I beg your pardon, but...

Not much of a post today. Been working on a problem with our refrigerator for about a full, eight hour day. 

We've been having a recurring problem with our refrigerator's freezer section dripping water which ends up on the floor below the freezer door. It has been happening intermittently and only small amounts we can see. I haven't kept a log, but it started out over a week ago, then has been a mix of sometime it's obviously bad while other days it can go all day without leaking. For most of the time, there has been a deposit of ice on the freezer's floor (inside), and I've tried to troubleshoot it to determine if it's the system auto-defrosting circuit, or just what's going on.

A complication popped up today and it stopped me for at least an hour. This refrigerator has a built-in icemaker and while it's a longer story than this, three years ago last month I moved its input hose from the input of a water filter we use on the kitchen sink, to that filter's output, so that our ice was made with filtered water, not "just" city water. Today, with everything off and cleaning behind the refrigerator, I heard a quiet little hissing sound. It turned out to be that hose was leaking - spraying a very fine mist of the filtered water out behind the refrigerator. I had to turn off the cold water under the sink to stop that spraying.

My fix was to disconnect the icemaker from that water filter, which means there is no connection to water for the icemaker at all. I had to go up to the corner hardware store to buy an adapter to not hook that hose back to the filtered water. The hose is just hanging there, not attached to anything. I think from now on we're just going to live without an icemaker. It's not like we use a lot of ice and neither one of us thinks it's an essential appliance. Since the water in the bottom of the freezer had to come from somewhere it was either the water going to the icemaker, or the refrigerator's defrosting system not working properly, and everything else in freezer looks normal, not too much frost. I think what comes next is I'm going to go buy some ice cube trays and we'll just fill 'em up in the kitchen sink and freeze them on a shelf in the freezer. Just like we used to do 60 years ago.

Anyway, that's why it's an odd day. Back to normal tomorrow.  

My other, other stuff going on

I occasionally mention other things I'm working on or that are going on. Life is the simplest word and I think it was Beatle John Lennon who said, "life is what happens while you're busy making other plans." That's what this little post will be about.

Let's start with something I haven't talked about in a long time: guitars. I've got a small batch of guitars, like pretty much all players who can get more than one instrument tend to. Unlike the vast majority of home players, I built two of those guitars; an electric guitar from a kit back around 2015, and an acoustic that was more like a very extended repair job. It was a sales demo model from some guitar store and before sending it around to show what their guitars looked like inside, the manufacturer, Breedlove, sliced most of the left side of the guitar off. 

The long story is I started playing in junior high school, around 1967 or '68, and like most others, I learned what was referred to as rhythm guitar. Playing chords. I played in a couple of garage bands, and at a couple of parties, but I never really fantasized about being a musician. It was enough to make some sounds that didn't hurt people's ears or they even actually enjoyed. 

In my "young adult" years, through my 20s, I hardly played at all and didn't get involved with guitars again until maybe 1988 or later. I picked it up with intentions to practice more often a few years ago. Those plans would get derailed every now and then by either an injury or osteoarthritis. Over roughly the last year, that second one had me unable to play for a year or more. The effect of the arthritis was limited entirely to one finger, my left hand's middle finger. The very last segment of that finger, the one that you press onto a string to change the note you're playing, rotated left or toward my pinky finger. 

How much did it tilt? I don't really know how to measure that with what I have, but I'll estimate 10 degrees. The rotation wasn't the issue so much as the insane pain it caused. I'm gonna bet you've heard someone say, "it hurts to look at it?" Exactly.

About a month ago, I realized that I had bumped that finger on something and it didn't hurt. After some time trying to figure out if it was really gone, I started thinking what I could do to test it. I picked up one of the guitars I'd regularly practice with and found I could play without the insane pain.  The problems now are the same ones everybody picking up a guitar and starting to practice faces. Building up calluses on my finger tips so the strings don't hurt as much, strengthening the muscles involved so when I go to play some chord that makes all six strings sound, I can apply even pressure and put all the strings on the neck. Most importantly though is agility of those fingers, and knowing what to play. As in re-learning a few songs. 

My tendency in here is to put up some picture that's relevant or related to the topic, but tonight, I was going to link a video from Woodstock that I've run before. The problem with that is YouTube has decided we're not allowed to embed it, but I think I can run a link to it HERE. And here's a picture:

A photo of my repair job on the Breedlove guitar. The guitar is on my workbench but just for posing, in the final post on this job.  

The finger has better and worse days, so it's not entirely being back to normal and I have no way to know if that's even possible.

Another Interesting Possibility

Two weekends ago, I did a small update on things going on in my shop, both Saturday and Sunday (with more pictures).  The two general directions are either a more complicated and functional internal combustion engine vs. something more beautiful but still functional, like an orrery.  

Much of the intervening two weeks have been spent doing grown-up or homeowner shit, as we're speeding into summer and there are things that need the attention.  The exception has been something I stumbled across while doing something else.  It's called the Antikythera mechanism.

I'd heard of the Antikythera mechanism before, but hadn't thought about it in longer than I can recall.  It turns out to be rather interesting for a few reasons.  It's an ancient Greek device recovered in 1901 from a shipwreck in the Greek Islands dated to approximately 100 BCE give or take 30 years.  The most succinct description of it might be that it's the world's oldest analog computer.  The "world's oldest" by a thousand years.  If you'd like to be more specific, the kind of analog computer that best describes it is that it's an orrery.  

To borrow from the online Encyclopedia Britannica:

The Antikythera mechanism had the first known set of scientific dials or scales, and its importance was recognized when radiographic images showed that the remaining fragments contained 30 gear wheels. No other geared mechanism of such complexity is known from the ancient world or indeed until medieval cathedral clocks were built a millennium later.

