[fuper]

Well the angle grinder did the trick! $29 for a Ryobi. I bought 3 different cut off blades just to be on the safe side. The thinnest one I tried first, and it worked perfectly.

I have a pretty big post on my blog Fuper | Home of the MillBox open source CNC MillFuper | Home of the MillBox open source CNC Mill

Here are some pics:

[RomanLini]

Cool! It's a good feeling when it finally looks like a machine instead of parts hey?

Is there any slop in the linear bearings? I noticed you used just one bearing block on each side. That might cause too much slop, which would make the top of the gantry wobble (as the bearing slop is amplified by the lever length). If you have a dial indicator it's worth putting it on the mechanicsm and seeing how many thou of slop or movement it might have.

[fuper]

Hey -

Been a few days. The in-laws are visiting for a few days. My father in law is a retired mechanical engineer. He worked at for for over 20 years and he gave me a surprise gift!



Anyway -

I got a couple boards made from batchpcb.com First time I had a board made. They are fairly easy to use but I don't know how they compare to everyone else. Takes a while for the board since they send a bunch of orders in at once, done in "batch", over in china.



My blog has a little more info: Fuper | Home of the MillBox open source CNC Mill

Will post in a few more days once my guests have left. I should be close to testing the gantry movement using the steppers.

[leaveme]

Nice simple build!

As Roman said, you need dual bearing blocks or a wide bearing block to avoide gantry wobble.

Cheers!!

[fuper]

Thanks leaveme!

I just got the two acme rod connector plates. I'll post an image soon. These have a through hole with the acme threaded flange. The plate screws into the vertical member of the gantry.

You and Roman are correct about the linear bearing slop. I was surprised but well it's a learning process right?

I will continue with the current design and get the thing to work then I'll go back and tackle the wobble. I agree two bushings/bearings will likely prevent the movement. I'm leaning towards using a single or couple of pieces of Delerin and just bore out the 16mm hole.

My problem is I cannot get holes of that size at all accurate with my Taig mill which is why I'm punting on the problem for now

[fuper]

Well I got the stepper driver boards built. They run hot, way hot. The TIP131 transistors I mean.

I asked around and come to find out the current output from the 4027 isn't enough to fully turn on the TIP131 so it runs hot. I measured it with a cheap infrared thing I got at radio shack, and it said over 150 degrees. I clamped a heat sink or two on them for testing but it's not shown in the pictures.

Anyway - after much fiddling I got the stepper to work correctly. I'm using an arduino to send a step signal for testing purposes. The sample blink program comes in handy!

Well turns out this may not be a great design going forward due to the heat issue and I still need fat resistors in order to drive the stepper at a higher voltage. I'm only running it at 5 volts which is pretty low.

Anyway with the voltage and current i'm dealing with here, I think these drivers will not suffice.

I suppose I could redesign and get the TIPs to fully turn on, using regular transistors I guess, or a quad opamp, but I don't know enough about electronics to know if this would work or not.

Here are the pics:

[RomanLini]

Hi again, are those microstepping drivers? If not you might want to check out the Linistepper; PIC Linear Stepper Motor Controller

It is an easy to build kit sold by one of the forum regulars (and designed by me, many years ago). It still gets hot (and needs a heatsink) but they turn very smooth due to the microsteps. They are also open-source if you want to make your own boards instead of buying the kit.

[fuper]

Hi again, are those microstepping drivers? If not you might want to check out the Linistepper; PIC Linear Stepper Motor Controller

It is an easy to build kit sold by one of the forum regulars (and designed by me, many years ago). It still gets hot (and needs a heatsink) but they turn very smooth due to the microsteps. They are also open-source if you want to make your own boards instead of buying the kit.

Hey RomanLini -

That's pretty cool you designed this! I know about the Linistepper. Did you do the circuit design and/or the software development? I'm a programmer during my day "job". Mainly C when possible.

Anyway does the linistepper do any current control, like chopping as I've heard it's called?

The circuit I built, which I took from elsewhere, does neither microstepping nor current control. so if you want to up the voltage you gotta get larger resistors. It's made out of two oldie but goodie logic chips, CD4070 a quad XOR and a CD4027 which is a dual flip/flop with master/slave reset.

Cheers

PS - are the eagle and brd files available for the linistepper?

[RomanLini]

Yeah re the Linistepper I did the linear current control concept, electronics, PCB and firmware.

