Just stubled across this thread . . .interesting project. The arm is a work of art . . .  I want one just to put on my desk and look at!

I don't see where a momentary 2k lb static load on the spindle bearings will do any harm, but the mechanism to prevent that is pretty clever. A sizzor arrangement would also work and not require an additional actuator. Just thinking out loud . . .

The Tormach setup is slick. First I've seen it.  Tormach most certainly does know the drawbar force (or range) required . . it's a liability issue to advise on that, I would imagine. Certainly, the quality and accuracy of the spindle taper and the collet are going to have a significant effect on the force required, as will the presence and type of oil on the holders.

2,000lbs seems like a reasonable number to me for R8.

It will be fun to follow this thread . . . very  project.

"A sizzor arrangement would also work and not require an additional actuator" - I considered that, but couldn't see a way to avoid the extra actuator, as the drawbar must be aboe to move down when not changing tools.  If you see a way to do without, can you describe it?

Here's my question to Tormach tech support:

    I have a Bridgeport-clone knee mill for which I want to make an automatic tool changer, using Tormach toolholders in an R8 collet.  Can you advise me on the optimum drawbar tension?

And here's the response:

Funny you should ask that question.  We have been trying to determine the optimum draw bar tension for a R-8 collet and have not been able to put an accurate description together for it yet.  If I had something I would be happy to give it to you but I don’t.  If you find something we would appreciate you letting us know.  It seems to be a mystery of the industry that no one can put a real number on, but develop a feel for it working with it every day.

I have been completely unable to find *any* credible information on drawbar tension for R8.  I have found several references to drawbar torque (typically 25 ft-lbs, or thereabouts), but nothing about tension, other than guesses.  For real ATC systems (CAT/BT/etc), the numbers are in the 1000-3500# range, depending on which one but that's such a different animal, it's really not helpful.  The only issue is getting enough tension to keep the tool from slipping.  Since I rarely work with anything over 1/2", other than face mills and flycutters which are pretty low torque, I doubt I ever exceed perhaps 1/2-1HP, which is pretty low.  I think about 1000# would be adequate for that.  So, if I can make 2000, I'll have plenty of margin.  First thing I'll do is take some really heavy test cuts, maybe up to tool breakage, and see if I get any pull-out.

Regards,
Ray L.


Regards,
Ray L.

I tried doing this exact thing a couple years ago. The TTS, a 4 piston stage air cylinder, and bellevue washers.

IT DIDNT WORK!

Might be good for a little mill but I could not a good grip on the tool holders and I had all sorts of pull out issues. And this was with well over 1000 lbs of pull.

Simply put, the whole TTS system is flawed. There is no positive retention. Side load on these things just starts pulling them out. Ruined several parts before I gave up. Even with a drawbar I had slippage.

I ended up getting a replacement spindle for mine in 30 Taper. It was only $300 from YCI-Supermax. What a difference!

R8 sucks. You can make it better with some of the other quick change systems out there. There are a couple that use a retention knob system that converts R8 to something similar to CAT or BT tapered spindles. They have a spring loaded drawbar and an actuator on top. They would be well suited to a tool changer.

My first notion was that liability was the hindrance, obviously it is knowledge:

"I have found several references to drawbar torque (typically 25 ft-lbs, or thereabouts), but nothing about tension, other than guesses."

This is equivalent of saying "I found the specs on voltage and amperage, but nothing on wattage"  It's not like bolts were invented yesterday . . .  the entire reason for torquing fasteners is to impart a specific tension (stretch) in the shaft . . . i.e. tension can be directly calculated from torque and vice versa. You're an electron guy and would not be expected to know this perhaps, but a mechanical engineer designing a tool changing system? . . . please.

As I mentioned, R8 quality varies greatly and the fit and finish has a significant effect on any tapered part like this. Also the presence of and degree of interference angle between the spindle and the collet will have a decided effect on the holding power per tension lb. There are many variables to consider, true enough, but stating that only guesses are available? Was there a development cycle at all with the tool changer setup? Did somebody just wake up one day with an idea and start manufacturing a product? They did no testing? Learned nothing?   That whole statement from Tormach is incredible, in my opinion.