The Antikythera mechanism was fabricated out of bronze sheet, and originally it would have been in a case about the size of a shoebox. The doors of the case and the faces of the mechanism are covered with Greek inscriptions, enough of which survive to indicate clearly much of the device’s astronomical, or calendrical, purpose. It is believed that a hand-turned shaft (now lost) was connected by a crown gear to the main gear wheel, which drove the further gear trains, with each revolution of the main gear wheel corresponding to one solar year. On the front of the mechanism is a large dial with pointers for showing the position of the Sun and the Moon in the zodiac and a half-silvered ball for displaying lunar phases. The drive train for the lunar position is extremely sophisticated, involving epicyclic gearing and a slot-and-pin mechanism to mimic subtle variations (known as the “first anomaly”) in the Moon’s motion across the sky. (See Hipparchus and Ptolemaic system.)

With something as famous as this, you might expect to find many pictures of reconstructions of it in everything from wood, to plastics (including in Legos) to intricate metal versions of it.  Some of those make the effort to recreate the final form and function of the Antikythera mechanism while others just emulate the functions.  This is a deep, deep rabbit hole to dive into.  

An example.  You can be sure the 100 BC Greeks didn't have plexiglass (or whatever that clear plastic is), but it's an impressive model.

Image Credit: Eternal Gadgetry  

There's this metal model on Instructables (design sharing), a 3D Printed model at Printables (Prusa), and another 3D printable on Thingiverse.  

And this illustrates only the gearing:

Image credit:  Communications of the ACM - figure 14 here.  Not that I could sit down with this and figure out what it should look like.  Besides, it shows you there are 65 gears but nothing about their sizes and numbers of teeth.  

Is this the next project?  I honestly don't know.  I like it, but 65 gears of unknown sizes strikes me as a lot for a first project cutting gears.  What's cool about this is the historical aspect.  That first orrery video I posted Saturday the 12th still seems pretty good (that links to the full video, not just the last few seconds).  I need to start looking closer at options.  Once I get my sprinkler system fixed.  And the other stuff.

Miscellaneous

A couple of random things to tack onto Saturday night's post about my choosing the next longer project for my shop.  

The first goes with the idea of hit and miss engines.  If you go to a home machinist show like Cabin Fever or a smaller, more local show, I can almost guarantee you're going to see one of these engines from PM Research:

This is a kit of the castings used to make the engine, and while they sell an "almost completely" built kit that's "fully machined" it's not like it's ready to just dump some gasoline, oil and water into it and hit the starter.  Yes, this is called a Red Wing engine and the casting reads Red Wing Minnesota.  I think. 

Before I started on the 1x1, I read about an alternative engine published in a magazine called Model Engine Builder in 2006 which is made from what's referred to as bar stock - chunks of various metals in standard sizes instead of castings like this one - and I bought a pdf of the issue.  "Bar stock" isn't exactly like going into your local lumber store and buying dimensional lumber like 2x4s or 1x10, but it's more like it than not.  The main difference is in the variety of shapes and sizes you can get in a wide variety of metal alloys.  

For the first time, I did a search on YouTube tonight to see if anyone had posted a video of this engine, and found this 17 year old video, which still isn't as old as the engine.  

To me, the Red Wing looks better, but more videos would be more better.

So Many Possibilities

Since the big story of the weekend appears to be sending Katy Perry and an "all woman crew" on a suborbital ride to the Kármán line, and how excited she is to be going on her trip, not much else is being talked about.  So time for a little side story. 

Within the last month or two, someone asked in a comment what's going on in the machine shop, since I haven't posted anything in quite a while.  I haven't posted much because I haven't been doing much with my machine tools, really only little repairs, or the occasional part or two to get something working better. 

As part of that, I seem to have dropped the engine I had started back in '21, called a one by one (or 1x1) because it's a one inch bore and one inch stroke internal combustion engine.  Putting it the way I did, ("I seem to have dropped the engine") underlines that this wasn't a carefully contemplated decision, arrived at by gathering lots of data, compiling that into spreadsheets used to figure what's the best alternative; it just happened.  There were several things that intervened and sucked up time on my big mill, notably tropical storm Ian in  '22 and Milton in '24.  The work after Ian involved totally redesigning the way my tower has been supported since I installed it in the early 1990s and was completed in June of '23.  

I've had to crank the tower over to work on antennas, most recently this past fall for tropical storm Milton.  It has all held together well.  

Again, these interruptions got me away from the 1x1 project long enough to forget where I was and what I had been doing.  It also led me to think more about other projects.  In overview, the 1x1 is much like my first engine, a Webster, in being a free running internal combustion engine.  I've started to get interested in engines that have RPM control - like a primitive cruise control.  The first such engines, called hit and miss engines, have a feedback mechanism that control the engine actually firing or not, so that when they're under load, they "hit" and fire more often than when they're under a lighter load - or no load - when they "miss" most of the time.  They deliver the required power when required.

I'm more interested in making one of those than another version of what I've already done. 

Then there's more.  Ever heard of an orrery?  These are essentially model solar systems, from the relatively simple to the unimaginably complex, whether machined metal like those two or 3D printed plastic like this one.  Include the incredibly simple - just the sun and one planet, like this.

This is a screen capture at around the last 10 seconds of an almost 19 minute video showing how everything here is made.  That's supposed to be the sun on the left and Earth and moon on the right. The motions of the Earth and moon are supposed to be scaled properly.  The big, light colored things are the thumb and fingers of the machinist, wearing surgical finger cots to keep from touching the brass gears and causing corrosion.  I'm guessing the entire thing is about a half inch from end to end.

So what's next?  I don't know.  The orrery appeals to my amateur astronomer and telescope maker side. The hit and miss appeals to my decades of designing electronic control systems or control loops. 

Just About Done with Milton Repairs

As I mentioned in my first update Thursday afternoon, we had very little damage from Milton.  Power was out very little, if at all, and there were only a couple of things that needed repair. The big one seemed to be that an element broke off of my HF Log Periodic antenna (a Tennadyne T6) on one side of the boom, very much like what happened back during Ian two years ago.  The second was that our east side fence gate broke, throwing the metal piece that latches the gate closed about six to eight feet from the fence. 

