You know, I told myself I’d take 2020 as a year to learn to relax, reflect, and stop building everything all the time because we’re all going to die soon anyway. And then I had to try and remember everything I did just since late September to write the last post. But there’s more, as in that post, I neglected all of the shop-building I’ve still been on a quest to do.
For one, I’ve been seeking a milling machine to accompany tinylathe (which does need its own writeup), but not needing one for business purposes, I wasn’t keen on buying a tinymill new. I kept an eye on Craigslist and Zuckerburg’s Emporium for good deals on small to medium sized mills – while I could have easily bought Bridgeport sized machines for days, that violated my rule for the time being of No Multi-Thousand Pound Objects That Can’t Drive Themselves.
My other constraint was no round-column mill-drills. I know they’d get the job “done”, but I can’t stand those things because of their propensity to rotate on the column and lose all your zeroes for you. So really I was just sitting on my ass waiting for “The One”, and was close to being able to get a few Grizzly mid-sized mills with square columns (and similar)… but damn, it turns out other people also want them, and they went quickly.
Luckily, fellow robot builder and machinery enthusiast Alex Horne made me an offer that I found very hard to refuse – on my way to Boston last November to obtain Overhaul 1 Among Other Things, I picked up this little guy from him.
Huh. Well that’s cute. It’s like a larval version of the classic American heavy manual mill pattern, like the first instar stage of a Cincinnati or Kearny-Trecker. I loved it.
The travel is about 12″ x 6″, which is pretty impressive. It’s in a similar size class as the Seig X2 type “tinymill” that’s sold everywhere, but built like a battleship. This was a difficult two-person lift, where as I alone can chuck a tinymill-sized machine onto the workbench back at MIT.
So I’m barely 2/3rds of the way to Boston and already have picked up several hundred pounds of junk. Well this trip is certainly going well! The mill came with this very heavy work table which itself was another hundred pounds or more of very dense and nicely finished Old Wood topping a frame made of 1/4″ thick steel angles.
We stopped by a local machinery dealer which I keep calling Hank Hill Machinery or Hillbilly Machinery to inspect their wares, and ended up finding a small treasure trove of full drill/mill/tap organizers. I spent even more money I didn’t intend to in order to swipe these – we made a “What if I took all of them” offer and split the goods afterwards.
So after I got back home and unloaded…. what on earth did I just buy. This is how I operate, as you know. Obtain first, figure out what it is you got after the fact. I’m literally the most advertisement-agnostic person on the planet. You can’t egg me on to buy something through viral targeted marketing, but you can set your product out so I trip over it and bring it home, then I’ll do research on how to buy more and subscribe to your services.
This adorable neotenic critter is a Benchmaster, made by a company called Duro that eventually just became “Benchmaster”. The product? Benchmaster. What does it do? Be the master of benches.
It was, as it seems, targeted at the hobbyist or a ‘second machine’ type situation. Sounds like a limited market, but they aren’t as rare as I thought they were, and an enthusiastic community exists around them where people have done comical swaps such as putting a Bridgeport M-Head on the damn thing.
If I haven’t beat this drum enough, I’d like to reiterate a point I made when I posted about crabmower: I bought an old, obscure device without knowing what it was, and someone had made an entire page on how they fixed up and modified it. Folks, this is why we’re here.
Alright, I now had to find a home for the Benchmaster so it can be the master of a bench. Ever since I built the benches, I’d already earmarked half of one of them next to tinylathe for the installtion of a mill. It had recently become occupied by random sanding/grinding tools and Overhaul parts, so there was a lot of cleanup and displacement to do.
Namely, all of my tooling (the stuff you need to USE a mill and lathe) had to be displaced. I therefore was forced on a hunt for new tool organization, which will come later. For now, it’ll just live in a pile on the floor like my soul.
I decided to disassemble the original heavy wooden bench to form a foundation for the mill. The 1″ of OSB my benchtops were made of felt just a bit too flexible for it to be a good anchor, so the plan was to secure the big wooden block to the bench, then bolt the mill through all of it.
