Back from the dead... with a new toy!

Been a nightmare couple of months, kinda fell off the face of the planet for a bit there. Still wicked busy (which is a good thing), but I haven't had time to tinker and such. Not having the time to actually DO anything left me at times with sketching out various ideas and little code snippets. Well, today I had a few hours to myself with no pressing emergencies, and declared it a Tinkering Day. Actually, I had a few hours yesterday also, during which time I completely rebuilt the Frankenbotic Thinganator with new stepper motors and ULN2003's instead of discrete transistors. I'll post an update after I set it up and work in a code fix that corrects for the radial distortion in the renderings. Way too much geometry was needed.

However, I have a need for something much more useful.. so I set out this morning nice and early for Home Depot for some supplies.

Now, those that know me know I am the master of epoxy and duct tape. I went into Home depot with a budget of sixty dollars, and came in at around fifty, because I was able to get the lumber I needed from a "cull bin", where they toss odd leftovers of whatever, and charge next to nothing for them. From the bin I was able to find a partial sheet of 3/8" plywood, which fit my needs and cost a whopping fifty cents. Next over to the cabinet hardware, where I got two sets of ball bearing drawer slides for twenty five bucks (the most expensive part of the project). Next, the hardware aisle, where I grabbed a couple of two foot sections of 1/4" threaded rod, two three footers because they were cheap, and a couple of handfuls of the various hardware odds and ends like some washers and screws.

By the time the kids got off the school bus I had the basic framework cut and laid out, then hit the box of those steppers I love so much, and pulled out the epoxy. It was fully assembled with epoxy curing by the time dinner hit the table.. wired up while watching TV and chatting with the wife.. and threw some basic testing code at around 11pm. Total time from concept to working prototype: fifteen hours.

For what?

My Arduino-driven, 13"x13" travel, slow but extremely accurate (screw drive, gear-reduced stepper) XY table. After a small amount of testing, I will be adding the wooden gantry and another screw-drive stepper to complete the project... to raise and lower my dremel on the XY table... poor man's CNC. Fifty bucks, not including the Dremel (off brand one at that).

This may be my most abhorrent-to-reasonable-engineers project to date. And yes, I am fully aware I need some more epoxy for those mounts...

It's VERY slow.. but it works and has an impressive amount of torque. When the photo was taken, it was running a simple sketch just moving the platform back and forth and left to right.. flawlessly. High resolution CNC router, anyone? With these steppers and their built in gearing, one rotation of the drive shaft is 2048 steps.. dang high torque ones at that. More than enough to drive the threaded rod and move the table. Then take into account the drive screw is twenty threads per inch, and you get well over 40,000 steps/inch. Sure, it moves like two inches a minute... but I don't care about that, really..and of course I know actual accuracy will be a miracle if it's within a sixteenth of an inch reliably. So far-- it's working, which is never a bad starting point... I should have something more to show on this in a couple of days.

Good to be back, doing all the things I shouldn't!

Are you getting much backlash on the threaded rods?


looks pretty primitive :slight_smile:
but if it works...

the x axis (left to right) seems to be at a slight angle
does that make the motor tilt as the carrriage moves?

what's the "thing" resting on the other motor?

Not exactly sure what you mean by backlash.. I'll assume you mean 'play' in the motion, looseness caused by the inherent meshing of gears, etc.. It's surprisingly small. Considering design and construction, it's downright amazing. If you mean mechanical springback from motion.. It's practically nonexistant. The motion is SO slow that there's very little mechanical stresses on it it seems. When I do some more tests later today, I am already convinced the biggest issue with the design is going to be speed.. That and because one motor is not located in the center of the platform but pushes from the side, there is likely going to be a bit of error introduced to one side.

The x axis is at a slight angle because that's how it ended up lining up when the epoxy set, the motor moved slightly. Since the carriage doesn't get much closer than you see here, there's enough flex in the rod to take care of it. Only a small amount of epoxy holding that one, so I will be able to possibly move it a bit should I decide it's a problem. The 'thing' you are talking about is just a scrap of metal from my junk box, bent and epoxied to the motor as a stablilizer or mounting bracket. Everything I do, I try to do cheaply, because I am on such a limited income... I 'recycle' frequently. Take my advice on this one, and keep a box of random hardware for this kind of thing.. Rip apart a floppy drive or cd changer and you'll have a ton of random screws, brackets, frames, gears and all sorts of goodnesses. Printers are particularly good sources for parts including stepper motors and their driver IC's, and they also have fairly decent power supplies in most of them, to drive the steppers and print heads. A junkyard bound printer is a minor goldmine, especially the older dot matrix printers.