ON further review, I noticed that the Tormach setup had no positive drive facility. I don't think this is a viable strategy for a tool hoilder. How difficult would it be to put a dog drive type arrangement on the bottom of the spindle to engage the holders.  Here again, not invented yesterday . . it's  been a feature of gearboxes for at least 100 years.

As to a scissors arrangement, I do not understand what you mean by the draw bar having to move down when not changing tools, unless you are planing to continue to use the quill (which I think is also doable with some extensions) but I'll have a go at describing an example mechanism anyway. Imagine a large washer between the bottom of the spring pack and the spindle top, not attached to the spindle. The top scissor jaw simply pushes on the top of the draw bar. The lower jaw does the magic by lifting the washer which is free to float above the spindle top. All you would need it a positive stop (attached to the spindle or head) to limit the upward travel of the lower washer and when that stop was reached, continued motion of the scissor would force the collet out of the spindle. All one motion from a single actuator.  

SImpson, With a BP type arrangement and using the quill (most of us do) the drawbar has to be able to move up and down with the quill. The mechanism has to be able to disengage the springpack when running normally so it can float up and down as the quill is used.

THe main problem is using a collet holder to holder the toolholder. The R8 collet holder is marginal at BEST to hold a cutter under heavy loads.

Even changing over to a qc30/nmtb30  does not help much as there is NO provision in the spindle for a claw to grab a knob. What you may can do is find an old BT30 BP spindle that will fit. It has the provision for tool changer machanism. BUT the BT30 tool holders get HARD to find. BUT the saving grace is the cat holders can be made to work.

NOW IF someone were to come out with a BP spindle with cat30 at a reasonable PRICE they would probably become $$$$$$$$$$$ overnight.

Just a thought, (;-) TP

I tried doing this exact thing a couple years ago. The TTS, a 4 piston stage air cylinder, and bellevue washers.

IT DIDNT WORK!

Might be good for a little mill but I could not a good grip on the tool holders and I had all sorts of pull out issues. And this was with well over 1000 lbs of pull.

Simply put, the whole TTS system is flawed. There is no positive retention. Side load on these things just starts pulling them out. Ruined several parts before I gave up. Even with a drawbar I had slippage.

I ended up getting a replacement spindle for mine in 30 Taper. It was only $300 from YCI-Supermax. What a difference!

R8 sucks. You can make it better with some of the other quick change systems out there. There are a couple that use a retention knob system that converts R8 to something similar to CAT or BT tapered spindles. They have a spring loaded drawbar and an actuator on top. They would be well suited to a tool changer.

I can tell you for a fact 1000 pounds is nowhere near enough for R8.  You need more like twice that.  With proper tension, TTS works just fine - I've been using it almost exlusively for a couple of years and never once had a tool so much as change position on me.  This drawbar *is* working, with retention just slightly below what I get with hand-tightening.  Once I rework it to provide just a bit more tension, it will duplicate the grip of the manual drawbar.

Regards,
Ray L.

SImpson, With a BP type arrangement and using the quill (most of us do) the drawbar has to be able to move up and down with the quill. The mechanism has to be able to disengage the springpack when running normally so it can float up and down as the quill is used.

THe main problem is using a collet holder to holder the toolholder. The R8 collet holder is marginal at BEST to hold a cutter under heavy loads.

Even changing over to a qc30/nmtb30  does not help much as there is NO provision in the spindle for a claw to grab a knob. What you may can do is find an old BT30 BP spindle that will fit. It has the provision for tool changer machanism. BUT the BT30 tool holders get HARD to find. BUT the saving grace is the cat holders can be made to work.

NOW IF someone were to come out with a BP spindle with cat30 at a reasonable PRICE they would probably become $$$$$$$$$$$ overnight.

Just a thought, (;-) TP

My first notion was that liability was the hindrance, obviously it is knowledge:

"I have found several references to drawbar torque (typically 25 ft-lbs, or thereabouts), but nothing about tension, other than guesses."

This is equivalent of saying "I found the specs on voltage and amperage, but nothing on wattage"  It's not like bolts were invented yesterday . . .  the entire reason for torquing fasteners is to impart a specific tension (stretch) in the shaft . . . i.e. tension can be directly calculated from torque and vice versa. You're an electron guy and would not be expected to know this perhaps, but a mechanical engineer designing a tool changing system? . . . please.

