Alexa, please turn on game room

After a kind Christmas gift from R, to get me started, (how did you know? <3) I have been dipping my toes into the Home Automation hobby. There are several places in the house where some previous owner made a variety of questionable decisions about where to place switches, and putting light sources in strange places. So, I installed a proof-of-concept, a switch and outlet that are paired together, despite being on different circuits. The parts go in pretty easy these days, although the in-wall units do take up a bit of junction box space. You also need to have neutral in the switch box, which means only certain runs can be automated. There are a number of ways to go about it, though; the switch, the outlet, the fixture, or even the bulb can be the "smart" bit. With that working (adding voice control is almost too easy), I took inventory of the switches in the house, and decided on the first set I wanted to swap out. There are a couple of 4-gang boxes around, and what with one thing and another, I realized I needed to swap them all at once. That is, there were a couple of switches (among the 8) that could have been left as-is, but I needed to replace the faceplate anyway... I'm still mid-build on getting all those put in. I was close, then decided to go the extra mile with this home renovation project, and replace one of the wall junction boxes. It was a whole thing. So, I'm off to the hardware store tomorrow, and there is still some more wiring-fu that will need to happen. But, it was pretty cool to walk through the floor and call out "Alex, turn off X" ... and *magic*. Alexa, thank you.

Posted in Electronics, Knowledge, Making, Technology | Tagged , , , , | Leave a comment

CNC – another satisfied customer

I love it when a plan comes together.

This is a 6″ wide, 3 foot long chunk of the side of some old barn that someone tore down and left the pile “free wood”. The planks are a foot wide by 10′ long or so, I have to pull all the nails out so it lays flat and also so I don’t mistakenly run the router bit into it. Things happen.

I couldn’t get all the nails out of this one (I’m not an expert, and a couple of the smaller nails were almost rusted through), so I chopped the naily bit off with the tablesaw, and found a good 3′ section, and screwed it down to the table (I just use decking screws in “known safe” areas, for now).

The board is not planed. That’s the original paint. Yes, it could be lead, don’t lick it. Did I mention how much I love the Dust Deputy?

A few action shots:

I cut the wider legs of the letters first.

This turned out to have an interesting and useful side effect, which I may put into my normal process; in a moment. More action shots:

… then fill in the rest of each letter using “Engrave”

Engrave follows the outline of the font, and can produce fine detail, down to the resolution of the bit. As a rank amateur CNC operator, I only have endmills that are the same diameter (more or less) as the stock collet on the router, 1/4″. This one’s a 15/64″, which are slightly cheaper, generally, than 1/4″ ones.

The original design had more lettering in a second row, but the uneven surface and my novice skills botched the second cut. I need a distance finder that’s a bit more accurate, if I’m going to cut wood that’s not as sensibly close to flat as I can possibly make it. The first row, cutting a little bit of each letter, then moving on, before committing to the entire cut… I think that helped give me confidence that the cut was going to work, with the depth set correctly, &c.

The second row was purpose-made to fit the 1/4″ bit. So no first pass was required. I probably would have noticed, and could maybe have fixed it, before fully committing. But, that’s the lesson for next time. This time, I had a present to deliver. ๐Ÿ™‚

I was once told, by a wise Jedi, that a good carpenter hides his mistakes.

One more trip to the tablesaw solved that little dilemma.

I like that I can produce something like this, with a few days’ notice.

The CNC performed pretty well. The new Z-nut performed as well as I could hope.
I think the Y table could stand to be re-cut and re-installed. That’s a decently big project, with all the teardown &c. Maybe run a few aircut drills. Maybe having a touchplate would help? Could be a small project, if the line is already laid in from the G540.

Posted in Knowledge, Making, Technology, Woodworking | Tagged , , , | Leave a comment

CNC – a much needed repair

I decided that I absolutely wasn’t going to cut anything else on the big CNC before fixing the Z nut. The hack I made to get it installed in the first place has never worked properly, but it was really starting to look worrisomely close to failure.

Why, no… it’s *not* supposed to be angled down 30 degrees… why do you ask?

For some reason, I was really stressed out about how to complete the fix for this. It seemed obvious to me that plywood was not going to be the answer. I mean, I know the original design called for ply, but I couldn’t figure out how to make it strong enough to work.

