View Full Version : YAHSS: Yet Another High Speed Spindle...

04-01-2009, 05:27 AM
This is for my plotter to engraver conversion. The main part, the motor, is a very nice little military surplus motor from a mechanical tuner mechanism for an old vacuum tube radio. It is rated at 28 volts which is an excellent match to the power supply of the stepper drivers. At 28 volts driving the spindle it is turning at around 25 to 30,000 rpm. It puts out plenty of torque while only drawing as little as 100 milliamps at 28vdc.

Parts view


Assembled. Note that I did a little more machining to shorten the nose a bit since I took the parts view.


It also makes a very nice battery powered hand held engraver/dremel style tool.

04-01-2009, 07:44 AM
I always marvel at what nice work you do Evan.

What kind of radio was it? Very fast tuning, no doubt!
28V at 100 mils, that isn't much power input but for engraving I guess it wouldn't take much due to the very light cuts.

04-01-2009, 08:12 AM
The tuner mechanism was an excellent source of nice little gears. It had about a 2000 to 1 compound gear train to reduce the speed.

100 ma is the no load current, about 3 watts. Under load it slows a bit but the current jumps to 600 ma, about 16 watts which is plenty for engraving and even micro milling or routing. I will probably have to fit another heat sink to the upper part of the motor as it is a sealed unit. The nice thing is that even on 12 volts it turns about 10,000 rpm and does not overheat at all.

BTW, I judge rpm by the pitch of the sound it makes. I have perfect relative pitch and once I calibrate my hearing from a pitch pipe or keyboard I can tell what pitch a sound is to within a fraction of a percent.

04-01-2009, 09:01 AM
Hi Evan
Nice project so far.
I have a plotter that I was thinking of adding a high speed spindle to but the z mecahnism on the plotter a roland 1100 is a bit on the heavy handed style more like a Karate chop when it so called lowers the pen.
When we used to use the plotters in an open office system there were no end of complaints about the noise the plotter made as it clunked its way through the plot.
How is your plotter on the z axis.


04-01-2009, 09:42 AM
I won't be using the original mechanism. I am going to build a new bridge with additional linear bearings grafted on from the remains of another identical plotter. That will give it a lot more rigidity and load capability. I will probably use a small gear motor to drive the Z although I may opt for a complete but short throw stepper controlled Z axis.

A.K. Boomer
04-01-2009, 09:50 AM
I have not been following the High speed spindle threads at all so if this is already covered please excuse,
After seeing how small Evans is - is it not possible (or about time) to integrate one into an R-8 design? (inside) How handy would that be, just chuck it up plug it in and go (talk about a heat sink -- forget the fins - try using an entire spindle thats not in operating mode) --- the leads could be ran out the bottom and off to one side no? (unplug your mills real motor so you dont make catastrophic "force of habit" mistake and turn on the real switch) You could do all this and lose zero Z axis space.

it also makes me think of the mechanics approach --- stick the quill down and lock it an inch or enough to get a cinch up with an outer boss that contains a planetary gear system, the mills R-8 drive is incorporated to drive either the ring or planets and the final drive (the sun gear) is the one that the tool attaches too --- its bearings could be directly incorporated into the R-8's driven piece to keep the rigidity as much as possible --- planetary systems can be designed to "rap out" pretty good in the RPM's --- This system would take a little more space than that of a chuck up R/8 electric but there would be no other power source other than that of the mill to run it...

Thoughts/ rebuttals / confusion? Talk amoungst yerselves...

I guess im growing weary of seeing all kinds of braces and chunks of metal hanging off of peoples milling heads to mount this stuff.

04-01-2009, 11:18 AM
MEW had an article 3 years or so ago on making a planetary speedup. IIRC the original reversed rotation so you needed to have a reverse on your mill to get normal rotation, but someone else showed a version he made with an additional gear to get forward input/forwar rotation. Cute little unit running in oil.

04-01-2009, 11:49 AM
This spindle has an OD of just under one inch so it will probably fit an R8 spindle just fine. I will have to try it and see. I also have an idea for making one much smaller. :D

04-01-2009, 08:04 PM
Can you offer a little more info about your spindles? I am accumulating parts to make one myself as I think you are aware. There are several things I know nothing about when it comes to this. Most importantly, how do your bearings install, and how are they preloaded? The idea I was toying with is having a lower bearing pressed into the cartridge, a shoulder on the spindle to register the spindle to the lower bearing, a spacer between the lower and upper bearing the od of the inner race, and then counterbore the top of the cartridge for the top bearing and retain w/ a nut. Maybe add a internal snap ring to the bottom of the cartridge to prevent the lower bearing from walking out of the bore?

I guess, what I am asking, is can you show us more pictures of the spindle before assembly? Both of the bore of the cartridge, and the setup you are using to retain bearings/set preload?

Another question is on your collet design. I have seen this same spindle nose design on at least two of your spindles now, and am curious how it is done. I recently ordered a ER32 stub arbor, and am planning on useing that for my spindle, as I have er32 collets down to 3/32" already, but I am interested in how you are making that spindle nose, and how it works.