The antenna fix ended up being different from the last one and while I thought I could just go find a piece of the 1/2" aluminum bar that I used to fix it last time, the unnoticed curve ball was that the tubing for this element is larger diameter than the one that snapped two years ago.  That one had an Inside Diameter of 1/2" while this one was 5/8".  I went looking for a piece of 3/4 or 1" bar I could reduce in diameter until it fit in the antenna tube, but didn't have one.  I absolutely didn't want to leave this one on the ground and order a piece of bar stock - that would take days to a week. That's when I found a piece I had originally bought to fix the antenna after Ian, a piece of metric-sized aluminum tube that was a bit oversized in Outside Diameter: 17mm OD, or about 0.670".  

My lathe has been obsoleted by Little Machine Shop, but it's SIEG SC4 type, 8-1/2 by 20", they sold as their model 3540. Yes, SIEG is the Chinese machine producer that makes the vast majority of the smaller home/"hobby" lathes in the 7x10, 8x12 and similar small sizes. This one is capable of doing automated cuts by using some controls on the carriage originally intended for threading, so to reduce this bar from that 0.670" OD down to 0.620 (or so) is a matter of setting up for repetitive cuts, put the cutter in position, throw a lever and just keep an eye on things. Pretty much. At the end of the cut, flip the lever, return the carriage to the starting point, move the cutting bit to the next cut's depth and repeat.

I turned it down to the point where the ends measured close to 0.625 and then pulled the tube out to check in the antenna element. It's not uncommon for a part to not come out to the same diameter over its entire length (at least not on my lathe ... or not on these "hobbyist" lathes), and while the ends fit well, I took off another .003" and it made a nice easy fit.  As with the last time I did this, I used a gun cleaning wire brush in my battery powered drill to scrub the insides of both parts of the element.  As with the last time, it was full of dirt that I believe was stuck in the tubes by mud dauber wasps.

Here's what it looked like put back together.  Before cranking the tower back up, I did a NanoVNA sweep of the antenna and it looked fine. Then I swept my other antenna on the tower, the VHF beam, and it looked unaffected, too.

The aluminum tube has a blue line drawn on it to mark the halfway point and you can see that in the gap where the two pieces meet. I smeared the (right) half of the tube with nail polish to glue the tube in place, then drilled a clearance hole for the 6-32 hardware: a screw, flat washer on the far side, lock washer on this side. Yes, that's a brass nut on a stainless screw. 

The other issue, the fence gate had much less progress. It looks to me like that gate needs to be rebuilt or replaced.  We have a "poly-something-or-other" fence and the gate latch was ripped apart.  I forgot to grab a photo of it, but it's this basic idea, except the highlighted part looks to be about twice the size of the highlighted one in this picture and holds twice as many screws on the side of the right angle closest to the camera. That's the part that ripped off the gate post and got thrown.

Where the screw holes are in this pic, ours has slots, 1/4" wide and 1/2" long.  Every screw was ripped out of the plastic post, ripping the screw holes in the door (4x4) wider.  I found the ripped off piece about 8 or 10 feet away from the gate.  The screws are self-drilling steel screws that are just under 1/4" major diameter (across the screw threads).  While a "standard fix" might be to go to larger diameter screws, the metal won't handle that and the slots would have to be widened, perhaps going to 5/16 or 3/8".  Not an outrageous job for the CNC mill, but not a one hour job either.  The issue is how much damage that the post this mounts to sustained.

The antenna is fixed and operating; the gate not so much. Our neighbors on two sides, backdoor and next-door, have trees in corners of their lots that overgrow our lot all the time. Where the three of our lots touch, I'm constantly pruning their trees and repairing next-door neighbor's fence. Milty brought more attention to that because excessive wind from the wrong direction could have put backdoor neighbor's tree through the roof of my shop. It turns out the tree came down in their yard. This time. Turns out I had trimmed those back between Helene and Milton.

Sometimes the Shop is a Silly Place

You know that saying, "when the only tool you have is a hammer, every problem looks like a nail?"  It's sorta like that.  When you have a machine that's accurate to thousandths at the worst, everything looks like a project for it.

At some point a couple of months ago, my favorite spatula broke.  This is not an exotic, gourmet chef kind of item, it's a silicone rubber, flexible blade on a plastic handle.  It just worked perfectly for ice cream making, something I do every four days.  The handle goes into the silicone blade to hold it in place; there's what looks like a rectangular opening in the silicone blade and a rectangular piece of plastic that goes into it.  It's kind of like woodworkers would do to make a blind mortise and tenon joint.  Does the tenon-like piece on the handle have barbs or something on it that holds the soft blade?  Maybe, but I don't really know because I never got to see it apart.  Instead, the harder (and more brittle) plastic snapped off where it went into the soft handle.  

Sure I bought a replacement, but I couldn't find one in the local store that felt like the one that broke.  The broken one was better, hands down.  I tried gluing the two halves together.  That broke again within a couple of weeks.  

The two parts sat on a counter where I saw them every time I walked by.  Every time I looked at them, the background processor in my brain thought of how to fix it.  Without telling me.

I suddenly had the idea to create a joint by putting something like dowels in one piece and drill matching holes in the other.  I measured the plastic piece sticking out of the silicone rubber at .630 wide and .180 thick.  I did some edge finding with my laser edge finder and made an XY coordinate reference system.  With the hard plastic broken off in the soft silicone piece being 0.18" thick, plain old wooden dowels wouldn't work.  All of mine are twice that in diameter.  I looked at toothpicks and they were too small, like 1/16".  Then I remembered I have a box of 1/8" dowel pins.  Instant tenons for 39 cents.  

 Setting this up so I could drill two parallel holes straight into the plastic, holding the position to within "a few" thousandths was a bit of a pain.  The two blocks with three holes on each end are called 1-2-3 blocks, because those are the side lengths: 2" wide, 3" tall and 1" thick.  Behind the pin on the right is a triangle sticking up.  That was a piece of scrap aluminum to help clamp the silicone rubber and keep it from moving in the vise. I should have cut it off, but it honestly didn't get in the way.

The handle was worse to deal with.  It was too long to put in this position to drill and far too curvy to get it to hold still while tightening the vise.  I had to clamp it to the side of the vise with a woodworking clamp and position it all by hand.  All to drill two holes in the handle that would allow me to glue the pins in both halves. It's not centered, but the handle holds the blade.