3/4″ plywood is strong, but I was asking it to hold in a direction it’s not strong in, and didn’t give it a whole lot of meat to hold onto.

The area this part needs to be installed is a little restricted on space, so I needed something strong and small, and I went back to my old standby, Simpson Strong Ties. While I was at the hardware store, I picked up an A24 (which had the benefit of having a 1/2″ hole in it already) and an A33, which ended up being the one I chose to work with.

I needed to match the holes in the Y axis table, which I measured to be 3/4″ apart. I drilled all the holes one at a time, but I was pretty pleased with how this pair turned out.

I often have problems with holes being “a little off”, but I was pretty careful this time, and nailed it.

At this point, I needed to measure for the hole where the 1/2″ allthread rod would mount, and realized that I needed to make another trip to the hardware store, to pick up large drill bits. Sigh. I gave up for the night, and came back to it the next day.

After a quick trip to the hardware store (omg, big drill bits are spendy — I picked up a 1/2″, 5/8″, and 3/4″, … they were $25 *each* ๐Ÿ˜ฎ ), I chucked my shiny new 3/4″ bit into the drill press, and with a little more “measure twice”, I popped the hole right in.

As an aside, the drill press did *not* like cutting mild steel with the 3/4″ bit at 640RPM. I used tapping fluid, but unlike the 1/4″ drills I’d been making, the drill was really noisy and throwing off hot (!) swarf and smoking and heated the entire part (I kept touching it by mistake while I was deburring it, ouch). The hole went in, but yikes.

I did a test fit, and the allthread went right in! woot.

A couple more holes for bolts to hold down the nut itself (there’s a whole thing there, but suffice it to say that 1/4-20 with washers is the order of the day), and I was ready to finish the installation!

Getting the Y table reinstalled onto the Y rails was a bit of a chore, but with a little help from a friend (thanks, L!), it got done. Re-threading the roller chain required a quick check of the original plans, but it went in nice and easy.

I was suddenly ready to test the CNC machine again.

Booted up, ran Mach3, ran the Z axis back and forth a couple times, and realized that I needed to adjust the washers a bit on the holddown bolts (they were pulling metal flakes off the axis rod, oops ๐Ÿ™‚ ). Other than that, things seemed to be working pretty well. I re-homed all 3 axes, made sure that the machine can still move in all three directions, and that the Z limits still work.

I didn’t do any cutting, but everything related to the recent repair seems fine, and the CNC is back online!

It remains to be seen how well this will work, but it looks and feels a lot more sturdy than it ever had before.

I need to put the machine through her paces again, but it’s very good to knock this one off the punch list.

Posted in Knowledge, Making, Technology, Woodworking | Tagged , | Leave a comment

Drill Press — measure twice, cut once.

With yet a little more care, I was able to produce a perfect set of pole clamps, and I got them all drilled and sawed and drilled and clamp bolt installed, …

and I’d cut them too big in the first place. The clamps need to be a little smaller than the target rings, so that they’ll have someplace to attach. :facepalm:

I was all ready with the process, just needed to re-set the splay angle on the drill press table… and I went another way with the whole thing, but that’s another show.

I think I have compound angles better documented this time. At least, next time I’m drilling one, hopefully I will find this more useful than what I’d written last time (grumble).

Posted in Knowledge, Making, Science, Woodworking | Tagged , | Leave a comment

Drill Press – found my mojo again (well, most of it)

This was one of those days where you spend 1.5 hours setting up for 15 minutes of cutting. I was really careful, and got the drill press table level, then installed a stop block parallel to the bit, then tilted the table, got my workpiece all set up, decided to move the stop block for more repeatable work placement, was 100% ready to go… and had to move the table up, because the bit couldn’t plunge that far. ๐Ÿ˜

Raise the table (which messed up all the alignment, but I could set the plunge depth properly), re-level, redo all the alignment, stop block, tilt &c, and then I was pretty much ready to cut.

Doesn’t look like much, but that took a *lot* of setup.

What with one thing and another, I decided that doing the rake angle at the same time as the splay angle was called for. So I gritted my teeth, added some indicator arrows to the block so I could confirm I’d done it right, and ran off a test block.

