New Machine Build- My 8020 Build

[ClockMaster]

The time has come for me to build my first machine from scratch. The goal for this build is minimal tooling required and cost effective. The machine will be desktop sized estimated area 32"x24". All parts will be designed in SolidWorks and assembled before actual ordering and construction. This gives me the ability to scale the size up or down as well as make improvements and changes before the machine is actually built.

To me the most complicated part of any build is the linear motion. Rails and linear bearings can be quite expensive and require custom fabricated parts. My initial idea uses only 8020 parts with some spacers and bearings.

Please have a look and let the journey begin!!!

[ahren]

Interesting design, and clever use of commodity components, especially for the bearing adjustments. I would caution that aluminum is a pretty crummy bearing material. It will wear significantly with repeated cycles, and also tends to vary significantly in thickness along its length, so your wheels may lose contact at points along the travel. This is why we went with cold rolled steel for our linear motion system, which is designed to work with 8020 extrusions.

That being said, I'd love to see how this works out -- it's got to be better than 8020's sliding system, and might provide good intermediate-level linear motion, especially in applications where loads aren't that high (foam cutting, etc).

Best regards,

Ahren
CNCRouterParts

[ClockMaster]

Thank you for the input. I do agree with the fact that the aluminum will wear. A possible alternative would be to use 1"x1" square steel stock. Another alternative to prolong the life of the aluminum would be to use nylon rollers insted of direct bearing contact.

Keep the input and ideas coming. This will be a fun project to evolve.

Couple small modifications and the first version of the frame.

[ClockMaster]

Quick revision before bed. I changed the frame some and extended the slide rails to maximize the cutting area to the footprint.

[ClockMaster]

I've been doing some brainstorming and searching. I have another idea for the bearings but I dont know if what I already have is better or not. Using a 3/8ths steel shaft and a steel pulley this seems like it would work. I dont know how it would affect the precision but it would allow me to use 4 bearings insted of 6 per rail.

[vtx1029]

I just found this today and figured you may be interested.

MakerSlide Open Source Linear Bearing System by Barton Dring — Kickstarter

[louieatienza]

I think you have some interesting ideas. I experimented with patio door bearings and 1/4" plate that I radiused on each edge. Was not too thrilled wiith the results, though someone might be able to better implement it. I also tried sash bearings, which have a nylon v-groove outer ring, but the bearings themselves proved to be a bit inaccurate. I ended up building a machine from Solsylva.com plans, and was happy with the performance for the cost.

I would start with material that might be more consistent in dimension; Ahren mentions cold rolled steel like used in their LM rail system. You could also get ground tooling plate for a little bit more, which is more dimensionally precise.

[ClockMaster]

I am going to mock up a small test rig using the precision shaft with a nylon roller. The parts should be fairly cheap and allow me to see how sloppy it will be. The 6 bearings seem like they would be more stable but I am not sure of this yet. The pulleys would simplify the Y and Z axis but I would rather have less slop and better strength vs simplicity.

[ClockMaster]

I have decided that after looking into parts and different ideas that the simple bearing method will be the cheapest and easiest to design and offer the best accuracy for the price. I will continue with this design and down the road look for improvements. Tonight I will start the Y and Z axis.

[lovebugjunkie]

Misumi offers milled extrusions. Might be an option.

Aluminum Extrusions with Milled Surfaces | Mechanical Components for Assembly Automation - Misumi eCatalog

[pencilneck]

Originally Posted by ClockMaster Using a 3/8ths steel shaft and a steel pulley this seems like it would work. I dont know how it would affect the precision but it would allow me to use 4 bearings insted of 6 per rail.

Yes it works, this is how I've done it and I can only suggest it if you can't afford linear rails (as was the case for me at the time). When I use a runout gauge on my mill, I can see a wobble in my X and Y axis due to this set up. Now for my projects, it hasn't been a problem, but I'm sure the runout will not improve with use.

Also, I've run into some binding issues. I've found that if I spray some dry lube on the rollers, that prevents it from happening, but I shouldn't have to do this in the first place.

So... yes, it can be done, it is cheap and it does work... however, there are better ways and you may wish you would of saved up your pennies first time around.

[ClockMaster]

I found a source for some very cheap ($4) acetal pulleys at grainger. My plan is to fit them with some bearings and do some tests. I am willing to deal with some runout as this is my first build. This also leaves pleanty of room for improvements. I am favoring the pulley/rail method vs the direct bearing method. The Y and Z axis with pulleys and rails is greatly simplified. The parts count is cut in half and the design is much smaller and simplified. Everything for now kind of relys on how test rig comes out. I know this is more of a band-aid approach but its also a cost effective learning expierence at the same time.