Is there anyone here that attempted a 2.1m x-axis? Use case: Engraving and routing on a standard 1080 x 900 mm solid door panel. I read the articles on long runs, where timing belts slip, about pinion and rack methods, double belts, etc. My humble opinion is that if I just use a deeper tooth (GT3) 6mm wide, and control the surface area on the pulley, it should not slip. Example: If the pulley and belt only have a 1/3 surface contact with the surface, it will slip under load. The real surface is even less, because the outer teeth are already leaving the groove. I'm not an engineer, but here is my logic: Figure 1 is issue, too little surface area for long runs and where the combination of stretch and torque worsen the problem exactly where the belt make contact with the pulley (Entry point - let call it A). Though, in real example, it will be applied like in Figure 1B, which reduce the problem somewhat. (still and issue, cause of the slag that exist at idler A. Figure 2, better, but the surface contact area is still less than 50% (in the sketch its 50%, but teeth to groove, is less because the entry and departing contact is not 100%. Still and issue, cause the torque will "lift" the belt at entry point A. (all the slag over the whole length of the belt manifest itself at A, making it worse the longer you go or the less tension you have on the belt. Figure 3. my first choice. This is not just idlers to increase the surface contact area, but dynamic tensioners. Off course they will catch the "slag" as it occurs, and on top of it, it ensure full surface contact all the time. I will guess, that in practice, there will be 100% surface contact over 50% of the radius. (in the 2D drawing, its at 66%) Figure 4 is another way of looking at it, but this time its higher up, but with less friction on the belt (less radiuses for belt to pass through). I did not put more thought into it, whether this will just be idlers ensuring contact of the belt with the drive pulley, or just tensioners. That leaves an open question, what effect the tensioner will have on the departure point B of the belt? If it does, then Figure 3 is the worst scenario for departure and Fig 4 is better. Maybe with the smaller rollers even linear to the centre of the pulley. What else: This could be little bearings mounted to the mounting plate itself, or something more complicated as dynamic tensioners (rollers pulling to each other, and the belt's own tension pulling them apart) I think it is feasible. Will people with experience in such use cases please comment. There must be a solution to the issue of belts slipping on long runs without going to expensive linear guides and modifying v-rail profiles, and bolting metal racks to the rail. I also would like to make contact with people with axises that is 1500 mm and longer and with builders doing >= 400 mm z-axises, and 4th axis with rotating beds.
We have a 1500 x1500 axis and have found with the gt2 belts we have we get some slip under load, the guys here have given me ideas for rectifying this and I will keep you posted Your drawing and post has given me more ideas, thanks
What if you turned things around : make your X the width of the door, instead of its length ? The length of the door would thus be Y and you would need to figure a way to slide the door through the machine or have the belt tensioning issues on both of Y ... What if the door is moved through the machine like the phlatprinter moves a sheet of foam : pressure rollers ? Would it be reliable, material for a door is a whole lot heavier to move back and forth than a sheet of foam ... I'm working on myOX, a 1500mm X axis, 750 mm Y axis and targeting a Z which in theory could reach down to the floor (24" or so). My 'need' is to have a Z of 6" or so, with the piece being worked mounted on some 4th axis beneath the open table surface. To start, length of Z axis is about 10". Net actual work volume is currently approximately 52" wide (X) by 19" (Y) by 5" or so - taking away all 'wasted' space (ex: width of plates, tool centre position, ...) I'll eventually figure out a way to slide the work piece through myOX for anything longer than the 19" or so Y. I could also change my Y to full 1500 mm length of 20x80. Possibilities are limitless, I'm just taking it one step at a time for now. Initially, my 4th axis can be manual as I'm only thinking of machining mold bucks for my R/C monster truck and such ... Of course, the X carriage assembly and the Z axis will need to be redesigned to have a firm grasp of the tool when going far down ... However, with the 4th axis, Z only has to go half way down as the work piece would rotate ... This would also help compensate for a little issue of tool/spindle/router collision with work piece if going much more than a few inches deep ... myOX is not yet complete and it will go through several adaptations before it reaches its goal. The aventure is on ...
I've just built a 2.3 x 1.5M Ox (actual X-dimensions are greater than 1500mm - that's the available travel of the gantry). To cope with belt stretch over such a long run, I glued a complete run of GT3 face-up in the channel. This works like a rack and pinion system, which means the 'tensioned' length of GT3 is only about 6-8 cm, not the full 2.3M. This also gets rid of belt slip entirely.