I don't currently have any aluminum round stock to make a cartridge, would I be seriously shooting myself in the foot to use mild steel? My plan is to make a cartridge that fits my NMTB#30 spindle, and drive it via belt drive from either a trim router, or my full size router. I am shooting for up to 30K rpms with it. I have lots of heat sink material, and I am not afraid to set up water cooling for the bearings, but would it be pointless to try to do it with steel instead of aluminum?


04-01-2009, 08:37 PM
A lot of questions. I'll try to answer them sometime this evening with a few pictures. There are some techniques I have developed and still am developing to makes these spindles. Each one I learn some more. The one shown here has already changed configuration somewhat as I attempt to make it as efficient as possible. My goal is to have it so well balanced that if you had it in your hand while running you wouldn't be able to feel it. I am pretty close but am having problems with coupling the motor to the spindle shaft.

More later.

04-01-2009, 08:52 PM
A lot of questions. I'll try to answer them sometime this evening with a few pictures.

Take your time. As for myself, I am in no hurry. I have read countless threads both here and at cnczone on DIY HS spindles. There are lots of people building them, a lot of people ending with great results, but very little explanation into the depths of the design.

It seems most threads take for granted that we all understand how to choose/setup/preload bearings. I understand basically what is going on, because I have had all of my machines spindles apart, but I would expect a HS spindle is a bit different critter. I don't want to take the time to make one, relying solely on luck and a little bit of knowledge...

On a side note, it seems your host is not working properly tonight Evan, at least not for me (in which case it would be my computer ;)). I could see all of your pictures this morning on this same computer, they are gone now. If I try to go directly to your website, I get the ...page cannot be displayed... message as well.

**edit** it seems to be working fine again now...


04-01-2009, 09:28 PM
It isn't uncommon for various parts of the net to be temporarily unreachable from other parts. It's called a "Net Split" and can happen when an important node goes off line for some reason. The most common problem is a failure in the DNS or Domain Name System of your service provider. Also, if your computer can't reach a particular IP address for some reason it stores that information for a while and until it times out it will automatically assume that the address in question is bad. You can force it to check again by right clicking on the network icon by the system clock and selecting "Repair". That will clear the DNS cache and might fix the problem.

04-02-2009, 04:54 AM
Here are some images that might clear up some of your questions. A lot of this is still experimental. Using aluminum is a possible problem if you don't allow for the difference in expansion with temperature compared to the spindle shaft. It does make a big difference to bearing cooling though. I have found for these small spindles that allowing for the changes in length by using some sort of compressible material works well. It also makes it a lot easier to set the preload. For these small bearings all you need is enough preload to eliminate free play. Unless you use some sort of compressible material it is basically impossible to do because of the tight tolerances. I use very thin washers as spacers to get the exact amount of compression to just eliminate any free play.

Balance is everything. I machine the spindle in a collet on the lathe with a center on the end regardless of how stiff it seems. Runout must be held to much less than .001 or it will vibrate unacceptably. Everything must be concentric.

A good source of material for small spindles is grade 8 bolts. They are strong and the steel is usually a medium carbon steel which is easy to harden. Just heat to red and dunk in water, then draw it back to blue using a propane torch. For something as small as the collet chuck that I put on the end of my spindles you can heat it enough to harden with a small propane torch.

I use a high quality solid carbide drill bit to drill the 1/8th spindle hole. It cuts steel as if it were cheese. I tried one for hand drilling when my arm was tiring and I could drill through 1/2 inch of steel in 9 seconds. Unfortunately they are very easy to break and I broke it after about 20 holes. Drill the spindle hole as the last part of the setup when machining the nose as it weakens the nose. I thread the nose to 1/4-28 and then make a closer using a small nut. Drill out the nut to tap drill size and the run a tap only most, but not all of the way through. That makes it tapered thread enough to close the collet jaws.

I grind the slots using the 1/64 thickness cutoff disks. It makes a nice straight slot and you don't need CNC to do it but you do need some sort of high speed spindle to do it. Hang a dremel style tool on the side of the head temporarily and use that to do the grinding.

Make sure you use a scatter shield between you and the grinding. I have a sheet of lexan on an arm that I can swing in front of the action.



Old disk drives are an excellent source of small motors and bearings.


04-02-2009, 07:47 AM
Sorry but my server is down at the moment. ETA to fix is unknown.

A.K. Boomer
04-02-2009, 11:31 AM
MEW had an article 3 years or so ago on making a planetary speedup. IIRC the original reversed rotation so you needed to have a reverse on your mill to get normal rotation, but someone else showed a version he made with an additional gear to get forward input/forwar rotation. Cute little unit running in oil.

TG, There's no need for an additional gear for forward rotation if a planetary system is utilized -- its why I made the statement; "the mills R-8 drive is incorporated to drive either the ring or planets and the final drive (the sun gear) is the one that the tool attaches too..."

