Need help brainstorming a solution for this mill turn spindle

Quote:

---

Originally Posted by handlewanker

Hi, I quite agree with your choice...….. if it works out OK it'll be quite neat.

The long job part is one thing you would gain advantage from a 4th and servo drive with a through hole.

I see you are opting for 5C collets...…..one problem with 5C is that it needs to have exact size collets for the material holding and a draw tube too tightened from the end which makes the 4th spindle diam and bearings a bit on the large size.

I think I would go the ER40 collet option....25mm diam material size is handy..... as it has a squeeze down factor of a mm and so can cater for raw material diams a few thou undersize.....I've never found any bar stock in cold rolled to be an exact size.

My personal preference for the 4th I'm building is an 80mm diam 3 jaw chuck with an option for a 125mm 4 jaw chuck too......the main reason is the large through diam.

One problem with those chucks is they are rated at max 4,000 rpm but as I won't be doing any turning for me it won't be a problem.
Ian.

---

5c has its ups and downs for sure. Like you mentioned, it needs exact collet sizes, although it seems to have enough give to handle standard stock sizes. They are supposed to handle +/-5 thou. One advantage of 5c over er is that they can grip onto short parts right at the nose. Er collets collapse at both the front and rear so if you try to grip something right in the nose, it will collapse at the back and not grip correctly. 5c is also nice because it's extremely versatile. Any attachment you can think of with 5c on the back of it. Expanding mandrels, large step chucks, 3,4,6 jaw chucks, flat fixture plates, emergency collets, list goes on, including a 5c-er40 attachment which i will most definately have for gripping any size I want. Bit more run out in the 5c-er40 attachment vs just an er40 chuck, but not much. The draw tube is also something that adds pretty huge capability once investing in a pneumatic closer. A bar puller attachment is simple to add, then you can crank out continuous parts unattended, as many that will fit in the bar stock.

Well, all you say is quite true....I'm coming from the armchair engineer of old angle and am a veritable newcomer to CNC machining with a degree from the school of KISS....…….although I have a CNC mill it's not for production and mainly from the learning side.....oldies need their toys to keep their marbles clicking.

I'll stick with the simple 3 jaw chuck option as I think it will suffice for what I intend to do.

BTW....with all that servo drive and coupling in line with the 4th spindle, where can you fit a powered draw tube for the 5C.....by my reckoning you won't get much bar length if the servo is directly in line behind the 4th...…..would it not be a better option to drive from the side with an offset and a timing belt etc?...….. much as I dislike timing belt drives.

Even then I cannot imagine a longish bar of steel, say 20mm diam, sticking unsupported out of the 4th spindle end and revolving at a couple of thousand rpm with a turning operation...…...turret lathes have a bar feeder to feed and steady the long bar outside of the spindle.
Ian.

Quote:

---

Originally Posted by handlewanker

Well, all you say is quite true....I'm coming from the armchair engineer of old angle and am a veritable newcomer to CNC machining with a degree from the school of KISS....…….although I have a CNC mill it's not for production and mainly from the learning side.....oldies need their toys to keep their marbles clicking.

I'll stick with the simple 3 jaw chuck option as I think it will suffice for what I intend to do.

BTW....with all that servo drive and coupling in line with the 4th spindle, where can you fit a powered draw tube for the 5C.....by my reckoning you won't get much bar length if the servo is directly in line behind the 4th...…..would it not be a better option to drive from the side with an offset and a timing belt etc?...….. much as I dislike timing belt drives.

Even then I cannot imagine a longish bar of steel, say 20mm diam, sticking unsupported out of the 4th spindle end and revolving at a couple of thousand rpm with a turning operation...…...turret lathes have a bar feeder to feed and steady the long bar outside of the spindle.
Ian.

---

Yep, you are correct on all counts. Pneumatic draw tube doesn't work with the direct coupler. Also I am limited to about 12in bar stock.
Basically here's my plan, I'll run it for now with direct coupler. Advantages are it was cheap and will give the best accuracy and rigidity with downside of 12 in max stock and no pneumatic closer. The pneumatic closer is another pretty large investment that I'm not ready for at the moment. 800 bucks for the cheap one, more like 1200 for the one i should get. Keeping that in mind, without the pneumatic closer there were basically no downsides to running direct coupling for now. Since I have to manually actuate drawtube anyway, it's also not a big deal to add bar stock more often. The direct coupler was built out of pieces of stock i already had aside from a 25 dollar plate of 3/4 mic6 I'm using for the base. So it was a pretty small expense to take the temporary direct coupling path. Also i like to experiment and see how different methods work. Once I get a feel for this mill turn setup and decide I want to invest in full automation, ill switch over to the dual idler 5mgt setup. At that point ill be able to complete multiple parts out of longer stock without stopping. The max length of unsupported stock without whipping is something I'm also really curious about, but I imagine it can't be that hard to build some kind of bearing support structure to stabilize longer bars with different attachments for different stock diameters.
Another thing I considered is that it will be easier to build a proper dual idler 5mgt setup if I already have a functional lathe. With the direct coupling setup, I'll be able to turn the needed parts for the belted setup like idlers exactly how i want them, the posts for mounting them on, the adapter for the pnematic closer, etc. All stuff that will need turning for proper accuracy.
At this point I'm not sure if thats exactly how it will play out. I have dreams of a much more serious machine. Dual opposing spindles on independent ballscrews capable of part transfers and a b axis head capable of machining on the faces of those transferred parts. Also something larger than the pass through of 5c. 3j or 16c would be amazing. Insanely expensive to buy spindles like this. In the back of my mind I'm wondering if my machine can be adapted for turning and grinding these larger spindles. Maybe, maybe not. Basically I'm trying to slowly upgrade my machine in different steps that give it capability of upgrading itself or building new machines.

