backlash

[Ken S]

Early in the KJ jigs online class, as well as other times, Wolfgang mentions the importance of making the final microadjust setting by raising the support bar. This means raising rather than lowering the support bar. Wolfgang is quite correct in this, although I have never seen him explain why.

The reason for this is backlash. This is certainly not a unique problem with Tormek. All adjustment screws have this problem, even with premium woodworking tools such as the depth setting screw on Lie-Nielsen bench planes or on top of the line metal lathe feed screws. We can feel it with our support bars. Lowering the support bar by just turning the microadjust usually requires tapping down the support bar. Raising the support bar by turning the microadjust is the more controlled method. The support bar is raised solely with pressure on the threads. This is called backlash.

This is an example of a small gain from doing something right. By understanding the comcept of backlash and making the effort to always do it right, it will soon become a habit.

Ken

[cbwx34]

Early in the KJ jigs online class, as well as other times, Wolfgang mentions the importance of making the final microadjust setting by raising the support bar. This means raising rather than lowering the support bar. Wolfgang is quite correct in this, although I have never seen him explain why.

The reason for this is backlash. This is certainly not a unique problem with Tormek. All adjustment screws have this problem, even with premium woodworking tools such as the depth setting screw on Lie-Nielsen bench planes or on top of the line metal lathe feed screws. We can feel it with our support bars. Lowering the support bar by just turning the microadjust usually requires tapping down the support bar. Raising the support bar by turning the microadjust is the more controlled method. The support bar is raised solely with pressure on the threads. This is called backlash.

This is an example of a small gain from doing something right. By understanding the comcept of backlash and making the effort to always do it right, it will soon become a habit.

Ken

Do you mean at this point?   

https://www.youtube.com/watch?v=JFHqQg5wZEg&t=830s

I interpret it as just being "easier". 

I agree that the USB can hang if you go the other way, (sorta binds between the two posts or sleeves), but I gotta wonder where you got "backlash" from in this scenario?  I looked up the definition and how it relates, but didn't see how it fits here.

[BeSharp]

Absolutely. The biggest upgrade to my milling machine was installing DROs - Digital Read Outs. It allows me to compensate quickly for the backlash in all the feed screws.

That's why I measure by shimming underneath using 1-2-3 blocks (actually, 25-50-100mm blocks), gauge blocks, and drill bits. Gravity does the rest.

I'm in the process of making custom "gauge rods" using the lathe and the milling machine, so in the future would rarely need to do that anymore.

[RickKrung]

.

[cbwx34]

One of you needs to define how "backlash" is used in relation to setting the USB using the microadjust...I haven't found a definition that seems to apply. 

It seems to me, reading Rick's post that if the microadjust "fed" the USB up and down, it might apply, but if I want to lower the USB and turn the microadjust, it doesn't feed the USB, just allows it to move.

Clear as mud?  It is in my head... 

[RickKrung]

.

[cbwx34]

CB,

I've not looked at any formal description or definition of "backlash", just know it all too well from many years of using linear motion tools.   It may be a bit of a distinction without a difference, but I'll give it a try. 

A difference between the Tormek USB and our examples that should be kept in mind is that the "feed screw" in our examples is the actual threaded screw, which rotates and the "nut" remains stationary.  It is the opposite with the USB, where the USB, while threaded is the "object" that gets moved by a rotating nut.  Another difference is that on the Tormek, only one of the two are captured whereas in our examples both are, discussed more below. 

Backlash, as I described it, when the object has been moved to a position with the feed screw, it will remain there due to friction and gravity, until moved again in either direction.  If it is moved in the same direction as before, it will move immediately and positively.  If the feed screw is turned it the opposite direction, due to slop between the nut and screw, the screw will rotate a small amount before moving the object.  That is the backlash.

In the case of the vertically oriented USB "screw", when the adjusting nut is rotated clockwise, it pushes the USB upward in a positive manner (assuming of course both locking knobs are loose).  But, due to friction and/or canting of one or both of the vertical USB bars, when the nut is rotated counter-clockwise, the nut may rise along the screw above the sleeve and the USB may not move (same as if either of the locking knobs are tight). 

In the nominal case Ken and I are referring to, the two objects, the screw and nut are both captured and can only move so much before before coming to the end of the slop between the two objects.  When that happens, it will cause the fixed object to move.  The distance or amount of radial movement the screw is considered the amount of backlash.  In the case of the Tormek USB, the nut is free to move an unlimited distance above the case sleeve since it is not captured.  This is in effect the backlash and this is where these situations differ. 

I hope this helps clarify.

Rick

Ok.  It helps clarify, I guess I just don't see it as really applying here, nor is it the reason given in the video that I could find.  (I know Ken said Wolfgang doesn't give a reason at all, but the part I found, he just says it's easier.)  If I move the microadjust in either direction, moves just fine unless it hangs up.  My .02, that's the lesson, if you move the USB down, make sure it isn't "stuck" (which if you follow the other tip in the video - push down before tightening, you'll get it).

But thanks for the explanation!

[RickKrung]

.

[3D Anvil]

Even if the concept applies, it seems to me that the tiny distance involved in the slop in the threads would be overwhelmed by the much bigger change that happens when you tighten down the set screws.