If your mill does not have a reverse the thing you would want to do is use the R-8 to drive the planets and have the ring held in the housing thats attached to the slightly extended quill -- This will not only ensure foreward rotation it is ideal anyways as it would make the design far less complicated as the ring is the furthest out diameter gear anyways and thats handy for the housing that needs to be -- I have a 18V 3 speed de-walt drill all ripped down and even though the planetary parts are plain bearings I believe they could be used in this application if a sealed housing was used with good lubrication ---- the final drive would be very precision bearings that were bored to fit into the center of the R-8 -- the outer part of the R-8 (which of course would be an extended end for modifying) would be powering the planets whilst like stated the inside would be the bearing support for the final drive High speed spindle --- this again would be an Ideal situation due to the fact that the planets (R-8 spindle) are traveling in the same direction as the final high speed drive - just not as fast, what this equates to is welcomed reduced bearing speed and friction of the final drive bearings - even though their rotating much faster their splitting some of the difference between the mills spindle drive bearings and the H.SpeedSpindle drive bearings, The final drive achieves its speed by being rotated while its rotating;)

I would have built this system long ago for my mill -- I just dont see any applications where I would need one ----------- Yet.

04-02-2009, 01:16 PM
Great work as always Evan.

I don't know if you have seen this thread on cnczone: http://www.cnczone.com/forums/showthread.php?t=74524

there is a nice little homemade spindle shown in post 8 using an outrigger RC motor. You can see the coupling he uses quite clearly. I don't know how well that would balance though.

04-02-2009, 03:51 PM
I have been talking to the hosting service. The server lost both hard drives of the raid array. They are rebuilding them from backups. It should be up and running this afternoon.

04-02-2009, 05:53 PM
It's finally back up.

Here are some stats from my web site for March.

Figures in parentheses refer to the 7-day period ending Mar 31 2009 at 11:14 PM.

Successful requests: 208,480 (3,449)
Average successful requests per day: 1,546 (492)
Successful requests for pages: 3,894 (265)
Average successful requests for pages per day: 28 (37)
Redirected requests: 21 (0)
Distinct files requested: 812 (80)
Distinct hosts served: 20,200 (65)
Data transferred: 11.98 gigabytes (185.83 megabytes)
Average data transferred per day: 90.97 megabytes (26.55 megabytes)

The majority of the traffic is from just this BBS.

04-02-2009, 08:08 PM
Thank you! That clears up a lot of questions I had. I like the idea of compressible washers-- is this why you sometimes find belleville washers in high speed spindle designs? Same concept, but with a spring steel part?


04-02-2009, 08:27 PM
is this why you sometimes find belleville washers in high speed spindle designs? Same concept, but with a spring steel part?

Exactly. I would use belleville washers if I had some the right size and strength. I may try making some from an old tape measure spring.

04-02-2009, 09:25 PM
And another question-- The stub arbor I ordered for a spindle is a Chinese unit, 20mm diameter shank. Originally I was thinking of using 6304 deep groove ball bearings, as they are 20mm bore. More recently I was considering 6002 ball bearings as they are more readily available (i.e. I have a pair of them on the shelf in my shop, nothing special, just electric motor bearings), however their bore is 15mm.

This means I will have to machine the stub arbor shaft to 15mm. Since the arbor has a ER32 collet chuck on the end, this kind of precludes machining the arbor in a collet on the lathe. I hesitate to clamp the ER32 end in the 3 jaw set true, as then my machining of the arbor will be concentric to the OD of the er32 chuck, not the bore. Would the better scenario be to use my largest ER collet (which IIRC is 3/4") and chuck up 3/4" dowel in the ER32 arbor and that in the set true chuck for machining the arbor? Or, would I be better off just going with the 6304 bearings?

The main reason I am now considering the 6002 bearings is that I can start building it right away, and can order 6002 ceramic bearings rated for 33K rpm's for a reasonable price (about $40 each) for replacements down the road. The few 6304 ceramics I have found are much more expensive.

I know I am asking a lot of questions, and I thank you in advance for your answers.


John Stevenson
04-03-2009, 04:00 AM
MEW had an article 3 years or so ago on making a planetary speedup. IIRC the original reversed rotation so you needed to have a reverse on your mill to get normal rotation, but someone else showed a version he made with an additional gear to get forward input/forward rotation. Cute little unit running in oil.

That article was rubbish,
If you study it there is only one bearing in it at the bottom, the top of the spindle is not only unsupported but is drive by one planet wheel and not an opposed pair.
All this will do is force the mesh apart and throw the spindle off because it's not top supported.

As regards going in reverse if he had held the ring gear like most it would have gone forward at 4:1 reduction instead of 3:1 in reverse.


A.K. Boomer
04-03-2009, 10:20 AM
Thats just plain crazy -- and in fact I wouldnt run a planetary system with two planet gears --- minimum is three, his "floating" single bearing spindle sounds like a nightmare also, how did this thing even get up to speed without self destructing?