Build Thread CNC rotary fixture , modular , belt driven

Hi guys ,
I need some sort of rotary fixture for my mill ... nothing fancy or heavy , I just want
to make hollow spirals in wood and some light cuts in aluminum .
So ... this is what I think ... 8,5N/m ( 1200oz/in ) stepper motor , 25mm wide timing belt ( HTD T5 ) ,
3/1 ratio pulleys , angular bearings ( 40mm I.D , 68mm O.D ) , 100mm lathe chucks , aluminum frame .
I made some drawings , modeled the fixture in Rhino , then some renderings in KeyShot ...
What I need is some inputs ... have I made any mistake in my design ? will it be strong enough ? Do you like
the design ? Are there any problems with it ?
P.S . I already made a 4"th axis using a 5/1 worm gear reduction , but the backlash was too big , I hope that
this aproach ( timing belt + timing pulleys + strong stepper ) will do the trick for me ...
Thanks .
pics .......

IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/1_zps4aedab69.jpg)


IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/3_zps96204905.jpg)


IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/2_zps21a1f1dc.jpg)


IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/4_zps659981ec.jpg)

.......................................................................................................

I love it! I could use one like this myself. My only concern would be the strength of the bottom rail, but if the mounting features are strong and repeatable then it shouldn't be a problem. You said light cuts anyway, which is NOT what we do with our 4th axis jobs...so this would probably not be rigid enough for us. For what you're doing, it looks like a really cool project. I especially like the chuck tailstock. We are looking into building an independent chuck tailstock for our 4th axis much like that one.
Good luck and happy building!

Thanks Japazo ...
I will use a piece of aluminum extrusion as rail ... 160mm wide , 28mm thick , about 800mm long ... for my needs I hope is strong enough ...
The chuck is spinning on roller bearings , mounted back to back , and then preloaded with a retaining nut ... as I said , nothing complicated .
Center of the chuck will be at 80mm + 28mm from the surface of the mill working table , maximum distance between centers 350mm .
The tailstok will slide along the aluminum profile ( rail ) , and it will be locked in place with 4 M10 bolts . All aluminum components will be machined
using 6061 T6 ( MIC-6 ) aluminum plate only .
pics...

IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/2_zps54b3f2cc.jpg)


IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/1_zps5e4d3157.jpg)

.........................................................................................................................

Very nice but I have a couple of points that you might want to consider.

Hi guys ,
I need some sort of rotary fixture for my mill ... nothing fancy or heavy , I just want
to make hollow spirals in wood and some light cuts in aluminum .

If you sure about that you may want to discount some things I will say below. The problem is things like this end up getting more use than intended.

So ... this is what I think ... 8,5N/m ( 1200oz/in ) stepper motor , 25mm wide timing belt ( HTD T5 ) ,

In your pictures you have the motor/stepper mounted in the wrong direction. That is I looks like it might interfere with stuff mounted on the chuck. The other issue is what happens if you want to use the drive spindle with something other than a chuck? You may say that will never happen but few can predict the future.
I have nothing against timing belts but do keep it as short as possible. 25 mm might be a bit narrow, I'm trying to place in my mind which profile is T5. You should be able to find some online engineering info to determine is you specific belt selection is good enough. Keep in mind the belt has to be much tighter than normal to keep sponginess as low as possible.

3/1 ratio pulleys , angular bearings ( 40mm I.D , 68mm O.D ) , 100mm lathe chucks , aluminum frame .

100 mm is smaller than I first realized. I'm not sure what machine this is going on, but if you don't have physical issues I'd go larger with the chuck. Honestly I wouldn't build the frames out of aluminum sheet stock if I could avoid it. Either use steel of make an aluminum casting. Why, lots of screws in aluminum makes me unhappy. Eventually this threads strip, come loose or allow parts to slip out of alignment. Aluminum castings would eliminate that issue as would frames made out of steel. Even with steel I'd be tempted to do a weldment for some part. Yeh welded steel structure are more work but done right the unit will last forever.

I made some drawings , modeled the fixture in Rhino , then some renderings in KeyShot ...

Nice.

What I need is some inputs ... have I made any mistake in my design ?

Nothing really serious here other than the motor placement mentioned above.

will it be strong enough ?

Probably. You didn't mention the machine it will run on but I'd be more concerned about size.
I'm assuming the base gets bolted to another table. If so that is a good thing as it should minimize twisting. Of course for light use that might not be an issue at all.

Do you like
the design ?

Looks very nice.

Are there any problems with it ?

It looks good but how are the chucks mounted? If you need to install a face plate or drive spur how would you do it?

P.S . I already made a 4"th axis using a 5/1 worm gear reduction , but the backlash was too big , I hope that
this aproach ( timing belt + timing pulleys + strong stepper ) will do the trick for me ...

Backlash in the Worm? Some of that can be dealt with by making the center distances adjustable. Belt drives have their own issues also. This should work good as long as you don't drive it real hard.

Thanks .
pics .......
.......................................................................................................

Thanks wizard .
I can not make the frame out of steel , is out of my capabilities ... so I have to use aluminum .
I am building the fixture for this machine :
http://www.factorydaily.com/forums/vertic..._log_pics.html (cnc vertical mill build log ( pics ))
25mm wide belt is all I could find ... weight and dimension of the fixture is an issue , I can not make it too big or to heavy . The ideea is
to make everything as modular as possible , stiff and strong , small and good looking also ...LOL .
I will make some sort of alignement/locking pins for the base , so I can mount the fixture on the milling table straight and true . Is not
complicated , so I will leave this at the end ...
It seems that I need to redesign the indexer ... move the motor , think of a smart way to tension the belt ...
This build is intended just for study ... I am vey new in 4 axis machining , I need to learn a lot . That's why I need to keep all things
as simple as possible , easy to understand and also cheap .
Later , maybe , I will build a far more serious device : harmonic drives , servo motors , dual set-up modes ( indexer / lathe spindle ) , and so on ...
pics.

IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/3_zps6b204fcf.jpg)


IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/2_zps78c04ae1.jpg)


IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/1_zps9c6c1789.jpg)

O.K guys ... keep it coming ... I really apreciate all the help I can get .

Hi Katran
Looks nice, as long as you can get that belt tight, and the bearings don't rattle. However, there are two details you should consider.
The first is that the surfaces of the Al extrusion are never perfectly flat. There is always some warpage coming out of the extrusion die. That means you should probably skim top and bottom surfaces very lightly before building. But there is a problem here too: see next.
The second is related to the first. When the outer surface of the extrusion is not perfectly flat, you do get some bending in the whole thing when you clamp it down onto something using nuts in the channels. This can make skimming the surfaces tricky if the whole thing is slightly warped at the start by clamping it down.
So ... you may need to skim top and bottom surfaces twice each, very lightly, before building. Pity - that removes the anodising, and the anodising can be rough on the cutters.
That said, if you are maching wood rather than steel, both problems may be minor.
Cheers

Hi Katran
Looks nice, as long as you can get that belt tight, and the bearings don't rattle. However, there are two details you should consider.
The first is that the surfaces of the Al extrusion are never perfectly flat. There is always some warpage coming out of the extrusion die. That means you should probably skim top and bottom surfaces very lightly before building. But there is a problem here too: see next.
The second is related to the first. When the outer surface of the extrusion is not perfectly flat, you do get some bending in the whole thing when you clamp it down onto something using nuts in the channels. This can make skimming the surfaces tricky if the whole thing is slightly warped at the start by clamping it down.
So ... you may need to skim top and bottom surfaces twice each, very lightly, before building. Pity - that removes the anodising, and the anodising can be rough on the cutters.
That said, if you are maching wood rather than steel, both problems may be minor.
Cheers

That's corect , the Al extrusion is not perfectly flat ... I will do the exact thing that you described ( skimming the surfaces ) .
Good call , thanks for the tip .
As I said , this project is for study ... precision is not critical , but I would like to go as far as I can ( maybe 0.025mm , if I'm lucky ... )
Thanks

Katran, did you ever get further on your CNC lathe (with aluminum extrusions)?

Katran, did you ever get further on your CNC lathe (with aluminum extrusions)?

Nope ... I have some other projects waiting , and that lathe is not one of them ... sorry .

O.K ... this is what I would like to achieve ... One of my other projects is building/machining a free float handguard for
a M4 airsoft rifle ( much like Madbull JP ) , and this fixture would be a great time saver ... Machining that piece of aluminum pipe
should be easy , just some slots and holes , nothing more ...
I don't have to machine the barrel , or the picatinny rail , theese are cheap and easy to find ... the front guard is about 100USD and
is worth building it DIY , if possible ... Not sure about the flash hider , but I guess it should be quite easy to machine ...
If I can do this , maybe the wodden hollow spiral wil not be a problem ... or anything made out of wood for that matter ... I hope !!!
pics.

IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/2_zps99890c5e.jpg)


IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/1_zps8aed89bd.jpg)


In your pictures you have the motor/stepper mounted in the wrong direction. That is I looks like it might interfere with stuff mounted on the chuck. The other issue is what happens if you want to use the drive spindle with something other than a chuck? You may say that will never happen but few can predict the future.

I see your point , but I would like to disagree ... if you don't mind . Look at the pictures and tell me how is possible
for the stepper to interfere with anything mounted on the chuck ... everything in that pics is at corect scale , by the way .
If I want to use something other than the chuck ... well , that can be a problem ... we'll cross that bridge when we come to it ...
pics.

IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/1_zps77e8c9bd.jpg)


IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/2_zps4ff523a5.jpg)

...................................................................................................

Good questions.

I see your point , but I would like to disagree ... if you don't mind . Look at the pictures and tell me how is possible
for the stepper to interfere with anything mounted on the chuck ... everything in that pics is at corect scale , by the way .

Maybe it isn't an issue for that chuck. However unless the indexer and chuck are a dedicated unit you might find something other than the chuck on the indexers spindle.
The orientation however just rubs me the wrong way. You have a picture of the unit mounted on your mills table, are you certain the motor and mounting extension will never interfere with the machine structure? I guess you could do a 180 on the unit but for some reason I associate the driver on the left hand side of the machine.
In any event this is really a minor problem.

If I want to use something other than the chuck ... well , that can be a problem ... we'll cross that bridge when we come to it ...
pics.

IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/1_zps77e8c9bd.jpg)


IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/2_zps4ff523a5.jpg)

...................................................................................................

It does look nice!
One other thing somebody above mentioned the issue with extrusions and flatness. That is a problem but skimming most extrusions would rub me the wrong way. Of course this depends upon the extrusion but you risk thinning to the point of structural weakness and compromise of rigidity.
Have you considered a piece of aluminum tooling plate instead? Im assuming you are using this sub plate due to a sort table. If so you only need a tail stock moveable over a short range on the sub plate, for real short distance you can mount directly to the mills table. Nothing is perfect in this regard as you are still overhung the mills table. I'm just trying to figure out what will cause you fewer problems in the long run. With aluminum plate it should be easier to key everything together, puting key ways into an extrusion can be a problem.

Its an incredibly simple and cheap modification that leaves only a few centimeters of belt unsupported.

So where did you get that image? I could not find it on the URL you gave. It looks very good.
Cheers
Roger

Hi Katran
A couple of things do puzzle me though.
One photo of the big toothed belt pulley shows it as a straight pulley. Another photo shows a large boss on the outer side. What gives here?
Some of the photos and renderings of the driven chuck show three bearings. Given the tolerances associated with any machining, we usually only use two bearings on any shaft. What gives here?
Nice spiral oak, but where's all the sawdust???
Cheers
Roger

Man , I cleaned the machine before taking the picture ... that's all . As for that big boss , I have done some modifications to the design , that's
why you see some different things there ... I will make some new renderings , to see in detail what changed .
cheers.

