Well I think gluing a belt that long and wide with a proper jig or the right equipment could be a pain in the ass so I'm thinking about what Roland sugested. To machine the rack in wich the belt will be running. That way I can guarantee zero strecth and perfect engagament.
I plan to make the racks (I may be can call it linear pulley? or infinite diameter pulley) out of 6061 aluminum. Then I can set them up like a normal rack but with the benefits of the servobelt system. I know this is not the quickest solution but I can machine the racks easily here in the shop so that's not a big problem. I'm also thinking about using the loop servobelt setup wich I think it's more convenient. Here's a picture of the setup: https://static1.squarespace.com/static/51396d24e4b01b4441c64c1f/51396d24e4b01b4441c64c2a/58f69bcff7e0ab3d0693dee9/1492556758738/2016-05-06+09.00.22.jpg. The only thing that I need to decide yet is the belt profile for the system, since I don't need the belts to fit into each other I can make the racks with any profile. I'm thinking about using HTD or GT2 but that's not decided yet. For the servo reduction I think the more suitable solution for me now are the planetary reducers. I found some that claim to have 5 arcmins of backlash so that's pretty accurate for my application. I plan to reduce the 750 W servos about 10 times to gain torque. Since the machine is going to be used to machine sheets of wood I think that will do the job. Also with that kind of reduction I can get rapids of about 45 meters/min wich is not bad. I still need to determine the final weight of the gantry with more accuracy (I estimate it will weight 300 kg aprox), I'll do that at the moment of purchasing the motors, and may be I'll increase the power to 1kw for the long Y axis. El mar., 23 abr. 2019 a las 5:05, Les Newell (<les.new...@fastmail.co.uk>) escribió: > I quite like the servo belt idea but I think it needs pretty accurate > clearance between the fixed and moving belts. Note that you can't clamp > the fixed belt. Clamps cover the teeth. If I was doing this I'd look > into polyurethane based adhesives. For instance the stuff they use to > bond in car windscreens sticks to just about anything and is very tough. > Devcon also make some impressive 2-part polyurethane rubber adhesives > that are very strong and adhere well to both rubber and metals. > > An unsupported belt that length will have significant stretch. I used > some kevlar belt specifically designed for positioning on a CNC drill > about 10 years ago. It was about 4m long and I saw positioning errors of > up to 0.5mm under load. As the spec was +/- 0.5mm on that job I just got > away with it after using a little bit of software compensation. On the > plus side it has been running 10+ years in a production environment with > no detectable wear. About the only issue I ever had with it was a pulley > that came loose. > > If you go for rack, look for 'precision' rack. The cheaper racks are > pretty rough. They aren't designed for accurate positioning. You > mentioned putting the feedback encoder on the pinion. This is a bad > idea. Any backlash in your gear train will make it really hard to tune > the motors and the motors will tend to oscillate when stationary or > moving slowly. This will put a lot of strain on the gear train. Gears > don't like repetitive reversal like that. I have worked on a lot of rack > driven routers and they all used encoders on the motors. Most used > direct drive to the pinion with huge motors or single stage belt > reduction. Are you looking at brushless motors? If so they will have > built in encoders. > > You are thinking of using 2 module 30 tooth pinions and your motors are > 2.4Nm continuous. Looking at the specs of other 750W motors you have > about 7Nm peak. It's not generally a good idea to use all of that but > you can use say 4Nm for acceleration. > direct drive: > 2.4Nm / 0.03 = 80N > 4/0.03 = 133N > That's not really enough. Let's assume a single belt reduction. About > the most you can sensibly get with one stage of belt reduction is 5:1 so > you now have 400N continuous, 665 peak. Double that if you are using two > motors. It's not up to the sort of forces a large industrial router can > generate but it should be plenty for most applications. It is unlikely > you will use any where near that even taking friction losses into account. > > What sort of work are you going to use this router for? If you are > processing sheet materials I would strongly recommend using a vacuum > bed. If you are using blanks, vacuum pods work well. > > Les > > > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