As I mentioned, R8 quality varies greatly and the fit and finish has a significant effect on any tapered part like this. Also the presence of and degree of interference angle between the spindle and the collet will have a decided effect on the holding power per tension lb. There are many variables to consider, true enough, but stating that only guesses are available? Was there a development cycle at all with the tool changer setup? Did somebody just wake up one day with an idea and start manufacturing a product? They did no testing? Learned nothing?   That whole statement from Tormach is incredible, in my opinion.

ON further review, I noticed that the Tormach setup had no positive drive facility. I don't think this is a viable strategy for a tool hoilder. How difficult would it be to put a dog drive type arrangement on the bottom of the spindle to engage the holders.  Here again, not invented yesterday . . it's  been a feature of gearboxes for at least 100 years.

As to a scissors arrangement, I do not understand what you mean by the draw bar having to move down when not changing tools, unless you are planing to continue to use the quill (which I think is also doable with some extensions) but I'll have a go at describing an example mechanism anyway. Imagine a large washer between the bottom of the spring pack and the spindle top, not attached to the spindle. The top scissor jaw simply pushes on the top of the draw bar. The lower jaw does the magic by lifting the washer which is free to float above the spindle top. All you would need it a positive stop (attached to the spindle or head) to limit the upward travel of the lower washer and when that stop was reached, continued motion of the scissor would force the collet out of the spindle. All one motion from a single actuator. 

Simpson,

Since the quill is my Z axis, I kinda need it....  I am a degreed mechnical engineer, and you reallty can't get anything more than a rough estimate of tension from torque.  To do that you'd need to know the drawbar stretch, and the exact material used.  I have no way of knowing either.  Otherwise, an estimate can be off by +/-50%, depending on the quality of the threads, the fit, materals used, etc.  I did do the calculation, which is how I initialliy determined the tension needed to be in the 2500 pound range, and my experimentation seems to have proven that estimate reasonable correct.  Remember, this is a knee mill, not a 30HP VMC.  In normal use, I doubt I typically pull more than 1HP from the spindle, so Tormach *should* be perfectly capable of doing the job.  Certainly, I've never had one move on me to date, and I've used them a LOT.  I just have to get the power drawbar to duplicate the tension I set by manual tightening.

Regards,
Ray L.

Vmax, Someone mentioned an attachment for an X2 working in 'exactly' the same way as Himmy's device. The X2 has no quill, so I was a bit confused as to how exactly the mill was set up. I've seen conversions that use the quill and those that use the knee. I can imagine advantages to both methods.

While I do agree with your assessment of the options, I think the discussion is how to make the current configuration work, and I see some potential that it can, to a point. Although as an entrepreneur, I have to admit that your idea of producing a proper spindle is intriguing. A forum like this one would presumably be the right place to message such an idea, yes?

Himmy, someone said that you were a programmer, my apologies. Still, I'm going to respectfully disagree completely with your statement about tension. The tension calculations are common and have been around forever and obviously as you pointed out, they are based on the size and pitch of the thread. The ONLY thing that is estimated it the friction. The material has zero to do with the amount of tension produced by a given torque. Weaker materials with stretch more, that is the only difference. In critical applications, if the physical arrangement allows it (or can be designed to allow it) the actual bolt stretch is measured to eliminate the friction variable. In such cases, in my experience, the actual measured stretch is very close to the predicted stretch based on torque.

The scissor mechanism could be made to ride the quill by utilizing extensions to the spindle and drawbar and having the actualtor mounted on and 'riding' the extensions as the quill went thru its travel. I think what you have is a good design that needs only some testing and refining, the scissor idea was just an interesting discussion point. It came to mind as I have used the principle in several designs over the years and I am partial to it.

Vmax, Someone mentioned an attachment for an X2 working in 'exactly' the same way as Himmy's device. The X2 has no quill, so I was a bit confused as to how exactly the mill was set up. I've seen conversions that use the quill and those that use the knee. I can imagine advantages to both methods.

While I do agree with your assessment of the options, I think the discussion is how to make the current configuration work, and I see some potential that it can, to a point. Although as an entrepreneur, I have to admit that your idea of producing a proper spindle is intriguing. A forum like this one would presumably be the right place to message such an idea, yes?