Hi Katran
Yeah, my wife regularly suggests I clean up my workshop. I do, but then I do some more machining ... :-)
The boss - interested to see the new design. I am learning a lot from this.
Did you have any trouble putting three bearings on the shaft? I have not seen that before.
Cheers
Roger

Ian, The rolling involute form is what makes involute gear teeth work. However, they do require some backlash. The flank of the tooth applying force (front) to the mating tooth uses the rolling action. Without clearance (backlash) the back face of the tooth would slide against the face of the tooth behind it. This would cause the same type of wear as in a worm drive. At least I think so.LOL
Dick Z

Hi Dick, I went deeply into the theory of the involute gear form drive, that is the one with the 14 1/2 deg pressure angle, generated by a point on a circle rolling on a level plane, and one thing came to the fore.....if you have two gears meshed together in intimate contact, when the first gear is driving, both gear faces are in rolling contact for a distance along the pitch circle and above and below the pitch circle known as the addendum and the dedendum, and when the gear train is in the over run position, the second gear now becomes the driver, so having a rolling contact once more but with the back face.
That interprets as both faces have a rolling contact once more, either individually or simultaneously.....the contact area does not change it's strategy when it drives or is driven.
That is the principle I have to make a backlash free drive......that is 100% backlash free.
With zero clearance there can be no reversal of the gears when they are in intimate contact, therefore if no movement can happen in the reverse direction on the over run mode, there is no backlash experienced.....by either gear, or any gear in the train.
I would be the first one to say that precision in manufacture is of prime importance, both in the support for the shafts the gears run on and the profile accuracy of each gear.....the hardness of the material is also important to maintain the integrity of the rolling faces to prevent pitting and surface deterioration.
The meshing of the gears need not be with force, but closeness of contact is a prime factor for eliminating backlash.
If you go to the design of a ball race, you have practically metal to metal contact between the outer race, the balls and the inner race, separated by a film of lubricant, and at the same time, with that small running clearance the ball race is able to carry a heavy load and rotate at high speed without deviating from it's designated path of motion or producing large amounts of heat.
Angular contact, tapered roller and thrust bearings all run with a design factor of zero axial movement, and preloads for this phenomena can be quite high without placing the bearings under destructive stress.
If you cannot measure the end float on a spindle, for all practical purposes you do not have end float.
This would never work for a close meshed zero clearance gear train drive that carried even a light load, but at some speed.......the friction of squeezing the oil film away from the faces of the gear flanks would make the system too hot to handle, but for a 4th axis drive where the speed of rotation is relatively slow, this is not a problem.
Lubrication would have to be of the thinnest of oils, probably using straight Diesel fuel, which is still an oil, so that the accumulated contact faces did not add up the oil film and produce even the smallest amount of backlash when they are squeezed.
Using gears compounded to give 1:60 reduction would not only give a movement that was extremely and constantly accurate, had a high resolution capability, but was maintenance free too, IE no adjustment for wear.
That's how I envisage it, so if the proof is in the pudding, time to get baking.....with the coming event on my advent calendar, I'm gonna need you baby...LOL.
Ian.

One thing I ran across when looking at timing belt based rotary tables was the servobelt system:
ServoBelt Rotary Stage
IMAGE(http://s3.amazonaws.com/magnifythumbs/C71CBT05FV81NV13-l.jpg)
Its an incredibly simple and cheap modification that leaves only a few centimeters of belt unsupported. Correctly chosen timing belt profiles have genuinely zero backlash, I believe the AT series has teeth that are carefully oversized as it is designed for high precision linear positioning. The GT series supposedly has zero backlash as well.

Hi, this looks to be a very simple solution to the rotary table drive.
Mike Everman did say that over a length of travel there was some migrating of the belt teeth in the pinion and driven gear when the tolerances for production produced belts varied and was not quite to drawing.
He did state that very good backlashless and positioning results were achieved when a prototype belt was being R&D'd initially, and applied with almost lab standards to get the drive to a high level of accuracy.
This is the same as saying that if you have two rollers rolling together in a friction drive you will have no backlash, but due to slippage the positioning will not be 100% accurate when you go forwards and backwards a number of times.
Ian.

That interprets as both faces have a rolling contact once more, either individually or simultaneously

Um ... I may be quite wrong here, but it sticks in my memory that there is always some sliding contact between the teeth. That may or may not apply to involute gears of course - does any one have any good text-book references?

With zero clearance there can be no reversal of the gears when they are in intimate contact

True, but what happens if there is 1 micron of interference? I worry.
Cheers
Roger

Hi Caff.....I Micron is nothing........that is practically a close encounter .....LOL.....but it is the type of fit you need to ensure the gears don't rattle back on each other with clearance.....there must be zero clearance on all the gears in the train,
Involute gears roll on a flat plane, as in a circle with a point that rolls to form the involute curve.
A circle of infinite radius is a straight line, as in a rack with an inc angle of 29 degrees,
You can generate this curve by rolling a cylinder on a flat surface and watch how a point on the cylinder starts at the bottom and describes a curve as it moves forward with the rotation.
The point I am pushing is that when the gears are closely meshed there will be a smooth motion due to the rolling but no slackness due to the close meshing.
You cannot get this with a worm and worm wheel.....maybe initially, but as the action is a sliding one you need a significant amount of clearance to prevent the two metal surfaces (bronze on steel) seizing to one another.
A hard roller on a hard surface is practically friction free.
Ian.

Hi Ian

Involute gears roll on a flat plane, as in a circle with a point that rolls to form the involute curve.

