2.14.2010

making chips fly

buried somewhere in the recent middle of my resume is the job i held as head CNC machinist for Industry Nine Componentry (link below in the "props" column) - a little company that makes some incredibly strong and beautiful bicycle wheels. i was fortunate to have been an integral part of bringing those wheels to market as the company was just starting up, an accomplishment of which i'm very proud. especially considering i had no - as in  zero - experience operating CNC machines before i finagled myself a job there.

i had had some machine tool schooling in the Navy, and i have good math skills. i also had many years in the bicycle industry by that time, which enabled me to help with the marketing side while i honed my craft as a CNC machinist. CNC stands for Computer Numerical Controlled, which means the machines are programmed in a language called "G-Code" to perform their various functions about the X and Z axes (on lathes) or the X, Y, and Z axes (in milling machines). it's a pretty simple language to learn, but the machines do exactly what you tell them to do. if you miss a decimal place with a 4-pound chunk of aluminum spinning at 5000 rpm, it is guaranteed to make some kind of loud, sickening noise and could get somebody hurt.

my minimal experience with "manual" machining was primarily on a small engine lathe in the engine room of the submarine. to be honest, i had forgotten how to even operate a lathe in the interim decade-and-a-half. my aptitude had not left me completely, though, so i was able to pick it back up pretty quickly. a large part of machining metal successfully is in the "feel." obviously, there are some rules of thumb to give a starting point concerning speeds and feeds (how fast to move the cutting tool and the workpiece, respectively), but in order to be more than merely competent, the operator has to take visual and aural cues from the machine, adjusting for tool sharpness, ambient temperature, type of lubricant, and a number of other factors as the part is worked. few things are as satisfying as being able to balance all the variables and produce good parts repeatably.

i had a boring job to do on our old Bridgeport mill the other day, and i only was going to get one shot to get it right - we have been tasked to modify a set of very expensive, fully processed tooling plates for a client. in my present position, i don't have access to all the CNC lathes and mills we had at i9, just a small benchtop CNC mill and a manual lathe, along with the aforementioned Bridgeport.

the first step in a successful machining job is workholding. the part must be firmly clamped in whatever machine is going to do the work, taking into account the required movements of the tool. i had an aluminum plate, 1-3/4" thick, on the face of which i had to make two 2"-diameter bores, each .125" deep. the bores were to be 13" apart on the X (left-right) axis, so i had to make sure the plate was clamped down dead-straight on the mill table. i used a dial indicator on the front face of the plate and ran it back and forth until i had less than .0005" of sweep from one end of the plate to the other.

i centered up on one of the holes and used a 1-1/4" endmill to plunge the first cut down to full depth. all the axes on the mill have graduated dial handles with resolution down to .001" (one "thou"). once the cutter touched the surface of the plate, i squirted on a little lube and dialed up the Z (up-down) axis .125". i verified the depth to be within .005" (although i will hold myself to a little closer standard - .002") by measuring with a set of digital calipers.

next i put in the 1-1/2" endmill to hog it out a little more prior to boring. to make sure i hit the same exact depth, i blued the bottom of the existing pocket. this way, when i make the next cut, i can see when the second cutter reaches the depth of the first. i then touch off the larger tool and dial up the Z axis again until the cutter just wipes the bluing. since i am returning the table to "Z zero" after each cut, i take note of the position of the dial for reference when cutting the bore. it shows .127", which is a little further than intended, but i feel confident it was a slight error made when touching off the cutter, and the depth is probably still on the mark.

next i set up the boring tool. this is a fairly rudimentary tool, with a single cutting blade offset from the center of the mill spindle. for my baseline setting, i'll dial it back 'til it's smaller than the 1-1/2" pocket i just made, then drop it in the pocket and slowly dial it out until it just touches the outside edge of the pocket, which should be about 1.500" in diameter. when boring with this tool, we can only take about 1/8" at a time on the diameter, so i plan on taking 4 cuts to the rough diameter, then a finish pass of something less than that - maybe .050" to 1/16". setting the bore is done with a relative dimension - note the baseline setting, then dial the diameter out 1/8" (.125) at a time using allen wrenches to set it and lock it down. once the first cut is locked in, i run the slide down until the tool just touches the plate, then fire up the mill. speed for this single-cutting-edge tool on a diameter this large will be much slower - 150 rpm or so.

this is where the "feel" comes in. i watch the chips spitting off the tool while i dial up the Z axis, looking for the balance between smooth cutting chips that break regularly and minimal chatter. chatter is a high-frequency vibration in the tool or the part, usually caused by not pushing the cutter hard enough. pushing it too hard will cause a rough finish or tool breakage and can itself cause chatter. by watching the shape and size of the chips, and listening to the sound of the cut, i can finesse the best, most accurate cut. these are good chips:

i repeat the process three more times, taking .125" at a time out of the diameter of the bore on the first two cuts, then a smaller cut of .075", to leave me room for a finish pass. the finer cut on the finish pass usually allows greater accuracy along with a nicer finish. my trusty calipers tell me the diameter is 1.943", or .057" away from my goal. i make the final adjustment to the boring tool. experience tells me the cutter will probably bite just a tad & open the bore beyond the desired size by a thou or two, so i go shy and dial in .055" instead. one more time up on the Z axis, and i think i'm done. a check with the calipers reveals 1.9995". i'll take it.

as far as the finish depth of my bore, i would be satisfied with +/- .005", which we can easily adjust out in the mechanism. my personal goal was +/-.002", and a quick check reveals a final bore depth of .1265", only .0015" off the nominal dimension. i'll take that, as well.

the CNC machines i used to make hub parts at industry nine could repeatably hold a +/-.0002" (that's 2 ten-thousandths of an inch) tolerance, once the machine was settled in to equilibrium operating conditions, running automatically. if it ran out of material, or an operator was a half-minute slower to load the next part, the metal in the machine would actually expand or contract enough to throw that dimension off by 5 tenths (.0005") or more. ideal conditions for any machine is equilibrium operation - think of it like cruise control on a long, flat road. the machine finds its cruising levels.

the challenge in manual machines is to be able to anticipate what the workpiece and the cutter are going to do in a given operation, and to be familiar enough with each piece of equipment to be able to feel what it's doing. being attuned to that is pretty damn cool.

No comments:

Post a Comment

please set me straight -