SX3/X4 High Speed Spindle (Community Project ?)

[Ed from NY]

I was wondering if there would be interest on a user-side community project to enhance the SX3 and the X4 spindle and power transmission.

We would design, build, share results, and share enhancements to arrive at one "mature" design.

All project documentation would be made publicly available in common exchange formats including: plans, specs, part numbers, suppliers in different countries, CAD files, all g-code for the manufacturing of any parts, etc.

Here is a first shot at performance objectives in proposed order of priority:

1) Usable speeds from 50 to 30,000 RPM (different sets of pulleys/belts okay)
2) Eliminate vibration
3) Eliminate noise

Here is a first shot at community oriented objectives:

4) Have "one" final design (with versions for the SX3 and X4)
5) Maximize common elements on the SX3 and X4 changes
6) Keep costs of the final retrofit below $1000
7) Use globally available parts and/or materials. If that is not possible, then produce plans to manufacture those parts locally.

Here is a first shot at a list of secondary objectives. These may or may not be met based on whatever compromises we may need to make, but they might influence the design:

8) Spindle lock
9) Power draw bar
10) Maintain the ability to rotate the SX3 head

-----------------

There are probably a number of implications like the support mechanism for the motor and the seating of the pulleys on both sides would need to be redesigned. The type of pulley/belt used would likely change You probably would need at least 3 sets of pulleys and belts, perhaps in a tree arrangement. The draw bar would need to be modified to at least be self centering. Speeds would probably not be reflected in the machine display. Some additional mechanisms to align and keep aligned the spindle, pulleys, motor and all the assemblies would be critical. You might want a good shield between the spindle/tool/workpiece and your body parts or valuables.

There are a number of complicated things. This is a big project that would greatly benefit from the knowledge and experience of a good size group of people like this one

What do you guys think?

[Ed from NY]

Some reading http://www.mmsonline.com/articles/hi...struction.aspx which was linked from this thread High Speed Spindle for SIEG X2 Mill - getting technical

[Ed from NY]

Well not sure if there is interest, but...

Variable speed pulley design

If I understood it right this would get us from 100 to 10,000 RPM. Based on having 2 of these VPS-50 from Torque Transmission (2.8 ratio each) and a motor speed of 400 to 3.600 RPM. Here is a link to the product http://www.torquetrans.com/pulleys/v...leys/index.htm . These are relatively small at 5" diameter.

Unsure of the implications of vibration and noise with this design.

Multiple pulley design

Going for a V-belt design the smallest pulleys depend on the number of grooves. If we pick 4 grooves the smallest pulley I found was 2.6" in diameter and 1.9" thick which puts its companion for that same 2.8 ratio at 8" diameter (standard sizes). TB Woods and Baldor Electric have similar sizes in this range.

Don't have weight data but the multiple pulleys design would probably be many times heavier than the CVT-like design as you probably would need 3 sets (2.8:1, 1:1, 1:2.8).

However, depending on how they are mounted we may be able not to have all 3 sets on the machine and slide them in and out... as needed. Assuming that does not compromise the alignment.


.

TB Woods and Baldor Electric also seem to have variable speed pulley designs but I could not figure them out from the web. Still I would think they are similar to Torque Transmission's given the physical constraints.

BTW, http://productsearch.machinedesign.com/ is a great resource to take a glimpse at what is out there.

.

[Ed from NY]

Going to higher RPMs would probably need a second set of pulleys driven off the first one and that would get us very close to the design objective at 50-28.000 RPM. But it adds a lot of complexity, cost and weight.

It also adds more demands to other parts with higher requirements for alignment, spindle bearings, etc.

[Edit]
It could be a somewhat cool design if used the variable pulleys as both sets could move in a synchronized manner. Not sure if they would be rated to the higher RPM's though.

:unsure:

[Ed from NY]

Bummer. The distance between axles is barely more than 4 inches.

All parts above would require around 5-6.

[Edit]

Oops, did not mean this to be a new post, was supposed to be an edit

[Edit 2]

Gates seems to have much smaller diameter pulleys with 5 grooves, which kind of indicates that belts can turn that small. See http://www.gates.com/downloads/downl...older=brochure a 4.92-1.32 combo could be interesting and it fits the space we have

[Edit 3]

Gates has a few design guides which can help with product selection. As just about everything, one you start looking into it it gets complicated. Before I get too deep on the wrong track I'll wait to see what people with more experience suggest

[Edit 4]

Adding some interesting attachements
.

[john_t_h]

I'm certainly interested to see where this goes. I'm not sure if I can offer anything much more than encouragement.

[Ed from NY]

Well I horsed around with their design guide and it seems that the 5M PolyFlex is a suitable choice for the 100-10,000 RPM design with 3 replaceable sets of pulleys (which people in the know seem to call "sheaves").

The "sheaves" I picked are 1.67" (0.72 lbs $60) and 4.92" (3.8 lbs $114) running a 5M PolyFlex belt 19.7" long with 5 strands ($48). These would work for 1:2.9 and 2.9:1 settings. An additional pair of 3.15" (1.5 lbs $99 x 2) sheaves and a suitable second belt ($45) would take care of the 1:1 ratio.

I do not have a torque/power curve for the motor, but if you look at the numbers it should be within the specs as a doubt that we have anywhere close to 1HP at the low RPMs. If this goes someplace we can give the guys at Gates a call and double check.

I did not calculate the bearing loads, but the requested static tension is 40 pounds.