It does not "chop" but it does regulate current linearly. That's a more "old fashioned" way to do it but it works very well for smoothness, if of course you don't mind the heat and having some large heatsinks on the Linisteppers. It would provide a big performance improvement over your full-step non-current-controlled stepper drivers. But lots of people build machines with full step drivers, so don't get put off the idea of using the drivers you already have. Probably the most important thing in first CNC building is to get od done, and working. There's always time for improvements and/or a second machine build later.

No my original Linistepper eagle PCB files are not released but I think someone in the "open source electronics" section of this forum made a Lini PCB design and released his files there.

[fuper]

Well I got the stepper driver boards built. They run hot, way hot. The TIP131 transistors I mean.

I asked around and come to find out the current output from the 4027 isn't enough to fully turn on the TIP131 so it runs hot. I measured it with a cheap infrared thing I got at radio shack, and it said over 150 degrees. I clamped a heat sink or two on them for testing but it's not shown in the pictures.

Anyway - after much fiddling I got the stepper to work correctly. I'm using an arduino to send a step signal for testing purposes. The sample blink program comes in handy!

Well turns out this may not be a great design going forward due to the heat issue and I still need fat resistors in order to drive the stepper at a higher voltage. I'm only running it at 5 volts which is pretty low.

Anyway with the voltage and current i'm dealing with here, I think these drivers will not suffice.

I suppose I could redesign and get the TIPs to fully turn on, using regular transistors I guess, or a quad opamp, but I don't know enough about electronics to know if this would work or not.

Here are the pics:

I thought I was going to scrap this design.

The nice guys over at Electrotech helped me out and said I should try pin compatible logic level mosfets instead. The reason is they are voltage controlled instead of current controlled. The Cd4027 doesn’t put out sufficient current to drive a TIP31 but it can drive a mosfet nicely.

I have more information on which parts and the math [ as much as I understand!] on my blog.

Swapping out the TIP31 for Power Mosfets | | FuperFuper

For more information from the guys at electrotech on possible part replacement and the logic behind this read it over here

[fuper]

Yeah re the Linistepper I did the linear current control concept, electronics, PCB and firmware.

It does not "chop" but it does regulate current linearly. That's a more "old fashioned" way to do it but it works very well for smoothness, if of course you don't mind the heat and having some large heatsinks on the Linisteppers. It would provide a big performance improvement over your full-step non-current-controlled stepper drivers. But lots of people build machines with full step drivers, so don't get put off the idea of using the drivers you already have. Probably the most important thing in first CNC building is to get od done, and working. There's always time for improvements and/or a second machine build later.

No my original Linistepper eagle PCB files are not released but I think someone in the "open source electronics" section of this forum made a Lini PCB design and released his files there.

Pretty cool dude! Are you an electrical engineer? I've had to self teach all my electronics. I had a few physics courses in college that covered the basics of electricity and I had a digital logic class but nothing really involved.

Linear current control - ya I'm going to add a couple of high wattage resistors to limit the max current to the stepper. I am assuming this is called linear current control. Its not power efficient but then again I'm not powering this thing from a battery so it's not a big deal for me. I might swap out the discrete components and code a small AVR to do microstepping, much like your Linistepper. I want the opportunity to learn more about microstepping. Not sure yet.

I am swapping out the TIP31 transistors for mosfets. If that works, and I suspect it will, I'll finish my build using these boards, like you said, and I agree fully, the most important thing is getting the project done, then going back and improving the design.

[RomanLini]

...
Linear current control - ya I'm going to add a couple of high wattage resistors to limit the max current to the stepper. I am assuming this is called linear current control.

Resistors won't give proper current regulation, but they will help compensate for motor inductance.

A better option than resistors would be a constant current DC PSU for each stepper motor.

Better again is a microstepping constant current (per phase) driver.

The standard reading is "Jones on Steppers" which is a little outdated now but does cover the different current control methods quite well;
http://www.divms.uiowa.edu/~jones/step/current.html

[fuper]

Well I got my Mosfets in the mail and inbetween looking for a new job [total pain] I got everything soldered.

I'm happy to report everything works! I would likely make changes to this board now that I am using Mosfets, likely using a pull down resistor on the gate in order to make sure the Mosfet is off Mosfets have small capacitor on the gate that can cause the Mosfet to stay on briefly even if you remove voltage to the Gate pin. If you have ever seen something with a motor that wiggles a little when you turn it on, this would be the kind of behavior that might be caused by not properly driving the Mosfet.