This was before I glued it.  I drilled the holes in the handle .003" oversized and got a generous amount of glue in there.

The overall view is this:

(That clamp in the background, wooden handle on the right, is how I held the handle to the left side of the vise.)  While I'd be completely unsurprised to find that the parts are created in CAD and CNC used to make the mold, this is the only CNC-repaired spatula in existence.  At least that I know of.  Now where did I leave that number for the Spatula Hall of Fame...?  

A Little Me Me Me

You may have noticed I had no post Wednesday the 9th.  The peculiar part of this is that the previous Wednesday (the 2nd), I also missed a post which I explained about the next day.  It was due to a few recurring car troubles that sucked up far too much of the day and attention.  

As Roseanne Roseannadana used to say, "It just goes to show ya. It's always something. If it's not one thing, it's another."  I picked up the car the Monday before that, July 31st, and it continues to start properly.  I suppose I'm more comfortable with it not even two full weeks after picking it up than I expected to be, but it has started without even the slightest balk or hesitation.  

It wasn't the car thing that made me miss Wednesday, it was another.  On Tuesday night, as I was writing that post, we suddenly were interrupted by an alarm.  Not instantly recognizable, but vaguely familiar, we soon realized it was from our kitchen.  The second time we had to have our kitchen rebuilt because of a leaking dishwasher, we added an alarm with a sensor under the kitchen cabinets so leaks from the dishwasher wouldn't have time to ruin the cabinets like before.  It didn't take very long to discover that the cold water valve was dripping.  You can see a drip getting ready to fall in the red highlight box.

There was a puddle on the bottom of the cabinet under the kitchen sink, but not a lot of water.  Since the alarm sensor runs several feet along the lowest part of the floor there, we don't have a way to see the whole length of it. 

Since it looked like it was going to be a simple replacement; unscrew everything attached to the old valve, unscrew the valve from the wall and replace it with a new one, I opted not to call a plumber.  Probably needless to say, it didn't go quite that easily.  For me, plumbing jobs never go as easily as I'd like.  I think most readers know that pipe names have nothing to do with their actual sizes.  Someone once told me that "there's nothing 3/4" about a 3/4" pipe."  It's worse than that.  There are two T connectors visible in this picture, one sort of horizontal attached to a braided metal jacketed hose going off to the right, and another mostly vertical T going to two "poly-something-or-other" tubes running down.  The braided metal jacket (tube) is called a 3/8" and the others are 1/4".  

I didn't learn until this job that there are three families of connectors in both sizes, and a connector from one family won't mate with any of the other families.  In the RF world I come from, adapters between connector families are bread and butter products, available by the barrel-full, by the pound, or however many you want.  In the plumbing world, this doesn't seem to be the case.  

In the radio world, this is roughly a 1/4" to 3/8" adapter like I needed.  (These folks are expensive, you can find them cheaper.)

The reason for the long explanation is because I bought the wrong valve on Wednesday and couldn't get everything restored.  The valve I bought had two 3/8" fittings and I needed one of each.  I couldn't come up with an adapter from 1/4 to 3/8".  Since it was approaching evening, it was all I could do to put the leaking valve back in place and search for a valve that would work for me.

Thankfully, the company that made the valve also made one with both 1/4 and 3/8" fittings. The True Value guys were willing to take the return and credit me with the price difference (the one I needed was cheaper than the one I mistakenly bought).  Within an hour of leaving to do the exchange, I had this replaced and everything working. 

Ahhhh...  Everything was fine at the homestead.  Peace and working plumbing.  For about 24 hours.  This repair was finished Thursday by about 11AM.  Friday morning, I turned on the air conditioner in the shop about 10AM.  Within 15 minutes it didn't seem to be cooling off as well as it should.  As Friday went by it was more and more evident that the air conditioner had broken down and it needed someone with the equipment to fix it.  Being the peak of summer, it's their busy season.  No repair tech until next Wednesday.

"It just goes to show ya. It's always something. If it's not one thing, it's another."

A Little Update

Since it's a slow news day in the space news I've been following, I thought I'd post an update to some work on my tower fix I last talked about in early April.  A few of you left comments about the bare U-bolts on the pipe not being secure enough.  I got the idea that the heart of the issue is the contact area of the bolt on the 4" schedule 40 pipe is simply a very thin line – infinitely thin in principle – and the fix would be spreading that load over a larger contact area.  Reinforcing the pipe is another option and remains available.

Over the course of the next few weeks after that post, I struggled to envision some sort of thing that could go in between them, eventually settling on cutting a sliver of a cylinder.  I looked at U-bolts that came with things to spread the load, and eventually thought of a chunk of aluminum that would go between the U-bolt and the pipe, about a quarter inch thick and at least a half inch wide.  The place where that U-bolt went would be milled out the width of the bolt – half inch.  Ideally, the stainless U-bolt would tighten down into the aluminum, creating (crushing) a groove the bolt sits in and really spreading out the contact. 

I don’t think that my CAD drawing is any easier to understand than this image of the part after machining the two big channels and as it’s ready to come off the mill.  

The load spreading portion of the part is the semicircular arch that starts just in from both the bottom left and right corners of the piece.  You can see it’s very thin at the bottom and much thicker in the top middle (about 3/10 inch).  Looking down you can see a large aluminum (almost) semicircle; that’s leftover (scrap) and the bottom of the arch surrounds and presses onto the schedule 40 pipe (4-1/2 inch actual diameter).  There’s another arch cutaway above the one the U-bolt presses onto.  That one is the one the U-bolt sits in.  The clamping force extends over that arch that clamps onto the pipe.  

You can see that there’s an opening in the top of the part there, the top of the U-bolt sticks out there, and on both sides of the channel you can see what are 1/2” holes drilled through, top to bottom.  That’s because my original idea of how to mill the part was to stand it up in the milling vise and mill it left to right, raising and lowering the cutter to cut out the shape of the channel and the half inch holes removed a lot of what had to go.  I gave up on that idea after a lot of simulations and attempts to get tool paths that would do that.  After that I milled it in the position you see it, lower the cutter into the work on the right end of the semicircle, cut the semicircle counterclockwise, lift the cutter out of the work and move (at high speed) back to where it started and repeat the cut a little deeper.