The interesting thing about the Benchmaster is that the knee leadscrew pokes down from the mill by a fair amount. That’s why they always have to be on stands. I decided to drill a 3/4″ hole through everything as the leadscrew sphincter
#OSHACrane was used to line everything up and set the machine into position…
…upon which I lined it up with the marker lines I had drawn, then drilled through and bolted in-place.
And here it is, the menagerie of miniature machinery.
Alex threw in the 4″ milling vise, which we both agreed was way too large for this machine. It used up a significant amount of the Z travel just for itself, and this mill doesn’t have a quill (A bit annoying, and a good excuse to do a head swap later on), so its usefulness is severel curtailed. With the thing finally installed in place, I gave everything the ol’ lube n tune, taking the axes apart to adjust the slides and leadscrew nut tightness.
But damn is this thing rigid. Being made back when America was Great, Men were Men, and Steel was Free meant it’s exceptionally smooth (once I tuned the gibs in and cleaned & oiled everything) and I dragged this 1/2″ endmill at 1/2″ DOC through my sacrificial aluminum test piece at the highest spindle belt speed, and it barely flinched. This is a suicidal cut on a Seig X2 class mini-mill, and even if you did manage to do so by feeding slowly, the finish would have been chattery.
For now, until I want to get a 3 inch milling vise, I bestowed upon it my old toolmaker style vise that usually held motors under testing.
The other downside is at the moment it doesn’t have digital scales, so I’m back to using a lot of my “vernacular machining” skills learned years and years ago. My “edge finder” is really just putting a 1/2″ drill rod (itself really a cut-down, destroyed 1/2″ drill bit) into the collet and bumping off.
It fulfills my current “mill” needs quite well either way: Flat this shaft, key this, shave that down, bore this out. Anything substantially complex right now I have enough contacts and favors to call in so I can have a part made. I’ll be planning to add digital scales soon, and I’d like to eventually see if I can get it a quill via head swap or severe head modification.
After much ado about a whole lot of things, this site is now at least in a working state where all my information is accessible… even if it doesn’t look quite all aligned, all my plugins are missing, things might not be in the right place, and so on. This website is still a van, just a newer one.
By the way, I noticed all of your 63 emails asking what happened to the site! Hell, I didn’t know people still had the patience to read blog posts in this era of Youtube subscriptions and TikTok follows. A lot of valuable info resides here, so I definitely had the incentive to get everything running again, just a matter of willpower (This will be a theme for this post…)
So I had to relearn a lot of “Internet Stuff” since the last real revamp of the site from 2009. The biggest challenge ended up being re-importing the database which actually dates back to 2007 (the earliest posts on this site now), which is why this site was a potato dealership for a few days.
First, I had trouble importing the 200-something megabyte database dump, and it took several retries in different browsers and different times of day. Not only that, but fancy hax0r Charles of 2006 named all of his WordPress databases fancy names, so the new WordPress install didn’t know ass from teakettle. Next, because all of my domains are now unified on one hosting account (Equals Zero Designs and Marconi Motors), I had to connect all the subdomain dots. I’ve also never seen cPanel in my life, despite it being available back then also – I did pretty much all of the setup and back end work through FTP and phpMyAdmin directly, so there was just button clicking to learn.
I’m still going from theme to theme, so the immediate appearance of this site might change in the next few days. I’m trying to keep it a dark and easily browsable theme. The one I have as of 1/11 also has a banner image like the previous rendition, but I haven’t reuploaded those yet. It also has a bad habit of displaying the past few posts all together making the front page infinitely long, and I have yet to find the setting for breaking it up into previews only! I also still need to get used to the visual editor that WordPress ships with now – I’m not a fan of it so far, since it’s more of a walled garden experience and it’s a little harder to use my historic file and photo structure. But alas, welcome to the Internet of Today.
Anyways, after all of my makeshift database adminning, here we are again – I’m sure I’ll make a post like this again in another 11-14 years. All of the old posts should be there, but I have not (and will not) check them for layout or importation mishaps, as I consider those pretty much static archives at this point. Look, my van posts are here for my own reference and that’s all that matters.
