I'm in the process of adding a toolchanger to my BP clone mill.  First step is a power drawbar.  Rather than use the very common impact wrench approach, I wanted something that would leave the collet in place, and simply pop out the tool.  I will be using Tormach TTS toolholders.  So, taking a page from the design of many VMCs, I've created a design that tensions the drawbar using a stack of Belleville washers.  The drawbar is initially tensioned in the normal manner, by tightening with a wrench, to compress the Belleviille washers enough to create the desired drawbar tension (about 2000 pounds, as it turns out...).  To release the tool, all that's required is to compress the Bellevilles a bit further, then push the drawbar down to pop the collet free of the taper, and the tool simply falls out.  The trick is to provide a means of compressing the Bellevilles that does not place that 2000+ pound load on the spindle bearings.  To accomplish that, a small arm is rotated in under the Bellevilles, to take the downward load.

Since, on my first attempt at this, I was largely shooting in the dark as far as the required forces, and some of the other side-effects of the design, I've found I need to make a few minor modifications to the design to get it working properly.  But, the first attempt, shown in the below pictures, does, in fact, work pretty well.  The biggest problem is I under-estimated the drawbar tension required, and the air cylinder in the pictures is not quite up to the task.  A larger cylinder will be here in a day or two, which should do the job nicely.  The other "gotcha" is that the long drawbar extension sticking up above the head "whips" at speed - not good.  So, it will be surrounded by a cylindrical guide, rigidly attached to the spindle, that will rotate with the spindle and prevent any whipping.

The assembly is fastened to the head, by bolting an adaptor ring to the flange that retains the upper bearing outer race.  The upper bearing inner race retainer is replace with the drawbar guide tube.  The adaptor has two guide rods rigidly attached.  The rest of the assembly "floats" on these guide rods.  The guide rods have concentric springs on them such that the whole floating assembly is spring-loaded downward.  To do a toolchange, the quill is run fully up, which causes the top of the drawbar to contact the plunger in the center top of the drawbar assembly.  This compresses the springs, placing a downward force on the drawbar assembly.  This is the force that is used to pop the collet free, and will probably be on the order of 100 pounds - hardly enough to harm the spindle bearings.  A small air cylinder then flips an arm in under the flange that lsupports the bottom of the stack of Belleville washers on the drawbar.  Once that arm is in position, the large air cylinder is activated, which compresses the Bellevilles between that arm, and the plunger.  Since the whole assembly is floating on the guide rods, none of the Belleville force is transmitted to the spindle - only the initial downforce of the springs on the guide rods is transmitted to the spindle.  Once the Bellevilles are compressed sufficiently (0.030-0.050"), the collet pops free of the taper, and the tool falls out.  Once the new tool is installed, the air pressure is removed from both cylinders, and the collet is once again locked by the tension in the Bellevilles.

This is a totally bolt-on design, which requires absolutely no modifications whatsoever to the machine.  Drawbar tension can be adjusted over a very wide range by simply changing the number and type of Belleville washers used, and, if necessary, the strength of the air cylinder.  Collet-popping force is adjustable by adjusting the pre-load on the downforce springs.  Initial position is set by simply adjusting the length of the link on the air cylinder.  All very simple and flexible.  Tool changes become almost instantaneous - literally less than a second to engage or disengage the drawbar.

For the toolchanger, I am taking a very simple route, since I only need a few tools - I will have a single-row rack of tools mounted to the T-slot along the front face of the table, holding perhaps 8-10 tools, or perhaps 4-6 tool racks mounted at either end of the table.  The tool pick-and-place will be handled entirely by moving the machine, with no additional mechanism required.  This does cost a small amount of useable table travel, but since I have a 49" table with 34" of travel, I am more than happy to lose a few inches of workspace in exchange for the huge increase in productivity I'll get from the toolchanger.

Regards,
Ray L.

Looks good Ray.

Hood

looks really neat .

Have you thought of just using a 4" bore pancake cylinder supported over the end of the factory drawbar?

looks really neat .

Have you thought of just using a 4" bore pancake cylinder supported over the end of the factory drawbar?

Considered it, but it wouldn't get the job done.  At 100 PSI, a 4" cylinder would be only about 1200#.  I need at least twice that.  Plus, it would have too much travel.  I only need a total of about 0.100".

Regards,
Ray L.

Ray LOOKS good.   Here is something to consider. The piece that holds UP on the spring pack, the lower plate with the finger that slides under the pack. THis will be tricky to get the spindle in the exact position so the lever can slide under the pack AND still not have any play. ANY PLAY will cause the spindle bearings to load up as the pack is compressed. You might want to consider a push me pull me type of compressor arrangement that floats and as it is compressed both sides of the PACK are compressed at the same time. AND it puts zero load on the drawbar.