[Ken S]

CB,

The link you provided shows the issue exactly. You guys all have more machining background than I do. "Backlash" may not be the best term to describe the issue. However, you all seem to understand the issue.

Yes, one can tune the adjustment by either raising or lowering the support bar. However, when raising the support bar, the weight of the support bar makes raising it much more controlled. When lowering the support bar, I find I often have to tap on the bar with my finger to lower it. This seems more prone to being "hit or miss". For me, whatever we call the problem, raising the support bar is more precisely controllable.

I think digital read outs are marvelous ideas for machine tools. I don't think even the most anal retentive of us would seriously think of equipping a Tormek with DRO.

Returning to my original reason for posting this topic, I believe Wolfgang makes a very good point in recommending final tweaking of the support bar height by raising the bar rather than lowering it. It will become a regular part of my set up discipline.

Ken

[Ken S]

Even if the concept applies, it seems to me that the tiny distance involved in the slop in the threads would be overwhelmed by the much bigger change that happens when you tighten down the set screws.

This deserves a test. I plan to use kenjig devices (the slot for setting the Distance). The kenjig works like a gage block. Readjusting the support bar height ten times and checking the height with the kenjig before and after tightening the locking screws will tell the tale.

I will post the results, whether the change is small or much bigger when the set screws are tightened.

Ken

[cbwx34]

Even if the concept applies, it seems to me that the tiny distance involved in the slop in the threads would be overwhelmed by the much bigger change that happens when you tighten down the set screws.

This deserves a test. I plan to use kenjig devices (the slot for setting the Distance). The kenjig works like a gage block. Readjusting the support bar height ten times and checking the height with the kenjig before and after tightening the locking screws will tell the tale.

I will post the results, whether the change is small or much bigger when the set screws are tightened.

Ken

I don't think the type and amount of change he's talking about will be detected in your test...

Yeah, that was an exaggeration.  Sorry about that!  I just measured my machine.  At 15 dps with 139mm projection, I have to dial the micro-adjuster +2.4 in order to correct for the shift from the set screws, which I believe equates to 0.6mm.  In this case, that results in a difference of 0.4° in sharpening angle, which isn't huge, but it's significant if you're trying to be very precise.
...

[cbwx34]

...
I just measured my machine.  At 15 dps with 139mm projection, I have to dial the micro-adjuster +2.4 in order to correct for the shift from the set screws, which I believe equates to 0.6mm.  In this case, that results in a difference of 0.4° in sharpening angle, which isn't huge, but it's significant if you're trying to be very precise.
...

So, I actually took a look at this yesterday, by marking a point on the wheel and on the USB...



... and then measuring before and after tightening the locking screws.  There is a bit of a shift, I didn't take notes, but after several measurements, seemed to be around 0.4mm.  But, if you measure to the casing it's less (around 0.2mm) since most of the "shift" seems to be horizontal.

BUT, whether it's 0.4mm or 0.6mm.... It. Does. Not. Matter.  I have a pretty good idea why some think it does, but it does not.       Suffice it to say, in part, that this test was done with everything being "static", nothing was changed other than tightening and loosening the locking screws, and even doing just this, there was variation in the answers.  I doubt there's any way someone could be this accurate in a "working environment" where different parts are moving during the process.  There's other reasons too, but there's simply no reason to account for this.  (Here's a thought, if you measured to the casing, you might not even have noticed, or would be compensating even less).

If someone (I'm not saying just you in all this) chooses to try and adjust for this, that's their choice, just be consistent (as important if not more so than accurate), and you'll be consistent whether you adjust for this or not.  But my .02, it's an unnecessary procedure, and if anyone honestly tests it, they'd see no difference in results. 

[3D Anvil]

I guess if the shift is consistent, and you lock the USB the same way every time, then in fact it wouldn't matter.  I'll have to check and see how consistent it is on my machine.

I agree that it is a relatively small shift that won't matter in most cases.  It just struck me as being rather large in relation to the degree of precision that I see some folks attempting to employ.  For example, Wootz's methodology calls for extremely precise adjustments down to 1/10th of a millimeter.  Obviously, that's problematic if there's 6/10ths of a millimeter of slop in the set screws.

Also, I'm coming from a Hapstone R1 fixed-angle system (also KME) where one has to account for slight variations in stone thickness.  With these fixed-angle systems, .4° in variation would be a huge problem because you would be hitting different parts of the bevel with each stone change.  For best results with these systems you really need to hold the variation to less than .1° variation.

By it's nature, the Tormek system is less precise, simply because the knife jig allows for more freedom of movement than is possible with a fixed-angle system, and that's fine, because it's obviously still capable of producing outstanding results!

[cbwx34]

...
Also, I'm coming from a Hapstone R1 fixed-angle system (also KME) where one has to account for slight variations in stone thickness.  With these fixed-angle systems, .4° in variation would be a huge problem because you would be hitting different parts of the bevel with each stone change.  For best results with these systems you really need to hold the variation to less than .1° variation.
...

I've had the same conversation in the old Knifeforums... even with those systems, to be most effective, it's better not to lock everything down so much that the system has no room to "maneuver" (for lack of a better term).  You do need to do things like compensate for stone thickness, but not tighten things so much that the stone can't "find" the bevel, and it will work better if it can.