I buy epoxy in contractor pack, four ounces of each resin and hardener to make eight ounces of five minute set no-shrink tinkering goodness. I think it costs thirteen bucks for that, I am never without that stuff.. Nearly as useful as duct tape. It sets in five minutes and is usable (2000lb bond it says I think) in twenty, full cure in a couple hours. I also keep two part epoxy putty, which may be used to beef this monster up a bit, frame wise.

I pride myself on the crudeness of many of my devices, but I think that might be a common illness around here... In fact, I know it is... Pretty much defines us Arduino tinkerers, doesn't it?

I am predicting the error will increase the further way from the origin it goes, even if the accurate area is only about eight by eight inches, I'll be delighted. I do have to say I am amazed at the smoothness of motion of the drawer slides as ball-bearing extensible rails.. Screwed into the plywood at several points on their length, there's almost no wiggle in any direction.. Simply enough screws to make it a rigid system, and the bearings are well made and are supposed to be able to take a few hundred pounds easily.. Way more than is needed. I can't see it becoming a problem unless I put way more stress on it than I plan to.

I am figuring this may take a couple of hours to do a detailed cutting.. But I am okay with that. I have no problem with going off to do something else. In fact, in that respect, I am more concerned with the amount of time the dremel will be running than anything else...

I am setting this upnwith the intent that a cut

Hey, Focalist.

Backlash is what happens when you reverse the screw drive. The threads don't mate perfectly with the threads on the nuts, so there's a bit of movement until the threads contact again. There's probably more shift in the drawer slides than what you'll get in thread backlash. But I have to say, that's a nice piece of work.

Red Green would doff his hat. :smiley:

To give an idea, one of the things I want to do is be able to create a few levels of depth, which I can use via photoshop thresholding to create a three dimensional-ish carving from my photography. Pretty big order I know, something that really would be tough on a professionally made machine costing tens of thousands never mind THIS contraption..

However, if I can make it work for even a handful of carvings, I can turn those carvings into enough cash to consider looking for a professionally made version, etc etc etc. CNC tables minus the motors.. XY and rotate- can be gotten off ebay for a couple hundred bucks. My LED 'eyes' project from the fall was used as chromatic lighting for a photo shoot a week ago.. Roughly the same light as a small incandesent spotlight, with infinite gels. Worked well, saved me from buying gels and spots... A solid hundred bucks. Was a paying gig, too, so that project is now in the Profit column. Like my photography income buying better gear, I call it 'feeding the disease'- in only the best ways.

And yes, if that's backlash.. There is very little, I think I will be able to accomodate for it in code without too much problem. Since there is such a tiny step distance, I can happily call a thousand steps on the motor a single step in terms of movement, there is plenty of fudge factor at this gear ratio. Like I said, it's going to be a dream if it works to within a sixteenth of an inch... It looks like that ought to be well within the possible given how it's going so far.

I'm having a ball with making it, anyway.. And I might even be able to do something cool and useful.. And profitable.. With it? That's my kind of project!

If I can get to it later, I might set up the gantry with the dremel just fixed in position manually up or down, and see if I can use the beast as is to do a couple of cuts.. Or at least a marker and some paper and have it do some geometry to see how linear and accurate it actually is. I think I am going to be pleasantly shocked at the outcome... happily the motors don't seem to even be working very hard, warm but not hot even after an hour of back and forth on the two axises. I've already decided to add in one of the opto22 logic-level solid state relays I have kicking around and only switch on the dremel every 'x' steps to avoid running it nonstop for hours.. that's easy enough. When I say this thing is slow, I mean it looks like it may have a travel speed of around an inch a minute. Really. heheh. In the end I think I may have to factor in cool down periods for the dremel.. anyone know what the recommended maximum "on" time for a dremel kinda tool is?

Did I mention Arduino Rocks?!?!

In any case, I'll be moving this over to a project in the Exhibition thing later.. Just been away from the smell of solder and you bunch of like-minded goofballs longer than I intended! Good to be doing everything the wrong way again, it so much more fun if you approach life in the same manner as Wile E. Coyote...