Well, according to several vendors technical notes I consulted, an involute spur gear has both rolling AND sliding contact. EPI Inc have a nice paper on the subject, complete with animation. See for example
Gear Design Intro, by EPI Inc.
So there has to be enough clearance for sliding.
Cheers
Roger

Hi Katran
Yeah, my wife regularly suggests I clean up my workshop. I do, but then I do some more machining ... :-)
The boss - interested to see the new design. I am learning a lot from this.
Did you have any trouble putting three bearings on the shaft? I have not seen that before.
Cheers
Roger

Here are some captures of the design made in Rhino3D ... That big boss is nothing else but a way to bolt that big timing wheel to the
shaft of the spindle ( if you see , there are 3 holes in there , with countersinks... plus a 4"th hole , perpendicular to the main axis , that is
for bolting the boss to the shaft ... ) . Is like this : the timing wheel is spinning freely on that shaft , with a minimum clearance ( about 0.1mm ) .
The wheel is bolted to the boss with three bolts , then the boss is bolted to the shaft ... )
Three bearings on the shaft ... well , it seems to me that this setup is far more rigid and sturdy than 2 bearings only setup ... The angular
bearing are there to eliminate any play ( axial and longitudinal ) , but is not a good idea to preload the bearings to much ... just enough so the
assembly turns fairly free ... Therefore , when you tension the timing belt , is possible to pull the shaft ... and that's bad news for
the TIR of the assembly . To correct that , I put a third bearing on the shaft , so all the force applied by the tensioning belt is much beter distributed ...
I hope that you can understand my explanation ... my english is not that good , is hard to explain technical stuff in a foreign language ...
P.S. You need to remember that the goal here is study only ... nothing fancy or expensive ( yet ... ) . I had to use whatever scrap material I had
in my shop , without spending too much ... I only paid for the timing components and stepper motor , everything else was scrap and salvage ...
In my next 4"th axis project , I will make a much more elaborate design , the budget will be around 1500USD , precision components and so on ...
cheers.

IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/perspective_zps8d180617.jpg)


IMAGE(http://i1187.photobucket.com/albums/z400/romatsa/2_zpse1a99115.jpg)

...............................................................

That big boss is nothing else but a way to bolt that big timing wheel to the shaft of the spindle

OK, understood.

Three bearings on the shaft ... well , it seems to me that this setup is far more rigid and sturdy than 2 bearings only setup

Oh, it probably is. Takes very good machining to make it work.
Did you think about just having one solid bearing at the chuck end?
Um - I have just worked out why you hqave 3: longitudonal play. Hum ... neat.

I hope that you can understand my explanation ... my english is not that good

Your English and explanations are better than I get from some 'English as a FIRST language' speakers.

I had to use whatever scrap material I had in my shop , without spending too much

Wish I had that much 'scrap' aluminium to play with! :-)
What alloy, if you know? We can buy 6061 easily enough, but it machines poorly. I have some 6082 plate which is nearly as good as 2011.
Meanwhile, my CNC has a jammed ball nut on the X axis. Sigh.
Cheers
Roger

Thanks Roger ,
I own a pretty good quality lathe ... SP165 Pinacho , made in Spain , toolroom precision ... so is not a very hard job to make a spindle
shaft at correct specs .
I do have a lot of aluminium scrap laying around my shop ... from my other projects . Mostly MIC-6 6061. My cnc router is capable to machine
this alloy very easily ( with a 12mm endmill I can cut at 2-3mm DOC with 30mm/sec feed , and stay in precision parameters -0.025mm/0,001inch ) .
By the way , this cnc router is also a DIY project of mine , finished last winter ... If you like , I can upload a series of pics of the machine ...
cheers.

What alloy, if you know? We can buy 6061 easily enough, but it machines poorly. I have some 6082 plate which is nearly as good as 2011.
Roger

I only work with 6061 T6 ... actually , is MIC-6 ( has the surfaces milled from the factory , the thickness is controlled and stable ) . I tried some 7075 also ,
but my endmills were not very happy ... In Romania is hard to find good quality aluminium alloy , I had to buy all of my plates from ebay , at a very
high price ( shipping to Europe is expensive ... ) .
If you use a little WD-40 when machining 6061 , you will get a very nice finish ... try not to go too deep ( max. 0.8-1mm DOC ) , high feed speeds ( min.30mm.sec )
and around 8000-10.000 rpm . A good quality milling motor is essential ( HF motor , 18.000 rpm , min. 1,5Kw , controlled by a VFD , with a TIR of max. 0.01mm ) .
cheers.

Hi Ian
Well, according to several vendors technical notes I consulted, an involute spur gear has both rolling AND sliding contact. EPI Inc have a nice paper on the subject, complete with animation. See for example
Gear Design Intro, by EPI Inc.
So there has to be enough clearance for sliding.
Cheers
Roger

Hi Caff, thanks for the very detailed explanation of the Involute gear principle, much better than the ones I have read about over the years.
Looking at the animated display, it appears that the gear tooth tip is sliding when it is almost out of mesh with the second gear, but as the main driving force at that moment is being exerted by the next gear tooth entering the mesh in a rolling motion, I don't think I will worry about that end bit.
It is present as a late event in any gear drive for any gear box having clearance between the teeth and running at high speed, and if it does no harm in that situation it won't in the 4th axis design I have.
There is not that much force present in a 4th axis drive, so a bit of tightness in the mesh is better than a smidgeon of looseness, and as long as the tightness is very slight I think the drive with compounded gears will give good service without any dire consequences.
This type of thinking would not suit a drive that was rotating at speed due to friction becoming a problem, but in higher speed drives the presence of backlash (gear tooth clearance) is not a design flaw......you cannot have both worlds with gear drives, and for a 4th axis the tightness of the gear mesh would be my preference, otherwise if you cannot control the mesh tightness you may as well use a worm and worm wheel set-up.
I got the results I needed to propose the design by removing the banjo gear carrier on my lathe and although the gears are slack on their shafts, when they were meshed fairly closely I could not detect backlash at the beginning and end of a 3 gear compound set-up, so some experimenting is needed to make a drive as I envisage it.
It is quite possible that a set of Nylon gears would be able to mesh tightly without even having any lubrication and as they are of a resilient nature and the gear teeth are mostly of a rolling contact, this may be the way to go.
You can cut Nylon gears very easily with a single tooth fly cutter ground to the tooth form, but proper involute gear cutters with the right tooth form are not all that dear to buy on EBAY.
Ian.

