!



Could not have done this without the Dynomotion Kflop/Kanalog boards. If you are considering a retrofit or a custom built machine, I cant give a higher recommendation. If the idea of c++ programming frightens you away from this hardware, dont let it!

In this application there was a considerable amount of C++ programming needed to get the machine running, but there are a considerable amount of example C++ programs in the installation package that can show you exactly what is needed.

You can use mach3 as your human/machine control interface, but its not necessary. I used the KMotionCNC application, and for both my mill and this lathe find it to be a very good solution.

The g-code interpretation is not what the C++ is needed for when it comes to the dynomotion, its all the I/O attached to your machine. This gets handled by the C++ routines. If you want to control your machine from a touch screen, then considerably less C++ programming would be needed.

In addition to all the servo definition parameters, all the buttons and switches on the front panel, the tool changer, the analog output for the RPM meter, the JOG handle, all need to be handled with a C++ program.

In my machine the LOAD meter analog signal came from the spindle drive, so no Kanalog programming was needed to implement the LOAD meter.

Its been a fun and hopefully a profitable build, hope it helps you with your project in some way. Now I need to reload the bar and make some parts, that took me DAYS on my manual lathe!

Check you later!

Quote:

---

Originally Posted by kn6za

The chucks are sold as ER16 motor shaft spindles extensions, but hey are perfect for this application.

---

Hiya Andrew, Merry Christmas to you & yours!:)

When you get a sec, would you measure from the back edge of the ER16 nut to the rear edge of the shaft? That would be with a collet & tool installed, nut tight. I ordered some of them from China & am now worried they're too short to clamp up properly in my gang tool bar. Couldn't find the dimension in the ads.

Thanks!

Merry Christmas to you Milton!

This info is from memory, the collet chucks are pressed into the turret and I have no way of measuring the depth of the press fit. I drilled a hole larger than the thru hole in the collet chuck, so there is no land to take a depth reading from. The drilled hole does not go through the turret, so I cant measure from the back either.

The seller notes the outside locating diameter for the ones that I bought as 18mm, but this is not correct. The ones I got have a 20mm outside locating diameter, and this is .700" long. It is relieved down to 18.xx mm for the major thread DIA.

Depending on the collet and how far it gets collapsed to clamp, as they have a 1mm clamp range or so, its about .160" of gap between the back of the nut and the start of the 20mm locating DIA.

This gives you a total of .860" max depth into your holder, only .700" of that is a locating diameter.

The .700" dimension might be .750" at most, but I think, if I remember correctly its .700" to locate on. The ad is a bit deceptive in that regard, as the photo looks like its a straight 18mm shank.

Hope this info helps! I wish I could get a direct measurement for you, but I cant get them out of the press fit.

Wow, that was fast..thanks! It's going to be close but I think they'll work. I may have to extend the flat though. Do you think it's hardened or is it machinable? I modified some CTC ER32 chucks (Hong Kong) & they were good steel but definitely machinable. As cheap as they're selling these things you'd wouldn't think they were heat treated.

You won't have any trouble modifying the flat, they are hardened but not super hard. I would say 48 to 50 rc no higher. Carbide cuts it just fine.

Hope they work for you.

Andrew

Thanks Andrew.:)

Man oh man, I was so busy making grandkid Christmas gifts that I just now watched your videos. Your machine is beyond awesome! I'll bet you are quite the proud pappa! I envy your engineering & programming skills, you make it look like a piece o' cake.:cheers:

Thanks Milton!

Its been a fun build, and its nice not making these parts on my manual lathe anymore. I ran into some problems along the way, but now that it is running, its working really well.

Happy 2017 Andrew,

I got my ER16 spindle chucks today & they look great for the price paid. I think mine possibly came from a different vendor than yours. Do yours have a 19x1mm thread & 22mm collet nut wrench size instead of the more standard 22x1.5mm thread & 25mm wrench size? Not really a problem, just curious.