Jet Mill CNC Conversion

Altough my Prazi is a fine mill, I found it was a bit to small for the kind of work I want to do. So I started watching for a bigger mill. My shop is on the second floor of my house. In fact, the mill is placed almost exactly above my bed, so I really was not sure I wanted a 2000 pound Bridgeport up there. I found a Jet JVM-840 knee mill at Brothers Macinery Exchange for $1,000. It was in very good condition, has a table about 8"x32" and an R8 spindle. After some consideration I decided it was right so I listed my Prazi on eBay and bought the Jet.

I rented a pickup trick to haul it home. We took the head and table off and laid the mill on its back. I have an I beam arm bolted to the side of my shop, and a double door arranged for easy acces of big items. I took the mill almost entirely apart before lifting the pieces into the shop. I could have moved it without total disassembly, but I wanted to give it a thourough cleaning and check over as I put it together. I forgot to take any photos of it until it was in the shop and partly reassembled.

I decided to use the knee for the Z axis in this mill. It does need a bit bigger motor to lift the weight of the table, but all the quill conversions Ive seen are very clunky, and for many operations I like the ability to use the quill by hand.

This photo shows the bracket to hold the Z motor- its a simple 3/4" plate of aluminum, milled and drilled to hold the motor. Note Ive attached it to the mill with the 2 screws used to hold on the graduated collar in the origianl machine.

I did not change the knee elevating screw. Its a large acme thread, with a pair of spiral bevel gears driving it. There is considerable backlash in the gears, but I expect the table weight will keep it down, and almost all Z operation I do are simple plunge the tool, or dill, and retract. If it turns out to be a problem I can always add a ballscrew to the quill and just use the knee as a power feed.

This photo also shows the Y lead screw in place, you can see the aluminum post sticking up that will go into the saddle casting.

A closer view of the Y axis ball screw. This is a .631 in dia screw, .2" pitch, from Thompson-Dahner. Note I have 2 ball nuts, mounted back to back on the center block. The block was made exactly like the original, except with a thread to fit the ball nut mounting instead of the lead screw. The first nut was screwed into the block tight, the second has a ring nut about 3/16 thick screwed on, then it is screwed into the block. After the ball screw is in place the second nut is turned to take out backlash, and the ring nut is tightened against the center block to lock it in place. Commercial double nuts use a bellevue washer in this position to spring load the nut.

The ball nut diameter is limited by the top of the Knee elevating screw. There is plenty of room above it.

In this photo the saddle is in place with the Y screw connected. The X ballnut is visible on top of the saddle. For the X screw I used the same type screw, but the 3/4" dia. Again, I made a replacement block to match the origianl, but with threads to match the ballnut mounting. In this photo only the lef the left side nut is in place. Note the brown item in teh ballnut- that is a cardboard tube that is shipped inside the nut to hold the balls inside. When you run the screw intothe nut you keep the tube snug to the end of the screw and slowly withdraw it as the screw rotates in.
This is a photo of the end of the X ball screw, ready to install. Sorry its such a lousy photo. The screws are very hard, and I had machined the ends when I used them on my router, but in this mill I wanted to keep the original end bearings, and they had a bore larger than the ballscrew. I welded a short stub of steel to the end of the screw, then put the screw into my rubberflex collets, which are my most accurate collet. I then machined the end of the screw to have a shoulder and correct diameters for the end bearings. Note the threaded area, it holds a lock nut to apply pre-load to the end bearings.

For the back end of the Y screw I turned a short stub to the inside diameter of the keeper that came with the ballnut. This kept the keeper in line as the screw was turned into the nut. I was worried about this operation because it had to be done reaching up under the knee, a very tight and awkard place to reach. I got it in without a problem.

On the X screw I machined the end to the inside diameter of the keeper tube, then made a bushing to fit over that to the inside dimension of the end bearing. There is no shoulder on this end, the bearing simply keeps the screw in alignment, it can float a bit with temperature change.

