Overall, this ‘frame’ works fine so far - without the car batteries it would probably lift the corners under the retraction force. However I gotta say - it’s quite frustrating that I was led to believe that having anchors that are all low & at the same height was OK, because after having lots of issues with the machine wiggling while moving and having WAY too much downforce adding friction to the sled (resulting in motor overloads near the workpiece corners when belts were at their steepest angle), I finally decided to design my own anchors that make all belts level… and sure enough the machine worked A MILLION TIMES better. (Shortly after this I took it apart since I noticed one motor was wobbly, and found a bunch of loose screws. Used unit, no loctite was used apparently.) And despite that it was still performing SO much better than when I had the frogmouth anchors directly on the frame, at the same height as if they were anchored into concrete and those boards were removed, which is essentially what I saw claimed over and over is “acceptable”. I’m glad I tried this before giving up and putting the unit back on ebay.
I would recommend to ANYONE making a horizontal setup, don’t make the mistake of having all your anchors at the same dang height. I’ll attach my own personal designs here. Keep in mind they’ll only be the ‘right height’ if your anchors are like mine, mounted to the floor OR on this type of ‘frame’ about 3/4" below the surface of the spoilboard. If you made a frame with reinforcement plates in the corners, which makes the anchor bases level with the top of the spoilboard, account for that and shorten all of mine by that amount.
I used super glue and #8 screws to hold the anchors down (3d printed pla plus with a ton of walls basically printed solid) and some like 6" long stainless 10mm bolts just dropped in. Seemed great on that first test run, tiny bit of visible flexing but probably negligible. Once I get the machine put back together tomorrow, will recalibrate (since motor mounting bolts and drive gear set screws were loose) and try again … hoping for PERFECT results this time!
@bar awesome machine overall but I’d be happy if nobody suggested ever again having belts low, pulling the machine into the workpiece, at least in a horizontal setup. Doesn’t seem worth the risk to me.
Final things I almost forgot - 1) I didn’t realize until I took the unit apart that at full retraction the belt is getting pressed into the idler gear. This is a fail because the teeth get flattened in the same spot every time it retracts, just sitting there and getting smushed. Eventually these teeth will skip when extending, and the encoder will have no clue, resulting in dimensional accuracy issues. CUT YOUR BELTS LIKE 18" SHORTER than stock to avoid this. Last one number 2) the belt guides, both old and new design, have the belt teeth being pressed into a small round boss that a screw goes thru. It seems to me this surface should have a wide, smooth, sweeping radius (or a roller) to avoid belt tooth wear.
Last note, if you do change your anchors to all be level with the machine’s belt spools, you need to change the Z values in the .yaml to ZERO. (in machine interface, in fluidNC, click config items, and change TLZ TRZ BLZ BRZ to zero. I think they are normally 100, 78, 56, 34 or something I can’t check at the moment) and probably wanna recalibrate even if you were trying to put the anchors in the same place they used to be.
riser 4.step (83.8 KB)
riser 3.step (83.9 KB)
riser 2.step (83.8 KB)
riser 1.step (83.9 KB)