It took longer to figure out how to machine it than to decide what it would look like or how it would work.  Along the way, the name morphed from the U-Bolt Load Spreader to the U-Bolt Thingy then to just plain Thingy or the UBT.  

This is a photo of the Thingy in place, with the U-bolt tightened down (but not torqued to spec.)

For scale, the U-bolt itself is half-inch diameter and the channel it’s sitting is just a bit deeper than that (about .050 in.)  The “wall” on the right is the channel stock used for the arms.  The other channel stock is on the left side, well out of picture.  The really big round thing is the 4 inch pipe.

Looked at from 90 degrees to the right, the channel and UBT look like this:

The slot in the top of the channel and the bolt sticking beyond the UBT are very obvious; so are the half inch holes that were drilled (the second UBT doesn't have those holes).  The holes look like they're a perfect place to attract mud dauber wasps, so I filled those with caulk.  

When I started to work on the tower to install these, it became obvious that the tower had been driven from vertical by some wind episode we had.  Part of the installation then became to temporarily clamp the tower to the old house bracket to pull it vertical, like my temporary fix described in that April update, then to torque down the bolts before taking the clamp off. We haven't had particularly intense winds in the two weeks since this was added, but the tower has remained vertical. 

Tower Damage Repair Update

Strictly my repair to one of the tower legs, since my last update, last Saturday.  As I said at the end of that post:

To the previous posts, several people suggested tapering a steel bar and driving that up that leg to both open it back up and straighten it.  I intend to try that with a back up of cutting off that bottom ~ 2" and putting a smaller, heavier wall tube into it's place. 

Later that evening I decided on the second option and ordered a 12" long piece of 1.375 Outside and 1.125" Inside diameter aluminum tube along with a piece of 2" on a side, 1/4" thick angle aluminum from one of the online sellers I use regularly.  I have one drawing from the tower company that gives dimensions for the upright tubes; it says the OD is 1.50", the wall thickness is 0.065" and those say the ID of the tower leg will 1.370."  With an ID that's .005 smaller than the tube that's supposed to go into it, that told me I pretty much needed to prepare a way to reduce the diameter of the tube I bought down to about 6 to 8 thousandths less than the tube ID.  That looks like this on my smaller lathe, the Sherline, where every tool is about at its limit for the largest things they can handle. 

That thing just right of the lathe's chuck is called a steady rest, and they're pretty commonly available for all lathes.  It's awkwardly positioned in this view, barely doing anything, but while most people think of a steady rest as more appropriate to skinny stock, like 1/4" diameter rods 6" long, they're really appropriate any time the stock's length is several times its diameter.  What they do is reduce the amount the work flexes away from the cutting tool when it applies force to the metal.  That reduces the chances you'll cut a taper onto whatever you're turning.

It may not be obvious, but no cuts reducing that tube's diameter have been made.  I started asking myself if the wall thickness on those tubes was actually .065" and decided that before I reduce the diameter of that tube, I should know what it needs to be.  I took out my jigsaw and cut about the last two inches off that bent tube to better inspect the area I need to work on.

If you look at the top of this cutoff piece, the most important thing about it is that the profile isn't circular.  The side closest the camera was still badly bent.  Most importantly, it allowed me to measure the wall thickness and see that it was very different than that one drawing.  Instead of .065" the readings were around .058 - making the ID of that leg tubing more like 1.384 (1.500 - (2*.058)). That should mean my 1.375 pipe (which I measured at closer to 1.373) would be a comfortable sliding fit. 

Except it wouldn't slide into the tower leg.  I went to measure it with my telescoping inside diameter gauges (like these but not these) on two diameters at 90 degrees to each other (top/bottom and left/right).  I could see the tube was still flattened out top to bottom. 

After taking off another half inch or so from that leg, it measured closer to the same diameter in both directions, but the test piece of tube still didn't go into the leg.  It seemed like it could be a burr, so I used a brake hone, which I first bought to hone the cylinder on my first internal combustion engine, along with files to reduce any burrs along the cut edges, inside and outside curves.  That did it.

That's the way it sits right now, but I'm not sure if it's ready to proceed.  There's more work to be done before the tower is ready to go back up.  That tube will be secured in the leg with two 1/4-20 stainless bolts and then a replacement for the square bracket that secures the tower to the slab will be made and  put in place. 

Again, this is not being looked at as a permanent fix, just a way to get through the next few months. 

Running Late, So Tower Repair Update

I had started working on an idea that's turning into a Smallest Minority-style "uberpost" and it's looking like there's no way I can finish that in time, so a couple of quick updates on the tower work, simply because I can write this off the top of my head.  The task is turning into a bigger, more involved fix than I really wanted to work on.  Thankfully, it turned into fall back in September - kind of unusual for us - and it's not that bad working outside.  

The majority of the work this past week has been completing the tool that I intended to use to clamp around the tower legs and try to make the one leg round again.  Or, at least, rounder.  

The finished tool.  That's a little over 2" long, 1-7/8" square block of gray cast iron that was machined there on the mill.  It was painfully slow to do, for two reasons.  First off, I find cast iron to be a bitch to machine.  It goes beyond dirty to filthy in the chips it throws.  Second, the sheer size of it made getting tools onto the milling machine far more tedious than usual.  When I used my boring head, I could cut about a .050 radius to full depth either by CNC at 1 inch/minute or cut by hand, as long as vacuumed out the cast iron chips often. 

Why the mill and not my big lathe?  I still have the piston from my 1 by 1 engine on it that I haven't finished (and haven't worked on for a host of other times sinks) that's a position sensitive setup.  I really need to finish that.  "All I gotta do" is finish the two channels for the piston rings - so that I can get it off the big lathe and get access to it.  

The tool clamped onto a round section of the tower leg.