I wrote a much longer write up about this on my blog: Home of the MillBox open source CNC Mill

Now having done all of this, I broke down and purchased 4 Polulu stepper drivers: Pololu - A4988 Stepper Motor Driver Carrier

I am happy I got my board built and running, it was a great learning experience but I think for the purposes of getting my machine built with the least amount of headaches I have decided to go with commercial stepper drivers for now.

Switching gears, I still want to get a through hole stepper driver easily available. I like the L297 because of the wealth of information and the fact its through hole. I just don't like the $10 price tag on the chip [digikey]. I looked around for a pin compatible replacement and found these guys. LSICSI. I'm trying to get pricing as they imply they are cheaper than the ICs made by ST. More information is here:

LSI CSI - Design and Production of Full Custom and Standard Integrated Circuits

They sell a pin compatible through hole part. I could not find any seller online so I just contacted them directly and asked for pricing of 250 chips. If the price is at all decent I will likely buy 250 of them especially if I can make them available for others without having to shell out $10 a pop.

[fuper]

This was a small adventure. I bought a computer with a parallel port since I've moved to laptops in the last 10 years and never looked back. I live in the silicon valley so getting parts is not a problem, although finding a parallel port cable kinda was.

I just got everything put together running windows XP and the Demo version of Mach3.

Once I realized that you have to disable EStop in order to get Mach3 to do anything, it was all a breeze. For those that don't know, there is an EStop pin that by default [with nothing connected to the parallel port] will report that the emergency stop button has been pressed so Mach3 will not do anything until you address this. Its a wise default for safety reasons but not necessarily obvious if you are just starting out.

At this point I'm just using my Oscilliscope hooked up to a pin and issuing a simple Gcode command. Sure enough I saw what I expected!

I purchased a few stepper drivers from Polulu and this weekend I'll wire them up and test my first working axis. I'll post more pics soon.

I've got a little more info on my blog but this post covers the gist of it.

Cheers,
Shane

[fuper]

Hi!

I was unemployed until December, then got a job and have been busy since. Things are quieting down at work now and I've had time to start working on my Mill again.

Now that I have a paycheck, I have less time, so I decided to speed up the process and purchase a Gecko G540 with steppers in order to finish the hardware design first.

Here is a video I took shortly after I got everything connected and moving.



As always, you can check out my blog for files, source and general posts - Fuper | Home of the MillBox open source CNC MillFuper | Home of the MillBox open source CNC Mill

Cheers!

[fuper]

Hi

I'm still alive. I've been updating the plans in Solid Edge and they are available for download at Fuper | Home of the MillBox open source CNC MillFuper | Home of the MillBox open source CNC Mill

Here is a picture of the z-axis, which I admit I'm not that happy with. Next time I'll design everything before I start cutting. At least that is what I keep telling myself

The picture of the z-axis is a shot from below. I will do a redesign of the z-axis at some point I imagine, to make it thinner. I used the same sized rails as it was more cost effective to buy multiple pieces that are the same.

[fuper]

Here is the z-axis as promised:

[fuper]

Millbox - jogging all three axis - YouTube

This is just a quick video of all three axis moving at the same time. Little wobble on the Y-axis but I'll fix that.

[fuper]

Well I've been busy with a new job however I managed to take my first cut with my creation. Given that I knew literally nothing about maching or metal working I'm pretty happy with my accomplishment. The video is a little jumpy, sorry I was really babysitting it as it was nervous about a crash happening! I drenched the thing with WD40 before I started the program but that probably wasn't necessary at the slow speed I was cutting at.

Everything was built with hand tools and a Taig manual mill. The spindle is borrowed from my Taig, so I cheated a bit there!

First cut. Cutting a hole in aluminum - YouTube

Not sure what I'll do next, likely use it to make a better version

Shane

[askjerry]

Wiping the chips with a brush is not a good idea. I was doing that one day with mine... in about a half second the brush was thrown across the room and sticking out of the sheetrock. I was using a 1" (25 mm) diameter bit at the time. Anyway... use compressed air... nothing to get tangled in the cutter.

Nice job for a first build. Looks like you have a pretty accurate design. On your next design hopefully by now you have figured out that the threaded rods are to be secured with thrust bearings and that the flexi-shaft connection is only for rotational transfer of power... not to keep the rods from moving along a lateral path. I have seen that mistake before... the best method is a captive set of two thrust bearings in a bored hole with a bolted down retainer. The threaded rod is passed through and secured with two nuts locked tight. This method will eliminate much of the backlash, but be sure the other end has freedom to move... if you lock both ends you can get other issues.

Good luck!
Jerry