Himmy, someone said that you were a programmer, my apologies. Still, I'm going to respectfully disagree completely with your statement about tension. The tension calculations are common and have been around forever and obviously as you pointed out, they are based on the size and pitch of the thread. The ONLY thing that is estimated it the friction. The material has zero to do with the amount of tension produced by a given torque. Weaker materials with stretch more, that is the only difference. In critical applications, if the physical arrangement allows it (or can be designed to allow it) the actual bolt stretch is measured to eliminate the friction variable. In such cases, in my experience, the actual measured stretch is very close to the predicted stretch based on torque.

The scissor mechanism could be made to ride the quill by utilizing extensions to the spindle and drawbar and having the actualtor mounted on and 'riding' the extensions as the quill went thru its travel. I think what you have is a good design that needs only some testing and refining, the scissor idea was just an interesting discussion point. It came to mind as I have used the principle in several designs over the years and I am partial to it.

I'm kinda hard to classify.  My degree is in Mechanical Engineering, but I've spent most of my career designing high-speed audio and video signal processing chips, the silicon kind, and writing software - real-time embedded systems, language interpreters and compilers, etc. 

Any calculation of tension based on torque has to contain a "frcition factor" term, which is tough to know with any accuracy without testing the specific fasteners in question.  I have yet to see any reference that states otherwise.  (http://findarticles.com/p/articles/mi_qa5357/is_200703/ai_n21284777/)  Doesn't really matter, since I have now more or less measured the required tension.  My comment re: knowing the material had to do with tensioning by stretch, for which you *do* have to know the material.  Obviously a rubber bolt will stretch far more under a given tension load than a high-tensile steel bolt. 

I chose not to make the whole assembly ride up and down with the quill, as that would require an additional 5" or so in drawbar length.  The drawbar extension above the head is already an issue I've had to design around.  (it wants to "whip" at speed).  I do have my knee CNC'd, but it is for tool length compensation only.  The knee would be too slow for use as a Z axis if I ever get into 3D work, as I plan to.

Now, if Terry made a QC spindle replacement....

Regards,
Ray L.

Vmax, Someone mentioned an attachment for an X2 working in 'exactly' the same way as Himmy's device. The X2 has no quill, so I was a bit confused as to how exactly the mill was set up. I've seen conversions that use the quill and those that use the knee. I can imagine advantages to both methods.

While I do agree with your assessment of the options, I think the discussion is how to make the current configuration work, and I see some potential that it can, to a point. Although as an entrepreneur, I have to admit that your idea of producing a proper spindle is intriguing. A forum like this one would presumably be the right place to message such an idea, yes?

Himmy, someone said that you were a programmer, my apologies. Still, I'm going to respectfully disagree completely with your statement about tension. The tension calculations are common and have been around forever and obviously as you pointed out, they are based on the size and pitch of the thread. The ONLY thing that is estimated it the friction. The material has zero to do with the amount of tension produced by a given torque. Weaker materials with stretch more, that is the only difference. In critical applications, if the physical arrangement allows it (or can be designed to allow it) the actual bolt stretch is measured to eliminate the friction variable. In such cases, in my experience, the actual measured stretch is very close to the predicted stretch based on torque.

The scissor mechanism could be made to ride the quill by utilizing extensions to the spindle and drawbar and having the actualtor mounted on and 'riding' the extensions as the quill went thru its travel. I think what you have is a good design that needs only some testing and refining, the scissor idea was just an interesting discussion point. It came to mind as I have used the principle in several designs over the years and I am partial to it.

Oh, there are people using this same mechanism on X2s, using only about 500 pounds of tension, and the same Tormach holders.  But, an X2 has a fractional HP spindle motor, and is not rigid enough to take heavy cuts, so the tool load is pretty modest, even compared to my little BP clone.

Regards,
Ray L.

(;-) Some very smart people have been working on this for about 50 years now and so far no apparent winner.

Making a cat type spindle would be a piece of cake, "BUT" making it work in the BP enviroment, NOT so easy. There are design interface problems working with the BP spindle design (;-).

BP did have a 30 series spindle with tool changer the BT30  BUT I don't believe it would work in a standard BP/clone machine

Running a spindle like an X2 (bed mill) would be  simple to do as well Just a mini type VMC you could probably use the iso 20 type holders(not cheap and hard to find)

(;-) TP