I agree with that. Looks like it's missing some parts. Should be a triangle structure to the front bearing that should be out in front of the gantry for a good "stance"
[That's a good idea; I'll have to mess with my CAD model to see how it looks, but it doesn't seem unfeasible.]
Hmmm, I can do some math on that if you can figure out what the gear ratio is and the inertia of the harmonic drive, and the servo (should be published in the manufacturer literature) and the torque vs speed curve for the servo. Off the top of my head, yes, 100:1 would not work, but these things come in all kinds of ratios, like 3:1 or 5:1 or 10:1.
So looking at the basics, if your servo was spinning at 3000 rpm, and you had a 100:1 gear ratio, that's 30 rpm for the ballscrew. I assume you mean 1 inch lead ballscrew, that's 30 IPM speed, really slow, but the big problem is the acceleration. Because the acceleration of the screw will be 1/100 of the acceleration of the servo, and the servo must now contend with the input inertia of a harmonic drive at the same time. Like I said, I can do the math on it if you want, I'd just need more info.
[Yes, I've resigned myself to using those actuators to position the 5th and 6th axes, where speed isn't much of an issue but holding torque is. I need to find something else to slow down the XYZ servos which has a more reasonable ratio, but doesn't add backlash. I'd like to get XY rapids of 1000 ipm if possible, and it should be, I think.]
If you're looking at the gear ratio to prevent backdriving, I think you should also consider wearing out your ballscrew.
[Backdriving is definitely a concern with the 5th and 6th axes, since they need to hold a heavy assembly with a lot of counter-leverage. But I didn't think the Y axis ball screw would have any particular wear issues, since it's a fast screw (1" pitch), and the Z axis screw will be similar. Why the concern?]
Your Z assembly will be quite heavy, I don't see a good way out of using a pneumatic counterbalace for it.
[I was resigned to that; the gas springs I've implemented work pretty well in a couple of my other machines.]
So what kinds of speeds do you want your Z to move at?
[It needs to have good acceleration, but ultimate speed is less of an issue, since Z axis moves are usually pretty short. 300 ipm or so should be fine, if it's responsive.]
It's pretty easy to calculate the deflection of a tube with a point load in the middle, once you know what your gantry assembly will weigh. I believe it will be very small, even with a heavy gantry. 10" high beams, what wall thickness, did I miss it?
[I'm dithering between 1/4" and 3/8" wall thickness on the X axis beams; I'll probably go thick, since they won't be moving once they're in place. The gantry beam can't be that thick, though.]
I believe the mori machines do some kind of pretensioning, some patented process, to eliminate any sag, I read that somewhere, never did figure out what it is exactly.
[I'm still considering using steel cables to pre-stress the beams, just hoping they won't add weird twists. ]
How are you planning to level the surfaces you mount your rails to on your two long gantry supports? Do you have a place that will machine them? If you use epoxy leveling on each beam, you know you could always mount a weight hanging below each beam (in the center) equal to half of the gantry weight (including Z axis, etc). Now when you take the weight off, the center will be higher by the same amount the gantry will deflect it downwards when it is in the center position.
[Another good idea, thanks. My only concern is that it will complicate the mounting of my racks, which is complicated enough as it is. I was planning on epoxy to make flats for the X and Y axis rails; nobody in town has a grinder that is long enough for X, and people tell me that machining on the hot-rolled tubes will distort them anyway. ]
Total overkill I think, just use the big beams and you'll be fine.
[I was hoping that would be the case...]
Yep, I would also try Jim's new system if using servos. Are you going to use linear encoders?
[No, it seems like overkill; I'm not making aerospace parts on this thing.]
Unless it comes flatter than mild steel, I don't see it being beneficial, but that's just my opinion.
[Stainless is supposed to meet higher straightness specs than hot-rolled steel - it had better, for the price...]
For other machine designs, It really depends on the weight and geometry of the machine as to how much it will flex under it's own weight if it's susceptible to changes in the floor. I think yours is particularly susceptible, but I'm not an expert on ground heave. You have designed some adjust ability into it, so, I really don't know. I also don't know how much flex you will get in the threaded rod. I like the adjustability but I'm not really keen on those rods for some reason. I don't really have a better suggestion though.
[I've seen freeways held up by similar systems; I don't see them flexing much in compression.]
It depends on the robot arm. I'm talking about an industrial robot arm. I just watched a video on Youtube of one cutting hardened steel. But it was painfully slow. I also watched one of a robot arm cutting aluminum. Also very slow. But it did a good job. Clearly there are advantages to traditional machine designs vs robot arms or they wouldn't still be making them.
[Most people aren't making very complex parts, so robot arms might not justify themselves in most cases. I admit to being intrigued by the things, but frankly they frighten me. I've had runaways on other machines, but while they tried to tear themselves apart, they didn't have the opportunity to whack me upside the head...]
What are the advantages of making a 6 axis vs a 5 axis machine? I would think the 5 axis (including rotary) would be easier to make and could cut any shape you could think of?
[It has to do with what the tool can reach. If you can only tilt the head back and forth, you can reach sideways undercuts but not ones on the front or back sides; that requires swiveling the tilted head around 90 degrees. I want to be able to cut natural shapes with random undercuts all over them.]
What do you plan to cut with this, what materials I mean?
[Wood mostly, also plastics and foam. I doubt it will be rigid enough for aluminum once the 5th-6th axis head is installed, although without that it might stand a chance. I'd also like to try cutting marble with it, if only to see if I'm right about the rigidity required for that.]
The way your 4th axis is mounted to the floor isn't adjustable for heave like you are talking about for your long beams. Perhaps it's not a problem anyway.
[I haven't actually noticed the concrete floor moving around, although I suppose it's possible. I have observed some seasonal shifting elsewhere in the neighborhood as the clay layer underneath gets saturated and dries out. It might not hurt to mount the rotary table on rods like the beams get.]
I like the simplicity of your design, there are many things I like about it. I see it being a big project, alot of work though to finish the design, and build. Very ambitious as was mentioned. It seems like uncharted territory, very risky to travel there, but also could be very rewarding.
[Thanks, and yes, it's definitely going to be a challenge pulling this off - but how else am I going to be able to have the capability to make whatever-the-heck-I-want?]
Have you seen any other DIY designs like this?
[Not really, although I've absorbed a lot from reading posts here for the past decade or so and incorporated some of those ideas. It's more similar to some large-scale Chinese machines I've seen, like this one:
http://image.made-in-china.com/43f34...es-Pillars.jpg
Oh, and here's another robot arm I found on YouTube: