Black 3018 CNC router

paul d wrote: Cor Klippy the name cincinatti brings back memories! first cnc mill I used was a cincinatti with moog hydrapoint control with paper tape memory!!

I made Tape emulators using my own design hardware and Sinclair Spectrum computers. You could take a new program to the machine on a Microdrive. I wrote the Machine code myself. It was so successful I took my designs to the UK in 1985. Never got a cent in Royalties! I later saw in a Machine tool magazine there were quite a few machines modified in the UK! Thieving bugger!

One of the reasons I chose this particular machine is its ability to use a laser, my thinking being it's easier to use a proven system than re-invent the wheel. So after checking dimensional accuracy (one of the previous pics shows the 4 holes I drilled) which is spot on however the diagonals differ by 1.5mm on a 150mm square, which means the carriage guides are slightly skewed.
I unplugged the motor and mounted the laser, plugged it in, and nothing! I could not get it to work using the supplied program. Finally, quite by accident I spotted the problem.

This is the control board without its cover, at the bottom left are 2 white sockets, the laser which has 3 wires, it's a PWM plugs into the 3 way socket. You can see that the shroud is displaced compared to the 2 way one on its left. The 3 way shroud was the other way up, when the laser was plugged in it was getting no power, much like a servo that's plugged in the wrong way round. I pulled the shroud off, put it back the way you see it and bingo, a working laser! The laser power is selectable from within the operating program, a low power setting is available for focusing, although from what I can see the focus seems to change when you put it on full power! tried 1 or 2 sample programs and the laser did what it was supposed to.

So, the firmware on the control board it seems is called GRBL (you can pronounce this Gerbil or Garble). Thanks to Simon I went looking for clues and found everything I needed on a place called Github. The first thing you need is a means of talking to GRBL. I initially tried Telnet but it wouldn't connect so I then tried a program called ZOC. I won't go through all the failures, there were many!
Firstly everything must be switched on, this may sound obvious but you can actually speak to the Arduino (the interface and main computer) without anything switched on except the computer.
So, start ZOC.

On the top left of the first screen you have a GUI (small screen with a green arrow) click on it, then click on options and you will get the above screen, these settings are the only ones that will work! My machine always connects to COM4, other computers may be different.
Click OK and you should get this.

Type $ for help right? I was stuck for 2 days here!

You can type $ until the cows come home but the keyboard doesn't work! I changed keyboards and rodents, nothing, no improvement, finally in desperation (actually I think it mind have been mind in neutral) one day I switched on, connected, typed $$ and pressed enter, bingo, this came up!

The !@#$%^&*() program doesn't echo to the home screen!
I found the rapid setting for X and Y at 5,000mm/minute! Totally unreasonable. I had already disconnected the lead screws and found that my problem didn't go away, reducing the rapids to 1,400/min gave me a reliable rapid (1,500 was a no no) so I came off that a bit and settled on 1,250mm/min for X and Y with and acceleration of 100mm/sec/sec these settings seem to be rock solid.
Now I could put in long distances without fear of the motors baling out, which threw up another problem!

I was considering investing in one of these but have now decided to buy a new fret saw instead ! :S

Having re-connected the mechanical side of things I could now re-assess rapids and acceleration. With a load on the stepper motors I had to reduce the acceleration to 60mm/sec/sec. I also had to reduce the rapid to 1,000mm/min. While ding this I noticed that the couplings between the motor and lead screws weren't running true. On the X it was so bad the motor was actually wagging the mounting! I immediately assumed run out in the couplings so made some new ones.

A common error in mass production is turning the work piece around to finish it. The coupling needs a 5mm hole on one side and an 8mm one on the other. The only correct way to make this is to drill through the stock with a cutter guaranteed not to wander off, then drill to size with a 5mm drill all the way through, then drill halfway with an 8mm drill, ie, all in one setting.
This sorted out my problems in the Y axis and 80% in the X axis, but, the shaft of the stepper motor is slightly bent! But I am now ready to cut balsa!

Great thread Klippy, really interesting. I have a 3040T CNC but never set it up. So its sitting there waiting for me. Your thread has made me plan to get it going over Christmas.

Yes drill through but an alternative is to then use a boring tool.

The two diameters will be in line to the accuracy of the lathe.

I know I've come over a bit negative on this but actually it's quite a nice machine. In fact I decided to look at lengthening the Y travel. The current side rails are 330mm which translates to 190mm of actual cutting travel, although because of the design there's no limit, other than practicality, to the length of your work piece, a 3' slab of balsa will pass right out through the back of the machine. It's an easy job to change the side rails, lead screw and guide rods. I initially thought I'd extend to 1metre, without P&P that was $84.35, not the way to go, I've ordered the materials for a 600mm slide, $36.12 without P&P, much better. This will allow me to cut up to 460mm parts, I can't imagine cutting parts longer at the moment, should the need ever arise, parts are available to go to 1200mm side rails or 1060mm part length.

These ar now on order, so much better than supplied. I think the Chinese should think about a 2 tier system, low end with crappy parts and high end with all the best stuff and no messing with downloading information and programs, just a CD with all the current stuff on it!

Just the job!
My super cheap 'bottom of the heap' Anet A8 has exactly those on its Z axis lead screws. Even so one of them is very slightly out of alignment. :woohoo: )
However the very 'flexibility' built into such a coupler means it is no problem.

Somehow my life isn't my own at the moment. But anyway, I re-aligned the Y axis slide and immediately ran into losing pulses in rapid, so started over aligning the way I should have instead of cutting corners. I finally made a cut with an engraving blade (you get 3 in a kit)

My first ClarkY!

I have a large archive of airfoil coordinates, turning them into a CNC program is proving to be a very time consuming exercise!

I thought one of the first jobs would be to cut some riblets for my Tiger Moth. Looking at the wings I assumed ClarkY, Wrong, a 210mm Y has a maximum thickness of 25mm, the Moth wing is 21mm. I tried Raf28, no, that's the upper engraving. A ClarkY reduced to 21mm thick is just not bulbous enough at the Leading edge, that's the lower engraving.


Any ideas?

Arrh! In a blonde moment I pressed the return to zero button but one of my holding bolts for my sacrificial board was in the way! It didn't stop the movement but bent the spindle shaft! I now have a very serious vibration.

The full size Tiggy, and most of the other Moth variants, use RAF 15.
A classic WW1 biplane wing section that is just 6.5% thick which is why many scale biplane models look 'wrong'.