Ed Nisley's Blog: Shop notes, electronics, firmware, machinery, 3D printing, laser cuttery, and curiosities. Contents: 100% human thinking, 0% AI slop.
Tag: Improvements
Making the world a better place, one piece at a time
The white jumper plugs into the single +5 V pin in the row and is soldered to a straight wire running along the entire row of header pins. I pushed the black plastic strip to the bottom, soldered the wire along the pins atop it, then clipped off the pins so they’re about the right height when flush against the PCB.
Use a two-row socket to hold the new row in alignment with the existing header:
3018 CNC CAMTool – Endstop power mod – alignment
Slobber on some epoxy and let it cure:
3018 CNC CAMTool – Endstop power mod – epoxy curing
And then It Just Works™:
3018 CNC CAMTool – Endstop power mod – installed
Well, after you install the switches and tell GRBL to use them …
Reminder: If you intend to put limit switches on both ends of the axis travel, you mustclip the NC lead from both MBI switches. One switch per axis will work the way you expect and that’s how I’m using them here.
If you regard your new CNC 3018-Pro Router kit as a box of parts which could, with some adjustments and additional parts, become a small CNC router, you’re on the right track.
In my case, the aluminum extrusions arrived somewhat squashed inside their well-padded foam shipping carton, which leads me to believe the factory responsible for tapping the bolt holes in the ends must be a fairly nasty place. In any event, the hammerhead T-nuts for the gantry struts simply didn’t fit into some sections of the slots, although they worked fine elsewhere.
So, file a smidge off the rounded sides of a few nuts:
3018CNC – 2020 T-nuts – filed
Which let them slide into place and rotate properly despite the bent channel:
3018CNC – 2020 T-nuts – trial fit
The assembly instructions used a word I’d never encountered before:
3018CNC – Gantry plate position
Turns out ubiety is exactly correct, but … raise your hand if you’ve ever heard it in polite conversation. Thought so.
I’ve not noticed any harm from rounding off the position to 46 mm; just position both struts the same distance from the rear crossbar and it’s all good.
The struts behind the CAMTool CNC-V3.3 electronics board were also squashed, prompting a bit more filing:
3018CNC – CAMTool v3.3 board – trial fit
The CAMTool board is basically an Arduino-class microcontroller preloaded with GRBL 1.1f and surrounded with spindle / stepper driver circuits.
As with the MPCNC, I’ll dribble G-Code into it from a Raspberry Pi. Alas, the struts behind the CAMTool board are on 75 mm centers, but the Pi cases on hand have feet on 72-ish mm centers. Pay no attention to the surroundings, just drill the holes in the right spots:
3018CNC – RPi case – drilling
Add more T-nuts and short button head screws, with rubber pads between the case and the struts:
Since the PiHole runs all the time, it now hosts an FTP server to stash snapshots from the cameras onto a 64 GB USB stick. I installed ProFTPD, which Just Worked with a few configuration tweaks:
UseIPv6 off
ServerName "PiHole"
DefaultRoot /mnt/cameras
RequireValidShell off
ftp_snapshot=true
ftp_host="192.168.1.2"
ftp_port=21
ftp_username=$(/bin/hostname)
ftp_password="make up your own"
ftp_stills_dir=$(/bin/hostname)
The last line uses a separate directory for each camera, although they quickly ran into the FAT32 limit of 64 K files per directory; reformatting the USB stick with an ext3 filesystem solved that problem.
My high hopes for the UHMW bushing supporting the impeller lasted the better part of a day, because direct contact between the impeller and the motor bearing produced an absurdly loud and slowly pulsating rumble:
Bath Vent Fan – bushing installed
My hope that the UHMW would wear into a quieter configuration lasted a week …
Back in the Basement Shop, some free-air tinkering showed the impeller produced enough suction to pull itself downward along the shaft and jam itself firmly against the motor frame. My initial thought of putting a lock ring around the shaft to support the impeller turned out to be absolutely right.
So, make a small ring:
Bath Vent Fan – small lock ring – c-drill
With a 4-40 setscrew in its side, perched atop the impeller for scale:
Bath Vent Fan – small lock ring – size
It just barely fits between the impeller and the motor frame:
Bath Vent Fan – small lock ring – installed
This reduced the noise, but the hole in the impeller has worn enough to let it rotate on the shaft and the rumble continued unabated. The correct way to fix this evidently requires a mount clamped to both the shaft and the impeller.