All in good time as they say.....the longest journey starts with the first step.....or, it won't happen today and it won't happen tomorrow but it will happen...….how many times have I tried to practice that philosophy....LOL.
Ian.

Well dammit, looks like im abandoning some more parts i made lol. My dunham pneumatic closer showed up on ebay for about half price so I grabbed it up. Going straight to full automation. That means I have to let go of the direct coupling idea and go with the dual idler 5mgt setup. I guess in the end it's less wasted time and money. I can integrate all the pneumatics cleanly. The unit will have a single air line coming in that splits to 2 adjustable regulators, then a 4 way solenoid for the closer and a 2 way for the brake. Independent pressure adjusting for both. Back to cad I go. One thing I did realize is that I can just reassemble the original big belt design for turning any parts i need for the 5mgt setup with closer. It worked fine, just too much backlash for precision indexing, but it turn some parts no problem.
For the idlers I'm kind of wondering if it would be easiest to just use 2 big bearings stacked together. I'm thinking 40mm od, either 12 or 14mm wide double stacked for 24 or 28mm wide idler. No need for flanges since they will be right up close to the flanged 5mgt pulleys.

Nothing wrong with stacking bearings like that for idlers. Sound like a good plan.

Hi, provided they're sealed radial bearings there should be no problem.....just don't make them a force fit on the shaft.....perhaps a simple circlip to hold them snug will suffice...………..that makes easy assembly/disassemble when the need arises.

You might make them a fixed entity with the shaft they're mounted bolted against a slotted hole on the backplate face without having to user a swing arm and spring etc as once tensioned they won't need to be resilient.

There has been a school of though that a jockey wheel tensioner pressing against the back of the belt hastens belt wear due to the back flexing as it goes around the pulleys etc

I would think that if you just had the belt across 2 pulleys and tensioned them tight that would give you more than enough drive...…….. the extra tooth engagement of a back flexing belt is hardly measurable.

However, if you have a very small pulley on the servo then perhaps the belt back bending to get more tooth contact is necessary...……...testing will prove otherwise.

At any rate, the effect of the tight belt will put a big load on the servo motor spindle bearings so that is something to consider in the equation......a breakdown in the servo motor due to bearing failure is a major disaster.

I think a design that had a separate bearing in the backplate that the servo motor spindle poked through would take all the tension off of the servo motor spindle...……...but that would only work if the servo motor was fixed and you had an idler/s on a slotted mount to do the tensioning.

At best you would get 1:6 reduction with pulleys of 10 teeth on the servo and 60 teeth on the 4th spindle.

As you have considered the direct drive with servo at 1:1 perhaps a bigger pulley for 1:4 reduction etc on the servo, while giving less reduction, would not need an idler as you have greater wrap around and more teeth in contact.
Ian.

Quote:

---

Originally Posted by handlewanker

Hi, provided they're sealed radial bearings there should be no problem.....just don't make them a force fit on the shaft.....perhaps a simple circlip to hold them snug will suffice...………..that makes easy assembly/disassemble when the need arises.

You might make them a fixed entity with the shaft they're mounted bolted against a slotted hole on the backplate face without having to user a swing arm and spring etc as once tensioned they won't need to be resilient.

There has been a school of though that a jockey wheel tensioner pressing against the back of the belt hastens belt wear due to the back flexing as it goes around the pulleys etc

I would think that if you just had the belt across 2 pulleys and tensioned them tight that would give you more than enough drive...…….. the extra tooth engagement of a back flexing belt is hardly measurable.

However, if you have a very small pulley on the servo then perhaps the belt back bending to get more tooth contact is necessary...……...testing will prove otherwise.

At any rate, the effect of the tight belt will put a big load on the servo motor spindle bearings so that is something to consider in the equation......a breakdown in the servo motor due to bearing failure is a major disaster.

I think a design that had a separate bearing in the backplate that the servo motor spindle poked through would take all the tension off of the servo motor spindle...……...but that would only work if the servo motor was fixed and you had an idler/s on a slotted mount to do the tensioning.

At best you would get 1:6 reduction with pulleys of 10 teeth on the servo and 60 teeth on the 4th spindle.

As you have considered the direct drive with servo at 1:1 perhaps a bigger pulley for 1:4 reduction etc on the servo, while giving less reduction, would not need an idler as you have greater wrap around and more teeth in contact.
Ian.

