Knurling. A skill that I have needed on many occasions, but one that still eludes me. Part of the problem is using the Sherline lathe to knurl anything beyond aluminum puts tremendous stress on the system. Even using my homemade scissors knurling tool (have never written up the construction of this tool, but it is similar to many other such tools on the web), the Sherline just has too much flex. The South Bend lathe has the solidity for me to feel more comfortable knurling with it, but brass and steel are both difficult with the standard push knurling tool I have.
Another driver for this project is a fidgit toy, seen above, 3D-printed by my son. I would like to replicate this toy in metal and see cut knurling as a possibility on the curved surfaces. The toy presents many other challenges for knurling, but I can't even imagine using a deforming type knurling tool to get the job done.
Cut knurls are readily available, but a tool to present them properly to the work is not something readily purchased for the funds available. The Home Shop Machinist magazine came to the rescue. In 2010 a series of five articles were published by Michael Ward in which he describes a complex tool of his own design. The tool, from the cover of HSM shown below, will be a challenge to make, but is within my skill level, plus Michael's instructions are clear and detailed.
Based on Michael's thorough list of parts, both metal required and purchased parts, I will only need to purchase ball bearings, cut knurls and 5/8" drill rod. Metal from my storage and cutoffs boxes will be used for just about everything else. One purchased part, a #3-48 screw, will also need to be purchased along with a matching tap, unless the plans can be changed and a #2 or #4 screw substituted.
The first step in Michael's process is making the spindle housings. To make these a fixture is advised, so that is where I will begin. A cutoff of 1.5" X 1.5" aluminum bar was located and 1 1/8" was cut off with the horizontal bandsaw. The aluminum was first clamped to the angle plate with a large C-clamp as seen below. The four narrow sides were fly cut with this setup. The block was moved to the vice and again using the fly cutter the two large and most crooked faces were cut. The block that was left is 1.502" X 1.486" X 1.043" also seen below.
Some time was spent deciding on the best setup for drilling the fixture. The part was marked out on a large face to provide some indication of what needed to happen. A line was scribed across the to be drilled face and its center located. A tiny prick punch mark was made at the intersection of the two lines.
The part was not held at 60° in the vise as suggested as this leaves too little meat in the Sherline vise. Instead the vise was set up on an angle plate raised to 30° and the block set in the vise on parallels. The spindle was then centered over the punch mark and the x- and y-axes on the DRO were zeroed. A two insert cutter (3/4") diameter was used to cut a flat on the angled face, a slow process taking 10-15 thou passes. A fast feed rate, low speed and 10 thousandths cutting depth eliminated the squealing chatter as the cutter neared the back wall of the groove.
Drilling presented another challenge. With the angle plate mounted there is minimal room with the large drill chuck in the spindle. The largest drill I could fit into the chuck with the head raised to the top of its travel was a #7, 0.201", drill. After spot drilling the #7 drill was used to drill completely through the block. With this hole in place the #7 drill will be used to align the vice, still on the angle plate, in the drill press in order to open and ream the through hole to 1/2".
Unfortunately, there is not enough headroom in the drill press to handle the vise on the angle plate and the long 1/2" reamer. A lathe needs to be used to provide the room. The South Bend can handle the large drills and the reamer. Its chuck is also large enough to hold the fixture securely, rather than the small chuck on the Sherline, a precarious option at best.
Two different indicators were used to align the hole and the inserted drill, when the part was held in the four jaw chuck. Then the nightmare began, drilling aluminum! Drilling was started with a sharp 11/32" drill. It immediately seemed to work harden the metal and drilling could not proceed. Smaller drills were then used to power through the aluminum, advancing 1/32" with each size increase. The aluminum seemed almost undrillable, but with sufficient cutting fluid I could force the drills through. The final drill, 31/64", was a challenge, but eventually worked its way through. Reaming out the last 1/64" was extremely difficult and the reamer froze up halfway into the hole. So far I have been unable to get the reamer out. The reamer can be turned back and forth with a wrench on the four-jaw chuck and the vice grips on the reamer shaft, but I can't get enough outward pressure on the vice grips to pull the reamer free.
