Machsupport Forum
General CNC Chat => Show"N"Tell ( Your Machines) => Topic started by: mikep_95133 on March 13, 2012, 07:18:52 PM
-
I have been awed as a student of Mr. Simpson's (Simpson36) 4th axis work. I've watched each of his videos at least once. Some several times. I am so inspired that I needed to try my hand at making my own 4th axis. I was previously considering making my lathe cnc, but the 4th axis added to my mill will fit the bill nicely. In fact I made a couple of items for a medical facility not long ago, and a 4th axis would have cut the machining time by 75% easily.
I've got one spin indexer in house that I need to tear down to see if it's true enough for this project. I'll have to order another one for the tail stock too. Since the servo motors can handle something around 2400 in-oz, then I think a steep reduction will hold difficult work very still and allow 300-600rpm for turning down material. I'm hoping that this design will not require a brake.
Here is the first prototype cad rendering using Rhino 2.0.
-
I tore down the indexer today and wiped it free of all preservative oil. Man that stuff smells. As it turns out this particular spin indexer is exactly the same make and model as Steve had in his early prototype days. That was a surprise. I hope the 5c spindle is in better shape than the one he had or I get to repeat the buying pattern too. Steve reported that the bearings did not slip over the shaft easily.
I semi carefully drew up the 5c spindle in this rendering. It interacts with so many other parts. I wanted to make sure it would clear everything. I downloaded the shaft collars in 3d iges format from McMaster. Same with the ball bearings. I am convinced that I should use 40 degree angular contact bearings instead of deep ball bearings. Angular contact bearings are the same that were used when I rebuilt my Bridgeport head. They could last forever since they are designed for thrust loads. But the 4th axis is not as severe duty as the Bridgeport. So we'll see. I just made up the pulleys. They are dimensionally accurate. The tooth patten was just made up quickly. I had to slide the front wall forward and that eliminated the flat washers, so that end of the spindle stayed flush with the end of the base. The updated spindle model caused these changes. At least if a chuck gets installed it won't interfere with the base.
I still need to draw up the collet knob and threaded tube for the spindle. I also need to slot the back wall to move the motor to allow adjustable belt tension. A cover, bolt holes and other details are still missing.
-
Bearings. It turns out I could not find 40 degree angular contact bearings that have rubber seals, anywhere. So I went with radial ball bearings with rubber seals after the SKF engineer told me that their line of 6200 series bearings have an axial load rating that is 50% of the radial load rating. I was surprised that it was that high, but happy. That makes these bearings relatively cheap since they are so common. Higher cost bearings are usually higher rpm and have increased clearances and special low expansion balls for the higher rpm he told me. These bearings are 6209-2rs1. They are 45mm id x 85mm od x 19mm thick. So far my rpm limit if I was lucky will be 1000 rpm. More likely 300-400rpm. I have to do some testing of these servomotors for top speed to see how they will do. Maybe I can back calculate the rpm since I use the same motors for the cnc mill. If usps is any help, the bearings should arrive this Saturday. That way I can see how well they fit the od of the 5c spindle.
Thinking out loud. 135ipm. 20,000 pulses per rev. 20 turns per inch. Hmmm. 135ipm x 20 turns/inch = 2700 turns per minute or 2700 rpm. Hmmm. Much higher than I thought. And that 2700 rpm is at the motor while moving a heaving table and long ball screws. Although there is no cutting load. The limit for rpm so far is the pulse train out of the parallel port. My break out board is only 3 axis output. I'll need something more complex for this project. With a 5:1 belt reduction, 2700rpm becomes only 540 rpm at the spindle. 3:1 the rpm is 900. I don't think I turn more than about 400rpm on my lathe anyway. I'll double check that. Anyway, the rpm at the spindle with steep reduction I plan is slow. But the first project will be cutting delrin. So high rpm is not a requirement. There should be enough movement in the motor mounting to allow several different pulley sizes.