Ian, I was employed by a gear manufacturer for a number of years. They were pioneers in the production of "plastic" gears. The plastic gears function was to be sacrificial items between more expensive metal gears. As you stated, they do not require lubrication. However they do wear easily and were replaced periodically.
Unless the axis is used for positioning only or very light machining, the gears will see substantial forces. This is especially true in milling things like steel contouring.
Personally, I'd like to see an epicyclical gear reduction used on an axis drive. See if it proved my theory of compensating variables.LOL
Dick Z
ps, I'm still involved with the gear industry, my cohorts in crime and I design an build gear cutting machines. One is the 40 year old "wizz kid" that wrote the software for 8-axis CNC hobbers, shapers, multi thread grinding wheel dresser for machine like Rieshauer type gear grinders. Fun stuff.

Hi Katran, in the 4th axis design you've made.....(very nice bit of machining too)....there was some mention of using only 2 bearings for the spindle, but as you have the two main bearings right at the chuck end this would not work with a strong belt pull as has been noted.
However if the box had been made with the two end plates bored for the angular contact bearings and just plain end caps used to apply the preload with a spacer on the shaft between the bearing inner races etc, then the 2 bearing method would work too as the only part under load from the belt tension would be the short piece sticking out of the end angular bearing cap that the pulley is mounted on.
The shaft would be retained in the bearings by a threaded collar at the very end of the shaft, pressing against the face of the pulley which in turn is pressed against the inner race of the end bearing with a short sleeve.
There is no need to go to conventional spindle design for a 4th axis with the 3 bearing configuration of 2 angular contacts at the chuck end and a plain radial at the other end as this is not a spindle per se, but a drive shaft with low radial forces from the side thrust of the cutter and practically no axial forces to need preloaded bearings, like mill spindles.
In this instance I think that just 2 double row sealed radial bearings, one either end, held in position with caps would still be sufficient to maintain shaft integrity both in the axial direction and the radial one.
This would simplify construction and also allow bearings with large bores and smaller OD's so that a shaft with a large bore can pass material through it for those spindles with small diams and long lengths, when only the end area is being worked on.
The bore in a 100mm chuck can pass about 35mm if my memory serves, so a shaft with bearings having 45mm bores can have a bore of 35mm too, and a 5mm wall thickness for the shaft is more than strong enough for a 4th axis.
One question, why was the box not made from steel all welded together and machined all over.....much cheaper?
When I made my lathe in the late 60's I used two pieces of channel iron welded together with two 12mm thick end plates welded on, then machined the whole lot all over.
The box was bored in the lathe to take two bearing cups.
Bearings were taper roller and held in the bearing cups inserted in the end plates.
I think I will use this method again as it worked so well previously and is relatively easy to make, but instead of channel iron I might use thick wall square tubing, about 100mm square, end plates welded on and a 20mm piece of steel for the base welded on too.
I prefer to work with steel than aluminium any day.....steel can be welded to give a composite structure, but aluminium needs to be screwed together, very time consuming, or be carved from a solid block or casting, both quite expensive.
Ian.

Ian, I was employed by a gear manufacturer for a number of years. They were pioneers in the production of "plastic" gears. The plastic gears function was to be sacrificial items between more expensive metal gears. As you stated, they do not require lubrication. However they do wear easily and were replaced periodically.
Unless the axis is used for positioning only or very light machining, the gears will see substantial forces. This is especially true in milling things like steel contouring.
Personally, I'd like to see an epicyclical gear reduction used on an axis drive. See if it proved my theory of compensating variables.LOL
Dick Z
ps, I'm still involved with the gear industry, my cohorts in crime and I design an build gear cutting machines. One is the 40 year old "wizz kid" that wrote the software for 8-axis CNC hobbers, shapers, multi thread grinding wheel dresser for machine like Rieshauer type gear grinders. Fun stuff.

Hi Dick.......plastic gears....hmmmm, I think that for the 4th axis this might be a solution as the forces are not all that great, and if a belt drive can handle the forces so can a plastic gear.
I recently made some gears from a green type "engineering" plastic and they work OK in the screw cutting train with the other cast iron gears, yeah it's an old lathe, 1920/30's vintage.
It's all very hypothetical at the moment, but experimenting will have to done if I want to enter the 4th dimension....LOL.
BTW, I'd hate to have to try and make an epicyclical gear box and get the gears to mesh tightly......the ring gear would be a nightmare.
I think it is bad enough designing the 1:60 ratio geared 4th axis I have in mind at the moment.
Ian.

SP165 Pinacho , made in Spain , toolroom precision
I do have a lot of aluminium scrap laying around my shop ... from my other projects

Ah, you're a lucky man! My gear is smaller, and the X axis on my CNC is currently jamming. Something has got into the ball nut, so I will have to strip it right down. Sigh.
Cheers
Roger

Hi katran
Just dreaming ...
Did you consider using plastic bushes instead of double ball bearings behind the chuck?I am just wondering as the chuck would not be spinning fast at all. I have been using nylon bushes on ground 20 mm steel shaft for a tool&cutter grinder I made, and that went fairly well. Of course, the loads would be higher on a 4th axis, but if the ID was something like 50 mm, the surface pressure on the plastic could not be very high.
Cheers
Roger

Just so the plastic bearings are configured to eliminate axial movement of the shaft. Maybe flanged bush bearings?
Dick Z

Hi, I would think that for a situation where both an axial and radial position is extremely important, and given the choice and all things being equal, twin angular contact ball races or tapered roller bearings win by a million miles over either plastic or bronze bearings.
Bronze bearings would work as well as plastic on a highly polished metal surface, both with lube, but the maintenance to keep the bearing free from any axial or radial displacement means you have to go the extra mile or two not only in the making but in the maintenance too.
Angular contact bearings can be pre-loaded to give a rock solid positioning while still moving freely, but plastic is not so forgiving as it tends to migrate with a slow steady force over time.
The 4th axis could be designed with plastic bearings to meet the requirements, but why go to alternatives when no problem exists.
A typical case of mixed bearing configuration, for no apparent reason, exists in the Vertex BS-0 dividing head, where the literature clearly states that angular bearings (plural) are a feature of the design for their model, but in actual fact there is only one angular contact bearing and that is behind the chuck with the tail end of the spindle being in a cast iron plain bearing with a plain bearing thrust washer and nut to apply the pre-load to the angular contact bearing.
You would have to go into the design deeply to decide if this was the only way the spindle could be mounted, but for the best layout the twin angular contacts, either both at the front or spaced one at each end of the spindle, wins every time, and the cost of the second angular contact bearing is not a factor for consideration for an alternative design.
BTW, anyone who can machine plastic successfully to get ideal bearings must have a reason to want to work with a difficult material when a better solution is at hand.
Ian.