Fast-forward a day …
A careful look at the impeller shows seven radial ribs, probably to reduce the likelihood of harmonic vibrations. After a bit of dithering, I decided not to worry about an off-balance layout, so the screws sit on a 9 mm radius at ±102.9° = 2 × 360°/7 from a screw directly across from the setscrew in another slice from the 1 inch aluminum rod:
Bath Vent Fan – mount ring – tapping
Centered on the disk and using LinuxCNC’s polar notation, the hole positions are:
As usual, I jogged the drill downward while slobbering cutting fluid. I loves me some good manual CNC action.
Put the mount on a 1/4 inch tube, stick it into the impeller, and transfer-punch the screw holes:
Bath Vent Fan – mount ring – impeller marking
Apparently, some years ago I’d cut three screws to just about exactly the correct length:
Bath Vent Fan – mount ring – test fit – bottom
I knew I kept them around for some good reason!
The 9 mm radius just barely fits the screw heads between the ribs:
Bath Vent Fan – mount ring – test fit – top
Some Dremel cutoff wheel action extended the motor shaft flat to let the setscrew rest on the bottom end:
Bath Vent Fan – mount ring – shaft flat
Then it all fit together:
Bath Vent Fan – mount ring – installed
The fan now emits a constant whoosh, rather than a pulsating rumble, minus all the annoying overtones. It could be quieter, but it never was, so we can declare victory and move on.
Dropping fifty bucks on a replacement fan + impeller unit would might also solve the problem, but it just seems wrong to throw all that hardware in the trash.
And, despite making two passes at the problem before coming up with a workable solution, I think that’s the only way (for me, anyhow) to get from “not working” to “good as it ever was”, given that I didn’t quite understand the whole problem or believe the solution at the start.
But it should be painfully obvious why I don’t do Repair Cafe gigs …
The Baofeng UV-5R radios on our bikes seem absurdly sensitive to intermodulation interference, particularly on rides across the Walkway Over the Hudson, which has a glorious view of the repeaters and paging transmitters atop Illinois Mountain:
Walkway Over The Hudson – Illinois Mountain Antennas
A better view of the assortment on the right:
Illinois Mountain – North Antennas
And on the left:
Illinois Mountain – South Antennas
Not shown: the Sheriff’s Office transmitter behind us on the left and the Vassar Brothers Hospital / MidHudson pagers on either side at eye level. There’s plenty of RFI boresighted on the Walkway.
Anyhow, none of the Baofeng squelch settings had any effect, which turned out to be a known problem. The default range VHF covered a whopping 6 dB and the UHF wasn’t much better at 18 dB, both at very low RF power levels.
We use the radios in simplex mode, generally within line of sight, so I changed the Service Settings to get really aggressive squelch:
Baofeng UV-5R – Improved Squelch Settings
I have no way to calibrate the new signal levels, but I’d previously cranked the squelch up to 9 (it doesn’t go any higher) and, left unchanged, the new level makes all the previous interference Go Away™. Another ride over the Walkway with the squelch set to 4 also passed in blissful silence.
If the BF-F9 levels mean anything on a UV-5R, that’s about -100 dBm, 20 dB over the previous -120 dBm at squelch = 9.
The new squelch levels may be too tight for any other use, which doesn’t matter for these radios. As of now, our rides are quiet.
[Update: Setting the squelch to 5 may be necessary for the Walkway, as we both heard a few squawks and bleeps while riding eastbound on a Monday afternoon. ]
I’ve finally had it beaten into my head: any public exhibition requires paper handouts, if only for younger folks who are too shy to ask questions. Paper may seem obsolete, but it serves as a physical reminder long after the sensory overload of a busy event fades away.
Hence, I made up cards describing my exhibits at the HV Open Mad Science Fair, each sporting a QR code aimed at far more background information than anybody should care about:
Because I planned to take my collection along to HV Open’s Mad Science Fair, I finally used a Round Tuit for some adhesive action.
The general plan was to punch a ring from double-sided tape, thusly:
Astable – Radome adhesive – poor surface
The OD required touching up the edge of a brass tube punch I’d made a while ago:
Astable – Radome adhesive – punch sharpening
It worked exactly as expected:
Astable – Radome adhesive – punching
Unfortunately, the 3D printed spider’s “spherical” socket has such a rough surface that the adhesive had too few contact points to hold the ball in place.
My fallback has become 3M outdoor-rated double-stick foam tape, so:
Astable – Radome adhesive – 3M foam tape
This leaves a small black ring visible between ball and socket. Recessing the foam tape by half its thickness should improve its ahem optics, although it’s probably not worth the effort with black PETG.
The Smell of Molten Projects in the Morning 