So! Onto the new content. Besides now the Summer of Venand Overhaul 3 Design & Build series posts I need to backfill, there’s some new stuff in the pipeline because I will somehow always find new vans to work on. I’ll just add this to the “List of Things I Still Have To Blog About”. Here’s the short story of, I dunno, since late September or thereabouts.
You know what? I miss having my own drone. I keep working on everyone else’s drones, but I haven’t had one truly of my own since all the way back in the Tinycopter days. Back then, I had the audacity to code my own flight controller, but these days most of my work is integrating Arducopter and PX4, flight controller firmwares that are….. less haphazardly put together. With safety and what not. Somehow I’ve built dromes for many entities since then, including KIWI of course, and my current place of employ, but what measure is a drome engineer if he doesn’t have any of his own?
And so I went to pray at the Altar of Lord Bezos and visited the Oracle of Jack Ma. You know the adage “Buy right, or buy twice”? My take it on it is “Why buy right when you can buy very specifically wrong and buy a lot?” It’s like getting a 0 on the SAT, since you have to answer every question incorrectly and can’t just shotgun it at random. You have to specifically know what not to buy, so your pile of parts has a minimal chance of cooperating, maximizing your chances of failure but forcing an exploration of the tradespace into places no sensible engineer would touch. Long time readers will understand this is my M.O. for everything – I know what to do, so why do it when you can try something dumb since nothing matters and we’re all going to hell anyway?
As such, crafted out of a tote of deprecated KIWI parts and my robot electronics bins, helped along by some deconstructed Seg-baby packs dating back to 2015 (RIP seg-thing), and with the blessing of the lowest-priced drone parts AliExpress could provide, I present Trashcopter:
This thing is…. a drone. There’s nothing special about it. I just wanted a beater drone to fly when I felt like it. It works fine, I went through the usual setup and tuning and fine craftsmanship associated with putting a kit drone together, and it is still in one piece as of this writing. It can fly autonomous missions, take off and land itself, follow terrains and avoid (large, visible to IR light) objects, and do a barrel roll in mid-air once. (Okay, it was for a brief couple of hours not in one piece). It ain’t a Skydio II, it’s basically a potato someone threw very hard, very controllably.
I explored the sub-basement steam room of drone parts on this build by purposefully trying to sort by price lowest and free shipping. What I found is an entire under the fallen log ecosystem of used drone parts, selling motors and ESCs and subassemblies for $1-$5 apiece. As expected, I now own like 50 motors pulled from XiaoMi drones, and the ESCs that go with them.
The frame is the cheapest, most terrible DJI FlameWheel knockoff I could find. The finish is so ratchet that I had to deburr everything before using it (and correct some of the heatset insert work, and open up some of the PCB chassis plate holes…), but I also now have 6 frames worth of questionably molded nylon arms. I mean you should see the sink marks on these arms. What I’m saying is, I can build as many terrible drones as I feel like now, for less than the cost of getting parts stateside for one single functional unit.
I furthermore went shopping for the crappiest radio I could find – the “Can I find something even cheaper than the 4 channel HobbyKing 2.4Ghz radio?” and that result is sitting next to it, the “MicroZone MC6” series. Like Trashcopter, it is “An Radio”. It has all the right shapes and tchotchkes in the right places, and Doesn’t Not Work. Hell, it’s even 6 (secretly 7) channels.
The build report for this guy will expound more on the process I took to get the parts, exploring some of the parts themselves including taking apart the cheapo radio, and just generally show the setup of a modern-day Pixhawk and Arducopter based multirotor from end to end for posterity.
But that’s not all.
I hinted in the original Robot Trap House post that I had unfinished business in the sector of Very Lörge Dromes that I still wanted to explore and develop, but which wasn’t relevant to the KIWI business needs at the time. One of these in particular is my strong belief that the “One motor per prop” multirotor architecture doesn’t really scale to large, flying van levels. You CAN make it work, and many companies have, often at great expense of either buying or developing cutting-edge custom motors and materials for airframe and propellers.