Another thought when building a drawbar is the machines always use a long 30deg angle tool holder that does NOT require any of the force of holding in the toolholder to be used to HOLD the bit as well. (;-)

Just a thought, (;-) TP

Ray LOOKS good.   Here is something to consider. The piece that holds UP on the spring pack, the lower plate with the finger that slides under the pack. THis will be tricky to get the spindle in the exact position so the lever can slide under the pack AND still not have any play. ANY PLAY will cause the spindle bearings to load up as the pack is compressed. You might want to consider a push me pull me type of compressor arrangement that floats and as it is compressed both sides of the PACK are compressed at the same time. AND it puts zero load on the drawbar.

Another thought when building a drawbar is the machines always use a long 30deg angle tool holder that does NOT require any of the force of holding in the toolholder to be used to HOLD the bit as well. (;-)

Just a thought, (;-) TP

Terry,

Not so.  Initial position is not at all critical.  It simply needs to move up enough that the drawbar contacts the plunger.  At that point, the latch plate is guaranteed to be clear, so it can slide into position easily, as the whole assembly floats.  There will be slack in there, by design.  The initial movement of the plunger will lift the assembly further until the latch plate contacts the underside of the flange under the Belleville stack.  It is impossible for any force to be transmitted to the spindle.  Further movement of the plunger will then compress the Bellevilles.  The only force that can be transmitted to the spindle is the initial downforce provided by the springs on the guide rods, which is strictly limited, and provided purely to pop the colet free of the taper (and to do so much more "gently" than whacking the drawbar with a mallet).

I would sure love to have a CAT30 spindle, but that ain't in the cards.  This design does appear quite capable of providing the same drawbar tension as hand-tightening, or using an impact driver, so should get the job done.  Hopefully, the new air cylinder and Bellevilles will show up today, and I can put it to the test....

Regards,
Ray L.

OK that makes more sense the picture did not tell the whole story as to if the lower plate floated or not. (;-)

GET-R-DONE (;-)

(;-) TP

I was under the impression the tormach  holderfs/r8 setup only needed a few hundred pounds of upward pressure on the draw bar to be secure. we are running a washer stack that generates 800lbs on the mill and unlocking with a .2" stroke bimba 4" pancake cylinder , cylinder was less than $100 and I havent had a tool slip in the past couple of years of operation . I hadnt thought about the load being applied to the bearings when unlocking , I would imagine they are probably holding up OK but I am now wondering about the severe load on the ballscrew assembly on my quill . I like your idea of the locking mechanism to handle that load .

Please keep us posted on your progress. I am always fascinated by tool changers and related components.

I was under the impression the tormach  holderfs/r8 setup only needed a few hundred pounds of upward pressure on the draw bar to be secure. we are running a washer stack that generates 800lbs on the mill and unlocking with a .2" stroke bimba 4" pancake cylinder , cylinder was less than $100 and I havent had a tool slip in the past couple of years of operation . I hadnt thought about the load being applied to the bearings when unlocking , I would imagine they are probably holding up OK but I am now wondering about the severe load on the ballscrew assembly on my quill . I like your idea of the locking mechanism to handle that load .

Please keep us posted on your progress. I am always fascinated by tool changers and related components.

What kind of machine are you running, and where did you get the 800 pound number?  What's interesting is NOBODY seems to know what appropriate drawbar tension is.  I asked Tormach tech support, and the response amounted to "Beats the hell out of us!  If you find out, please let us know!".  So, what I did was put a 1/2" tool in the collet, tightened as I usually do manually, then measured the torque required to make the tool slip in the collet - came out to about 30 foot pounds.  I then put the new drawbar in, and played with different stacks of bellevilles, and different amounts determined the drawbar tension was about 2000-2500 pounds, so that is what I'm aiming for.  The only place I've seen any numbers tossed around, it was on smaller machines, like X2s, which certainly can get by with a lot less than my knee mill.  The 2500 pounds is also in line with my initial esitmates, based on torquing the drawbar to 25 foot-pounds, which is what most of the impact-driver drawbars do.  I have seen figures for drawbar stretch (0.015"), but don't know how reliable those are.

I do know my current setup can apply about 1500 pounds, and that is NOT enough - I can make the tool slip pretty easily at that tension.

Yeah, 800# seems like an awful big static load to be putting on the bearings or the ballscrews....

Regards,
Ray L.

I am running a full size bridgeport clone , 3hp , 10x54 and havent had any problems . I mostly work in aluminum though and rarely do any drilling with that machine so mine may not be a good comparison

The 6-800lb number for r-8 collet came from an articlle in HSM about various collets somewhere in the past 15 years , I have seen 600 lbs mentioned in a thread somewhere in the past also. I would guess your 2500+ lb number is probably more accurate . My manual list a torque on the drawbar of 15-20 lbs which calculates to over 3000lb of clamping force  for a 7/16x20 bolt

look forward to updates when your new parts arrive