I’ve been using MakerSlide for a while now, and I’ve gotten my ShapeOko working well enough that I finally feel comfortable in doing a review. As with all things, some are good, some are bad, and I’ll do my best to state all sides,
For MakerSlide — I both really like it and really dislike it. Here’s why I like it:
- It’s cheap. Cost is just fantastic for the rails. Orders of magnitude!
- It comes pre-cut to length ( a big deal ). Having cut T-Slot myself, this is a huge timesaver.
- It’s lightweight. It makes for very portable devices/
- It’s strong. Boy howdy! It’s probably stronger than a 2×4 or even a 2×6.
- It works. It really does. You can make things quickly and easily with it.
- DXF files are available — it’s easy to design with. I literally draw up everything first now, then build.
Here’s what I don’t like:
- Delrin wheels have poor axial load characteristics. The wheels flex, allowing cantilevers to shake, and allowing CNC mills to tip. I hear rumors of a heavily anodized, steel wheel system in the works, which should make it good at CNC as well.
- It needs adjustment after transporting in a vibration-prone environment ( aka a car’s trunk ).
- It’s very hard to get.
- Shipping can cost more than the material, depending on length.
- Carriages cost a lot. While the rail is cheap, the carriage plates are not. $2.85/wheel. $8.00/plate. A single direction of motion = 8 + 3(rounding up from 2.85)*4 = $20.00. If you include the plates and wheels, MakerSlide is not the cheapest linear motion solution. LM12UU or SBR12 systems are much cheaper, especially if China sourced. Not as easy to work with — MakerSlide is wonderful to work with — but cheaper.
For the most part, I’m at the point where I can’t be sure to recommend MakerSlide. There are some use cases where it’s the bees knees! Some where its not appropriate at all. You need to examine your usage and determine for yourself if it is the right solution for you.
Update — It’s no longer cheap or cut to length by default. It now comes in 1.2 or 2.4 meter beams, at $25.00/meter for the 1.2 meter beam. The cheapest you can do now is a $25.00 beam + $30.00 carriage — $55/axis. There are times when it’s worth the cost — like when strength/weight ratios are important. It probably has the best flexural strength/unit mass of any linear system. It’s also really easy to design with. So, I both like it and don’t like it at the same time. I do recommend it — it’s great to design a prototype with. But be aware — it may not work well for you. You have to really examine it yourself to see when it’s right to use it.
My ShapeOko review:
For the most part, the ShapeOko is a decent CNC mill. Here’s what I like:
- lightweight. Using makerslide really kept weight down, and its easy to carry around.
- Stiff. It’s much stiffer than a Zen or something along those lines. Slight modifications give you the ability to do steel, although slowly.
- Expandable. You can make it any size you need cheaply and easily.
- Modifiable. I’m using a Proxxon micromot as my spindle, and can do so thanks to how versatile the mill is.
- It’s a very fast and easy build. A “first timer” to CNC could have this built and working in a weekend or less. A person who is handy with tools could have it up and running in a few hours.
- Works with RAMPS at 12 volts. This is big, as you can use the electronics from a 3d printer, and drive a Proxxon Micromot spindle live from RAMPS using #2 butt connectors and a 1k resistor on T0. RAMPS is also very easy to source, so you don’t need to worry if there’s an outage of the recommended electronics. RAMPS can also run @ 24 volts or higher if you want — you’re not restricted to 12 volts. I recommend using the OSM N1728D63 motors instead of the SparkFun Lin engineering motors in the BOM. The OSM motors can take more current, run cooler, are faster, more powerful, and have a second shaft for an encoder. RAMPS has a lot of unused inputs — enough to put encoders on your motors and create a closed loop system. Cost is nearly the same. They’re 1.8 degree steppers, while the recommended steppers are 0.9 degree. The Lin motors are more accurate. However, I’ve found the Lin motors tend to stall out easier, and skip steps more often, at the same settings. The Lin motors are quieter — the OSM motors are louder, but still quiet. Whatever motors you use, use them everyplace. It’s no fun to re-flash your arduino when switching between a RepRap and the ShapeOko. If you use the same motors, pulleys, belts, and limit switches on all your bots, then you can keep the same software.