Hi Ian

I would think that for a situation where both an axial and radial position is extremely important, and given the choice and all things being equal, twin angular contact ball races or tapered roller bearings win by a million miles over either plastic or bronze bearings.

Well, this is certainly the conventional thinking. But I note that the old cast iron ways and bronze nuts (with huge backlash) have given way to linear bearings and ball nuts (with much less backlash), and in some cases these in turn have given way to engineered plastic surfaces on the ways and even engineered plastic nuts (with essentially zero backlash). It is interesting that it is the high-end machines which have been getting most of the make-over, not the cheapies.

plastic is not so forgiving as it tends to migrate with a slow steady force over time.

Well, yes, but how much and how fast they deform is another matter. LMWPE would not be my first choice, nor would PVC (just as examples). But something like PEEK, or even acetal, can give very fine performance, not to mention ones like Vespel, Torlon or more propietary ones like Frelon and DryLin.
More technically, while many cheaper thermoplastics (eg PE) will flow under pressure and at slightly elevated temperatures, some of the more crystalline plastics are far more resistant - due of course to their crystalline properties. It is a mistake to lump all plastics in one bin!

The 4th axis could be designed with plastic bearings to meet the requirements, but why go to alternatives when no problem exists.

Ah, but there are problems with ball races: brinelling for a start. Plain bushes have a much higher shock load capacity since their contact area is relatively huge in comparison.

for the best layout the twin angular contacts, either both at the front or spaced one at each end of the spindle, wins every time

Easily asserted, but is it really true? Since engineering plastics are creeping into some parts of machine tools, giving superior performance, what other bearings could bear (pun, sorry) investigation?

BTW, anyone who can machine plastic successfully to get ideal bearings must have a reason to want to work with a difficult material when a better solution is at hand.

It's not that hard to do with the right plastics and the right tooling. Some people will say that aluminium is easier to machine than steel (for instance), but that is a ridiculous generalisation. Compare a nice free-machining steel (eg silver steel) with 1000-series aluminium for instance: the latter has all the lovable consistency of damp chewing gum and is death on cutters through heat and clogging.
On the other hand, I can machine (good) acetal or PET to about 20 microns on a production basis with sharp HSS tooling.
I am not saying that plastic bushes would work; I am just asking whether anyone has tried them for slow speed use, given their increasing use elsewhere in high-end gear.
Cheers
Roger

Rule of thumb - 1% per 1000psi volumetric-ally
http://www.machinerylubrication.com/...-fluid-perfect

I'm not sure why you guys have been posting all this none related stuff in Katran's thread
Start your own thread & spill your BS there, as you have messed up Katran's 4th axes build thread

Hiiiiyyy Mac, belt drive not being the very best solution generates a desire for a better solution....etc.
Sure, the build by Katran is neat and tidy and I think even he would tire of the superlatives describing the build quality now it's up and running, and eventually it gallops off into the sunset as another thread dies never to be revived again, meanwhile back at the ranch the faithfull of the 4th axis god debate all the other possibilities that could be an alternative drive to the general design Katran is using.
He's already been encouraged to keep the third bearing for the spindle end support, so it does pay to chew the rag sometimes, and we all can benefit from a general discourse as long as it stays more or less on topic.
I'll be going into the 4th axis build scenario soon so I want to get all the info pertaining to a successful solution that has been used to date.....ie, plan twice, cut once is my motto.
Katran has already made his choice in the drive method and will tolerate any inadequacy in the drive mechanism if or when it occurs, but some of us are leaning on the side of perfectionism and would like to have a better way.
Did you have something to say, or are you just saying something?
Ian.

Play nice, guys.
Dick Z

Nobody participates in the chat area, they had to go somewhere and "help", if that is what you call it.

help???......whom?
This is an area where ideas are exchanged hopefully to gain new ideas.....you can also show off your latest creation and answer queries pertaining to the design or useage so that others can see how clever you are and admire your work.
Apparently someone thought it was purely a one way window to show off an idea and not attract questions relating to it....re post #98,
I hadn't realised that was the case, my apologies for being inquisitive.
Ian.

Hiiiiyyy Mac, belt drive not being the very best solution generates a desire for a better solution....etc.
Sure, the build by Katran is neat and tidy and I think even he would tire of the superlatives describing the build quality now it's up and running, and eventually it gallops off into the sunset as another thread dies never to be revived again, meanwhile back at the ranch the faithfull of the 4th axis god debate all the other possibilities that could be an alternative drive to the general design Katran is using.
He's already been encouraged to keep the third bearing for the spindle end support, so it does pay to chew the rag sometimes, and we all can benefit from a general discourse as long as it stays more or less on topic.
I'll be going into the 4th axis build scenario soon so I want to get all the info pertaining to a successful solution that has been used to date.....ie, plan twice, cut once is my motto.
Katran has already made his choice in the drive method and will tolerate any inadequacy in the drive mechanism if or when it occurs, but some of us are leaning on the side of perfectionism and would like to have a better way.
Did you have something to say, or are you just saying something?
Ian.

