You convinced me of cutting down on the size of my first build. Which I am about to start. I really want this only to cut wood or metal. Not for 3d printing. I got a local quote for .125 thickness welded seam for about a dollar a foot. Seems like that would be a good option to just cut wood and aluminum. I’m thinking 26 x 14 inches work area. And I am not dead set on the thickness. I may just go for the EMT anyway. One thing that got me curious is hardening. Could it work? I know that chilling with liquid nitrogen can further harden metal too. My thought is that hardening may bend the tubes. But I don’t know. Any comments will be appreciated.
They use concrete-filled steel tubes to hold up buildings for a reason. It definitely makes it stronger. Whether it makes it more rigid or not, which is what we care about, is another question entirely. Anyway, filling EMT with concrete would to make it too heavy for our use. To accommodate the extra mass you would have to upsize the motors, and then the belts, and then the stepper drivers. Overall it’s a bad idea. I think that Ryan has found a sweet spot with this design. Inexpensive, imminently hackable, and works well enough to push a router around at a fair pace.
Figure out a simple/reasonable/DIY-able way to make a torsion torus, and maybe you can fake a thicker tube/conduit wall to cut down on flex without all the weight…
isn’t the reason because they are stronger under compression? I can imagine a concrete filled tube being very strong as a vertical support.
There has to be a graph out there somewhere showing the diminishing returns of wall thickness in a horizontal beam, simply supported.
It comes down to balance, as always. thicker wall= slight improvement, filled, same thing. to what end? The more rigid it is the faster you can go the heavier it is the slower you have to go without adding a ton of cost in larger steppers, larger drivers to run them, Larger endmills, etc.
So yes, get thicker walls, fill them with whatever you want, just remember to lower your accelerations and you will be fine.
These machines are cost driven, 3D printer parts are mass produced and are extremely inexpensive, the second you move to something else the price move up stupid fast. You can chase improvements as long as your wallet permits.
This thread is exactly how these things should be handled. If you think your improvement works why not just try it. It took me 10 minutes and $15 to do it. As you can see it does not have to be fancy. I used a piece of string and a chunk of steel. I laid the rails over my chair and a table.
This. Concrete is awesome in compression, but shit under deflection.
They use CFTs in horizontal and vertical orientation. Think about what happens when you hang a weight on one end of a horizontal steel tube fixed on the other end. The top side is in tension and the bottom is in compression. Concrete is useless in tension but very resistant to compression. Steel otoh is better in tension than compression. So by filling the tube with cement you are essentially creating an inside out (and more costly) steel reinforced concrete beam, and these are used for most all horizontal load bearing members in construction.
Here’s what the guys in the white lab coats from Northeastern.edu have to say about it:
”The CFT structural member has a number of distinct advantages over an equivalent steel, reinforced concrete, or steel-reinforced concrete member. The orientation of the steel and concrete in the cross section optimizes the strength and stiffness of the section. The steel lies at the outer perimeter where it performs most effectively in tension and in resisting bending moment. Also, the stiffness of the CFT is greatly enhanced because the steel, which has a much greater modulus of elasticity than the concrete, is situated farthest from the centroid, where it makes the greatest contribution to the moment of inertia…”
Having said that, I agree with what Ryan said in another post. The simplest way to get a stiffer machine is to build it with stainless EMT, which is much stiffer than the mild steel used for conduit. But the mild steel conduit is probably good enough for machines of reasonable size, and it is a LOT cheaper.
I found a deflection calculator for various beam types, including round tubing. For a 48” (1.2m) tube 1 inch (25mm) diameter with a 10 lb (2.2kg) load, the resistance to deflection drops off rapidly after you go past 0.1-0.2” (2.5mm-5mm) wall thickness, or about 15% of the tube’s diameter. This makes sense since the greatest tension/compression is at the outside surface of the beam.
Also a hollow tube will tend to fail by buckling, not by reaching the material yield strength. Filling can help with that a lot, even if it’s just getting the most from the steel.
I’m skeptical about stainless being significantly stiffer than plain steel. Young’s modulus is not much different according to a quick google search, unless you can cite a better source.
Well, this guy did an experiment:
Ok fair enough. 0.049" stainless at 1 inch is substantially better than vs 0.049" emt at 23.5, and the diameter is close enough that apparently the modulus doesnt mean as much as I thought.
