hi there, any one else suffering x axis twist on their 2x 2060 profile 1500mm axis? I'm getting different depths of cut in one direction than in the other direction on x due to twist In the x axis. Really frustrating, I've made corner brackets to support x on y gantries to shorten the span. Still same amount of twist. I've bolted them together. But you can still literally twist axis with a small push on the z axis. Completely out of ideas now. Somehow I don't think this design is up to being this size. Thickest material I've managed to cut accurately with it is 3mm ply in 2d. Even adding 2mm heigh tabs makes x twist and z kick out. Adding 6mm thick ali triangular corner brackets to the internal corners below the x onto y gantries, are solid. The twist only seems to happen in the span between the brackets. Which tells me it's simply not up to it. Any ideas??
Did you put any pictures of your build on the Forum earlier? A look at them may give an idea as what is wrong. Generally, I would say any twist would initiate at the junction of Z and X, this being the weakest point in the set up. As the torque is being transferred from left to right, there could be a variance in the depth due to lateral twist at the junction. However, this would not explain clearly why the different depth of cut would be maintained whilst going in the other direction. Indeed, if the aluminium was bending, there would certainly be a spring back response once the new direction was established, and normal depth returned. I would definitely look towards the Z carriage and that general area for some movement, and a degree of "twisted fixture" in that area. Bring the Gantry right to the front, and maybe set a program to shunt the X axis, back and forth, over say 6 inches. Get right in there, with a bright light and maybe a video camera, and watch for that lateral movement. Perhaps try holding the non running router or spindle, and see if you can accentuate the problem. Maybe you will see it. Gray
Hi grey, thanks for response. Yes there are pictures here: http://www.openbuilds.com/threads/ox-max-build-with-mods.626/#post-6949 Initially I thought the same thing and noticed some slack on the top leadscrew collet which must have slipped. Something I have found with all the collets and pulleys supplied with kit. For which I have had to tap bigger grub screws and file flat spots for them to hold up. The ones on the leadscrew I have now replaced with two thicker versions top and bottom. On putting the z back together I made sure there was no play between collets, and tightened all eccentrics in x and z to their max setting. In hope it may fix the issue. I then ran a facing program at a depth of 0.5mm on a sheet of mdf using a new trend two flute straight cutter, and played with feedrate to find it had no effect however did notice the difference in depths was more in the center than near the edges. which says its definatly twist. is it my bed is too low., I already have 3 18mm sheets that make up the bed, any heigher and im going to be restriced to only using short bits. so not really a solution. am I asking too much from this cutter? its pretty clear it strugles to be stay rigid over 1500mm. I see robocutters have now modified this version to have double the number if profiles. am I really going to need to shrink it to get any results out of it?
Oh yes, I remember your build. Especially that excellent support framework. Anyway. How about, Still using that 6 to 12" left to right program, set at exactly the same cut depth to create the effect. Then, Move the Z as far to the right of the Gantry, and run the program. See if the same effect is experienced. Do like wise on the far left. Then the middle. See if the fault is repeated across all the gantry. I need to think about what that would prove. If the fault is only in the centre, it would prove rigidity at the ends of the X axis when near the Y axis, and maybe prove the twist is in the aluminium. If the fault is present at all points of the Gantry I would say the Z axis as before has got something to do with this. Did I understand you right, and that the new depth of cut is maintained in the new direction, or is it momentary? Remember, many others have built to the same spec, and not had this effect appear, so don't give up just yet, it must be something specific to your machine. I know the latest flavour at the moment, is talk of putting another length of aluminium along the X axis in a Horizontal position, creating an "H" crossbar Gantry. This deals with many of the problems brought about by Torque. I think the actual width of that particular piece is not the important factor, but the fact that it is counteracting twist of which you may be experiencing. Try those ideas out and get back to us. I'm sure others will have other useful ideas. Gray
its bafling me gray, its only in the middle half of bed in one direction. whats even mire bafling is if I were to face in a squre using climb milling ie working inwards, with the stock in the center of the bed. exactly through the middle of the square will be a step. now I thought as it set to climb mill, the cut will be going in the same direction against material irelivant of its actual direction, so the forces on the bit should remain the same. definatly a good idea to see if the issue exist onky in places or if its exactly central or not I will retest and let you know. cheers
Keep the direction of cut consistent to start with, when running the test. Lets test the stress factor on one cut direction, but two travel directions. Then alter the cut to see if anything significant occurs. I know this is really a stupid question, but it needs to be asked. Your work bed is definitely locked down in the middle? Gray
By the way, you are only going to get simple questions and simple answers from me! What you might need is the help from the Techies! You know who you are. So, come on Guys, from what you have read so far, what do you reckon? Gray
Another good idea, will give it a go valid question. its 3x 18mm sheets of mdf glued together and screwed to frame from underside. so solid as a rock. top has had 0.2mm faced off so is flat to cutter. once again has the dreaded steps in it grrrr
Have you re adjusted the exsentric spacers on the bottom side of your x-carriage, after mounting the z-stage and putting the whole router together. The v-wheels on top of the carriage would sit thight as you show in the video, just because of the weight of the carriage.... Johan
A suggestion on a fix for your twisting problem. Try using a "Fletch Beam" That is where you bolt a 1/8" piece of flat steel the same size as the extrusions between the 2 extrusions. Thus you gain the strength of steel from twisting and the lightness of the extrusion. Bob
There is a technical term for it! Then it's got to be a winner. Had thought about this. Does mean I will need to adapt the holes in my y gantry plates to allow for extra width and add longer bolts and extra spacers to the x gantry assembly. But bob I think your right. I can see any other way of doing it. Still again I ask the question, am I really the only person suffering with this? Have I got some dud profiles?