I tried fitting some 5c collets to the spindle last night. Not much success at first. The indexing pin was far too tall. So I reached in with a dremel and knocked it down some then deburred it the best I could. After some more fiddling it tightly fits a 5c collet now. The id of the spindle is the collet area seems to need some smoothing with scotchbrite or equivalent.
Just added the 5c collet knob and tube shaft. The knob was easy to design. I took the main diameter furnished with the indexer then subtracted 6 1" diameter rods inset just .125". This gives a nice gripping feature to the knob. But without super sharp edges. I'll had a hole for using a lever for added tightening as I have seen in the past.
-
Bearings, pulleys, and a belt are on their way! Hope to have the bearings Saturday so I can see how well the spindle fits them. Now I'm off to eat and hunt for aluminum stock.
-
Picked up a 25lb 1" aluminum plate just a bit ago. Their saw was down. Cutting the parts from this will be fun. Will be enough stock for the head unit. Probably need a bit more for the tail stock, that's not designed yet.
Mike
-
Added a few details to the head stock. The tail stock is started.
-
If anyone knows the code to make these pics smaller and clickable, I'd really appreciate it. These pics are just taking up far too much thread volume.
Mike
-
:)
Hi Mike, nice work !
For the pics., use the REPLY button then Additional Options then browse to your file (of the types listed) and attach it instead of embedding. Best to resize them to about 800 x 600 first.
I did your earlier ones, hope they are OK.
Regards,
Russ
:)
-
Thanks Russ! That's so much cleaner!!
Mike
-
Nice work, Mike.
50% axial load for a deep groove ball bearing... sounds too much to me too. Hope he knew what he was saying ;)
Dan
-
Nice work, Mike.
50% axial load for a deep groove ball bearing... sounds too much to me too. Hope he knew what he was saying ;)
Dan
The SKF bearing engineer sure sounded competent. He did specify that is was these larger 6200 series bearings that handle 50% of the radial load. Smaller bearings only handle 25% he said. I ordered the 45 x 85 x 19 bearings. Some of Mr. Simpson's bearings were 45 x 100 x 25. The ones I'm using are well within load and rpm specs. I just hope they are accurate and also that my spindle fits them well. Fingers crossed.
-
The bearings arrived today. The one I tried appears to be a press fit on the spindle. I could push it on by hand about 1/8". So tomorrow I'll go into the shop and see how much of a press fit it really is.
-
I tried to press one of the bearings onto the spindle. After I got it on there straight, it went on nicely. I pressed it on about 40% of the way and realized I need to probably press it into the front wall first, then press it onto the spindle. I have yet to machine the walls. At least the fit is good.
-
Check bearing fitting tables. The bearing fit in the housing doesn't have to be as tight as on the shaft. It's not an interference fit for that matter in most cases.
Dan
-
Hi Dan,
I've not heard of bearing fit tables. I'll look into them.
-
Here, for your reference:
http://www.ntnamerica.com/pdf/2200/brgfits.pdf
http://www.promshop.info/cataloguespdf/a049-057.pdf
Dan
-
I never knew there was so much science to press fitting bearings. I've always done it more or less by feel with success. Thanks for the links Dan.
I will be cutting up the pieces from the 1" stock here soon. Then we'll test my larger bearing press fit skills!
-
I don't have any photos to post of it, but I added a 4th axis to my machine about a year ago. Basically, I just constructed an upside down U shape out of 2" aluminum blocks, and mounted the headstock shaft to the assembly using 2 pillow block bearing assemblies. (With eccentric collars on each). For the shaft itself, I used 2" diameter stock, and turned the ends down on my lathe to fit inside the eccentric collars of the pillow bearings. This created a shoulder on the shaft, which the collars seat against.