Those are some in-line truss poles ๐Ÿ™‚

With a success under my belt, I cut 5 more blanks (5″ chunks of 2×4) on the table saw, and got started drilling the holes.

I really love the huge wood flakes produced by Forstner bits.

Here’s what’s sitting under all that swarf.

Pretty easy, once it’s all set up properly.

I did all the right-hand holes first, so I didn’t have to shift the jig around.
Then I swapped the jig (you need to rotate 30ยบ for one hole, and -30ยบ for the other), and cut the other side.

Looks pretty good

With all the holes drilled, I needed to cut them lengthwise with the table saw, to turn them into clamps. Let me be the first to say that a cross-cut sled is not the right tool for this job. But, I don’t have a working fence at the moment, so I had to make do. I still have all my fingers, but I hated this part a lot.

There is just no good way to clamp the workpiece in this configuration. Grr.

I was pretty excited to see whether my clamps would work.

With even a couple clamps installed, the truss will stand on its own.

And, after scouring the shop for enough clamps to finish the job (you can never have enough clamps), suddenly I had a fairly sturdy 6-pole truss!

Top and bottom blocks parallel, I call that a win.

Setting up to cut pole clamps is still a real pain. I really need to figure out how to cut them using my CNC’d jig (or a variant thereof), or else pole seats are going to continue to be a barrier to entry for future builds.

I decided not to take a photo of the slight misalignment of the pole holes. I think it’ll still work ok once I get them installed. And if not, I’m not moving the drill press table anytime soon. ๐Ÿ™‚

Posted in Knowledge, Making, Science, Woodworking | Tagged , , , | 1 Comment

How to set up a drill press for compound angles, an investigation

So, I’m stuck trying to get the drill press set up for truss pole seats (They have both a “spread” angle and a “rake” angle that need to be set). I couldn’t figure out how to make my CNC’d jig work properly (yes, I probably need to remake it, but let me get the show on the road here), so I reverted to the way I’d done it last time, which was to set up the bevel angle on the drill press table (a very manual and error-prone process, if you don’t know what you’re doing), and then getting the cut depth and spoilboard and jig set up.

The idea is to position the workpiece so that repeatable cuts can be made; brackets come in sets of 6, 12 holes, compound angles — read the 2011 rebuild for lots of detail.

This drill press has the ability to rotate its work table around 3 axes (swing around Z, roll around Y, and yaw about the center of the table. I’d like to set up a jig to cut the proper spread and rake.

First, level the table. Use a bubble level, then calibrate the iPhone level, if desired.

Then, clamp a stop block in place, parallel to the Y axis (so it won’t change under roll).

Spread angle can be set by rolling the table. On my drill press, there’s a single 18mm bolt holding this axis in place. loosen with a socket wrench, and set the angle with a level (I used the one in my iPhone again). Holding it in place while tightening the bolt is a task. Feels like I need a block and tackle. ๐Ÿ™‚

Rake is added by yawing the table, or, in my case, I cut a 30-60-90 triangle, and clamped it to the jig. (Rake is always 30ยบ for this build).

Get the workpiece set up to drill the two holes, and get it centered under the drill bit (mine has a laser crosshair, you want both holes to be the same distance from the end of the workpiece).

So that’s the procedure, as I understand it.

I’m not there yet with my drill. I’ll work on it some more tomorrow, now that I think that I have an idea what to do next.

Posted in Knowledge, Making, Science, Woodworking | Tagged , | Leave a comment

CNC – Put the horse to work

I’ve been trying to get a CAD solution that doesn’t require me to be sitting at the computer in the shop, which for me means it has to run on Mac. I tried out LibreCAD, and I’m starting to get the hang of it, but I still ended up rebuilding a majority of each part in CamBam.

I had 3 parts I wanted to build for my telescope, and what with one thing and another, I ended up finishing two of them.

The first part was built mostly from the DXF I got from LibreCAD, a bunch of circles and rectangles. I did have to maneuver the bits around a lot, but that was expected.