Actually, its a "flitch beam". (Fletch is a movie character.) Flitch beams only help with vertical load. The best way to improve torsional rigidity would be through bonding the two extrusions together. Some recommend JB Weld, some don't. I'll let you do your own research on the subject. The real key to success is surface prep so read up on that too.
Well actually I was proposing epoxy but if you think you can weld them without distorting them go for it.
Good point. I wouldn't be doing it myself. Will have to speak to an engineering shop about that one. Cheers
Welding will always leave a difference in the project. Since we are trying for sub-mm accuracy, I would think epoxy or JB-Weld might be a better first solution. David "The Swarfer" propose splitting the rails and putting the router in between the rails. If heavy enough, it should halt the tipping you are seeing at the Z axis. Iw ould require the gantry plates to be reengineered, but shpuldn't be too hard. For no more that a piece of V-slot costs (compared to ruining a piece of work), you could double up on 2 20x60 or 20x80 on each side of the router mount. Strong, and the router would be able to straight plunge rather than tilt. It might screw up anyone with plans for a 4th or 5th axis, but that is not most of us.
Could you not drill through both extrusions and bolt the thing together? The t slots on the front and rear faces of the X gantry aren't really used for anything and will clear any fastener with a head smaller than about 6mm. You could bolt on from one side and use T nuts on the other. It possibly won't be quite as consistent as the glue bond, but it should allow a level of adjustment if needed that the glue won't give. As far as the issue being present on other peoples machines, I'm sure it is there (I checked mine and it does flex a bit). The main issue with your machine as you pointed out earlier looks to be the large moment arm that is created with the long Z axis. Normally the standard part is about 200mm from memory, with the support V-wheels spaced at around 100mm. So at the extremes on the standard model, the Z is roughly 50/50 supported to unsupported. Your machine looks to have a Z axis around 300mm so roughly 30/70 supported to unsupported which gives much more leverage on the v wheel supported section and also your X axis extrusions. Buuuuuuuuut I'm sure this is all old news (I'm very new to this cnc thing).
Thanks for the advise. After second thought, and heeding ricks advise about not welding as it may cause warp. I bolted and epoxied the profiles together. Which seems to have done the trick has reduced the flex by around 75%. Wish I knew to do this in the first place!! I would say 20% of remaining flex is in the z gantry assembly now. I'm thinking adding extra, vwheels at 90 degrees to the existing ones, in the v slots on the front and back of the z profile would help. Just need brackets to attach to the existing extrusions. Cheers
It would be nice to be able to buy beefier profiles like 60x60, 40x40 and 20x... wait - scratch the last one ;-). Is there a small chance that you made photos of the process?
I haven't yet, because I haven't completed my build, but looking at your axis with a (relatively) light Kress router, I suspect I am going to have a whole lot more deflection than you What I have done, though, is to mount the router flush with the Z-axis, whereas yours is quite a bit off that, giving a significant amount of leverage. I have reinforced the two 60 extrusions with some 20-series extrusion plates that slide into the track. I have drilled and counter-sunk and bolted these steel plates through the two extrusions and straight away I noticed an improvement in rigidity. I have a couple of litres of boat epoxy which I could use for further stiffening, but it takes a few hours to set and 24 hours to properly harden. I would have to waterproof the two sections first - not so hard, just use silicone sealant. It will peel right off after I am done. It would add significant weight, though. Something I suspect my GT3 belts won't appreciate. I am seriously looking at the dual-belt system to reduce the chance of stretch.
we have a 1500x1500 machine and it was supplied with twin x supports perhaps there was a problem,perhaps your supplier can see about upgrading your machine with new beams and end plates
Afraid not, I simply clamped the two together, drilled a 4mm hole through both profiles every 200mm in center slot, tapped and bolted with flat low profile head hex bolts. They don't fit into the slot but enough to locate the profiles together, and enough of the head is touching to get a decent amount of Torq on them and they clear the x gantry plates. You could notch out the profile so they sit inside the slot. I then split the profiles, mixed up a whole syringe of metal epoxy and applied a thin coat on one of the profiles and bolted back together. Using clamps between bolts. Is very strong now. Did notice the metal epoxy stays fairly flexible when set. I would say a boat epoxy may set too rigid and may not have the same adhesion.
I am currently experiencing the same problem you had. How, if you corrected the problem, did you correct it?