That clashed with my general philosophy of “Don’t custom unless you want to make a project out of the custom thing”, and consequently the direction of KIWI, where every aerospace engineer we tried to hire dropped to the floor and foamed at the mouth as soon as they witnessed our extremely BattleBot-like building approach: COTS and easy sheet metal and extrusion weldments.
The magic sauce to me when it comes to electromechanical hardware startups lies not in exotic in-house cooked and served materials and genetically-evolved one-piece structures, but getting out into the field with a working, reliable robot in front of the customer and a practiced means of getting there many times. I’m a bad CTO – I don’t like technology.
So how do I aim to demonstrate an alternative? Well, I reached just a little bit back into history, like a few years, into the domain of the Variable-Pitch Multirotor. Also called “Heliquads” or “Collective Pitch Multirotors”, they trade a little bit of mechanical complexity (The collective-only rotor head) for, in my soon-qualifiable opinion, a broad increase in the maneuverability space and control bandwidth.
My still-in-progress entry into this design tradespace will be what I affectionately named “Wigglecopter“:
In short, for a minor increase in thrust for vehicle attitude correction, a conventional multirotor has to spin up and down the propellers. Your torque to inertia proportions really, REALLY matter. Everything needs to be as light as have as little MOI as possible, and your motors need to be as torque dense as possible, to get a high enough control loop bandwidth to keep the vehicle stable.
Conversely a VPM/CPM can issue corrections by adjusting the pitch of its propellers. Single-degree movements will induce variations in thrust corresponding to possibly hundreds of RPM of motor speed. There is a lot of literature in the advanced aerospace controls scene pertaining to these, and I’ll collate and dive into a few papers I’ve taken a liking to in its build reports.
I actually tried to buy one of these, as they were sold for a while in the Early Teenies by a few hobby vendors with models such as the HobbyKing Reaper 450, WLToys V383, and the CJY Stinger 500. They’ve pretty much all died out, so instead of hunting around for used or new-old ones, I decided the mechanical problem was simple enough to just put together and get the point across.
If you look closely, Wigglecopter is just made from the same pile of garbage that Trashcopter emerged from. I just ordered a few DJI F450 quad frame cards from Amazon to make it a quad, and had to gently re-engineer the motors to accept the collective pitch mechanism and propellers. I’m going to put some more legitimate gear into this thing from the flight control and sensing side, as I’d like for it to be a development platform.
Notice that it still does have four independent motors? Well, you can still do that with a CPM, provided you now keep the motor speed constant so your thrust output is not a multivariate surface of sagging motor speed and flexible propeller blades…. just one of them, as much as possible. I decided trying to make a serpentine belt drive was just going too hard the first time out, and will just bypass this issue with inertia rings pressed onto the motors if need be, and with the ESCs set to speed govern. We’ll see what it does!
My LTE plan for Wigglecopter is to finish and validate it, then start getting larger and larger. I’m going to need to modify the firmware a little for myself, as I would like to make a collective-pitch Hex and Octo down the line. Wigglecopter itself should be all done and ready this spring, and its bloodline is completely unplanned except for daydreaming of lifting Kei vans in the air.
Overhaul 1 Restoration
A very exciting new development in my life is that I now have Overhaul 1 in my possession again. In November, I made a speedrun up to Boston to collect the remainder of the several hundred pounds of life I left in the ol’ vape shop. At this point, I was able to extract Overhaul 1 from its dormant state. For the past few weeks, I’ve been going through it (there’s not much, mind you) and getting it back in running order.
There’s no intention of putting it back in battle except a few token matches with Sadbot, Overhaul 2, and Overhaul 3. Yes, somehow I will soon have four operational heavyweight Battlebots. It’s like vans, they just keep spawning. Everyone I know agrees that it would be incredibly funny if Overhaul 3 loses to every preceding generation of Overhaul. I mean, it’s never won against Sadbot, so this is a distinct possibility.