My approach was to slide that tool closer to the obvious kink in the tube and clamp down on it, reshaping the tower leg a few thousandths at a time.  It worked to some degree, but when I went to get a bigger, apparently stronger C-clamp, I did something I've never done; something I didn't even know was possible to do.  The clamp you see there opens to 3" and I replaced it with a 4" clamp.  

I bent it.  By hand.  Just by tightening the clamping action bare handed - not with a "half inch breaker bar."  

Then I went got the even bigger 5" C-clamp.  For that one, I got a short length of pipe to help me tighten the clamping screw.  That bent the handle on the clamping screw.  I took my jigsaw to the that clamp and cut it apart because with the screw bent to 90 degrees, it was impossible to loosen.  I might have found a way around cutting the clamp screw in half, but I had pretty much burned up every nanogram of patience I have by then.

The smaller clamp on the right is bent down close to the handle.  The top of the clamp, lower right, is no longer aligned with the other screw; it's about 1/4" off the table and it closes with the two clamping surfaces out of alignment.  You can see the damage on the one on the left. 

As I've mentioned before somewhere, the tower leg is bent over it's bottom two inches.  

In the second picture you can see a slot in the tube below that bend. I'm sorta following a procedure here that's like an anonymous comment to my last post.  To the previous posts, several people suggested tapering a steel bar and driving that up that leg to both open it back up and straighten it.  I intend to try that with a back up of cutting off that bottom ~ 2" and putting a smaller, heavier wall tube into it's place.  

If end up cutting off the bottom two inches and attaching a smaller tube inside this one, I can use the existing fastener in the concrete slab with some luck.  If I'm using this leg, I'll avoid that slot and drill a new bolt hole perpendicular to the axis of that one; across the tube from this viewing angle.  Either way, I'm replacing the small piece of angle aluminum (2"x2"x1/4" thick).  If I drill perpendicular to that slot, I might put identical pieces of angle aluminum on both sides of the leg and two new fasteners in the concrete slab it's sitting on.  

One Out of Two Isn't That Good

But thankfully it wasn't worse.  

As a quick followup to the post on repairing my antenna, I ran into a problem trying to complete the work today.  The bracket went up fine and appears to be secure.  I ran into a problem with the tower itself. 

I was too busy to grab a photo of the damage, but I found a six year old picture from when I installed my new counterpoise pipe that supports the weight of (and torque from) the tower as it gets cranked over.  This will help explain the problem.  

The tower, like most towers, is a triangular design; it has three main vertical members and those are connected by a matrix of smaller tubes.  The three vertical elements are the interface to the concrete pad.  The base is aligned with a factory installed hinge on the north end of the pad so the north side of the tower aligns with east/west.  On the south end, there's a single leg and it has hardware built on that attach it to fastener in the concrete pad. 

The vertical members are approximately 1-1/2" OD; the smaller tubes are approximately 1" OD.  Highlighted in red is the anchor to a bolt in the concrete pad.  The damage here is that as the tower blew in the winds with the house bracket broken, that south leg with the right angle aluminum piece bolted to it bent.  That round tube is now kind of oval in cross section, and points a little to the left.  As a result, the tower doesn't "want" to stand vertically on the mount.  It leans left in front of the house bracket.

So now what?  I'm not really sure.  All of this repair work has been a stretch for my skill set, and this is a bigger stretch.  I have some steel stock that I can put into that leg to try to restore more roundness and straighten it out, but I've never tried to do something like that.  I don't have an anvil or anything that can support the metal while I bang on it, so I'm still trying to figure out alternatives.  I could put something under that leg that keeps it from being bolted down but lines the tower up with the bracket better.  Extending that bolt in the slab, so that I could secure that point with the leg standing on something, doesn't seem feasible though. 

An Odds & Ends Night

A couple of things that have nothing to do with each other.  First something that just seemed funny - strange to me.  

Yes, it's real, and no I don't have one.  I don't end up caulking things often enough to justify it.  Mark 1 finger works well enough, and I've never gotten splinters. 

The hurricane cleanup was largely in a back corner of our property where my next door and back yard neighbor's trees had dumped on the workshop's roof.  The hazardous one is the back yard neighbors because they have an enormous tree that dropped a branch on the roof that was hard to get rid of, not just because it was 5' long and heavy but also because it was still attached to the rest of that tree.  I had to chainsaw a couple of chunks off it while standing on a ladder before I could drag it off the roof, then chainsaw that log into smaller pieces to carry to the curb.  The roof is cleared off, though, and there are no indications of roof damage.  It could have been much worse.

Next door neighbor has different trees, but the one that always causes hassle is her bougainvillea.  Pretty flowers but thorns big enough to kill a rhinoceros.  Mrs. Graybeard and I came in with matching scratches on our right arms from brushing against a branch.  You have to pick up branches very gingerly to drag them to the curb.  The surprise there is that I had four broken off thorns in my shoes after we finished and they didn't go through the soles of my running shoes and jab into my feet.

Cutting back the branches in that area took all day Friday.  Saturday I sat down to fix the antenna elements and ran into a supply issue - as in I don't have any metal I can machine into the parts I need to fix it.  I need to describe the situation and how I intend to fix it.

The antenna had one of its dipoles snapped off - the before picture is in Thursday's post.  The metal tubing that snapped was 5/8" outside diameter and the remaining broken-off parts sticking out of the antenna's boom are about 6" long.  Here's the after picture with the broken-off parts highlighted.  This was almost as soon as I cranked over the tower and it stopped raining. I've since sanded those broken tubes to shiny metal and sawed off their jagged ends.

Think of a broken pipe plumbing problem.  A standard fix for a copper pipe is to clean up the ends and then silver solder a sleeve over both ends (quick example).  Since every joint in the antenna's elements is two pipes slipped into each other and attached with a self-tapping screw, I thought I'd machine a sleeve made of aluminum around 6" long to 5/8" ID.  As for the OD, 3/4" seemed natural but only the ID matters.  I figured each of those ends on the antenna boom and the ends of rest of the broken off elements would get about 3" of sleeve and I'll use two screws on each side of the break. 

