[HuFlungDung]

H3ndrix,

RE: the internal grooves, etc in your spindle. If you gave me that job, I'd be scratching my head to do it in one piece like that, because working in a long slender bore to produce those shapes is a nightmare, as Ballendo implied

Perhaps there is no reason why you could not make that section of your ball retention, spring seat as a drop in cartridge. This would make construction much simpler, and you could also harden the hell out of the ball seats so that they don't wear out in short order.

Re: singular angular contact: not recommended, although I am not sure where you intended to make this change. What useful rpm range do you expect to run in? Have you contacted any bearing suppliers for a catalogue? Even an outdated catalogue will give you some numbers to mull over. I would recommend "The FAG Standard Program" from FAG bearings. There's lots of info about clearances, limiting speeds, etc that you should puzzle over.

[twombo]

This is the cyber version of a group of old codger master machinists chewing the fat about the best way to prototype a design in the old days.

I grew up around a crew like that, all wearing their blue shirts and pants and wearing those little hats with no bills ... (front or back). Absolutely amazing the development and improvement of a prospective part. You all have done is just a few days.

I am very impressed.... y'all are good!!!!

H3ndriX... You OWN Solidworks. Great renderings!

I'm not even interested in building an ATC spindle, but this thread is compelling!!!

[HuFlungDung]

H3endrix,

I don't see how the backstop (one-way clutch bearing, or sprag clutch) would be practical here. It would add a lot of bulk to the end of your spindle, because it would be a pretty large unit, considering it would have to have a large enough ID to grab the flange of the toolholder. I'm not even sure it would grab anything unless the toolholder did start to slip around in the spindle, which is not something you want to have happening, even if it is only a tiny bit of slip, it will spoil the tapers, eventually.

Further, I don't know if it would release easily, once it did lock on. You would have to initiate a bit of rotation to unlock it, and that is going to be tough to do when everything is already tightened up.

[h3ndrix]

Yes Hu, thats what keeps me thinking mostly these last couple of days, I am trying to come up with a stop that is very precise for the drive keys (drive dogs etc...) to engage the tool holder collets for the automatic tool change. There must be a better way, I just havent found it yet

edit: twombo, thx man

[ballendo]

Hello,

I'd sacrifice the 1/3 of the taper anyway...

At this level of machining and accuracy, the additional length just makes it MORE likely that the taper(s) won't fit each other. And the drive dogs take care of the drive, don't they?

This means your spindle has one though hole, with an enllarged upper bore--for the belleville stack--with another enlarged bore ending at a 45degree ramp to the through diameter--for the tool lock balls to ride upon-- with a taper at this same end. Now that's not a hard part to make! (start with some DOM 4140 and it's a cakewalk)

And you'll want to use some pre-hardened strees proof kind of steel IMO, so you won't need the cartridge Hu mentions. But you "could" put a "ring" at the upper end of the lower bore--think outer race of a timken bearing-- for the tool lock balls to ramp on. This would make the spindle even easier to make, and would deal with Hu's proper concern over the balls indenting their "runways"...

Now to the subject of the current message:

Why not use the time-proven, field standard... Solenoid or air actuated shot pin?!?! It's simple, works well, easy to make, small in size, etc. If your solid model of the headstock is accurate(what machine is this designed for again?), it would appear that you could use the front or side wall of the HS casting to mount a solenoid or air cylinder. A tapered hole in the spindle is then all you need...

Enough skunkworks Let's get this thing built

Ballendo

Originally posted by h3ndrix
Yes Hu, thats what keeps me thinking mostly these last couple of days, I am trying to come up with a stop that is very precise for the drive keys (drive dogs etc...) to engage the tool holder collets for the automatic tool change. There must be a better way, I just havent found it yet

edit: twombo, thx man

[HuFlungDung]

H3ndriX,

What kind of tooling are you going to use? The tool flange type has an influence on the type of tool carrier you will use, too, and whether you have to pay attention to the drive dogs in the tool carousel.

Any chance you would be using an AC variable frequency drive on this spindle? If so, perhaps you can use a solenoid plunger that someone mentioned, to lock into the tool drive slots. You would simply have to add a low speed spindle rotation logic to your tool change command, something like this:

Spindle stop
solenoid on, plunger slides out and rubs on tool flange
Spindle jog
Plunger engages slot
Spindle stop
Tool is released

For this type of mechanism, you would have to have a limit switch on your solenoid plunger, to detect when it has engaged the tool drive keys.