4 1/2 years later we are getting some real engineering done!
Anyone have links to some sort of international EMT standards? Maybe this is when we offer the 1" EMT? Now 3D printing is not very common, lasers would hate the reduction in accelerations but it looks like we might be on the verge of some firmware fixes for that anyway.
It is pretty clear from Ryan’s test that stainless steel EMT of the similar wall thickness has less deflection under load. But it is crazy expensive.
Young’s Modulus be damned? No, but there is “steel” and then there’s STEEL. The stuff they use for EMT, which after all is designed to hold up wires rather than buildings, is the former variety.
“According to the World Steel Association, there are over 3,500 different grades of steel, encompassing unique physical, chemical, and environmental properties.” (Here, if you’re interested).
So, when you google Young’s Modulus for “Steel” vs “Stainless Steel” you are not likely to come away with any useful information. Of course, Google doesn’t tell you that
I imagine EMT is actually designed to bend. isn’t that part of it’s designed purpose?
Here is another (probably bad) idea. I was thinking about the way they pre-stress steel cable in concrete to allow it to handle tension. I wonder if the same thing would work here. Get a length of 1/4” (6mm) steel rope cable. Swage on a rope-end 1/4-20 (6x1.0mm) stud on each end. Get two metal caps of sufficient diameter to fit over the end of the pipe. Drill a center hole in the cap, clearance sized for the stud. Make the cable long enough that the studs extend just far enough out the ends of the pipe to get the nuts on. Run the cable through the pipe and attach nuts to the studs, tighten to pre-stress the EMT.
Might be worth a try, anyway. Though in the end I’m not sure it’s any cheaper than using stainless steel EMT
I was just thinking of that too last night. Might set up some weird resonance with the compression load.
I have always wondered whether or not insulation foam sprayed in would help. At least dampen vibration.
But I have a 32"x32" build and it does wood superbly. I’m content with the design at the moment.
Sorry this is not good enough for me. I would believe that EMT is especially low for some reason but commonplace mild steel and commonplace stainless have nearly the same modulus unless you can cite something even a little bit authoritative. I am open to being convinced.
I appreciate there are lots of types of steel but we are not talking about HSS or O1 tool steel in the hardened condition.
Also a spring that is loaded keeps the same stiffness (spring constant) as a spring that is unloaded. This is true for belts too and is why high tension does not improve stiffness. I don’t see how stressing the EMT would be an improvement. It is not weak under tension like concrete, and even then it only matters if it is loaded to failure.
I just did my own quick and dirty measurement of the flex in my stainless steel tubes.
All my stainless steel tubes are already in my MPCNC, and I did not want to disassemble it. So I placed the gantry against the Y max end stops, and at the center in the X direction. The router bit was not touching the spoil board. I mounted a clock gauge right next to the gantry, so in the center of the gantry tube that is parallel to the Y axis.
The distance between the ends of this particular tube is 950mm. The tube is 25mm O.D. with 2mm wall thickness, 304 stainless steel.
Then I placed a 750g bottle of water on top of the Z stepper (with the stepper turned on). The clock gauge measured a sag of 0.040mm. With a 1500g bottle, the sag was 0.080mm.
When I put the 1500g bottle on the steppers that are on the outer frame, with the clock gauge in the exact same place in the center, the stepper on the side of the gantry showed 0.020mm sag and the other stepper showed 0.040mm. A 1500g bottle on top of the Z stepper would put roughly 750g on each of the 2 steppers on the frame. So between 0.010mm and 0.020mm of the 0.080mm of sag I measured using 1.5kg of weight could be attributed to other parts of the frame sagging.
So my best estimate is around 0.065mm for 1.5kg of weight, which would be 0.04mm to 0.05mm (0.0015" - 0.002") per kg for a roughly 1m long tube.
EDIT: the side with the larger sag was not firmly supported by a table, which explains the larger sag value (doh!). My final estimate still holds I think.
Is it really “crazy expensive”? My 304 stainless steel tubes are $4.20 per foot, for a total of $95 for my build (100x75cm frame size).
The EMT at my local hardware store would have been cheaper, but it was so crooked it made a banana look straight.
Even with the $95 tubes I could keep the total cost around $500 for the whole build.