I’m still figuring out how to lay out parts on the material — I find large enough spots manually, and re-zero the machine in between each cut. I think this is called a “fixture” and is completely normal, but I’m a bit new at it.

I hate when I’m off by a little and leave a divot in the side of my part…

I could also use a better indicator of where the zero point is; I tried various things today, but either a laser crosshair or a camera, or at least a light… there is a lot of hit-or-miss work to laying the parts on the board.

The second part I ended up rebuilding from scratch in CamBam; I just couldn’t get all the angles set up properly in the LibreCAD.

The slots are to hold the jig beds — thinking in 3D bends my brain a bit.

LibreCAD has a bunch of “line orthogonal to target line” and “line parallel to target line” tools that make this kind of build easy to lay out, but I am still learning how to set lines at particular angles.

I had to to a bunch of mental gymnastics to get this to work in CamBam, but once I figured out the Transformation Matrix, things started coming together pretty quickly.

I will someday learn how to do holding tabs, but for now I put hold-down screws in everywhere.

I also went back to putting in an “outline” MOP at the beginning of each cut, so I can see where the pieces are going to be (helps me make sure that the screws are out of the way).

Oh! And I figured out how to get CamBam to insert a machine pause after the drills! I have a chunk of G-Code that I stick into the CustomMOPFooter of the Drills MOP, and that does it. I’m sure that there’s a way to make a “style” to handle it, but I didn’t bother to figure that out.

I updated the “default” style in CamBam to use 15000 spindle speed, and -1 default depth (because -0.75 wasn’t cutting all the way through).

I think that hold-down screws are my favorite part, although I really like those pocket cuts…

The Z axis nut is really a problem; I need to fix it first thing. I can see that the bit isn’t plunging all the way in, because the Z nut moves up and down by about 1/2″, yuck. I have a couple ideas for metal L-brackets, I just need to carefully measure things and get one installed.

I like the way a complicated cut looks on the monitor.

With the exception of one drawing that had an error in it (I had a piece rotated 90ยบ in the finished part), everything went off without a hitch today.

The inaccuracy in the Z nut is the biggest problem with the machine right now. I have a feeling once I get that squared away, my cuts will be a lot more repeatable.

One way or another, though, I really feel like I’m able to treat the CNC as a “tool” and not so much a “project” now. If I could get my CAD workflow a little smoother, I’d be able to start producing parts at a steady rate.

Posted in Knowledge, Making, Technology, Woodworking | Tagged , , , , , | Leave a comment

CNC – Hello, G-Code.

The scripting language that CNC machines use is called G Code.
There are lots of commands, said Google. But the first few are about moving the machine.

G0 – rapid move (linear)
G1 – controlled speed move (pass Feedrate in addition to coords
G2 – clockwise turn (pass point to rotate about)
G3 – counter-clockwise turn ( ” ” ” )
… and so on.

Draw any circle, A.
Draw any second circle with a center on the circle A. This is B.
(Note: circle A’s center is on B, also. Circles, yeah. ๐Ÿ™‚ )
You can only draw two circles that can cross these in the same way.

Or, sweep a circle from each point on an equilateral triangle to each of the other two.

In any case, it’s just 3 arcs, from one vertex to the next, centered at the third.

I wrote a helper script to calc sqrt(3) for me.

python 20.0
G2 F100.0 X0.0 Y0.0 I 20.0 J 0.0
G2 F100.0 X10.0 Y17.3205 I-10.0 J-17.3205
G2 F100.0 X20.0 Y0.0 I-10.0 J 17.3205

< insert fiddling with CamBam imports and machine ops, reminder to self, update the machine parms in CamBam. ๐Ÿ˜ >

But all’s well that ends well.

I made this.

The Z nut assembly is super janky.
The table is not flat. You would not believe how many spoilboards.

I had to set the machine to dig to 1″ in order to clear the 3/4″ MDF ๐Ÿ˜€

I have a feeling I will be seeing a lot of MDF. I bet it seals up pretty nice under spar varnish.

I love that I could just slam one of these out of a Sunday morning in the summer.

Here’s the machine just hobbling through the job. I was very surprised how little sawdust is around with the Dust Deputy on patrol. Yes, I went there.

It’s just running … a Lot. Better.

