UPDATE to Building a CNC

[Diane Kauffmds]

I love using the desktop cnc, but there were limitations on size. So, being the tightwad that I am, I've build what is basically, a new cnc.

She's a pretty hefty thing now. I've put it on a cabinet with wheels, so it can be wheeled out.

A very special thanks to Steve Wellcome (MaineGeezer) for his engineering skills and milling. I sent him 4" x 36" x 3/8" aluminum, which he milled to my specs for the new sides and spindle holder plate! He did a great job!

My issues were table travel length along the y-axis, and z-axis travel. I would like the capacity to cnc the inlay pockets in a fretboard, after I design the inlay. The machine already did a good job cutting inlay pieces. I'm also looking to the future (a very long way into the future) to possibly roughing in my necks by cnc.

Both of these things require a travel length a lot more than the 12" that I had. I also needed a good 6" of travel height for eventual necks.

The original spindle was a whopping 75 watts. I've installed a 500w spindle with separate power supply. It's huge compared to the wee spindle that I had. I'm using a 200m (8") linear stage. Originally I had 40mm travel up and down. Now I have 7.5 inches.

I've replaced all original motors. Since I'm still using the 24v, 5amp control board, I made sure all motors fit within those parameters. All 3 original motors were .5 amp. I'm running 2 nema 17, 1.5 amp on the xy-axis, and a nema 23, low current, high torque motor, using 1.8 amps.

The only original parts are the z-axis guide bars and screw. I made the front and back. Steve made the sides, plate, and delrin bearings for the screws.

It was a project designed for Murphy's Law. I'd plan things that in the light of day didn't work. In the end, it's come together. It runs smoothly with almost no detectable vibration, and it's quiet.

So here is Beasty:

[Diane Kauffmds]

Wow. I'm starting to really test this thing, and I'm blown away. It's so quiet that unless you're sitting next to it, you won't hear it running. Normal conversation level that you'd use sitting next to someone on a living room couch, is louder than this cnc while it's engraving in wood. The other required hearing protection. This does not.

It sounds like R2-D2 from Star Wars!

I've been running various programs for types of possible logo's. All of the motors are cool. The spindle feels Luke warm.

I'm running the spindle with its own power supply, right now manually. I'm debating whether I want to get a solid state relay, to wire it for computer control. In many ways, I like having manual control. I'm not using the power supply that came with the spindle; I bought what I think is a better power supply.

But honestly, I had no idea that cnc's could be quiet!

20221028_130345.jpg

20221028_132117.jpg

20221028_132520.jpg

[BEJ]

Looks like a great project, will be interesting to see what you can make with it.

A lot of the fun of making stuff for me is making tools/jigs to make things easier/better. Must have remade some of my guitar tools 2-3 times, you start out with what you think should work but after using the stuff a few times see there are some improvements that could be made.
Been working on a duplication machine the past few days, 1st attempt was junk got a better plan now and should be able to test it out in a few days.

Bruce,

[TEETERFAN]

The ingenuity displayed on the forum is awesome!!

[Diane Kauffmds]

I ended up building a new table. It's as solid as a rock now. Yesterday, I totally finished the cnc after installing limit switches on all axis'. I shortened some of the wires so they weren't in the way. I also found a control board that allowed me to tie in the 48v power supply and spindle. So, everything is back under computer control.

I had to calibrate the y and z-axis'. The new screws apparently turn at a different rate. I was wondering why the new machine cut too deep, and sample inlay was out of proportion. When the computer called for 1mm, it was moving 2mm. After preliminary default calibration, I moved each axes 100mm, took the actual measurements, did the math, then reprogrammed with the new numbers. I've got it to within 0.00001mm. (Yeah. I'm that particular). That took a couple of hours, but the recalibration was worth it.

The first thing I did was an acid test. It was a simple pocket cut into a piece of plywood, to lay my dovetail template into. I tried this on the old cnc, but it totally stopped the spindle and stepper. This cut through it like butter. The plate fits precise.

The second was the logo inlay for my headstocks. I've never been able to cut the inlay without losing the middle of an "a" or "e". I'm able to cut it now.

Third was recutting the vine for the Irish guitar fretboard. This was most telling. The difference is startling.

20221107_141346.jpg

20221109_082953.jpg

20221109_113133.jpg

[Diane Kauffmds]

This is going to be a complicated post, so bear with me.

Love the new cnc, but I wanted complete computer control over the spindle and power supply. I had found a control board with integrated 48v, and 500w spindle support. Unfortunately, the x-axis driver went out on the board. I found another board that had the support, but frankly, I didn't like it too much. My feelings were verified when it didn't work.

What I really wanted was for my existing control board to control everything. But, the existing was only 24v, so it ran the spindle at 1/2 power. I didn't mind too much operating the spindle manually for short projects, but when it comes to 2 hours +, I like being able to leave the room.

So, I bought a solid state relay, and some 100k resistors, and went to work.

Also, I wanted more useful travel length. I define useful as being able to use the bit, not total table or travel. So, I redesigned the table.

I'm going to explain each of these things in 2 separate posts.

[Diane Kauffmds]

First, the table...

As I explained above, I designed a linear slide system, so there is a piece of bar aluminum underneath each side of the table. Rach rides on an aluminum angle.

I looked at my table again and decided to move the bearings more centered. I cut new 1/8" thick aluminum, drilled and countersunk them, one for each bearing, as nice solid spacers. These are larger than the first ones I made.

I had originally spread the 2 pieces of table apart, but with the rail slide system, I figured it was well supported. I closed the gap. I cut more plywood for the table, 18" x 12".

