Knurling

[knifer-one]

is there a certain pitch knurling wheel to use for different size workpieces? what is the formula for determining the pitch? I am planning on purchasing a scissor type knurling tool and would like to get the proper wheels with it or buy the most used type. I have used a straight knurling tool in the past with not much success.

 

[Conrad Hoffman]

There aren't a huge number of knurl pitches available- I don't even know what mine are. You'll usually just fiddle with the part OD a bit so the knurls track correctly and don't double track. I like the Eagle Rock scissors tool. Even with a scissors tool, the forces on the carriage can be very high. QCTPs will rotate. If you have backlash in your crossfeed, it can be difficult to stay over center because the crossfeed slips forward or backward. Not trying to scare you but good knurling doesn't happen instantly; you'll likely have a learning curve.

 

[Conrad Hoffman]

My Machinery's is #23, old but plenty adequate for anything I ever do. A guy at work had the full book size one with large print and IMHO that's the one for us older guys to have!

I recently knurled some knobs for a small precision instrument in 303 stainless. Though it was possible to get a full pointed knurl, I just didn't like it. Too tough on the fingers and too tough for the lathe. I settled for just a mild crosshatch indentation and that both looked and functioned well. I also look for antique knurls and have a few nice ones. There's a convex straight knurl used on old Stanley planes I find attractive. There are also some odd reverse knurl patterns you rarely see anymore.

BTW, always apply a 45 degree chamfer after knurling.

 

[Frank Ford]

Personally, I think knurling is just a little bit of magic. Press the knurl into the work, and it eventually sorts itself out and produces a nice pattern. The trick is to try it a bunch of times and pretty soon it works, just like it should.

I use a scissor type knurl and find that it's a bit of a chore to squeeze the knurls hard enough to get a good pattern, so I cheat just a bit. I set the knurls so the both wheels just touch, one above and one below the work, right on center. Then I back the cross slide out, and close them down a little. Once that's done, I turn on the lathe, and feed the tool into the work. The upper knurl never goes directly over center, but it comes close, and I get a nice job every time.

For straight knurls, it's a single wheel press-in tool, of course.

 

[Pacer]

As usual, Frank describes a simple approach to something with out getting brain strain. Knurling does indeed at time seem like "magic" - on one try beautiful knurls will seem to 'magically' appear with little effort, and on the next just be a pain to finally get them.

Franks description of backing off the scissors and closing them a little is the method I've adopted from watching another talented machinist (Lane) do it that way.

 

[Tony Wells]

I also use a scissor knurl. I have found that setting the rolls to full contact at about 165° apart, front of centerline. In neutral, I rock the spindle by hand forward and reverse about 30° while applying pressure with the crossfeed. When I am satisfied with the depth of the pattern, I roll the spindle around dry, and see if the pattern is acceptable. It almost always is. The rolls never are exactly on centerline, as they would have a tendency to pull over. Slack in the screw would work against you in that case and spoil the knurl. I always start with the tool square with the work, and fully on the work. Flood coolant always, rich concentrate.

Knurling is somewhat an art. I have had to knurl to a specified PD to attain a surface speed between a matched set of knurled rollers in a PE pillow bag sealer. The surface speed had to match a pair of pinch belts pulling the bag through the sealing machine. It was quite a feat to get it by accident. Much better to document efforts and develop repeatable procedures when it is a functional knurl, rather than a cosmetic or ergonomic knurl.

I've wanted to do some straight convex knob knurling, but haven't made the rolls yet. That's just one of my tinker ideas.

 

[mikew67]

Knurling is "interesting", both at an experimental and theoretical level. Search the web for some really high-end tooling for CNC lathes (Swiss and German, IIRC) and tech info behind them. Choices include knurls that purport to work by forming and some that work by cutting, wheel axes parallel to the lathe axis (most of the HSM tooling) and wheels tilted at 45º, wheels that are cylinders and wheels with convex faces (like a bicycle tire). Some of the vendors says certain materials should be formed, others cut, and provide tooling for both.

Home Shop Machinist magazine even ran a 6-part(!) series on how to build a tilt-wheel knurler. Took them all six installments before they revealed how it was supposed to work -- by cutting with the leading edge of the wheel. Only problem is that even if the top wheel does cut, the bottom one doesn't -- its leading edge is dragged, not pushed. Just look at the geometry (Note: it could actually be fixed by moving the lower wheel forward along its tilted axis so that its right face/corner is the contact point. But then the tool isn't very compact laterally and can't be used in tight quarters.)

So for you inventors out there, there's still plenty of room for a creativity in knurling tools

Mike

 

[Conrad Hoffman]

As a result of this thread I'm trying to get a better understanding of stock sizes for form knurling. It's confusing! Apparently there are knurls standardized at 1/DP stock increments and others that aren't. My old Eagle Rock knurls have 50 "teeth", rather than 48, so the DP is 66.66666 (devil knurls?) rather than the expected 64. It seems that a good starting point would be a stock size equal to the nominal diameter of the knurl, so that means that the stock size might increment up and down from that number, not up from zero. IMO, what happens initially when the first impression is made is more important than anything else, so it seems logical to work from the knurl OD, rather than the PD. The data in Machinery's, like much of their math, consists of a group of formulas with circular references, but no instruction on how to proceed to get a useful answer. I'm working on a spreadsheet to better tie this stuff together, but exact answers are elusive. Any good technical links would be appreciated. Oddly, Eagle Rock, who make great tools, has no info at all and seems to suggest that if you just knurl to full depth in a single pass, anything will work. That has absolutely not been my experience!

 

[Conrad Hoffman]

OK, I've got a page together that's pretty simple. No photos yet, nor any actual link on my site, but GO HERE and tell me what you think. I'll try to tune it up depending on what everybody says.

Thanks!

edit- OK, added a couple photos and a link from my shop page. Now all I need are comments so I can find the typos or just erase the thing if everybody laughs.

 

[GerryR]

I had a little more time than usual on my hands and have wanted to try this for awhile. I'm sure that most people have noticed the resemblance of the "Kant-Twist" clamps and the scissor knurling tools. Well, with a little cut-and-paste and a 2-1/2" Kant-Twist clamp:





Conrad, I found that info very worthwhile. Thank you.