I had requested drill points for me to put in hold-down screws, far from anywhere the endmill was going to be.

Figured out how to edit in a machine pause. Aw, yeah. Notepad FTW.

I didn’t put in a pause between the center cut-out and the Reuleaux. I admit that I like all the vowels in that word. But, I digress.

I just waited for the machine to get to a safe part of the cut, and I reached in and unbolted the circular scrap piece, those are fun, but later. Jr. Astronomer was duly impressed with my circley roundness.

The cut was for an astronomy project, here’s one I cut by hand 9 years ago and the one from today.

They were pretty close. Like, sandably close.

And, result!

Trixie gets a new hairdo.

The machine has a backlog, but it performed beautifully today, nearly flawlessly.

Yeah, so 3 lines of G-code. It blooms to 3k bytes or so, with repeating at different depths &c.
That works.

Posted in Computers, Knowledge, Making, Science, Technology, Woodworking | Tagged , , , | Leave a comment

CNC Build – And that’s how we do that.

Today, I decided that I was ready to start making some sawdust again. I had gathered and sorted all my endmills last night, and I found the one that had tried to kill me all those years ago. By the way, if anyone reading this knows a good way to clean endmills, I’d like to hear about it. I remember when I was cutting out the parts for the Big Machine, I’d have to swap in new cutters almost every new piece of ply, because the glue gets gummed in the teeth. The cutters are still sharp, but they can’t reach the wood anymore. Probably the lower speeds of the SuperPID will help a bit, but I have some really nice bits that just need to be de-glued. I digress.

I started with the speed-n-feed test. I ran it on MDF once and on dimensional lumber once. In both cuts, it seems like 15k RPM made a pretty clean cut. The 100ipm feedrate looked like it would work, too.

It was around this point where my wife turned up, and I allowed as to how I was going to start in on the Pumpkin I owed her, and she said that what she really wanted first, was a new set of house numbers, for the sign at the end of the driveway. We have a lot of problems with packages getting delivered to our neighbors, and although there’s a sign already, it would be helpful if it was a little larger.

We had scrounged some free wood when a local farm set out a bunch of scrap, and we’d earmarked a couple pieces for the CNC specifically, so that’s what I used. It is some really cool old lumber, rough sawn 1x12s. I guessed that 6″ high digits should be visible a long way off, so then it was a simple matter of choosing a font and we were off!

With a couple of rounds of “which one do you like better”, K helped me to pick out the one she wanted. I found that my CamBam skills are a little rusty, but it came back pretty quickly.

I set the tool up, did a full “air cut” to make sure I understood what was going on, and then set the machine to cut its first “real” project in years and years. I think it turned out rather well.

I love the smell of sawdust in the morning.

I chopped the board to size in the tablesaw (I swear it’s easier to set up than the chopsaw), deployed it to the end of the driveway, and started thinking about what’s next.

As an aside, the Dust Deputy / vacuum shoe worked like a charm. There was a faint smell of sawdust as the machine ran, but nothing on the table or in the cuts or in the air, to speak of. Nice!

For the next trick, I decided to take on a new sign for the observatory; the “Hilltop” sign it’s been sporting for the past couple years was actually the first thing I ever cut on the Big CNC (it’s pretty easy to do text, don’t even need a CAD program, just do it all in CamBam), and it’s cool and everything, but it was cut on a nasty piece of scrap wood, so I’d been looking forward to replacing it.

I considered just cutting the same design as before; I’d picked the font because it has rounded serifs (which work well with router cutting), but I ended up re-creating the project, because I wanted it to say “Hilltop Observatory” instead of, simply, “Hilltop”. I looked at a bunch of fonts, and landed on one that I thought might look cool with rounded serifs. My first attempt to cut the project went sideways, because I hadn’t set the cut depth deep enough, and the board is decidedly not flat. When I tried to re-cut to the new depth (in the same spot), that failed (pretty much as suspected), because I had forgotten to “Ref All Home” in between attempts, and the machine is not quite tuned enough to be repeatable, yet. Note to self: remember to re-home the machine! I wonder if there’s a way to do that in G-Code? hmm…

Anyway, so I ruined a chunk of 1×12, but there was still more than enough left to cut the sign out in a different spot. This time, the cut depth was fine, and the job ran without a hitch.