For my efforts, I now have 16 1/2" of useful cutting length. I had hoped for 15". I best it by 1 1/2".

I bought UHMW tape, impregnated with ptfe, and installed it on each surface of the rail system.

The end result is a table that glides quietly, and is rock steady with zero deflection on the entire cutting surface. I can literally lean my weight on that table with no movement. The stepper motors don't have much work to do to move the table.

20221202_122544.jpg

[Diane Kauffmds]

I'm sure you noticed the tangle of wires. I'll address this tomorrow. I know it looks like a Rube Goldberg contraption. All were necessary to wire the spindle so that it's under computer control.

This has Murphy's Law written all over it...

I had a controller that did it for me. It failed and Amazon didn't have any an exact replacement. I ordered a different board, which didn't power up out of the box...got a replacement, same thing.

I had replaced the stock controller, a long time ago, with a Genmitsu Sainsmart Prover controller, a very nice, reliable, grbl controller. I started searching for a way to use it. I found a couple of sites that mentioned the use of a solid state relay. Basically, the spindle is still powered by the 48v power supply that I bought. But the power supply itself, is controlled by computer. The cnc controller is 24v.

I soldered the appropriate wires and added spades. The 110v power supply comes into the reply, and I connected the other side to the live place on the power supply. On the vdc side of the relay, I contacted wires to the jst connector, which goes to the spindle connection on the control board. The idea is that when the spindle is turned on in a program, the 24v triggers the relay, which turns on the 48v power supply.

It didn't work.

I quadruple checked everything. The wiring was right, but the relay wasn't triggering. I thought the relay was bad. Then, I thought of connecting a 12vdc Transformer that I'm using for a marble run, to test it. The red light came on. So, I knew the relay was good. I ran the spindle directly off the power supply, so I knew it was good. I also ran it directly off of the 24v controller, which ran it at 1/2 power, so I knew it was good.

SO...the only thing I could think of was a failed solder. I redid everything. I was certain it had to be something I did wrong. Nope. Nothing.

I saw someone had put a 100k ohm resistor on the positive side in the wire coming off of the relay. This was to act as a pwm interference filter (which I had replaced too, 3x). I took it out.

Damn if that wasn't the problem all along! The computer took over running the spindle for me, even turning off the 48v power supply.

But, the spindle only ran at almost full power.

1. I wanted full power
2. I really wanted full computer control over spindle speed.

I unplugged the rheostat on the power supply, and using the existing connection, made a wire with a jst to run to an empty 5v with ground connectiom on the controller. I also piggybacked a wire, with (cringe) 100k resistor on it, connected to a + terminal on the power supply.

I now have complete computer control over the power supply, spindle, and I can control the spindle speed from the computer.

I have the whole enchilada, as the saying goes.

However, I now have a mess of wires to contend with. Tomorrow, I'll put them in sleeves, and get them under control!!

I put the relay on the aluminum t-slot side. I've read they can heat up, and although the relay has a heavy metal back, I figure the aluminum will act as an additional heat sink.

[TEETERFAN]

Inspiring, you sure did battle! Wow. I admire your problem solving ability. Sorting out electrical problems is very challenging, but you stuck with it, you didn’t have to compromise your vision. And you have a wonderful addition to your shop!

[Diane Kauffmds]

Just when you think it's all over...Murphy calls.

Well, the cnc was performing great, then all of a sudden it decided to drill a hole clear to China while engraving my logo into a headplate. I thought maybe I'd put the wrong parameters into the CAM program (maybe inches instead of mm). Tried again and it happened in a different place. Then the really bizarre started happening, like the z-axis moving when the computer told the x-axis to move.

On top of this, my beautifully rounded lettering suddenly looked like a kindergartener had run riot over it. Long story short, I had a couple of things happening.

First, I found backlash in all axis'. The z-axis in particular had severe backlash. I totally disassembled it and found it needed adjustments in several areas. Now it's as steady as a rock. I had anti-backlash nuts on the x and y axis', but the springs seemed awfully weak. I replaced them with better springs. After recalibration, everything looks 5000% better.

Secondly, the bizarre behavior was elusive, until I realized that it coincided with the cold weather, the solid state relay that I installed, and I saw dust in the dustshoe that was full of static. I had severe EMI/RFI problems.

The first thing I did was attach a copper wire to the bare aluminum side of the machine; I attached the opposite end with a ring clamp to a copper waterpipe, thereby grounding the machine. I had soldered a 100 ohm resistor to one leg of the machine side of the SSR. I was going to solder a 1uf capacitor across both legs, but was unsure of what kind to use. I opted to wrap all wiring with EMI/RFI copper tape.

The tape, made to correct emi, has conductive adhesive. It's extraordinarily thin. I carefully wrapped all wiring, taking care that it came in contact with all of the wires that came off of each jst jack. Doing this also cleaned up the tangle of wires from the controller. I shorted a lot of the wires too.

For good measure, I soldered a thin wire to each copper wrapped wire, attaching it to the bare aluminum sides. So all wires are wrapped in emi tape and grounded. The tape abraids easily. I used spiral conduit to cover the wires to protect the tape.

I'm happy to say the cnc is up and running better than ever. It's even quieter. I cut a couple of pieces of inlay today for the first time. They came out beautifully.

So, God willing and the creek don't rise, the cnc build saga is over and I can finally figure out how to program 3D projects on it.

20221222_182726.jpg

20221220_124100.jpg

20221222_183055.jpg

20221224_125430.jpg