Intex Air Mattress

Last summer I ordered this mattress off of the supplier of all things reliable. For those of you not clicking links today, it’s a full size air mattress with a “pillow” on one end, and an integrated fill pump.

Air mattresses are constructed with ribs inside them so that they hold their mattress-y shape. Without these ribs, they turn into a bag of vinyl smelling air.

About two months into sleeping nightly on it, the leftmost rib ruptured. After the feelings of terror and paranoia subsided, this was actually and improvement to the design of the mattress. It made it slightly squishier.

Fast forward to this summer (6 months in storage, in a van, through cold and heat). About two months in, two of the center ribs ruptured. This caused the mattress to have a W shaped profile, which is no good at all.

So I cut it open!

First: one cut parallel to the ribs along the length of the mattress. Here we see two ribs. The first one is ruptured, and the remainder of it can be seen on the upper part that I’m holding. The second one is in good condition, and we can see how it is heat bonded to both the top and bottom layers of the mattress. Note also the holes through the rib.

rib fail parallel

rib fail parallel


Continue reading

Rear Hub Failure

For the search engines: 2000 Toyota Sienna ABS light Tire Pressure light

For the humans:

On the drive back from college last summer, the left rear rotation sense hub cover fell off of the hub. This triggered the ABS and Tire Pressure lights to go on, and a mad scramble to attempt to Do The Right Thing and replace the hub in its entirety ensued.

It turns out that expoxying the cover back in place is much easier and more economical than attempting to remove the hub from the frame of the van. JB Weld is your friend.


hub cover pair

hub cover (interior / exterior)

I don’t have pictures of this, but the left side failed in this manner: the cover began to rotate with the wheel instead of staying in place which twisted up the sensor wiring harness until it broke. Due to the magnetic strength of the sensor, the cover stayed in place. The total dollar cost to replace it was about $20 for a new harness from the local auto parts store, but it took about three afternoons of messing with it to get there.

So when the ABS and Tire Pressure lights came on one rainy morning, I knew what it was. Full of confidence, I reached under the van and wiggled the cover only to find it still securely in place.

hub interior


From the center outward: the hub (see the fins? they work with the notches inside the cap as part of the rotation sensor), bearing race cover, a slightly inset ring from with the cover broke off ( apply expoxy here), the outer edge of the hub, and then the plate that aligns it all on this side of the frame (if I recall correctly). Note that this picture is taken from underneath the van, looking outward and at what would be the inside of the wheel.

Turns out the cover on the right side had broken off. Slightly different failure- the cover came off and fell off, but was retained by the sensor cabling which was undamaged by the excitement. I applied the same solution as before: knock off some of the rust, apply epoxy carefully, squish in place, turn a little to make sure it’s not jammed, and wait.

It is very important to make sure the wheel rotates freely of the cover- if it is possible to jack up the corner of the van being worked on safely, I recommend it so you can just spin the wheel to determine that it moves freely. Else, practice aligning the part before apply epoxy so your hand knows how it should feel, and twist it slightly into place to make sure the fins do not engage the notches in the cover.

hub cover in place

cover in place

The slight rotation off vertical is correct. I don’t know that it means anything, but both the right and the left covers were rotated off vertical from the factory, so I attempted to recreate that alignment when fixing the covers. Also, you can see here that the wiring harness (white and blue parts) have a tab the lifts off of the cover to release it.

Update: Industrial Velcro

My hope that all of the velcro adhesive would melt out and leave a one molecule thick layer that strongly adheres the velcro to the vinyl without leaking further is failed.

Instead, the velcro lifts off with no effort. Especially when sun-hot.

Interestingly, only the velcro in the driver’s and front passenger areas melted. The velcro placed along the edges of the sliding door windows has not melted, and remains in place as it should. I believe the stronger tint on the sliding door windows, as compared to the windshield and forward windows, has protected the velcro.

I suppose that if it couldn’t be removed with force or chemistry, using a hair dryer to warm it until it peels easily, and then using goof-off to wipe up the remaining adhesive would work really well.

Aside: the driver’s side door lock mechanism jammed briefly. The lock tab did not disengage the lock completely, and when the interior button was pushed to open the door, it jammed in, causing the door to be locked shut. Jiggling the outside handle did not effect the situation. The gestalt of the Sienna forums suggested: percussive maintenance, disassembly of the door, a shot of silicon lubricant to the mechanism, in that order. Percussive maintenance worked after letting it bake in the sun for about four hours, which will absolutely solve the problem forever.

Note to self: shot of lubricant if you’ve ever got the door taken apart


Industrial Velcro

It said, right there on the box, “do not use in direct sunlight”.