A few others mention that approach as well. I might have to do same for my X axis. How is the 'top' length of GT3 fixed given there is probably no more room to use t-nuts ? For now, I use two T-nuts at either ends.
The lower belt doesn't need to go right to the limit of the extrusion, just further than the last wheel, if you see what I mean. I left enough room to get a belt anchor in place. http://kayutek.com/gallery/_data/i/upload/2014/11/18/20141118053726-1a24c469-me.jpg
hi all, new here.. i've been looking at the system in that foto, and realised which ever direction the moter travels along the axis, at leas one wheel is actually turning in the wrong direction against and onto the belt, would that be the cause of the belt slip? the loger the belt span on one side, the more slack there would be bouncing around enaballing an actual tooth jump at very long lengths? i.e. use two pulley wheels to feed the belt clear of the 2 delrin wheels.. be patient i'm new lol. (posably 2 other pulley wheels to make a sort of anti slip trip..two at the bottom fed from the motor, then 2 mor higher up to make like a 'U on eather side so the belt can flop all it likes, it has to wobble around two corners befor slipping over a tooth..)
I'm a fan of figure 3, more tooth engagement + proper tension = no slippage. Simple.... Not sure about spring tension to take up slack, your accel would need to be fairly soft in order for the carriage or gantry not to jerk on a start or stop using the spring tensioner like a shock absorber, cuts wouldn't be to accurate that way. My opinion, 2 horizontal slots for the idlers and just tension them once in awhile. Again simple..
I was considering - on the 2 belt system - milling 2 additional wheels down - width wise - so that they can sit within the slot and press the 2 belts together directly. In theory, that should mean any stress should be contained within the distance, from one pressure wheel and the relevant teeth, over the Pinion, and onto the next pressure wheel and relevant teeth. Gray
I have been thinking of doing exactly this. What glue did you use and could you post a picture of the way you mounted the pinion onto the "rack"? I'd love to see a picture of your build as well.
I get a 404 error when I try to view: http://kayutek.com/gallery/_data/i/upload/2014/11/18/20141118053726-1a24c469-me.jpg
Like a Bell Everman Servo Belt. correct? I read somewhere that if it is for own hobby use then the patent is not applicable. There is a whole discussion about this somewhere else. I also feel that is a good approach.
Bell Everman don't own the patent to two belts sitting on top of each other - their system is completely different (and about $30,000). Don't over-analyse things... just build them. Patent protection doesn't ever stop you making anything - just stops you selling something you stole from another designer.
Whooha... $30K, crazy, with that one buy a PCNC1100 with all the bells and whistles, or granite tables or miniral cast base. Thanks for the price feedback.
Seeing that the teeth of the belts facing each other has to mesh, what type of belt would you guys try first. (Its not that GT2 or 3 pitch profile has a match if they have to mesh with each other.
You'll find that while GT3 doesn't mesh perfectly, it meshes more than adequately when trapped beneath wheels in the trough of V-slot.
Here is Everman's patent publication US 20090301237 A1. https://www.google.com/patents/US20090301237 Everman's patent does not cover the $30,000 machine Zootalaws references, it just covers this $30 part of it (the type of gantry used to guide the apparatus along the rail is inconsequential to the patent) The purpose of the design is to reduce the area of the "tractor belt" (top belt) that is subjected to tension and stretching. In this design that area is 33 & 34 only, instead of the entire length of the belt if the "rack belt" (bottom belt) were not there. The patent does also cover a "tank track" concept wherein the top belt is a closed loop belt As with ALL patents, you are free to use the design in your machine as long as you do not sell the machine. You can also use your machine to produce products that you do indeed sell without infringing on the patent. A patent grants monopoly to the inventor only for sale of items that actually contain the design. There is a beautiful long (27 page) discussion of this design at http://www.cnczone.com/forums/linear-and-rotary-motion/59570-forum.html. Mike Everman actively participated in the discussion urging DIYers to build the design - for their own use. The discussion covers which belts are best to use (few belts are designed to mate with themselves), the need for multiple idlers, and weight/speeds that can be attained and contains many pictures of DIY implementations. GUS: The Everman design is used to move heavy loads, long distances, at lightening speeds. Your need for a router will cover only the long distance problem. To eliminate slippage between the pinion and belt you might try raising the idlers on drawing 3 to the drawing 4 position, thereby forcing/camming tooth engagement between the pinion and belt - While that would eliminate slippage you still have the stretching/hysteresis problem of the long belt (which is what Everman solves). Because of your lighter load and reduced speed, consider using a 20mm or 30mm wide GT3 belt instead of a 6mm one.