I also recommend the variable frequency drive, because I don't know how else you would easily get a safe, low speed jog from your spindle motor.

[h3ndrix]

Ballendo,

I dont know about sacrificing the taper, I would like other people to comment on this aswell. There must be a reason that the engineers figured out the engagement of the taper to be that long (I am not worried at all about my machining skills to be able to make that taper accurate and precise ) Although I will definately take your suggestion into consideration and make a model in solidworks and find out what it will look like.

I have also thought about the balls making their runways on the spindle bore, as Hu mentioned it before, and I am thinking about using a harder steel to take of that.

the selonoid idea is interesting, unfortunately I have never seen one of these auto tool changers to understand how it would work exactly, but I think I have an idea of it. My only worry is I did not want anything that worked with air on this CNC, but maybe I can find another way to do that.

Thx

Hu,

The tooling will be all NMBT 30 type collets or tool holders (the flat flange, not grooved).

My intention on the spindle motor was to either use a AC brushless servo or to use a DC brushed servo with a high HP, but I guess using servo's for spindle motors is not so practical. The only reason I wanted this was to have speed control from the PC rather than manual, so after doing tool changes for example, the machine could adjust the spindle speed according to the tool that it picks up. I am still trying to find out if this is doable (I know you told me to look at vector motors but man are they expensive )

Thx Hu,

H3ndriX

[ballendo]

Hello,

A solenoid is an electrical coil with a metal plunger. It can either pull in or push out upon power application. The "other" direction is usually accompished by a spring. So, no air required!

Drill a hole in the front wall of your headstock casting, in line with the spindle centerline, between the bearings. Tap it. Use a solenoid with a threaded coaxial mount. (You already have the indent/tapered hole in the spindle itself, as I mentioned in my last post.this indent is in line with the hole in the HS casting.

Now when power is applied, the coil attracts the rod AWAY from the spindle, allowing it to turn freely. When you want to index for a tool change, you simply slow the spindle (or use timing if you're not controlling spindle speed--waiting until it is slowly turning) and then turn the coil off.

The spring pushes the pin towards the spindle, where it rides until it falls into the tapered hole. That's it!

In reality, you'll likely want an accurate bore and a pin CONNECTED TO the solenoid plunger, rather than using the solenoid plunger itself.

Hope this is clearer,

Ballendo

P.S. As someone (Hu?) mentioned, you'll want a microswitch (or2) to let you know where the pin is! Don't want to try and turn the spindle on with the pin in place...

Originally posted by h3ndrix
Ballendo,

The solenoid idea is interesting, unfortunately I have never seen one of these auto tool changers to understand how it would work exactly, but I think I have an idea of it. My only worry is I did not want anything that worked with air on this CNC, but maybe I can find another way to do that.

Thx
H3ndriX

[John S.]

Just to clear a few points up.
I'll go will Ballendo on the shorter taper. No reason why not to do this and what he says makes sense. The taper is only there to locate the holder. Classis example which many people don't realise is that the short taper on an R8 spindle is the same as the NMTB 30 taper but shorter.
I also like Ballendo's shot pin approach, simple and if you use a soliniod you don't need air on the machine.

Using an invertor drive or VFD as they are called will give you the range of speeds you need and most of the better vector ones have 3 or 4 presets that you can call on.
It's also very easy to control the spindle speed from the PC by getting the PC to output a 0 - 10v analoge signal, via a convertor, that the VFD will understand.

To get back to the roller clutch idea.
I have a Beaver CNC mill, about the same size as a Bridgeport Interact 2 but long bed.
Mine has no tool changer on it and just uses standard drive dogs but I bought another scrap machine for spares that had a ATC on it and this uses the roller clutch arangement so it does work.
I still have the spindle but got rid of the tooling that came with it.
The Beaver incidently was built in 1985 and had just a two speed high / low speed arangement by air operated clutchs on tooth belt drives. This had an 7 hp invertor drive as standard even for this time. The invertor has been replaced by a modern unit and the PC handles the air clutches and the speed range.

John S.