I used a wood & threaded rod based cnc pcb drill for many years & got about 0.010" accuracy in the x/y plane which was a lot better than the laser printer I used to produce the pcb artwork.
I spent a whole day debugging the code driving the steppers looking for errors before realising it was the printer causing the problem.

Nice project. Lucky for you it was cheap. Good idea to use the "metal slider".

I am not to found of using glue for the motors. I will probably use metal brackets or make my own brakets. You may use those motors again...

Most of my piece of woods I need came from the garbage. Just need to saw and cut.

I pride myself on the crudeness of many of my devices, but I think that might be a common illness around here... In fact, I know it is... Pretty much defines us Arduino tinkerers, doesn't it?

Yep...same for me. Thinkering and building the "McGyver" way. :grin:

It's tight for woodwork, that's for sure!

Backlash is thousandth of an inch kind of thing. If the rod threads and nut threads had no room between, they would grind up just turning. So you turn one way, one side pushes and the other has that gap. When you turn back, it takes part of the turn to close that gap before the rod begins to push on the nut in the new direction. How much is the backlash and it's so small you need a micrometer to find it. It don't matter for wood!

$#!+ man, you could sell those!

I make some goofy crap. I tossed a pic of the beast and a basic description of it up on Facebook, and half a dozen friends were emailing me within an hour.

About the only thing I may change is the motors. That gearing is SO slow and that accuracy level just isn't necessary for wood and the like. If you know Terry from around here (he runs he actually stocks the motors I am using and even a little premade driver board with a ULN2003 on it.. seriously, these are really cheap. They are mass produced for the purpose of turn vanes in air conditioners and things like that.. slow but high torque, for that kind of thing. Though I didn't buy mine from Terry's shop, it's the same beastie. I bought a carton of them (eight I think it was) from an ebay vendor- just purchased, no auction- and I remember it was like twenty dollars for the CARTON including shipping. Must recommend them for playing around.

Adding the screw reduction makes it just silly slow. I have half a dozen salvaged stepper motors about, ones that need no gear reduction to spin that screw.. Seriously, a printer or scanner will always be disembowled for the precision steppers by me at the end of it's days. I've got a pair of drive stepper motors I got from a Lanier (or maybe it was lexmark) commercial dot matrix printer a few months back that would work very well.. and can be mounted in much more standard ways. Still even if you bought them, they are only Unipolar stepper motors, 12v, I think they are rated for half an amp per winding.. you can get them online for probably under twenty dollars brand new. 500ma is actually more current than I want to pull through one, so I'll piggyback another ULN2003 on the first for headroom on current if I need it. They really are not under any real load, I'd probably get away with it.. the Darlington Arrays are rated for 500ma per channel, and I enjoy tempting the gremlins.

We'll see how it goes. Really must restart this tomorrow in the projects section...

What thread screw are you using?
In general: The smaller the diameter, the more turns it takes to move an inch.

1/4-20 UNC (coarse thread), 20 turns of thread per inch. 1/2-13 UNC, now only 13 turns.

If your only cutting slowly, you might be able to use a flexible shaft (often come with a dremel kit) and a high speed electric drill.

Great idea with the draw sliders.

Some of the new inkjet printers have like and index wheels instead of steeper motors, attaching an index wheel could let you monitor the Backlash and adjust for it.
Below is a step motor with optical sensor from an HP printer.

I think you could make a linear drive with stepper power. Get a timing belt and gear, cut the belt. Use one end of the belt to pull a movement axis. Use pulleys to attach the other end of the belt to the other end of the axis. Be sure there's no slack.
Or cut the belt, epoxy it under the table to drive with the gear on the stepper.

I think you could make a linear drive with stepper power. Get a timing belt and gear, cut the belt. Use one end of the belt to pull a movement axis. Use pulleys to attach the other end of the belt to the other end of the axis. Be sure there's no slack.
Or cut the belt, epoxy it under the table to drive with the gear on the stepper.

Such a linear drive is fine if your linear loads are low - for small-scale CNC, you would be limited to probably wood or plastic, and none-too-high feed rates. Aluminium would probably be out. Although I could imagine using a metal chain drive instead (that or wider cog belts).

I was actually looking at the Microrax product line @ LynxMotion - that would make a great and fairly inexpensive (but strong) frame for a small CNC machine (certainly more expensive than what Focalist is doing, though).