The steel platen behind the belt has been a bit problematic since the start. As you can see in older pictures, it was a piece of angle iron that was on the left side attached to the belt-grinder arm with two screws. The problem was that when I exerted pressure on the right side, it bent ever so slightly and the belt started to wander off and the whole thing behaved a bit unpredictably. I learned to work around the problem, but it was a problem and it needed solving. So this week I finally solved it.
© Charly, all rights reserved. Click for full size.
I bought two hinges and screwed them on a piece of 3×40 mm mild steel at a distance slightly bigger than the thickness of the belt grinder arm. Then I drilled two holes through the arm and put in them two carriage bolts with winged nuts. Then I cut slots into the hinges that I could slide under the screw heads and the nuts. Because the hinges are slightly more apart than the thickness of the arm, when I tighten the screws they work as flat springs with a slight bend and it all holds together very nicely.
To serve as the platen I have screwed a piece of 10×50 mm mild steel on. It works perfectly. It is stable regardless of which edge of the platen I work on and I can exert as much pressure as I want, although admittedly, I have only ground bone on it and not steel. I might either harden the platen by carbonitriding it or I can weld a hardened steel plate on the face, but mild steel works just fine too and with my rate of making blades it will hold for a long time as it is.
There is one last problem that I need to solve, and unfortunately, I do not know how. It is also difficult to describe, doubly so in a foreign language, so if you do not understand what I say, the problem is probably not at your end.
I need a reliable and stable way to adjust the belt’s traction. The spanning arm was a bit wibbly-wobbly, and when I fixed that, I got new problems. When I optimized it for a forward-running belt and tried to run it in reverse, it wandered off to the right and fell off. And when I optimized it for a reverse-running belt, it wandered to the left and bit into the arm. After a lot of faffing around I managed to get the belt stable in both directions – but then it was off-center and the sideways position of the spanning wheel had no influence on it whatsoever. The spanning wheel is crowned and that should make the belt tend towards its apex and thus I should be able to move the belt sideways, but it did not have any influence – the same thing happened whether it was near the arm or as far from it as it could go. Only the tilt of the wheel had an influence, but when I fixed the wobbliness, I lost the ability to tilt it.
I have managed to get the thing running by bending the screw on which the spanning wheel is fixed. When I turn the screw now, the wheel tilts, and I can adjust the belt. The problem is that it is too responsive and not very stable, just like it was before. I am wracking my brain for solutions, but so far I have not come up with any that are doable with my equipment.
Usually, belt sanders have a “blapper” (not real name) that is hinged so you can tighten it with a screw through the base-piece -- that changes the angle, which adjusts the running position of the belt. Most of the current sanders have a tensioner with an actuator.
This is typical
Could it be that one of the axes on which the wheels run are too much out of parallel?
Do you have the ability to adjust the angle on one of the axes?
In this video where Matthias Wandel is building a strip sander out of wood, you can see him building and operating such a mechanism for the strip sander. Around 3:20 you can see him operating it, and you can see the change in makes in how the belt runs.
Note that this is not the same wheel with which the belt tension is adjusted.
He has another video with a complete build of a similar machine, also out of wood.
@Marcus, thank you for the link. It looks similar to the system I had before, only better.
@rsmith, the problem is almost certainly because the wheels run out of parallel. My belt grinder is not very precise. Thank you for the links, I did watch Matthias Wandel’s videos about building his wooden power tools in the past. Unfortunately, this exact thing is not doable for me without building a de-facto completely new machine, but it is something to consider.
The machine works for now, I will think about it and maybe I will get an idea of how to implement these concepts into my existing machine.
Measuring parallelism of multiple axles accurately requires expensive tooling.
However, if you could find a flat reference that’s long and wide enough to cover
the two axles with the longest distance between them you have a reference
surface to get a good impression of what’s wrong.
Anything flat and stiff enough to not deflect under it’s own weight will do.
E.g. a piece of wood planed flat, a piece of thick plywood, an aluminium
extrusion or a rectangular steel tube.
Lay the reference over two axles and press it against one of the axles. If the
other axles lies outside the plane of the reference, you should see a gap.
If you don’t see anything then at least these two axles are in the same plane.
That does not necessarily mean they’r parallel though; ideally you’d also measure
the distance between two axles at the base and the tip which would require a big vernier caliper.
But by measuring with the flat reference between all the four axles, you should at
least get a decent idea of what is crooked or not.
And then its time for some “percussive maintenance”. :-)
Interestingly, I suspect that Wandel’s machines have some advantage in being made out of wood. I don’t ordinarily think of all wood construction as being the way to go, but he is obviously very good at this, and the stuff he’s using is very rigid.
I’ve always found sanding belts touchy and critical, and prefer a disk or a wheel if I can get away with it.
Thinking on this a little further, I thing it’s important to remember that every pulley must be prevented from flexing in all axes, so even if it’s well trued up and down, it might not be so well trued forward and back. This was a problem I had with an older Grizzly 6 x 48 sander I bought years ago (fortunately used and cheap, probably because its former owner found it so annoyingly deficient).
I was just looking at some stuff on the web, and decided to look and see what Grizzly is offering these days. It turns out they are still offering essentially the same machine 25+ years later, but….with a big change. I feel vindicated. The one I have had only a single thumb-screw attachment at the top, and because it could twist on the flat springs that serve as hinges on the bottom, it never tracked well. I added a wing to the bottom, with adjusting bolts, to prevent it from twisting. Note that the current model now has two adjusting screws on the top, which effectively do the same thing.