Here’s what I don’t like:
- Missing stuff. Minor stuff was missing in my kit. Screws, spacers, etc.. Nothing I couldn’t just grab a spare from at the hardware store for a few cents, but the kit felt incomplete. I also had a badly manufactured Z plate. It was too warped to use, and had obviously been CNC milled at too low a feed speed for the material. ( Lot’s of melted plastic “strings”, thick cut lines, visible chatter marks, crooked tool marks. I’d venture it was made on an early ShapeOko, before most of the bugs got worked out. )
- Tipping is a problem. This may be a function of my machine, as I had to modify it to fix some shipping/handling issues.
- Speed is a problem. I’m torn on this one. The machine is capable of very high speeds. 293 IPM, 7,500 mmPM@ 12 volts, and even faster at higher voltage. But the tipping means you can’t go very fast or very deep, lest you run into lash problems. I’ve found a good speed for wood to be 6 mm/s. At 8 mm/s, I notice tipping. Edward has it moving very fast while milling without tipping issues.
- Lots of space lost to the carriage plates. By default, the beams are about 14-” long, but only 7″ of that is usable. The carriage plates are about 7″ long.
- Weight. The carriage plates on mine weigh about a pound ( 450-480 grams ). In comparison, a 15″ long 1/2″ thick steel rod weighs about 380 grams. For milling, this is not a big deal, and I think this will be addressed very soon. for printing, it may be a killer. One problem is that all three carriage plates are part of the gantry, as are the motors — making the gantry very heavy. Inertial mass is a problem, and I don’t think the machine can function as a 3d printer @ 12 volts. I think it might work @ 24 volts. You’ll need to add a an 18 ohm high wattage resistor to all hot parts, otherwise no changes are needed. I’m still creating the test case, so I can’t answer this question for a little while longer.
- The workpiece is not supported on the machine. This makes the machine “lift off” or slide around, unless you clamp it down.
- Software issues. I use Marlin, and have found positional accuracy bugs in some cases ( Many large, fast, Z motions ahead of a single Large, fast X motion, will cause the X position to be substantially off, even though the steppers are unloaded ). You may want to stick with GRBL and modify the pin-out for RAMPS. This is not a ShapeOko bug — their recommended stack will work fine. It is very easy to work around this issue — use smaller Z motions. Printing uses very small Z motions, so it should not be a problem for printing. It is not a problem for skeinforge generated G-Code, as that uses small Z motions. It is a big issue for hand-written GCode used in milling. You’ll have to really test/modify your hand-written GCode to compensate for the bug.
Overall, I’m a fan of the ShapeOko design, although there are some issues, especially in my use case. I want it more for printing than CNC milling, and I no longer think it is possible. Someone else may get it working, though. The machine may be capable — I’m not sure. I think the designer is capable. Edward’s pretty cool, and pretty good — I think he’ll get it to perfection with time. The other problems can be solved or worked around, and I hear that it does make a decent mill. I still like the machine, and I still recommend it. Just be aware that you’ll need to spend some time adjusting the wheels, clamping things, you may need to have lower feed rates, and making it a 3d printer may be impossible. People are still working on these problems, and its entirely possible that changes under way right now will solve all the issues. My ShapeOko is an older design and has experimental components. A modern, stock ShapeOko may perform significantly differently.
UpDate — I recently got to see the design fixes made to the ShapeOko. The Z threaded rod is now on the side, so it looks like tipping is a problem for everyone, as moving the Z rod is an attempt to compensate for the makerslide wheels flexing. The right solution is to move the wheels further apart on the Z beam, or to heavy anodize and use steel wheels — or even angle grind a strip of O2 steel and use steel V624 bearings instead of makerslide. I got so frustrated with the problems of my ShapeOko, that I’ve given up on it, and started my own design. I really want a machine that doesn’t tip at speed, can print, has end-stop sensors, etc… I think some people will get great use from their ShapeOkos — it’s a great entry to CNC milling — especially if you’re the kind of person who likes to tinker with it unti its perfect. But I have spent too many hours trying to get good results, and never have. The closest I got was some very slow milling of wood, under the ideal feed speed (with very rough edges ). By slow milling of wood — I had to run it so slow, that it would take 5 hours to mill a small plate the size of my hand with 12 holes in it from 1/4″ plywood. I hear of happy people with good results — so I think it is possible. I just never got it working right, and I’ve hit my, “I don’t want to mess with this any more” limit.