I built the tool that built the tool that made this sign.

So. The machine works, it cuts big projects without killing anyone, and I even have some ideas about what to cut next. I do need to fix the Z transmission nut bracket. I may make one out of metal instead of wood, though, in which case it’s not the CNC that will be cutting it.
I do still owe my wife a pumpkin, which I think she’ll love, once it’s done. I have some rings to cut for Trixie, a door sign for my daughter… We are into “what project are we going to cut” and (with the exception of a short todo list) out of the “build punch list”.

These posts get to officially change from “CNC Build” to just “CNC” from now on.

Posted in Knowledge, Making, Technology, Woodworking | Tagged , , , | 1 Comment

CNC Build – Back in the game!

I decided that I was going to figure out the sensor cable today, once and for all, one way or the other.

And, spoiler alert, I did.

I cut the zip ties and looked at the way the port is installed, and decided that Sugru wasn’t going to help. I reinstalled zip ties, but in a different configuration, and that seemed to work a bit with the new sensor cable.

Then, I started ohming out the cable, to make sure that I hadn’t messed up the DB9 port while pushing it around.

and I noticed that none of the top-row pins in the cable were registering. Yikes! I ohmed the cable all by itself, AND THE STUPID CABLE DOESN’T HAVE THOSE PINS HOOKED UP END-TO-END. Are you kidding me?!

So I ohmed out the original sensor cable (of course it works), and ohmed out the DB9 port (that worked too), and decided to just try the system with the old cable in place.

As I was poking around inside the PC case, I happened to notice that the BIOS clock battery was a 2032 (a variety that I keep in stock around the house), I popped a new one in, and then promptly failed to get into the BIOS upon next boot, so XP booted into Aug 2005 ๐Ÿ˜ But, there’s a fresh battery in, aw yeah.

I booted the PC again (setting the BIOS time this iteration), and was also able to see 5v on the pins I expected, nice. Then I brought up Mach3 and was able to see the “Home” pin flashing when I touched GND to the Home pin (I have all 3 home switches wired to the same G540 input — can’t remember why I did this, but I do remember it’s a thing).

I stupidly decided to plug the sensor cable into the Super-Pid at this point; I know I’m not supposed to plug stuff in when it’s up and running, but I decided to go for it. That turned out not to be a problem. The SuperPid turned on as soon as it got 5v, and I was able to hand-turn the router collet and see the spindle sensor working.

I spent the next little bit getting the home switches all registering again; I was barely to this point in the initial machine calibration process (kind of like polar aligning a telescope mount) when I stopped, but I do recall vaguely that the X axis was not homing properly, and this turned out to be the case again (or still). I retrieved a spare magnet from the kids’ whiteboard, and the sensor tripped right away. I stuck the new magnet to the already-installed one, and they clicked right together (because Magnet), and all of a sudden, I had home positions on all 3 axes! What?!

I turned off soft limits temporarily, clicked “Ref All Home”, and watched in wonder as the machine went and touched all 3 home switches. That was Very Satisfying.

I spent the next little bit remembering the relationships between “Machine Coords” and “Table Coords”, and what happens when you say “goto zero” if those aren’t the same, &c.

I ran the feed-n-speed test, and it of course wouldn’t run right at the origin, because of a soft-limit violation. But I offset the tool by 3″ in X and Y (and -0.5″ in Z), and it ran flawlessly.

I then plugged in the 110VAC mains side of the SuperPID, and made sure that it could start and stop the router, and change its RPM.

When I ran the test again, I was a bit surprised the first time to hear it changing the router speed in between each cut (although I shouldn’t have been — that was the whole point of the test ๐Ÿ™‚ ).

I captured another video of the accomplishment (again, too big to post).

I need to get the CNC table cleared off so that I can actually put in a endmill and run the test “for real”, but I am happy to declare that the CNC machine is officially 100% back to where it was when I left off.

3 upgraded motor cables, and a bunch of careful testing, and I’m ready to roll again.


Posted in Computers, Electronics, Knowledge, Making, Technology, Woodworking | Tagged , | Leave a comment