[epineh]

I love printers, they have many uses, one of which is the inspiraton of this build - cheap parts !!!

I acquired a probably working Canon printer today, and got to work making sure it will not work normally ever again

The goals for this build is are follows :

1 - CHEAP ! This is going to be the cheapest but hopefully not the nastiest build log around, I do not apologise for this, rather I am embracing the idea. Parts will either be second hand, or so cheap it won't really matter. I will have some things to aid the build. Hopefully I will have my main DIY router going in a couple of weeks and will use it to fabricate a lot of parts. Also I have built a couple of servo drives, one based on the Elm Chan design, the other the DSPIC servo design, this machine will be a test bed for both.

2 - I need this machine to drill PCB's ONLY ! I will muck around doing some isolation routing once it is finished but this is NOT the design target. I want to be able to load a 300mm by 300mm sheet of copper clad fibreglass, hit the button then let the machine drill out a full panel of circuit boards at a time. Then I will be using either the direct printing method or most probably toner transfer as it will align much easier in larger numbers.

3 - Simple. I hope to provide some basis for the never ending stream of people wanting to do PCB milling. This machine is not designed for this exact purpose but with a little extra work should be able to do so.

4 - Fast. Direct drive timing belts driven with DC motors and linear encoders from printers should do the trick, the motor's may not be from the printer though.

5 - Cheap ! Just in case you missed it

Don't expect any nice pretty CAD renderings of what it is going to look like, this is a "flying by the seat of my pants" type of build.

Next up some photo's...

Russell.

[epineh]

Here are some of the parts I have "found"...first up the printer "before shot", nothing exciting here.

Next up the internals ready to be removed. Next photo is the DC motor/timing belt and linear encoder optical strip that forms a DC servo.

Last photo is the print head assembly (well some of it) and the rail, the rail is nice but the print head slide is just plastic so it will not be much use.

[epineh]

Now it gets a little more interesting, first up is the optical sensor used to pick up the lines on the strip and convert this to quadrature, this just happens to sit inside the print head assembly (what is left of it) nicely, one of the first jobs is to put this together somehow, hack the PCB the sensor is on to power it up and take the quadrature signals from. Then plug this into a servo drive and try jogging the thing around. I will be looking for smoothness and control, if it isn't good enough it will be onto plan B.

Next two photo's may be part of plan B. This is a rotary encoder, same sort of thing as any servo motor will have stuck to it, but judging by the diameter and very fine slits along the reading section it is a pretty high count job. I won't know until I can hook up a drive and calibrate it but I would guess around the 2000 Counts per revolution ! Not bad for free There is a little DC motor that connected to this via several plastic cogs, probably not much use.

Last photo is what is left of the printer, this will go into a bucket just in case I can rob more parts, there is a little stepper on one side, very tiny, not much use for this project, only thing I can thing of is to convert it to a MPG for a pendant.

Russell.

[chich2]

Hey Russ,
Looks like you are on you way to another HUGE thread. This will be an interesting project. I will be watching.

Chich

[epineh]

Hey Chich, welcome to the madness

I am thinking of sticking the print head and sensor PCB back into the printer "as is" to test the servo function, this will be the first constructive work to be done...

Russell.

[epineh]

I had enough of the print head left to be able to fix the PCB in place, it could use a little hot melt glue to keep it in place but that is no problem. It also slides along the bar nicely, though the bar is held in place by two bits of springy wire and I am not too sure how they go back together, I hope it is good enough to test.

Took some pics of the optocoupler device, not the best but they will do. I am pretty sure I can get it to work, I am going to assume 5V logic supply and hook it up with that. I would think in the photo showing the underside of the board, in the vertical row of pins on the left, the bottom two pins would be the quadrature output, next one up is the Gnd and top is +5V logic supply. The two pins at the top towards the right power the LED, left is + and right is Gnd. Too tired to warm up the soldering iron up tonight, will pick this up tomorrow.

Russell.

[epineh]

Well today had mixed results, didn't work properly but didn't fail completely.

Mounted the DC motor, belt, encoder strip and print head PCB back on the printer, connected up the servo drive to the DC Motor and wired up the +5V supply to the print head PCB and took A and B encoder signals from the PCB to the servo drive.

Powered up and played with the servo drive settings a little, strange results, I can feel the motor wanting to turn slightly in one direction and as I force it back it fights a little harder, a good sign that the position loop is doing something...

I brought the torque setting down and the drive faults as soon as you try to move the print head, but still no holding action. With no step/dir to the drive it should hold in position, with response according to the drive settings.

After a while of this I checked the PCB and I had dropped one of the encoder pulse wires from the board, unfortunately I have to leave it for now and continue tomorrow...

First photo is the servo drive, logic supply, H Bridge and print head PCB connected.

Middle is the print head PCB mounted in position with some "extra" wires added

Last photo is all of the above, I think I will have to work out some way of connecting the encoder wires a little better, I just want to test the idea for now, eventually the encoder itself will be removed and a dedicated PCB made for it.

Russell.

[epineh]

No need for me to type anything here... check out the video



Russell.

[Crevice Reamer]

Nice project epineh! Great video!

CR.

[KTP]

Very nice! I love the way you have tapped off of the print head pcb to get the linear encoder signals

If you run across any HP 7550 plotters in your area for free or dirt cheap, they have a pittman DC servo motor with 500 line (2000 count) quadrature encoder mounted on the motor. I did some tests with hooking one up to a servo drive awhile back before I went in another direction. Pretty strong little motor, should move a dremel tool OK. In this vid, it is running over 2600 IPM. http://www.skyko.com/videos/HP7550A.wmv

I will be watching your thread with interest. It seems a very *green* thing to do...making a cnc mill out of printers destined for the landfill

[epineh]

Cheers guys

I hadn't thought about the recycling factor, I was just going for cheap

I will keep an eye out for those 7550 plotters, sounds like a good thing.

I am not too sure how large a servo motor I will need, I am planning to make the table an X Y version, like a standard milling machine, this way I can keep the weight down a little if I am careful, lighter means smaller motors needed, and hopefully still nice and fast. Means a bigger desktop area but too bad.

Next is to scrounge for some kind of slides, I have a couple of ideas, just need to try some out. I will make the X Y part of the machine next, and move the linear encoders from the printer to the table, and test again...

My plan for the Z is for small travel, maybe about 12-25mm (1/2-1 inch) will be plenty, this will only be drilling 1.6mm PCB material, but I may as well allow for a little room. I am thinking of using a BLDC motor from a R/C plane and associated speed controller, they are cheap and should be more than enough for what I need.

Russell.

[chich2]

Very Cool Mr Russ!!!!

Chich