Thanks Tweakie! A couple of gems I got from that cnczone.com article worth mentioning for those who plan to use the Everman concept: Everman uses 3M double sided tape to afix the bottom belt to the rail (after trying every kind of glue out there). Duck brand also make an extremely thin double sided tape. There must be 2 idlers on each side of the pinion pulley. With only one the compressed contact area between the two belts is only 1 tooth and the belts will wear. Mike Everman suggested spacing each pair of idlers to span and compress at least 10 teeth. Also true for GUS's #3 drawing, the inside idler should be as large a diameter as your design can take. Belts are not engineered to withstand "backbending". A small diameter idler will cause separation of layers inside the belt itself. Everman found he needed no more than a 60 degree curve leading up from the idler to the pinion, with 50 degrees being what Bell-Everman products use. Note the larger diameter of Everman's inside idlers to combat this backbending Everman uses T type belts. A T type is a square tooth in a square slot rather than the curved tooth of the GT type. The problem Everman ran into with GT belts is that they wear simply because the curve of the tooth is not actually the same curvature as the curved slot. But, the teeth of the T type do not actually fill the belt's slot. While that appears to be a no brainer failure for T types, it's actually an asset. The fact is, no belts truly engage themselves. What Everman does is engage only one lateral side of the tooth/slot for positioning. He does that by making the pinion pulley/motor the adjustable part of the design rather than the idlers. As the pinion is moved upwards to tighten the loop, one lateral side of the top belt tooth is pulled against the corresponding lateral side of the bottom belt slot. The mating of those two sides is what holds the top belt in lateral position with the lower belt. Since this happens throughout the ten tooth length of the belt between the two idlers on each side of the pinion, this effectively engages and positions the two belts. Using that mating Everman can move hundreds of pounds at 3 meter/second with .001" accuracy for hundreds of hours without belt wear (another plus of using the 3M double sided tape is that the bottom belt CAN be replaced by scraping the tape off of the rail). That said, routers do not weigh hundreds of pounds and routing is not done at 3 meter/sec. For that application, GT belts appear, from the above comments, to give adequate long life and accuracy. GRAYUK: Suggestion, instead of milling down the width of the idler wheel to fit inside a v slot to engage the two belts, consider laying a metal or plastic insert on the floor of the slot to raise the floor so the wheel can engage the two belts. That way the bottom belt can be afixed to that insert - surely an easier task than afixing it to the slot bottom itself. That way the wheel can act as both an idler and as a v-slot guide wheel because the beveled edges of the wheel will still engage the bevel of the rail's slot. OK, I admit that I've not found such an off-the-shelf insert the right height and width - but it's out there somewhere.... That said, this cobble-together does work - Using thin doublesided tape, afix a GT2 belt to the slot bottom with the teeth up - then, using tape, afix two GT2 belt smooth sides to each other to form one piece with teeth top and bottom - engage that bonded set to the teeth of the belt earlier afixed in the slot - then use a 4th belt with teeth facing down as the pinion driven top belt. In otherwords, a GT2 belt doubled back on itself is the right height to act as an insert on a v-slot extrusion and still engage the bevel of an openbuilds solid wheel with the v-slot bevel for guidance.
Very nice Hussein. How did you determine where to drill the holes? Is it possible to move the wheels in closer and skip the idle pulleys altogether?
Here is some to scale drawings, but in PDFs, couldn't unzip the DWF's to upload. I have the OX kit from OOZNEST. So worked on the plates that came with the kit. Nothing extra, only drilled the extra holes needed. Also modified the Z axis, so there was no need to purchase extra wheels etc. Only longer bolts needed.
The new Z axis, and how to fix the new axis is how I ended up with where to add the extra wheels and pulleys. I used t nuts on the end of the wheel bolts to also secure a length of vslot. And then mount the new z axis kit(which is slightly larger than vslot and none the existing holes on the z plate would line up) on the vslot.
There is a thread somewhere on OB by the guy who patented this concept for high/speed high/weight pick-n-place machines.
I really have no idea. Theoretically it's unlimited as all that matters is the area between the compression wheels but success comes down to how well the two layers mesh and how well they are held together at the compression wheels.