“It can’t be that bad.” I said to myself, “What could go wrong?”

It turns out that the adhesive on industrial self-stick velcro liquefies when sufficient heat is applied.


Oh dear.

The adhesive only loses it’s neat rectangular shape- all of the stickiness is retained.

Bus Route Complication

My school runs a shuttle bus service through all of the major off campus (but very near) apartment complexes. In prior years, the busing was contracted out to RTS, the regional public transit service. These were full size city buses: 4-5 regular buses in service at any time and one bendy bus as well.

Two of the routes have particularly difficult corners for a 40′ bus to make. One is illustrated below.


Perkins gate entrance

The route proceeds from the bottom right corner, makes a right turn between the sign and the curb, and then an immediate left turn into the complex.

The RTS drivers handled this turn with aplomb, and occasionally flair. Shoutout to the driver I saw drift an evening bus through that turn in the snow.

The service is now contracted out to First Transit. First Transit is running approximately the same distribution of buses that RTS was, although the buses now wear RIT livery, instead of RTS blue and advertising.

The First Transit drivers do not handle that turn with aplomb.

About a month back, a bus got stuck making the right turn. Single section buses are rear axle drive: the right wheel pair rode up on the curb, and then dug into the mud on the grassy side of the curb. The bus was unable to drag the wheel pair back over the curb. It’s hard to tell because of the water, but the depression is about 8-12 inches deep, which is about the depth of the tire on a bus (about one third of the total radius of the wheel).

cut left

going into the turn

cut middle

halfway through the turn

cut right

and back over the curb!

There is more than one bus with the right rear completely covered in mud which indicates that 1) buses change routes and 2) this has happened more than once.

Notice that the curb is shattering.

curb detail

curb detail

That is presumably the suspension of the bus dragging back over the curb.

There’s now a stake placed in the corner to enforce the bus not going over the curb. The drivers now creep around the corner. It’s probably very stressful.

There’s also a pretty intense mud+gravel smear going on there now.

mud drag


The articulated bus seems to have less trouble making the first turn, instead clipping the embankment with the nose in the immediate left turn.

additional scrape

additional scrape

Those poor buses, that poor tree.

Posit that only one bus runs this route (despite evidence to the contrary). 3 mile route x 4 trips per hour x 10 hours per day x 5 days per week x 16 weeks per semester x 2 semesters per year equals 19200 miles per year, and 6400 trips. This discounts all of the weekend, evening, intercession, and summer term routes. Three or four failures isn’t bad but it’s still a problem.

A standard city bus has an expected lifespan of 12 years or 500000 miles, with at least one complete service/replacement/rebuild of brakes, suspension, engine, etc during that time. There have been new developments in low floor buses, for accessibility and passenger comfort reasons. I think RIT/First Transit is running that style of bus. However, the Wikipedia for articulated buses reports that the failure mode is

 relatively rapid disintegration of the vehicle’s superstructure.

due to not having enough suspension for uneven roadway. It seems reasonable that single section buses could experience a similar issue.

Oh boy.

RIT’s planning horizon is 10 years. This one hypothetical bus should last 26 years by optimistic mileage.

Maybe the curb isn’t a problem because the bus will outlive expectations simply by being low mileage.

Maybe the curb should be moved to facilitate the turn so the turn is easier, the buses don’t experience the wear (from going over the curb or creeping very slowly around the corners at the limits of their turning radius), the roadway infrastructure doesn’t experience this abuse, and the drivers are less stressed.

I propose a new curve to the curb: about 6 inches inside where the tracks in the mud are.

(If you want to know about the mileage buses receive, this article and this article have information. RIT is running some eco-friendly bus type, but I can’t find what specific kind it is.)

Squeaky Chair

In my classes, I’m learning about meeting the needs of the customer. It’s a primary objective, the One True Goal in lean systems.So now I look around me and can ask what needs was this thing designed to meet? and what needs is it meeting? and the most important question what needs is it not meeting, but could be? That’s how a customer is truly made happy.

The way the system is taught, there is the SIPOC

  • Supplier
  • Input
  • Process
  • Output
  • Customer

and the customer is typically not the average consumer, but the entity that sells you or me a thing or service.

Which is how I come to have a squeaky chair.

The chair was designed to meet certain specifications: material of the seat, a certain height, durability, to look a certain way, to cost a certain amount.

It’s got

  • black fake leather seat made of two sections that are (probably) glued together. The bottom section has an internal board so that the bolts have a place to secure and a layer of plasticky fabric covering the board.
  • black fake leather back rest made in a similar style as the seat
  • a single piece bent tube leg/rocker assembly
  • bolts to secure the seat pieces to the frame

Most processes are supposed to have an evaluation phase- is the produced thing the thing we thought we were producing? and does this thing actually meet the needs it was designed for?