[avsfan733]

Hey heres a thoguht i would like to throw in the ring.....how about a magnetic tool release? coul you say for example, permanently magentize the spindle, and then run a some wires around it so when energized it functions as the reverse of an electromagnet releasing the tool? I also have a what i think is pretty cool thought for the toolchanger that i will try and throw into solidworks as soon as i find a few spare minutes...specifically a way of orienting the tool exchange mechanisim that depending on my understanding...or misunderstanding...of the way these normally work, might be easier. Hendrix i love the idea and it has totallyt got me thinking

[avsfan733]

two other comments...sorry didn't get to the edit button in time...is there any need for a thrust bearing especially when doing things such as drilling or plunge milling? and im curiosu abotu the mention of drive tang in earlier posts...to drive a 30 HP or whatever mill, how solid must those tangs be...especially as they can't possibly be very big and still maintain the taper or am i thinking about it wrong?

[John S.]

Avsfan,
I think the amount of power needed to hold a holder by a magnet wouldn't be possible.
Added to this all the chips flying up to stick on a rotating spindle is a definte no - -no.

As regards driving dogs with high powered spindles the taper will increase accordingly. Something like a 30 Hp spindle will be running a 50 ISO taper with drive dogs about 1" square.

Hendrix has stated he wants this for a Mini mill, at this scale and the fact that he's probably limited to about 1/2 Hp I feel with a decent fit and finish taper he'll get away with no driving tang at all.
Many Bridgeport operators refuse to have the small locating peg in the spindle for ease of loading.

My nephew has an Italian wood router with ATC thats' on 30 Int taper with no provision for driving dogs, this just relies on the taper for driving.
I must stress it's not a toy as it cost upwards of $90,000 and really moves some material.

John S.

[ballendo]

Hello,

Yes, on most machines this side of a cat 50 the real reason for the "drive" dogs isn't driving at all...

It's simply a good way to orient the tooling in the ATC turret, AND it DOES provide a backup if the taper tries to slip (or does).

Lots of operations in machining with cnc will need to "know" which direction a tool is "facing"-- boring being the chief one.

Lastly, it also confines any runout error to being the in the same "direction" so that as a tool is used and re-used, its cut size is repeatable. (Think about the stackup errors of spindle runout, toolHOLDER runout and tool runout...)

Ballendo

[HuFlungDung]

Here is an idea to play with for the function of the drive keys in the spindle. Since it is easy to get stuff lasercut nowadays, the actual complexity of the shape is almost irrelevant.

Anyways, this is kind of a "finger-stop" idea, such as a person would use when locating the flutes of a tool on a tool and cutter grinder. Only, in this application, the fingers would engage the notches in the toolholder. This thin "spring plate" would be mounted permanently onto the end of the spindle. It would be thin enough that most of the fingers would bend back out of the way, but at least a couple of them would stay straight, and thus engage the drive key slots in the toolholder.

You might even consider modifying your standard toolholders to open up one side of the drive key notch to permit better, or more sure engagement of the fingers.

[h3ndrix]

Hey guyz, I know I ve been away for a while but real life interfears with my persistent CNC project every once in a while.

Ok, Here is (probably) the final version of the spindle design.

-It has belleville washers, stacked in parallel stacks of 2.
-It has a labyrinth design on the nose that attaches to the spindle body with fine threads.
-It has a grease (or oil) nipple that is on an angle for the bearing lubrication.
-The four bearing design that you see is what I was able to get from some info I found on high end spindles. It looks alright.
-Got rid of the mini mill casting as it was creating more problems than it was solving. This casing is either a steel piece or iron. I got scraps around that I can do this with.
-I cleaned the inside of the spindle to make it as simple as possible to make as per recommendations.
- The top has a timing belt pulley.
-Got rid of the cooling through spindle as it was making stuff complicated at the top, might do that at a later time after the spindle is done and working.



Let me know what you think

Oh, I think I will be using a 1 1/2 HP Baldor motor with a Sensorless Vector VFD to drive the spindle, as that is much more economical than a vector motor and drive.

BTW, I found this brake with positioner that might be just what I am looking for for positining the spindle for automatic tool change instead of using a solenoid, check it out, seems nice:

ROBA-stop Positiong brake

Thx ppl,

H3ndriX

[HuFlungDung]

H3ndrix,

I don't really understand the 3 bearing design. However, if you put the radial ball in between the two angular contact, it would make a great spacer

In normal circumstances, the balls in that radial bearing won't even touch the races, because the angular contact bearings are so much better of a fit, they will be carrying that end of the spindle anyways.