It turns out that 5/8" ID aluminum tubing isn't particularly easy to find. Another possibility would be to start with a 3/4" aluminum round bar, drill it on axis and then bore it to 5/8" ID on my lathe.  That's throwing out almost all of the bar, but sometimes you gotta do what you gotta do.  Unfortunately, I don't have pieces of aluminum that could be turned down to 3/4" and 4 to 6" long.  A 12" long bar is on order and will be here tomorrow.  It's possible I'll get the elements repaired in the next day or two. 

As for mounting the tower and the house bracket, I will go to 1/2" x 3" long lag bolts for a short term fix just to get the tower up and out of the way.  The "real" long term fix (which hasn't been imagined yet) is going to be in a month or two; "when it cools off."

Update on the 1 by 1 - Part 28

Update on making the piston.  It's done except for one cut.  Possibly two, but I'll explain in a minute.  The engine plans call out a Vyton plastic piston ring, but when I look at sources online, I need to specify which blend of Vyton I want and that isn't called out in the drawings.  I emailed the designer two weeks ago and never heard back.  Since the last postings I saw from him was that he was taking a break, it's possible he's on a summer vacation trip somewhere.  

So I did the next best thing and went to Otto Gas Engine Works who makes many sizes of cast iron piston rings and ordered three of the recommended size, last weekend.  He shipped immediately, Tuesday after Memorial Day by USPS.  They said delivery by Friday but the rings now appear to either be lost, or somehow, somewhere they're still in the postal system, just not being picked up by any scanners.  Meanwhile the guy at Otto said to wait to cut the ring slot until I had them for best sizing results. 

Every feature is cut except for that.  I'm not 100% sure if I'm going to use two cast iron rings instead of the one Vyton, so that's the only real question.  The Webster I did last year had two cast iron rings with a smaller bore and stroke.  Sounds to me like two might be better.

Nothing to brag about, just to show the oval counterbore made by moving a 7/16" (0.4375") end mill back and forth 0.282" is cut, the circular counterbore is cut, the through hole for the wrist pin is visible as well as the threaded hole (#5-40) for a screw to grab the pin.  All of these things were done by moving the blank to the mill, finding where X and Y zero were and using the CNC (manual entry) to go to specific spots to cut.  

Before that, when I went to check the piston and cylinder for a sliding fit, I found that the cylinder had a narrow spot in it, closer to the top than the bottom, about .001" too narrow.  Hard to call being small by 1/1000 a wasp waist, but removing that gave the piston an easy sliding fit.

In this view, you can see a groove near the wider stock in the vise jaws.  That's where the piston gets cut off the rough stock.  If I make one ring slot positioned where the print says, it's 0.170 from the end of the piston to the closest edge of the ring slot.  If I cut slots for two rings, I don't know where they need to go just yet. 

The last time my package of piston rings was scanned was Thursday, June 2nd at 11:23 PM in Linthicum Heights, Maryland.  Since it's two days later and it hasn't shown up anywhere else, I suspect it fell off a bench, a truck, or something.  A year or two ago, I had  package shipped here from South Florida that instead of coming here from post office facility near Orlando, went to Pittsburgh,  PA.  That one never stopped being in updates; it was clearly still in the system.  This one isn't inspiring confidence.  

---

Potential tropical cyclone one update.

Longtime readers will know my standard line about tropical storms, "it's a poopy day with a press agent." This one never even became a tropical storm.  I think every forecast since they started calling it "Potential Tropical Cyclone One" has predicted it would become a tropical storm by the next forecast.  So every three hours from Thursday until now, when the forecast was issued, they essentially were saying, "this time we're sure!  It will be a tropical storm in the next update!"  

We had rain all day - actually from mid-afternoon yesterday until mid-afternoon today.  Much of the time it was small droplets, kind of extra chunky humidity.  The wind wasn't noisy, didn't seem strong, and I'd be surprised if we saw a 25mph gust here.  It was simply a poopy day with a loud, trumpet-sounding, press agent

Update on the 1 by 1 - Part 27

There's honestly not much to update here, my time has been split between a big project that I have to wait on someone else to do their part before I can work on it, and the engine.  The time I've spent on that big project has been learning some new skills in the radio technician arena, and that has taken most of my weekly spare time.  I've only spent a little time on the engine, starting work on the next part which is the piston that goes on the end of the connecting rod I just finished.  The piston is a fairly simple shape - flat-topped rather than a some sort of complex curve like real, high-performance engines have.  Like the rest of the drawings I've excerpted, I've left out certain dimensions but not all.

It's essentially 1.000" diameter now but needs to be sized for a sliding fit in the cylinder, which is close to .003 under (or 0.997").  I would have preferred to have started with a just slightly oversized bar, but I had some scraps of 1-1/2" bar in my stock buckets, and all I've really done is cut this down to 1.000 diameter.  No features have been machined into the 1" diameter end and the next step is to finish the piston to that "sliding fit" the print calls out.

I have to say that one of the things I find hard to deal with is terms like "sliding fit."  What exactly does that mean?  I mean, a one inch diameter piston would slide in a cylinder a few thousandths of an inch bigger, but it could also "slide" in a two inch diameter cylinder - along with rattling a lot.  It would just be really tough to get any compression out of that.  I'm more comfortable with numbers and a little investigation showed that numbers have been assigned to this over the years.  There are categories of fit and online calculators that will tell you the limits given a specified size and a class of fit.  Given the descriptions of running fits for an engine, I thought the most reasonable choice was one in the middle, what's called ANSI B4.1 RC4 or RC5.  RC4 says my piston can be between 0.0008 and 0.0016 smaller than the cylinder. RC5 is looser; from 0.0016 to 0.0024 smaller.  Since the prints don't distinguish those, I'll shoot for the smaller number and if I overshoot RC4 and hit RC5, that's probably OK to live with.