This chair probably satisfactorily answers both of those questions, if they were asked at all.

It absolutely fails to answer does this thing meet the needs well?

The seat squeaks. The plasticky fabric shifts against the steel(?) aluminum(?) frame when the user sits on the chair, or moves around on the chair. There’s three things the user can do

  • ignore it
  • tighten the bolts
  • place a strip of non-squeaky fabric between the frame and the seat

I took the second two options: solve the problem in such a way that if it tries to come back, it can’t.

The blue strip is the bit of fabric peeking out.

The blue strip is the fabric peeking out.

There’s three customers in this situation, so the whole system looks like this

[Supplier]         -> [Customer 1]                -> [Customer 2]                      -> [Customer 3]
[manufacturer] -> [furniture supply store] -> [student housing complex] -> [the user]

I won’t ever know who generated the specs for the chair (probably the manufacturer), or if they ever tested a chair.

But I do know that my chair is not the only one to squeak egregiously and that means there’s something wrong in the production process.





My Transmission Lines professor jokingly mentioned that it is possible to build your own capacitor. Little did he know…

My friend and I used this post for guidance and settled on a aluminum foil and waxed paper rolled capacitor. Based on the values in the post, we made ours about 15 feet long, about the usable length of floor space in my living room. On the first attempt, we tried to roll the entire length. After the layers became misaligned to the extent that the plates would have touched, we unrolled it and tried again with a new method. We folded it in half, with an extra layer of waxed paper to insulate the inner plate from itself. Then we rolled it up, starting from the folded end. This worked much better.

folding order

Folding diagram

srsly tho

srsly tho

On the PVC core, the entire thing is about 1.5 feet, with the actual capacitor body being about 12 inches long.


Plot of the capacitor displaying capacitor like behavior

Using a capacitance meter, it was determined that it is a capacitance of 290nF.

Monster Dolls: Additions to the Family

Since my last post about the monster dolls, I’ve made three more and have started giving them to friends.

still can't hold their heads up

A delightful little family

The two on the left are “twins”- they have the same style of hair, horns, flower crowns, face design, and the same number of extra arms and legs. Their dresses are made from old arm warmers and socks. The mermaid is based on the pattern that I got from TheDenTree, but I changed the body to flow seamlessly into a mermaid tail. The pattern is otherwise unchanged and the tail pattern is available here. The scales are embroidered on.


They’re so happy!

The flower crowns are based on this flower crown tutorial. I discovered in the process that some glittery and textured nail polished will also work but didn’t use any of those flowers here.

note the paired horns

Note the paired horns and the confused expression

The mermaid got a new style of horns. Each “horn” has two separate horns in it, the forward one being much smaller than the back one. The yellow stripes were created by cutting the horns into layers and splicing in a very thin layer of yellow.

Cabin Fever 2k15: The Adventure

For those of you who don’t know: Cabin Fever. I went with my dad a couple of times in high school. This time I dragged my friend from this post to it. He likes tiny engines and is willing to drive, it doesn’t get better than that.

My friend drives a little shtbox with amazing gas mileage and some personality quirks. By the time we arrived in York, we’d: taped the windows on the drivers side shut to reduce wind noise, fixed the radiator, and discovered that the hex wrench used to open the windows was actually missing.


oh dear

We saw a guy with the best job ever:

Everything I ever wanted out of toy trains

Everything I ever wanted out of toy trains…

It sounded just like a tiny train should, and there was another guy giving tiny train rides.

Continue reading

Multi-strand Necklace

It’s occurred to me that I can’t wear my scarf the entire year and so I needed to step up in the neckwear department. A bit of research showed multistrand necklaces and certain types of statement necklaces looking good with collared shirts so I decided to make a multi-strand necklace first.

This helpful chart suggested that the shortest strand should be about 18 inches and the longest strand about 22 inches. I checked that with some yarn that I cut to be the lengths I wanted the necklace to hang and it agreed. About 60% of the strands are regular black seed beads, the rest are the iridescent black kind. All of the strands have some of the other kind of bead in them. The larger beads are used to set off the strands a little bit.

goodness I am pale



Toggle clasp

I made each strand individually on doubled over beading thread. Each end of a strand is tied and glued to a bead that fits onto a jump ring. There are two clusters of strands, about five on a jump ring. The jump rings are joined to the clasp with another jump ring.

And somehow I still wound up with more beads in my stash than when I started.