We would need to redo the calculations for the 50-30,000 RPM configuration but given space considerations it is likely that for those RPMs we would need to have to special order sheaves and belts as it would greatly exceed the maximum recomendation of 6500 FPM for the stock items. As it is, with 10,000 RPM and the selected sheaves, it already puts us at 4500 FPM. We could check, but that would probably blow the budget... just the standard sheaves and belts above put us at around $500 and you still need bushings and stuff.

[Edit]

It may be possible to squeeze in a larger 6.3" (5.6 lbs $154) and 1.4" (0.46 lbs $57) sheave pair and that would get us to 15000 RPM. But that would be getting closer to the limits of those parts. Also at high speeds the sheaves become flywheels and there are safety considerations there too. For example, if the cutter were to jam it may shatter more readily because of the extra energy stored -- given the weight distribution the 6.3" sheave probably stores 2-3 times the energy the 4.92" does --baby oh baby... make sure to video that 6-pounder wheel flying off at 3500 RPM

Tempting though

[Edit 2]

An "elephant man" design where 2 sets of pulleys are always in use may get us to 50-30,000 RPM for nearly the same cost as the 100-10,000 RPM design. We would use two sets of 1.67" and 4.92" in the following combinations:

1:2.9 + 1:2.9 => 1:8.4
1:2.9 + 2.9:1 => 1:1
2.9:1 + 2.9:1 => 8.4:1

One draw back is that there would probably be torque gaps, for example at 3500 RPM in the spindle we would be using 1:1, but at 4000-6000 RPM in the spindle the motor would be doing only 400-700 RPM. Then at 7000-8000 RPM in the spindle it may start getting more juice again till 30000. We really need a torque curve for that motor

However, one advantage of the "elephant man" is that we can design the assembly so that tensioning the belts is done only once and works for all gear combinations... till of course we need to re-tension due to wear/replacement.

[asuratman]

Hello,
It is interesting. Will you change motor and spindle? Are you going from 4000 rpm to 50 to 30,000 RPM with the pulley design. Thanks.

[Ed from NY]

Hello,
It is interesting. Will you change motor and spindle? Are you going from 4000 rpm to 50 to 30,000 RPM with the pulley design. Thanks.

Same motor. Possibly only change spindle bearings. That is the basic idea, not sure, I do not understand much of this.

Right now considering plain poly v-belt sheaves. The CVT type sheaves would provide the same basic ratios, but I'm not sure they would hold as well and what would be the vibration and noise charactristics. If we find a design guide from one of the manufactures we can take a look.

This picture might clarify things a bit.

[Edit]

Option 3, depending of the assembly mechanism could also allow for the same connections in option 1 that give you 1,150-11,000 and 140-1,200 eliminating any torque gaps. Again a torque curve for the motor would help in making the decision of when sheave changes are needed. That would need to be commented in the g-code somehow because the CAM only knows of 400-3500 RPM, all reductions and so forth need to be accounted for manually.

[Edit 2]

There was a mistake in the drawing which has been corrected. The RPM for option #2 is 15300, not 12900 as it was in the previous version of the drawing.

[Ed from NY]

Called Gates.

For option 3:

At 12,000 FPM they would be able to fashion the sheaves from 1144 steel (http://www.matweb.com/search/DataShe...db26ac24a2ba0b) or 2024-T3 aluminium (http://www.matweb.com/search/DataShe...12964a1821fb61). I would feel a little better with steel. If they are bored to size they probably would not need to be balanced.

For option 3 we would probably need only one pair to be made to order, the pair driving the spindle. The pair driven by the motor could be the regular stock item.

I did not get a quote. My guess is that the special order is probably around 2.5x as expensive as production items. This is based on 5 hours of design, setup, and machine time and a run of 5 big and 5 small sheaves. This would get the sheave/belt cost for option 3 at around $700 leaving $300 for spindle parts and other materials... we kinda blowing the budget already and we only looked at one element.

[Edit]

Another concern with option 3 is that it will add 9 lbs of pulleys and probably 25 lbs total to the head with half of that being off axis at about 5 inches to the right. We could try aluminum sides, but in all likelihood we will need top and bottom plates on the sheave assembly to hold the axles true with bearings above and below which would also diminish the load on the spindle bearings. We can mill pockets for weight reduction. Depending on how the pockets are milled we may get the center of gravity a little more towards the column at the cost of a heavier assembly.

The quill might need to be gone, when we look at the spindle more closely we will see if we can still have that kind of loose fitting we have today around the top sheave. That may or may not mean that we need a brand new spindle. Not sure

Still, one advantage of option 3 is that it would simplify the motor assembly, making it static and allowing the spindle and the motor shaft to run exactly parallel.

[Ed from NY]

It seems that for speeds over 10,000 RPM you need balanced tooling.

Waiting on a call back from Lyndex to confirm the limits on their regular R8 collets. But it seems that at 10K you start to hit cost barriers in many fronts.

Found some comments that R8 seems to be a design more suited to speeds below 8k.

Finding that high speed holders are mostly CAT40 or CAT50 and BT or HKS. There are probably companies that make balanced R8 holders, but that is probably telling us something.

And... some of the cooler tooling is not available in R8

http://www.lyndexnikken.com/prod_cus...4-115&status=1
http://www.lyndexnikken.com/flash/te...video.asp?id=4

...and have not looked at spindle bearings yet.

[ataxy]

well allready 8000 would be a major improvement over 3300