Thing 2: don't forget that you need to clamp all the outer races of the angular contact bearings together and against the shoulder of the spindle housing. Maybe add a spacer between your labyrinth cap and the outer bearing race. When the spindle grows hot, you want the excess length to go out the top as much as possible. I can't determine if you have some axial preload applied to that top bearing or not? It should not, as a matter of principle, I guess.

Don't forget to seal the top of your spindle with some kind of a cap to keep the dirt out of that bearing.

You might want to add some kind of a grease retainer underneath the top bearing, so the grease doesn't just run out of it and down the spindle.

[h3ndrix]

Hey Hu,

The 3 bearing design, which I took from the design of Gilman spindles, is for this reason: The The angular contact bearings are not supposed to carry loads which are radial, although they are back-back with medium preload, their lifetime becomes shortened by radial loads. For this reason, an additional deep groove bearing (which is also precision made) helps take on most of the radial load and thus lengthing the life of the angular contact bearings. This theory must be working, I guess thats why Gilman spindles uses them.

You are right about the top cover, but that bearing is double sealed, so I was hoping it wouldn't need a cover, but it wouldnt hurt to have one.

The clamp for the nose bearings is missing, because it was in "hide" mode when I took the screenshot and I didnt realize it until you mentioned it, but there is a bushing there which is out of PTFE (that looks like gear so it can let the grease pass towards the labyrinth) to take up expansion when the bearings want to expand.

Thx for the reply ,

H3ndriX

[turmite]

h3ndrix,
I just wanted to thank you for this thread. I haven't posted but once, but I have been quitely watching. I don't know if you have any idea what you are doing for the diy group or not, but this will
eventually help a ton of them....er..us. I am not currently in the diy other than having done a retrofit to my existing machine but I am planning a build this spring/summer. I hope to use your design as a basis for a larger spindle, if that is ok with you??

h3ndrix you mentioned you are in school. College?? Keep up the good work.

Hu now for the serious stuff. What would something like this cost for a bt30 holder?? or maybe even some sort of 40?



Mike

[HuFlungDung]

Radial loads are bad for angular contact? That's news to me Could be, but that would be the first I've ever heard of it.

Take a ballscrew: the proper points of contact in a good screw are tightly defined, too. I think they are pretty similar to what an angular contact bearing's points of contact are. The balls do not contact at 90 degrees to the axis. A ball screw is a thrust device, as well as a small radial load.

No bearing will work properly if the opposite points of contact are not on the correct radius of the ball, so that ball circumference velocity = the race velocity at the contact point.

Balls do not make more than line contact with the race, because there are different circumferential velocities to every chordal contact circle on the ball. If the race matched the ball perfectly, it would have to wear immediately, because smaller chordal circumferences would have a different velocity than the maximum ball diameter. Enough rant

There are several different internal contact angles for angular contact bearings. Spindle bearings (which are a higher precision class of angular contact) generally have only a 15 or 25 degree contact angle, whereas "regular" angular contact have 40 degrees. Make sure you get what is right for your use.

So I would really be stuck to tell you what "radial" ball bearing these guys would be using, because I can find nothing that is anywhere near the speed ratings or preload rating of the angular contact bearings. I still cannot visualize how there is going to be sufficient movement in the preloaded angular contact bearings to allow the radial ball bearing to even pretend to accept part of the load, especially when it is that close.

Now, if that were a ball thrust bearing, that would be a different story with some merit to it.

[h3ndrix]

Turmite,

Thank you and you are more than welcome to make it for yourself . I could probably make money of this design, but that is not my intention. Yes, I am in college, Pratt Insitute, in Brooklyn NY and I am a Industrial Design student. Now, dont ask me how a design student gets so interested in maching, because I cant really answer it, it just happened waaaay back.

Hu,

I dont really know if its true or not, but that is the explanation of the Gilman guys for putting that 3 bearing setup. In my head, if the angular bearings are back to back, then there cant be any radial play, but I am no engineer and their engineers must have figured out somehow that this is a optimum solution. You may very well be right, I am not at the knowledge level to argue with you about it though As long as there is no downside to this setup, it will take me an extra half hour to bore that hole for an extra bearing and it will take me about $15 in terms of cost to make it like theirs. If you really think that it will impact the design in a negative way, then I will take it out.

Thx guyz

H3ndriX

Oh forgot to say: The price for now, for the whole spindle design that you see there is around $300. That includes the belleville springs, stock metals (8620, 4130 and grey iron class 40), bearing (ABEC-1 unfortunately ), timing pulley, 6 bearings, etc.