---

I'll be sticking with 1:1 ratio with the belt. As I figured out earlier, I need to give 8 to 1 electrical ratio to get 3k rpm within 500khz which brings my resolution down to .04 degree step, but those full steps will land exactly on binary indexing, 90, 180, 45, etc., so there's really no need to gear it down. I can also quickly change the electric gearing and steps per unit for more resolution with lower rpm limit when I need to hob gears or anything else that requires higher resolution. As good as .005 degree per step with 325rpm.
I agree that the idlers aren't going to significantly improve my accuracy and rigidity, especially since I'll be using a brake with indexing, but I do see a couple other advantages. First if all with no idlers, the tension of the belt is going straight to radial load on the servo and spindle bearings. With 2 idlers pulled in close to eachother almost touching, the direction of the tension is directed inward, cutting down the radial load on the shafts by a huge amount, nearly eliminated. Another advantage is cutting down on any possible vibrations from unsupported belt. With the dual idlers, there is almost zero unsupported belt. Also much easier to set tension with the servo and spindle being fixed.
I argued against the idlers before, but now I'm seeing the advantages.
I plan to do it with the slots like you mentioned. There will actually be dual plates that capture the idlers, slots in both plates to allow adjustment of the idler positions. The dual plates will extend all the way over to the servo, so I do also gave the option of using the bearing support for the end of the servo shaft.
My original belt setup had the servo right next to the spindle with both shafts pointing in same direction. This worked and was nice and compact, but after spending some time modeling different gang tooling setups, it will open up a lot more options if the servo isn't next to the spindle. Much better if the spindle is in front by itself with clearance on both sides. The new setup will have the servo in back with shaft pointing opposite direction. You would thing that would make a really awkward configuration that takes up a lot of space, but now that I'm adding the pneumatic closer, it's basically the same footprint, the servo just sits next to the closer in back instead of next to the spindle in front. I'm gonna do some modeling this weekend and get it all layed out.

Good thinking on the 45, 90 and 180 indexing and and 1:1 drive.

If I was going to cut gears with CNC I would use a conventional form gear cutter and just index the blank under CNC as usual.

In my opinion having idlers that pull in close to one another does mean you need to have a longer belt to give you a gap between the servo gear and the 4th axis gear...…..the idler diams also need to be big enough to not stress the belt due to a pointy back bending and if the belt slackens from stretch etc the adjustment also needs to allow the idler to get closer in.

I think that you only need to have more belt radial contact on the driver and driven pulleys if they are small like as in the EBAY 4th axis offerings at 1:6 reduction with a very small drive pulley and a large driven pulley...….that setup would benefit from a largish idler or two between the pulleys......you wouldn't want to attempt milling even aluminium with one of those......wood carving perhaps.

At the end of the day how much tightness is sufficient to just cater for indexing accuracy.....I don't think the belt will stretch in just turning the 4th spindle while indexing......the brake you favour will no doubt hold the spindle once in position against milling cutter forces...…...but when you mill on the move the brake won't be functional so the spindle can kick back.....probably not a problem on a finishing pass.

I considered the aspects of milling with a 4th axis for quite some time and decided on the resilient worm drive as it covers all aspects of both indexing and milling without having to add a braking force etc.
Ian.

couple revisions later, its working pretty well. built a pneumatic disk brake for locking the spindle during index milling, very rigid now. it was pretty useless for index milling without it. A servo simply cant hold a rigid position when geared 1 to 1. Today is the first time running some parts continuous with the bar puller and m98 sub. still need to setup the multi channel coolant. With this setup ive turned aluminum, brass, 303 stainless, and titanium. Wont handle crazy removal rates without chatter, but plenty fast for the size of parts i work on. Think i was peeling off about 5 cubic inch per minute in aluminum. Faster than i can go with my milling spindle. Tolerance on diameters is pretty impressive. A mostly aluminum machine in a sealed insulated enclosure is very controllable thermally. When running continuous, the enclosure stabilizes at 82F, so thats where i keep it 24/7 with just the lights. Holding diameters at about 2 tenths continuously. Sorry for the crappy video quality. need to figure out how to lock the focus on my phone

https://www.youtube.com/watch?v=1SJS...AWUnfsmKtlYoto

https://www.cnczone.com/forums/attac...d=425260&stc=1

Well that's just fantastic. You've gotten that thing functional in a remarkably short amount of time considering all the features.

Care to share photos of the whole thing? I'm particularly interested in the brake and collet closer.

Looks like you did a great job. Quite a project. Would love to see a wider view of the whole machine.

Yeah I'll post some more pics in a minute here. Hard to show the inner workings of the spindle assembly so I'll post a couple cad pics too

heres a few more pics and cad views. 2 cross sections of dual action brake and the whole spindle with closer and brake disk out of the way.

https://www.cnczone.com/forums/attac...d=425268&stc=1
https://www.cnczone.com/forums/attac...d=425270&stc=1
https://www.cnczone.com/forums/attac...d=425272&stc=1
https://www.cnczone.com/forums/attac...d=425274&stc=1
https://www.cnczone.com/forums/attac...d=425276&stc=1
https://www.cnczone.com/forums/attac...d=425278&stc=1