Once I mounted everything up, I used a dial indicator the set the shaft to absolute center- then locked the set screws on the eccentric collars of each bearing. A tooth belt and gear assembly was used for installing the stepper motor to drive the headstock. I threaded a round-head bolt into the side of the shaft (between the two pillow blocks) which strikes against a microswitch to index the A axis to zero.
I used a similar construction in building the tailstock assembly. Both the headstock and tailstock assemblies are cradled between two pieces of angle steel. The headstock is of course bolted to these angles, while the tailstock slides along the length of the angles. The tailstock assembly itself is free to slide along the angle rails, and is adjusted by use of a lead screw threaded into the tailstock mounting blocks. The opposite end of the lead screw is inserted thru a second aluminum block upright which also slide along these rails. Two bolts pass through the aluminum block upright, (at the knob end) and thread into a piece of 1/4" flat stock suspended below the horizontal planes of the angle rails, which acts as a clamp mechanism to secure the position of tailstock.
This allows for easy (rapid) adjustments of the tailstock position from the top side of the machine. The lead screw has about 14" of travel, so plenty of adjustment capability. However, adjusting the position by turning the lead screw knob is a slow process. This was the reason I added the 1/4" threaded flat stock to clamp the assembly to the rails. For rapid adjustments, I simply loosen the two locking bolts, and slide the entire tailstock assembly to the desired position. Re-lock the two bolts, and then turn the lead screw knob to load the tailstock against the work piece.
The rails were added alongside the table, for the 4th axis assembly. Which allows use of the machine in 3 axis mode. I later added an adjustable limit switch (the positive Y axis limit) to allow for rapid setup for 4th axis use. I simply change the Homing from Negative, to Postive for the Y axis when preparing Mach3 for 4th axis work. This took some trial and error the set the Y axis positive limit switch in the correct position, so that the router's center is moved to top dead center of the 4th axis during Homing. Because of the distance Mach3 backs off the switch, once contact is made during Homing.
This assembly being a direct-drive system requires a LOT of torque! I originally had a 900 oz in stepper, which just wasn't up to the task. So I upgraded to a 1200 oz in stepper to drive the 4th axis -which worked.
After adding the 4th axis, and experimenting with it a few times, it has largely remained unused since. I do not have an encoder setup, but rely on the limit switch for indexing to zero on the A axis. Which of course, is not reliable enough for cutting gears, and such. But does function for doing things like small columns such as chair legs, etc. up to 30" long and 8" diameter. I use CNC Wrapper to create the 4th axis tool paths.
-
Hi adprinter,
You really should post pics and a write up in your own thread. Great projects like that are a help to many folks.
Mike
-
Found some 1.25" aluminum rod for the motor shaft adapter at a local supplier. Originally I was going to use steel. But a lower rotating mass is a better choice. And another piece of 1" flat stock for enough material to have enough for the tail stock. I'll lay out the parts on the flat stock to make sure I have enough material. I don't think my horizontal band saw is tall enough to hold a 9" piece. The standard back up plan is to use a hollow ground saw blade made for cutting aluminum in my circular saw.
-
The 80, 32, and 16 groove GT2 pulleys came in today. The belt was the wrong length. After looking at the calculator I am suspicious of it's accuracy. So I have drawn the belt path in cad. That showed I need about a 90 groove belt. Maybe longer would be better since I have 2 drive pulleys of differing diameters. The pulleys came from Econobelt. And the belt came from SDP/SI. At least I have lots to do before I need the belt.
My horizontal band saw takes at least and hour to cut about half way through the 1 inch al plate. I may have to use the skill saw with the metal cutting blade.
-
heat the bearing in hot oil and it should slide on
-
Certainly a temperature delta would help the press fit.