All three motor brackets are in place in this view.
A close view of the X bracket, with the servo motor mounted.
A view from the left side showing the bracket and motors.
Far left shows the Z motor in place.
For the control box I bent up and old PC cabinet to make a chasis. I bought a torroidal transformer for the power supply, and used the PMDX power conditionig module. This board has all the power supply components on one board, except the tranformer. You can simply use a bridge rectifier and capacitor for a cheaper solution, but I think the PMDX module make a much better supply.

You can see the Geckodrive G320 servo driver modules, and the PMDX breakout board.

The box mounted on the side of the machine, where it originally had a transformer for the light, and a power switch for the motor.

The wires from the servos are being connected to the box in this photo. About this time I realized the box was a bit to small, as things were starting to get crowded.

I got everything together, ready to run just an hour or two before I was to leave for a Thanksgiving trip. I fired up the box, hoping for an instant success, but no such luck. In fact, it appeared I had blown the servo drives, because they a were showing the fault led, and the motors had zero resistance to movement by hand. I was hooking a jumper wire to the breakout board for my scope ground, and I hit something and blew a capacitor on the breakout board- a nice firecracker sound!

At that point I threw in the towel and left for Thanksgiving.

When I returned I got to work checking carefully. I first discovered the G320s were fine- I had simply turned the limit pot all the way DOWN, instead of all the way UP. So any movement immediaetly triggered an over current fault. Dumb, and I had orderd replacement G320s while I was away so they would be on hand when I returned. Fortunately Gecko accepted them for return since I had not used them.

Once I got everything setup the machine ran great. I could move the X and Y axis at 130 in/min with no trouble. Even the Z would move -up and down the whole table- at 130 in/min. I purposely de-tuned that to about 65 in/min because it just seemed to fast.

After about a week of good luck things took a turn for the worse. My servo motors were found in a surplus place, and they had integral brakes. The brakes require 90 volts to keep them released. I wired the power box with two switches, so I could first turn on the brake supply, then turn on the servos. One afternoon I left the machine with the brakes on, but servos off. The next morning when I turned the machine on to continue testing I soon had the smell of burning motor- The Z axis motor brake had locked up, and the servo drive was trying very hard to drive the motor. The motor was VERY hot, the G320 was on a big heat sink, so it was just warm.

I considered buying a new servo. I could not find anything appropriate for under $200. I was also a bit put off by the Geckos- I was never able to tune them to stop the growl when at idle. Apparently that is a "feature" of the Geckos that comes with their low cost. I considered buying just one new stepper for Z and keeping the servos on X and Y, but then Id have a power supply problem- I was running the servos at 80 volts, the stepper would need only about 50v.

About this time Marris, of Geckodrive, wrote one of his frequent messages about the selection of steppers vs servos. He pointed out that a servo ought to be geared to run at a reasonable speed- for my mill it would work out to about a 5:1 ratio. My pulleys were only 2:1, which is good for steppers.

I finally decided to change all 3 motors over to steppers. I was able to sell the Geckos and the transformer so my net loss was only about $50. I bought 2 740 oz-in and one 1200 oz-in steppers from Home Shop CNC

My servos looked like NEMA34 size, but while I was waiting for my steppers to arrive I had a closer look and found they were an odd size. The mounting screws were off just enough that drilling new holes would overlap the old. I had to make a complete new sset of brckets!

Oh well, having made one set I made the second set better, and took the time to add tapped holes for a belt cover.

The new brackets and motors from the left side
And from the right side
I built a complete new control bax as well. This is larger, and has room for an index pulse board and maybe someday a speed control board.
Here is an overall view of the entire machine. The PC is in the cabinet to the right. I used a flat pannel mounted to a door on the PC cabinet.

I have experiemented with a couple controls for the PC. So far, I like the simple Xkeys pendant best. Actually, I would prefer a Genovation Keypad, as I used on my router, but when I went to order it the name X-keys stuck in my mind, so that what I ordered. The Genovation has an extra 4 keys, and is cheaper.

I have just started to use the machine- so far Ive just made a few test cuts. I will end this conversion page here, maybe do a page after I've had some experience with its use.

Back to my home page