There's alreeady a thread on backlash-free rotary axes in the linear motion section; of which you've been a great participant. Why hijack Katran's build thread, when these great ideas can be shared there, and just direct others to that thread? Also using "underhanded" comments like "will tolerate any inadequacy in the drive mechanism if or when it occurs" may not be so constructive. Katran has been receptive of possibly implementing some ideas, but do we need to beat him and others over the head as to one's idea of "perfection?"
The "art" of engineering is designing to meet the specific needs and tolerances of the end-user, within that end user's means and budget. This is the practical aspect; and is something most mere mortals still working have to consider, as opposed to those who are retired and have the time on their hands. As to perfection, there is no such thing, as we can only build to a certain level of tolerance that our tools will allow. The rotary as Katran has designed it is simple and practical in its design. If the need arises for upgrade I'm positive he is up to the task!
As to hydraulics, yes they can be pretty "stiff" and exhibit zero backlash, like "hydrostatic" lead screws (although their cost must be stupendous...)

Hi Ian

The hydraulic motor can quite small for tons of force and can also be held stalled in any position, and with an encoder give accurate positioning in both directions without missing a step or encodec line.

You know, this represents a significant change in strategy, one which has not been examined very closely. Let me explain.
Case A: Use stepper motor or servo motor with encoder on motor for both drive AND position sensing, and gear down. Spur gears, worms, belts, whatever - it does not matter: the position sensing is being done at the motor end of the chain.
Case B: Use servo motor for drive, gear down (any way you want), and sense final movement/rotation at the rotary table.
We are all very familiar with Case A: most CNC machines are built that way, despite the probelsm with backlash. But Case B does not have anywhere near the backlash, and in some cases it has essentially zero backlash. By way of example, we do have some small ultra-precision CNCs which use laser interferometry to measure the X, Y & Z movements of the table. You are talking sub-nanometer precision here.
It's a bit more tricky with rotation: you have to use something other than interferometry, but it is possible. I do have some old Selsyns somewhere ... put a 20-bit sigma-delta ADC on one and you have 1 part in a million or ~1.2 arc-secs resoluion. Hum ...
Thing is, once you start sensing the rotation of the table directly, how you drive the rotation simply does not matter. To be sure, a smooth low-backlash drive helps, but it does not affect the measured rotation.
Just a thought
Roger

Hi Caff, I bow to your superior knowledge.......I'm not very up on sensing mechanisms, being a purely mechanical orientated person, but I'm learning at a rapid rate of knots how movement is sensed and the outcome of high technology down at our level.
Having worked on mills and horizontal borers for a good part of my working life, one thing you don't do with those machines is climb mill, not unless you have a hydraulic buffer to resist the grab of the cutter etc, but with ball screws, climb milling, (providing the set-up is rigid enough) is just another way to get the cutter to do the job.
My point is, you can drive the table around or in a straight line with whatever mechanism takes your fancy and sense the movement with whatever sophisticated methods are currently available to an extremely fine degree, but if there is any self feeding aspect in the drive train without any form of hold back mechanism, like a cutter in a climb milling mode, it will play havoc with the set-up and overload the cutter.
That is the reason I am pursuing the alternative drive methods and consider a belt drive as having a limited although quite suitable application for most needs.
Good enough is not good enough.
A toothed belt drive, without argument, can be considered as a very flexible rack and pinion, but with a compromised material and tooth profile for accurately moving the table.
The point on that score is, if you can have a linear expansion from rest to tension you don't have accuracy as it's randomly variable.
It's possible that the toothed belt is 99% of the story, but only a test program would reveal it.
No matter how it is applied, there has to be a better way.
If it's true that the sensing mechanism senses the ultimate position for the input stimulus, then a hydraulic drive with sensitive check valves will provide the best of both worlds, IE, move from A to B and back to A without loss of resolution, and move without having a forward or reverse reaction under any load.
As this thread started out as Katran's project I won't debate further on alternative drive methods.
Ian.

[QUOTE=louieatienza]There's alreeady a thread on backlash-free rotary axes in the linear motion section; of which you've been a great participant. Why hijack Katran's build thread, when these great ideas can be shared there, and just direct others to that thread? Also using "underhanded" comments like "will tolerate any inadequacy in the drive mechanism if or when it occurs" may not be so constructive. Katran has been receptive of possibly implementing some ideas, but do we need to beat him and others over the head as to one's idea of "perfection?"
Hi Louie, no "underhand" aspect intended or posted.......if the drive has slackness or lack of resolution then Katran will have to live with it as that is his choice of method to do the job.......he's not going to change the set-up in the near or distant future unless it becomes too inefficient for his needs.
As far as he's concerned, it works to his satisfaction and exhibits no problems that are apparent.
The fact that the test piece was a carved wooden object where resolution is of a very low order is irrelevant.
Ian.

Hi Caff, I bow to your superior knowledge

Dangerous, very dangerous! :-)

Having worked on mills and horizontal borers for a good part of my working life, one thing you don't do with those machines is climb mill, not unless you have a hydraulic buffer to resist the grab of the cutter etc

Can I interest you in a very cheap, almost new, 20 mm Dormer square-ended milling cutter - currently in two pieces? And I was milling plastic at the time. Actually, the plastic grabbed out of the vice, but no matter.

but if there is any self feeding aspect in the drive train without any form of hold back mechanism, like a cutter in a climb milling mode, it will play havoc with the set-up and overload the cutter.

No argument at all!
In fact, that is precisely why I hesitate to use a simple toothed belt drive on a rotary table. Even at 20:1 reduction, I am sure you could back-drive it some day. So, ... my conclusion is that for a reliable rotary table I will have to use a worm drive. We know they do not back-drive.
Over to the other thread: this is Katran's.
Cheers
Roger

CIAO for now.....I've always wanted to speak French.
Ian.