I had a tough time getting the basic parts cut from the 1 inch thick stock that I picked up from the recycling center. The horizontal bandsaw only cuts about 48% through the width of a piece. So when I flipped it over they were 1/4" apart. The band saw blade broke. My spare was broken when I slid the material in from the side and banged into it. I ordered a pair of blades from Amazon. They were from Starrett. The new blades cut quite a bit faster. But they now curve through the material instead of cut a straight line like the band saw used too. The blade maybe thinner as I noticed there was a gap between the bearing guides that seemed excessive. This saw cut dead straight for 10 years. So an alignment is in order. Also the weld on the Starrett blade was thicker than the blade itself. It would thump every time the weld passed through the material.
Now at least all of the pieces are rough cut. Each one just needs to be milled true then I can start machining the 4th axis parts.
-
The pulleys came from Econobelt.
Mike, my google-fu is usually not lacking, but for the life of me I can't find GT2 pulleys on the Econobelt website, nor in their downloaded PDF catalog. Do they call them something else? Their prices are very attractive.
Thanks for any enlightenment you can provide,
Randy
-
One of the tags that came with a pulley I ordered says this on it:
Quality Transmission Components
Division of Designtronics, Inc.
55 South Denton Ave., New Hyde Park, New York 11040
516.437.6700
One of the pulleys was part number QPMT5A16032F08 T5 Timing Pulley.
I don't remember which web site it was affiliated with.
-
Thank you for the quick reply, Mike. That's interesting--Designatronics is the parent company of SDP/SI (used to be called Stock Drive Products) so Econobelt is essentially the same company as SDP/SI. But you answered my question in that they're calling the pulleys "T5". Are you entirely sure they are GT2 pulleys? The Econobelt catalog page for the part number you quoted says "T5 profile per DIN 7721-2 Rev 6/89" which is a regular trapezoidal-tooth contour. A 5mm-pitch GT2 pulley would be called "5MGT".
Randy
-
Yikes! You found an error for me. The tooth profiles that I have here are trapezoidal. I'm short of time this week to do much about it. At least the belt has the right tooth profile, but it was too short. Sigh.
Man, it's not like I haven't done this a time or two....
Mike
-
I would use the trapezodial profile if you want tight backlash tolerance from the drive. I use the Gt2 profiles for Kart racing final drives and they are not as tight on backlash as one would want for a 4th axis drive.
Just My experence here your mileage may vary due to local conditions,
Just a thought, (;-)
-
Man, it's not like I haven't done this a time or two....
You and me both, brother. But it's a lot easier to spot other peoples' mistakes ;) I see that SDP/SI only goes up to 72 teeth on their 5MGT pulleys. And, they are inconveniently recessed on the sides, unlike their T5 pulleys. (Not to hijack your thread, but I snagged a Harig 5C spin indexer that I'm motorizing, which led me to your thread) BTW, the patent that defines the GT2 tooth and groove profile is http://www.google.com/patents/US4515577 There's a guy with a wire EDM machine who owes me a favor, and I'm thinking of drawing up a large 5MGT pulley for him to carve out for me...
I would use the trapezodial profile if you want tight backlash tolerance from the drive.
BR549, that is an interesting observation. From everything I've read (from the above patent to all of Gates' literature on the GT2 belt/pulley system) the GT2 has lower backlash in biderectional/reversing drives than the trapezoidal system, when the pulleys are properly made for the belt (i.e. the belt's tension members are on the pitch diameter and the teeth are only "cleats" for traction). Is your experience with off-the-shelf pulleys?
Randy
-
Look at the drawing in the patten app. Notice the air spaces between the tooth and the sprocket. With the Height of the tooth that space is required to reduce the friction on the belt as it slides in and out of the sprocket form. The Gt2 is a derivative of the older HTD and is a modified curvelinear design with MORE contact angle on the circular tooth than the old HTD but not as much as a trap design.
Minimum Backlash is a relative term(;-) Yes it IS less than the HTD that it replaced
Just my experience, Your milage may vary
(;-) TP
-
You make a convincing argument, BR549. I need to do more research (though your real-world experience counts for a lot! :))
I'm attaching a screenshot (highlighing mine) from the Gates PowerGrip GT2 Design Manual which has helped color my opinion (though since GT2 is proprietary to them they have a vested interest in pushing it...)