Weekly Update on the 1 by 1 - Part 26

And... it's done.  By the time I was done with last week's post on the version 1 connecting rod, I had decided to do a second one, using the technique I had used on the Webster's rod.  By Monday, I had found a sacrificial piece to screw the second blank to, made the initial screw holes in that blank (to hold it down to the "fixture" and got it zeroed in the Mill's coordinate system.  That looked like this:

My plan was to cut a wider rough contour cut - that's the kind that just cuts the outline of the part - leaving 1/8" more metal at every point around the part instead of .025" like I did on the first version.  Then I'd cut the circular bosses and finally cut the finish contour taking away the extra 1/8" of aluminum.  

The rough cut went perfectly, but when I went to cut the circular features, Mach3 wouldn't run the G-code, so I called it a day and thought I'd double check code later.  This is code I wrote (manually) so some sort of dumb mistake is almost guaranteed.  Sure enough, it turned out that I had left off a minus sign on one semicircle I was cutting.  With that fixed and running, I cut to depth around the circular features, taking away the 1/16" of depth on the rod between the circles. By Thursday, that side was done. 

The second side took more time to fixture than to cut and was done Friday. Since it looked almost exactly like that previous picture when finished, I'll leave that picture out.  

The next steps were done today.  I took the rod off that little piece of scrap, stood it on the small end, making sure the extended sides were vertical with a machinist's square.  That took several operations: I had to re-zero the part's coordinates to use the mill as a precise drill press.  For two holes, each of which gets five operations.  First, they get spotted with a center drill, then drilled through, tapped for screws, drilled halfway vertically with a bigger bit to clear that screw's threads, and finally, touched with a 1/4" end mill to make sure there's a flat spot where the screw cap sits.  After all that, it looks like this:

You might notice a couple of horizontal grooves on either side of the circular boss; I had the idea that I'd make scratches to lower the slitting saw blade down to and align it visually.  After I cut those grooves I decided that wasn't a good way to do it - parallax and other visual errors because the saw blade can't be right next to the engraved scratch, and set the depth of the cut by making the top of the part Z=0 and lowering the bottom of the blade to a depth I could get from the drawing.

All this to prepare for cutting the cap off.  Which has to be done before that small hole you see can be replaced with a drilled and reamed 3/8" hole for the crankshaft.  Which required yet another setup and another set of coordinates.  Now it looks like this:

So there it is, another part to go in the "Done" box.  I see it's almost a month to the day ago that I posted about getting ready to start and how I wanted to do it.  I don't know how many parts are left to do, and I don't really want to know.  If I make one part per month I'll be working on this engine for years.  When I started the cylinder (in December!) I figured the next parts would be the matching piston because it gets sized to the cylinder, which means I need this connecting rod and the pin that holds it in the piston (called a wrist pin).  Those are lathe parts so the adventures in CNC milling are over for the moment. 

Update on the 1 by 1 - Part 25

I ended up the last update saying I had poured epoxy into the finished side of the chunk of aluminum I've been working with.  I concluded with a picture of the epoxy-filled side after it had cured a while and said it would sit overnight.  

Based on the experiences with the first side, I was going to do four passes: two rough and two fine, but very different from what I had done on the first side.  Essentially, I broke the one rough pass from the previous side into two.  There are two similar algorithms that my CAM program use to create tool paths, called waterline and contour.  The major difference is that waterline takes a series of layered cuts down the vertical extent of the part while contour makes one single path around the part.  For waterline, we (the user) defines how thick those layers are and how close they're allowed to get to exact size of the part.  For contour, we have some control over how deep that one path around the part is - and we also set how close the cutting tool gets to the exact size of the part.   

I chose to do a contour cut around the part and then do a waterline just deep enough to ensure the top of the connecting rod was the proper distance (1/16" - .062') below the raised, circular bosses. 

Cutting a single slot, the depth of the part can be a problem because it can be hard for cooling fluids to get into the slot.  In the case of aluminum (which this part is) that can cause the stock to melt onto the cutter.  In addition, on one side of the slot, the action of the cutter is to pull itself into the stock while on the other side, it can push itself out of the stock.  This can set up a vibration in the cutter and the mill, depending on how rigid the machine is.  

As usual, there's a possible trick to get around this: cut several thin layers and not one pass to full depth.  It turns out it's simple to change the single pass file using a plain old ASCII text editor.  Simply change the depth of the pass from full depth to what you want (in this case it was change a single 0.220" deep cut to seven consecutive .031" passes) using block copy and paste in the tool path file, then editing every block's line where the cutter drops vertically to the right new depth. 

It turned out even doing those thinner cuts I had a vibration and resonance issue that caused my machine to vibrate so badly that a motor fuse shook loose and shut the spindle motor off.  It didn't mess up the position badly, though, and using cutting oil instead of water mist helped.  

Both rough passes: contour on the left and the waterline for three layers on the right.  For the fine paths, I went back to one contour pass at full depth, just trimming away the margin around the part. That made the cut in the stock a little wider and had no vibration.  For the waterline file, I just trimmed the margin around the part.

This went smoothly and by Tuesday the second side was done.  Time to soak the epoxy in my shop toaster oven.  It broke apart easily.

Those yellow chunks in the front left are the epoxy, and the leftover, machined metal is in the back. The first side machined is facing you.  The dimensions I've checked are all within a reasonable tolerance of the intended size (.005" or less), and while a long way from done, if not for the machine screw-up on the first pass, it would be usable. The machine screw up is visible in this photo, at the lower right end.

Another view from the second (back) side.  It's about .075 extra cut away along that one (top) edge.

I've done only one other piston connecting rod; the one on my Webster engine.  I did that one a bit differently - more conventionally.  I held the blank rod onto a tooling plate with two #10-32 screws and ran contour cuts around the rod.  Then, to cut the equivalent to the circular bosses on the 1 by 1's conn rod, I cut circles around the ends with a few lines of G-code I wrote.  For grins, here are the two rods, side by side, Webster in white.

Looking at the whole process of cutting this practice connecting rod, I'm not sure that doing the epoxy trick really bought me anything.  I'm thinking of going back to the approach I used on the Webster's rod.  The advantage to doing it with the epoxy method is I've already debugged the tool paths.  That should speed things up considerably.   No matter what else, there are a couple of fiddly little operations that need to be done on the rod after all this so those will be done the first time on the real one.