A little research indicates aluminum work hardens just like brass and copper. Learn something new everyday. Heating the aluminum to about 600° should temper it and return its machinability. First, the reamer was removed by holding the reamer with pliers and heating the aluminum block with a torch. The block was easily slid free. The block was set on fire brick and heated for a prolonged period with the butane torch. It never turned color beyond the gray of aluminum oxide. After four or five minutes of heating the aluminum was allowed to cool to ambient temperature.
The fixture was returned to the four-jaw chuck and set up as before with the same bit of shim. The reamer easily completed the job. As the photo below shows a lot of deburring was required. With this completed the fixture is now ready for completion.
The block was returned to the lathe and once again the hole aligned within 0.001". The hole was bored to 0.626" for a depth of 0.375". The finish was less than desired. The block was set up in the vise on the angle plate set to 60°, precariously high, but sturdy enough to cut a flat with a 1/4" end mill. A hole was drilled in this flat with a #25 drill and tapped #10-24. After a bit of cleanup a set screw was installed completing the spindle housing fixture.
Two spindle housing blanks were turned from mild steel. Both had a 1/2" section 0.751" long and a 5/8" section >0.75" long. The latter gets faced at an angle in the first operation. The two blanks were made from different steel rounds, with the second part coming from much more machinable steel. This part is seen on the right below.
A spindle housing blank was inserted into the fixture and the set screw tightened. This set screw shouldn't be loosened until all work has been completed on the exposed end of the housing. The block was held in the four-jaw chuck on the South Bend lathe and the end of the housing faced. The photo below shows the part almost completely faced.
The part was faced to within 0.020" of the fixture. After marking out the desired hole center the fixture was aligned in the four-jaw chuck to put this center in line with the spindle center. Drilling began and ended with a center drill, whose tip broke off in the part, even though there was cutting fluid and I was carefully peck drilling!!! Maybe I'm not supposed to make this tool? I walked away and will tackle the broken drill and potentially ruined part tomorrow.
A quick inspection revealed that the drill fragment was completely buried in the part and nothing could be grabbed by needle-nose pliers. The part was judged a loss. The steel of unknown provenance used to make this spindle housing was very gummy, resulting in a poor finish (as seen in the above photo), even with care to speed and depth of cut. The center drill was also old and one cutting tip was already broken off.
Unfortunately, I don't have any mild steel of the correct size to remake the part. Off to my favorite metal supplier, Speedy Metals, for some mild steel, 12L14, in a variety of sizes. 12L14 is one of the most machinable steels, and finishes nicely, thus my choice. It is not a great steel to use if hardening is required. 12L14 has added sulfur, phosphorus and lead, so it is probably advisable to utilize cutting fluid liberally and minimize high temperatures during cutting. Steel is so cheap and so heavy that shipping is almost as much as the price!
While waiting for the steel to arrive, the second housing blank was put into the fixture and faced to about 0.040" above the fixture. The fixture and housing were removed from the chuck and the hole center was marked out as before using the Starrett height gauge, what a nice tool. The center location was punched and returned to the chuck, where the divot was centered with the spindle. The photo shows the marked and punched hole location.
I don't have a way to bore small holes, so the blank was drilled with a 0-2 center drill and then drilled with a 1/4" drill. This drilling went into the aluminum of the fixture to make sure the hole was through the part. The hole was enlarged to 19/64" and then reamed 5/16". To accomodate my smallest boring bar the hole was opened to 3/8" for 3/16" with a drill. The bearing seat was bored to 0.500" diameter and a depth of 0.196". The spindle housing was then faced so it sits 0.027" proud of the fixture. (This number is important as the second spindle housing needs to be finished to the same length.) The photos below show the makeshift boring setup, which is why the South Bend tool posts are being made, and the completed spindle housing.
After an eight month delay the cut knurling tool is back in focus. Today a spindle housing blank was made from 3/4" round steel. A 1 3/4" length was cut from stock and held in the three jaw chuck on the South Bend lathe. One end was faced and then this same end was reduced to 1/2" for 0.750". This reduced end was held in a collet and the opposite end was faced to about 0.80". The diameter of this end was reduced to 0.625". The part was a tight fit in the aluminum fixture, requiring light tapping with a hammer to seat it. The set screw was then tightened.