In lightly-loaded mechanisms I have traditionally used 80DP belts and pulleys with good result. But this thing will need to withstand machining forces and will be very low speed (quasi-static) vs. power drives that GT2 is aimed at.
Thank you for making me dig deeper!
Randy
-
HIYA RANDY we put a little more load on the systems than a 4th drive would. OVER 20 hp and 13000 rpm shaft speeds (;-) Have not broken a belt yet BUT have stripped our fair share of teeth off the belts over the years.
You used to be able to get the tooth design drawings from GATES. I have the 8mm series of the HTD. BUT they hold the new GT2 series TIGHT and you have to KNOW someone to get a copy to cut a new set OR LEASE/BUY the hobbs from GATES(;-).
WHAT you see in the latest CAD drawings is a generic profile.
Please keep us posted as to what you learn. I am all for learning something new. (;-)
(;-) TP
-
...they hold the new GT2 series TIGHT
Ah, but that's where the patent I linked earlier comes in. They show the dimensions for an 8mm pitch belt tooth and pulley groove, and you can get the other sizes just by ratioing the pitch. I drew up a 90-groove 5mm pulley last night. Just for educational purpose to verify their numbers, of course.
I wouldn't even think of attaching a DXF--given the patent situation that's something you'll need to draw in the privacy of your own house. :)
Randy
-
D'oh, I realized that patent 4515577 expired a decade ago. Based solely on the information in that patent, anyone is free to make appropriate pulleys...
Randy
-
To complete the design reference, http://www.sdp-si.com/D265/PDF/D265T016.pdf gives pulley O.D.'s for the various belt types and sizes up to crazy groove numbers. Those O.D.'s and the above-referenced patent give all the information needed to draw up any GT2 pulley you need.
Randy
-
I don't come here much, but I was posting a Q in the VB section and was going to add a few more pics to my 'mini machining center thread'. I tripped over this thread :o
Unfortunately I do not have time to really read the entire thread at this time, but in skimming, I found a couple of items that I may be able to clarify.
1) Bearings side loading: I also spoke to SKF engineering long ago and got the OK and the specs to preload their deep groove precision plus bearings. This makes something of a 'po boy' angular contact bearing although you benefit only from the zero backlash and do not get additional precision. In fact, the lack of precision can cause slight 'tight spots' in the rotation which would not be present in preloaded ABEC7 bearings.
2) There was mention of bearing install specs. This is a valid discussion, but the large OD of the 5C spindle results in such a friggin huge bearing, only the alignment spec (which is pretty wide, especially for the typical C3) really needs to be considered. If you have a 4,000lb armature spinning on these size bearings (a more appropriate use) then heat transfer and so on become important considerations. I am now completing an all new prototype 4th axis which will have for the first time a precision option with ABEC7 AC bearings. To meet that spec, the assembled frame will be line bored. I am going to do them all that way, but it is not needed for the deep groove ball bearings.
3) Belt Type: Technically, the trapezoid form is tighter and is still 'good practice' spec for positioning as was mentioned in this thread. How . . ever, other than being an easy tooth form to make pulleys for, it has nothing else going for it. They are very noisy and have far less power capacity that HTD or GT. Part of the good positioning performance comes from the tension spec, which is sort of 'cheating' in my view. There is a 'secret' to tightening up the GT that I will share here for the first time; cut the teeth a little bit (the spec I will not share) shallow. This compresses the teeth into the groves and eliminates the tiny amount of backlash that is intentionally built into the form. You can probably duplicate this effect in an of-the-shelf pulley by machining a very small amount from the tops of the teeth. Since this is likely to reduce the overall belt life, I am not recommending it . . . . I'm just saying . . . ;)
(http://www.thecubestudio.com/4thAxisSpindleProject/SpindleComponentsAll01WEB.jpg)