Hi Caff, you're quite right on most points, with R&D anything is possible with anything you can name.
I'm not a great fan of plastics, but I appreciate their uses even if I don't fully utilise their potential......I have a couple of plastic gears on my lathe and they work well, but it was a matter of expediency at the time and a certain curiosity with the material being available.
A 4th axis is probably one area where you could have many choices, and provided the bushes met the requirements offered by a conventional bearing arrangement, why not use them.
When it comes to spindles and bearings, I'm a dyed in the wool conventionalist when it comes to choice of bearings.......the design of the housing would dictate what the bearing lay-out would be, and as labour and time are important, doing it one way means you only want to do it once.
I have the same attitude towards plain bearings generally, and I would never use a plain bearing if a needle, ball or roller could be used.
Old habits and methods die very hard in my school, and that tends to slant your imagination when it comes to being inventive.
The biggest leap of faith is at the moment......I'm going to be a complete unrepentant heretic and go CNC with a vengeance, the whole nine yards and a bit, even though I do not anticipate any production in the air.......embracing the "new faith" is a compulsive addiction that needs to be completely catered for or the end result is a cataclysmic failure.......imagine carbon steel cutters on most materials at high speed as opposed to a carbide one.
The same goes for any form of plastic for current design needs........maybe in the future the plastic revolution will fill the needs, but not as I see it for the present.
I couldn't tell one plastic from the next, so maybe I'm overlooking another technology through complete ignorance.
Nobody would dream of a lathe built from wood as a viable option, but they do make routers from Bamboo that work very well, provided the material doesn't get damp and start sending out roots and take root.....LOL.
Ian.

Plastic really has not infiltrated machine tools in the way you imply.
Plastic deforms very easily. Turcite/Rulon machine slideways require low loading per area, and use a single thin layer to reduce stick-slip friction.
A ground ballscrew/rollerscrew outperforms plastic nuts in every category except cost (especially in machine tools where rigidity is king).
I can certainly imagine a working plastic bearing, but it would use only a thin layer of plastic to reduce friction between large metal surfaces (just like how it is done for machine tool ways). A reasonably sized plastic bushing simply cannot hold the work-piece rigidly under normal cutting forces.
Plastic is like jello compared to most metals. Even the new engineering plastics can't compare to steel, and most of the really hard ones don't slide that well anyways.

Hi Ian

A 4th axis is probably one area where you could have many choices, and provided the bushes met the requirements offered by a conventional bearing arrangement, why not use them.

That is what I am wondering. Mind you, 'wondering', not 'doing', at this stage. Maybe 'wondering why not'.

I'm going to be a complete unrepentant heretic and go CNC with a vengeance

Enless fun. I did that too.

imagine carbon steel cutters on most materials at high speed as opposed to a carbide one.

Well, that's the funny thing. I have gone back to HSS for some materials as it works better. You see, the edge you can get on carbide, even the best, is not as good as what you can get on HSS. It's a function of grain size. On a micro scale, the surface of carbide is rougher due to the matrix structure. So aluminium sticks to carbide much worse than to polished HSS. Plastic machines better as the edge gets sharper. Much testing done to confirm this. So my lesson is don't use a cutter previously used on steel for Al or plastic.
Cheers
Roger

Hi 691175002;1367676

Turcite/Rulon machine slideways require low loading per area, and use a single thin layer to reduce stick-slip friction.

Well, yes, that is true, but so what? Fact is, high end machines are using the stuff. Yes, they do use big areas to keep the loading down, but the amount of wear (especially compared to old cast iron ways) is reduced fantastically. In fact, there are firms out there whose main business is to resurface old cast iron ways with plastic. Done properly the results are excellent. At the right surface loading a good plastic does not scrape away like cast iron, and holds tolerance better with far less maintenance.

I can certainly imagine a working plastic bearing, but it would use only a thin layer of plastic to reduce friction between large metal surfaces (just like how it is done for machine tool ways). A reasonably sized plastic bushing simply cannot hold the work-piece rigidly under normal cutting forces. (My italics)

Oh yes, exactly. You use thin layers of plastic, not thick ones. The idea that there is anything wrong with this in comparison with 'reasonably sized plastic bushing ' is simply a misunderstanding of the material. You do not treat it the same as (say) a bronze bushing.
Fact is, engineering plastics are being used in high-end machine tools. Fact of life. And I don't think those guys are stupid.
Cheers
Roger

Plastic really has not infiltrated machine tools in the way you imply.

I don't know about that, synthetic bushings are very common in the custom machine world especially if a clean environment is required.

Plastic deforms very easily. Turcite/Rulon machine slideways require low loading per area, and use a single thin layer to reduce stick-slip friction.

Exactly. Igus makes such in their DryLin series. You can also look for what are called PTFE liner bearings. There are lots of options actually and plain bearings may hold up better in some situations

A ground ballscrew/rollerscrew outperforms plastic nuts in every category except cost (especially in machine tools where rigidity is king).

Cost is always a consideration. Beyond that sometimes it is easier to fit a plain nut in a tight location.

I can certainly imagine a working plastic bearing, but it would use only a thin layer of plastic to reduce friction between large metal surfaces (just like how it is done for machine tool ways). A reasonably sized plastic bushing simply cannot hold the work-piece rigidly under normal cutting forces.

What you have imagined is pretty common in the industry. Beyond that you can spread the load out.

Plastic is like jello compared to most metals. Even the new engineering plastics can't compare to steel, and most of the really hard ones don't slide that well anyways.

That is a bit over the top. Like all engineering projects the task at hand and current technology determines which way you go with a bearing solution. Old fashion oilite bushings can sometimes be a reasonable solution.

Hi Wiz, I have a DIY power hacksaw, using standard hand hacksaw 12" blades, built in 1982 for a special purpose job and the main swing arm had Oilite bushes in it.......they only lasted 31 years and now have to be replaced due to some ovality........needle bearings or ballraces would still be going strong.
Ian.

Add new comment

Images
More information
  • Files must be less than 2 MB.
  • Allowed file types: png gif jpg jpeg.
Documents
More information
  • Files must be less than 2 MB.
  • Allowed file types: zip rar.
By submitting this form, you accept the Mollom privacy policy.