I designed up a retrofit for the drive motors on the shuffle pods, implementing a design idea we should have done but didn’t have the time to execute. Right now, the electronics bay is a small plastic tote bungee-corded to the frame, but I’m going to design up an integrated battery case and electronics deck so I can close it up. It won’t be as (unnecessarily) fast as it was before, as as a bot I’ll probably reserve for demos and showings only, doesn’t need to be anyway.
I also had to straighten out a lot of bent parts. You know what – my adventures in Big Chuck’s Auto Body came home to roost. There were a lot of fun rednecky processes involved in straightening the welded unibody-ish frame and the pointy plow.
So, hopefully Overhaul 1’s “Rebuild Report” will just read like one of my many other hundreds of “I fixed this stupid thing that broke because I was stupid to begin with” titles.
all of the ven are piles
As of right now, my entire treasure fleet is in disarray. While everybody runs and drives, I wouldn’t characterize any as “particularly competent”. It’s winter, and they’re not in danger of being towed or fined for the first time, so in a way this little return to form with me building robots again has been at the expense of the ven.
Why are they so derelict? Well, I think in part it’s due to me continually throwing them up and down mountains.
Now that I’m only about 3 to 4 hours from the very vannable mountain roads of northern Georgia and the North Carolina/Tennessee border, it means I go…
I do think at least once every month so far I’ve ended up somewhere in the area with vehicles nobody expects to ever witness in general, much less on a mountain. I’ve gone with groups (typically composed of SPROTS CARS) and when I damn felt like it.
The downside is obviously that the exercise is very strenuous for tired old ven. Here’s the lockout tag captions for everything as it stands:
The entire exhaust path from the axle-clearing bend back fell off in late May when I was on the Tail of the Dragon. Yes, fell off. As in the person behind me had to dodge it. Straight-piping 3 hours home was hilarious, albeit dissatisfactory for hearing longevity. I replaced the exhaust in my first fully welded/fabricated custom exhaust job in June. In fact, look at it ratchet strapped to the roof rack above, as a victory trophy.
Complete front brake caliper and rotor replacement in November – it’s had one mildly dragging caliper for a while, and it was tolerable until some amount of smashing on the mountain caused it to seize even more.
Now it’s slowly leaking brake fluid from the master cylinder/booster assembly – while it stops fine, the fluid loss is gradual and both faster than I’m comfortable with and want to deal with the mess.
The power steering pump is now making absolutely terrific sounds and leaking at the shaft seal, so it’ll be on the chopping block for replacement.
There is a cable harness that the cruise control computer intercepts the transmission overdrive solenoid with which has failing pins. This has manifested in sporadic loss of 4th gear, meaning I’m either going 55mph tops or absolutely revving it flat out to hit 70. A kick or tug on the harness will often resolve it – I’ve tried various methods of biasing and restraining the connector pigtail over the past year or so, but outright repair/bypass is now a necessity because it’s getting too annoying.
Developed either a misfire or bad exhaust leak from the right cylinder bank, so while it will drive fine, it sure sounds like an old rattly diesel when it isn’t one (yet…). I’ll need to do a full heuristic debug before commenting on it more – it got worse lately as the weather cooled down.
It’s recently began emitting blue smoke out the exhaust intermittently. I’d attribute this solely to something like worn/crispy valve stem seals or sticky piston rings, but what was more worrisome is that the oil pressure gauge began to not register pressure. Now, in this era of Ferd, the oil pressure gauge appears to be a fake one – really an on-off scenario. I haven’t correlated the two symptoms by physically measuring the oil pressure yet, and really cannot say I’ve paid enough attention to said pressure gauge in months past for it to even have been symptomatic of anything. It could be a coincidence. Either way, out of an abundance of caution, I haven’t been driving Vantruck around the past few weeks.
Rear drum brakes have something going on, probably just excessive wear. If I set the parking brake, the rear brakes will drag for a while after releasing them. If I brake in reverse, then drive and brake forward, there’s a palpable clunk as something with just a bit too much slop pops back into position. Sounds straightforward, just willpower-limited for dissection.
