Sometimes it hard to believe how difficult the path can be to get to a simple idea. The engraver on the right is an example. It is such a simple device that it's embarassing how long it took to develop.

We wanted to engrave part numbers and port identifications on the manifolds we were making. Tool pathing the letters was a big enough challenge (I'll post something on that at a later date) but being able to cut the letters turned out to be way tougher than we thought. Actually cutting in letters and markings isn't that difficult, it's doing a good job of it that's tough.

Our original attempt was to grind a standard engraving tool out of a piece of 1/4" drill rod, heat treat it and run it at top speed. It worked okay but because the tool diameter is so small (theoretically zero), we needed very high spindle speeds. Our Okuma's 8,000 RPM worked but it really is way too slow for the job. Feed rates were limited and tool life (hence, burr control) was quite unpredictable. To do a decent job we needed to bring the speed way up.

So our next attempt was a high speed air spindle. The 22,000 RPM from this tool was more suitable but it was tempermental to get consistent results. For making nice fine text, it is best to keep the plunge shallow (I liked 0.005" below the surface best). Deeper means bolder and this forces larger text to keep it legible. The problem with very shallow cutting is that you must have very precisely offset tools (which we now do with a Kelch presetter but didn't have at that time) and you have to know the exact deck height you are writing to on the part... the whole deck height. If the part is off level at all it will cut deep at some places and not touch at others. Going deeper is only a partial solution as the text get rather ugly when the depth is inconsistent.

This lead to our next invention. We took a Snap-On reciprocating air pencil and machined it to fit a CAT 40, 5/8" end mill adapter. Feeding air to this device we got 20,000 hits/minute. To operate it, we lowered it to about 0.090" above the part, turned on the air and ran the tool path. Compared to the rotating tools, it was not nearly as sensitive to height variations... so that problem was solved. The only real problem with this device... well actually there are several come to think of it: The main problem is that it hit too hard for aluminum. A burr would swell along the edges and because we need to use exact stop, everywhere the machine made a momentary pause a little divot would be created. As you can see in the photo, the rotary tool did much nicer looking engraving.

The other problem was with the tool itself. Occasionally we would turn on the air and it would stall and do nothing. Usually you just had to touch it with a finger and it would start. Other times it would partially stall while we were running so we knew it was going to be a reliability problem.

The last problem I can't really say that I know was a problem but I didn't like it. And that was the 20,000 hits/minute of impact acting on the spindle bearings while they were sitting stationary. It seems like a little pencil shouldn't be able to do much damage but that's because when we use it by hand, our hand makes a lot of compliance. When it is mounted firm in a tool holder and spindle with no give, it actually impacts very hard.

As you can see from the image to the right, the impact engraver did do a very nice job on steel. This piece was done on our 4th axis. I really didn't want to get into the Text-to-G-Code convertsion utility that we wrote but note that software can set the text at any angle. It can also scale and work on various planes as well as cylindrical surfaces on the 4th. We set up the utility to work with AutoCAD shape files, so it can do the same with logos and use any SHX font available for AutoCAD. Also note that you can see some stalling on the lowest test text.

In an attempt to reduce the impact we came up with a spring chamber head that allowed the air pencil to float against a relatively light spring. This really did nothing but make the system complicated. The impact was so fast that the spring made no difference to the beating we gave aluminum. By the time we had this made up and it wasn't working I had lost my patience. While were testing, we started playing with the air pressure again - we played with it when the pencil was solidly mounted but it made no difference. It still made no difference. We dropped the pressure ridiculously low to where the tool could barely run and then it happened. The tool stalled but this time with the carbide out instead of in. And instead 20,000 hits/minute we dragged it around. When I saw the result, I slapped my forehead and was ready to scream. Instead we laughed... KISS.

In the end, this is what we came up with. We stole the carbide out of the Snap-On pencil and we built a housing and installed a spring. It still fits a 5/8" end mill holder. There is about 0.180" travel on the tip. Generally, we plunge about 0.090" down and drag. The parts can be very out of level and the deck height is not critical. We use a light spring for aluminum and a heavier one for steel and iron. It raises a bit of a burr but is more than satisfactory for our manifold work. It seems like a lot of work for such a simple tool but machine design is always that way: once you are onto a good idea it's easy; if you are onto a bad one, it takes forever and the results are way less than what you are looking for. Never confuse an innocent problem with a simple one.

By the way this is what the engraving looks like in aluminum:

...and here is a close up of 0.060" tall letters. Note that the end mill was run a second time to clean off the burrs raised by the engraver. Normally we don't do this but the letters were so small, they just finished nicer. And I threw it in because I just like this photo!