With the part in the fixture it was placed in the four jaw chuck and the angled face was cut until it was completely parallel to the face of the fixture and then further such that 0.027" remained proud of the aluminum. The fixture was removed and the center of the just cut face was marked out as above. The fixture, returned to the four jaw, was adjusted until the center divot was centered relative to the spindle. At least that was the plan. No matter how I adjusted the fixture in the four jaw, it refused to be centered relative to the tailstock.
I finally realized the tailstock must be the problem after repeatedly moving the fixture on center from one direction, rotating 180°, and seeing the point off by the same amount just moved. A scrap of steel was held in a collet and machined to a point. This point when next to a pointed tailstock center revealed the tailstock was not centered. This was quickly remedied with the two tailstock adjustment screws. Returning the four jaw chuck to the spindle with the fixture still in place allowed for a quick alignment with the same tailstock center.
The part in the fixture was then center drilled, drilled through with a 1/4" drill (of which many were found to be quite dull!), drilled 19/64" and then reamed 5/16" as done with the first spindle housing. The housing was drilled with a 3/8" drill about 3/8" deep. It was then bored to 0.500" to a depth of 0.196". After a bit of deburring the part was compared to the first. Very similar, but too tall! I assume the first housing was not seated as deep as the second in the fixture because the bottom sections are 0.750" for both parts. I will return it to the fixture and pick up the bore with an indicator. Then bore it a bit deeper and face a little more off. All of this will be done after carefully determining the excess relative to the first spindle housing.
Comparing the two spindle housings to the plan indicates that the first spindle housing needs to be remade. It is about 1/16" too short compared to plan. It must have been a tight fit in the fixture and not inserted deeply enough. Oy! Let's hope the rest of the parts in this project are simpler to make.
Yet another spindle housing was made following the same procedures used above. This one is very similar to the last, but it is still a bit shorter (1/16"!) as seen in the photo below. I will attempt to bring them both to spec as opposed to making another. The shorter spindle implies the spindle was not completely buried in the fixture, though it had been tapped into place with a hammer until it "rang true".
The longer spindle housing was returned to the fixture, where it was faced closer to the length of the shorter version. The bore was centered crudely with the use of a tailstock center and then completely with a dial gauge. The bore was extended the same amount as was removed when facing. The two spindle housings are shown below.
To finish these two parts they need a bore and threaded hole in the 1/2" end. To this end the first spindle housing was placed in the four jaw chuck in the Sherline and dialed in. The end was centered drilled and drilled with a #38 drill for a 5-40 thread to 0.75" depth. The hole was then opened with successive drills to 0.25". From there it was bored to 0.355"producing a straight and centered hole. The depth of the bore was 0.375". The hole was then reamed 0.375". This left a rounded corner on the bore, so the boring bar was reinserted and the corner squared up.
The final operation was to thread the hole behind the bore. Unfortunately, a 6-40 tap was selected instead of the 5-40. Tapping was a challenge, but eventually succeeded! I will need to adjust the mating part accordingly. This part was labelled with a "6".
The matching spindle housing was treated similarly to the above producing a 0.375" hole, 0.375" deep with a threaded, 5-40, hole behind it.
The next pair of parts to make are the spindles. The first spindle began with a 1" length of 3/8" O-1 steel in the four jaw chuck. The steel was centered with the dial indicator, seen below. The dial indicator holder is a budgie job and should be replaced with a fixture that has builtin handles, omitting the need for a screwdriver and wrench. The rod was faced.
The four jaw chuck was removed from the lathe spindle and attached to the hexagonal indexing block held in the vise. The block sat on two parallels and the back of the chuck was firmly pressed against the side of the vise, locating from three directions. A 1/8" end mill was used to cut each face of the hexagonal stub after locating the end of the work piece with an edge finder and locating the top of the rod with the end mill. The cuts taken were light, 0.010". Cutting 0.125" deep into the rod for each of the six faces. The first photo shows the setup and the second the completed hexagon. Distances between opposite faces were measured: two were spot on and one dimension was 0.004" shy of 0.125", good enough for a screw head.