It came with a diesel leak around the left bank of injectors – old and crispy return line fittings, and the cold weather has made it worse to the point where I’d prefer not to drive it. Less due to the fire hazard and more because it stanks of diesel, costs me money by leaking it out, and is rude to others for leaving dribbles on the road. Willpower-limited repair, as I have the fittings and hoses sitting in it right this minute.
Thrashing about the mountains has caused a power steering system leak. I haven’t dug into it to find out where from, but it’s actually not from the gearbox itself this time (a known failure mode of many a Ford truck), so it’s probably a stiff hose or loose fitting. In fact, I had to abandon a day on US Route 129 a few months ago because the power steering leak became dramatically worse all of a sudden, a small puddle per power cycle. Luckily, the system was filled with transmission fluid and I had a quart to keep topping it off on the trip home.
You notice it’s all turning and stopping related problems, more or less? Well, in order to not fly off the side of a mountain, it’s imperative that you be able to turn and slow down. Vans, while imperfect at this, can be coerced into doing so somewhat gracefully, but they’ll only put up with it for so long.
Oh, yeah, where’s Murdervan? Spoiler alert – I sold it back in September after shoring everything up nicely and writing a Facebook ad that, in light of current events might get me Investigated. It was sold locally in-town to someone who seemed enthusiastic and knowledgeable of old Ferd diesel trucks, and will join a small business fleet that does urban gardening and landscaping work. A very fitting end to its brief story with me, as it was always just too normal for my misfits. I’m sure I’ll see it around the city more!
So there’s also a lot of Ven to write up, besides the Summer of Ven series itself. I better get used to loving this keyboard and its probable timely successor once the keys start falling off.
Cute little robots
A few weeks ago, I was skulking around knick knack stores in the farthest reaches of Georgia (my latest habit, finally checking out all those antique and flea markets I keep blasting by on the way back and forth from the Smokies and Blue Ridge). A lot of these stores have vintage tools and hardware, which I enjoy perusing. However, at one of them, I found this little guy:
That, if you’re not familiar, is a Dr. Inferno Jr. Well, not really. It’s a Tomy Omnibot, a little robot toy of the 1980s that was probably pretty badass for its time, being programmable via cassette tape and all.
Needless to say, I made off with it because hey, it has some relation to BattleBots history as well as the history of programmable smart toys. It was in good physical condition, though the proprietors said they couldn’t locate the remote control at the time but would keep mining their stocks for it.
Without the OEM remote, it seems rather static based on my research, and so I decided to perform a unique restomod. I’d do a mechanical repair and restoration to get it in driveable first, but I had an element I wanted to add.
That is an old Futaba T4NL Conquest I got for free at some Swapfest at MIT many moons ago, and have just had sitting in one of my Electronics Mystery Abyss totes since. What better to control your 80s robot with than an 80s radio!
What you can’t see from the outside is the MicroZone MC6 transmitter that I organ-swapped into the T4NL. Yup, I done did it – a restomod of the transmitter with a modern day, albeit potato, 2.4G computer radio. This was a fun adventure, and I think I approached it in a unique (but harder) way than just tapping the PPM summation point and feeding it into a 2.4G radio module. I fully embedded the MC6 using the original Futaba gimbals, added the MC6 servo reverser switches to the back side, and wired in new switches to turn the 4 channel T4NL into a full fledged 7-channel radio.
And of course, this photo of my 80s robot that I drove around with my 80s R/C radio was taken at a car show I took my 80s van to. This, as I called it on the Facesphere, is #Radwoodbait for whenever those shows come back up.
I’d definitely love to write up the whole restomod of both the Omnibot and the Conquest T4NL radio, because it was just a fun distraction project over the holidays when everything was closed and I didn’t feel like going outside.
Remember, even while I’ve refrained from fixing this web-van (HEHEHE WEBVAN) up to post content, I’ve been taking my usual excessive amount of photos of every step or interesting happening. The content exists, I just have to find the willpower to write it up – and I hope finally having the damn site operational again will motivate it.
Also, I have so much to remember what I named “Potato”…starting with the title of this site. I’ll take care of it soon, I promise.