After a break to forge a knife this project has returned to the top of the list. Happy Independence Day to all. The rod with a hex end was checked to make sure it was still centered. The end with the hex was turned down to 0.1875" for a length of 0.462". After deburring the opposite end was faced so the 3/8" part was 0.435". The process was repeated on a second 1" length of 3/8" drill rod. The photo below shows the two partly complete spindles resulting from this work.
A spindle was held in a 3/16" collet in the Sherline lathe. The wide end was reduced to 0.313". The terminal 0.375" was reduced to 0.250" for threading. After repeating and removing the burrs the parts at this stage are shown below.
The spindles were threaded. The part was held in a collet and the Sherline was set up for threading a 28 TPI thread. The end of the part was blued and a line was marked with calipers at 0.156", the limit of threading. A thread was cut to the line. Testing was done with a purchased 1/4"-28 nut and the nut fit long before reaching the theoretical 0,049" thread depth. Both spindles were threaded and quickly deburred. The completed spindles are shown below.
The spindles were finished by heat treating. A scrap of steel round was held in the lathe and drilled 1/4" for a depth of 1/2" to serve as a heat sink for the threaded end of the spindles. A spindle was placed in this heat sink and set on firebrick. The small end was heated to cherry red for thirty seconds and quenched in motor oil. The second spindle was treated similarly. After cooling the bearing end of the spindles was sanded lightly to reveal the metal under the scale, seen below. The two spindles were then tempered. The grill was heated to 425° and the spindles set on a metal plate in the grill. They were tempered for 15 minutes and are shown in the second photo below.
After two days of having fun with the oldest grandson bearings were made. One inch of 1/2" bronze round bar was cut from stock. Held in a three jaw chuck it was faced and the diameter carefully reduced to 0.32125" a bit under the desired 0.3125" but close enough with the planned use of retaining compound. The rod was center drilled, drilled with a #18 drill and reamed 0.1875" to about 0.35". The end was chamfered and a 1/8" length was parted off. The parting off was stopped about halfway through to chamfer the new corner. This was repeated producing the two bearings seen below.
The spindles vary by a quarter of a thousandth from 0.1875" with one nearly spot on. Chamfering both sides of the bearing holes allowed the spindles to slip in. The reamer probably produces a slightly oversized hole, but the fit is good.
Two nuts were made next from 3/8" steel hex. A 1" bar was faced, center drilled and drilled with a #3 drill to about 1/2". The hole was tapped 1/4"-28. The nuts were parted off and after some clean up are shown below. Both nuts fit nicely on the spindles, also shown below.
Spent 1.5 hours seeking the desired cut knurls: 5/8" OD, 1/4" ID, and 40 TPI. They were nowhere to be found. Looking closely at a photo in the article the letters, G, K, & S, could be discerned. Seeking GKS cut knurls successfully pulled up these and two were immediately ordered. I was almost ready to abandon the project!
Not until the knurls were installed did I realize that 128 TPI wheels were purchased! These are the correct 40 TPI knurls.
While waiting for an empty house to complete the noisy milling (described in Part 2 of this series), the spindles were tackled. The spindles were a press fit into the bearings. A bronze tube was used along with the vise to press the bearings onto the spindle ends. A drop of Loctite 648 was used to insure a permanent bond. The bearing seats are spot on 0.500" as measured by a hole gauge, so a press fit is in order here as well. The bearings were pressed into the spindle housing using the vise and the original fixture for their manufacture. One spins freely and the other is tight. Possibly the bronze bearing was not quite straight. The spindle housing was returned to the fixture and held in a four jaw chuck. Oil was applied to the bronze bearing prior to putting all in place. With the spindle somewhat centered, its end, wrapped in aluminum, was held by a Jacobs chuck in the tailstock. The lathe was turned on and left spinning for ten minutes in hopes this would loosen things up a bit. It is easier to turn, but not free spinning like its partner. Test assembly with the two, too fine knurls is the last photo below.
Part 2 of this series begins with the manufacture of the holder arm.