Author: Gerrit Coetzee

A Workshop in an SF Studio

“I should just build a floor on my floor!” I joked to my friend. “Yea, I could just put some plywood on top of the carpet. Then it would be safe from metal shavings and other workshop mess.”

We laughed at the ridiculous idea.

However, my subconscious wasn’t laughing. “build a floor on my floor…. build a floor on my floor” it echoed in the depths. It waited.

A long weekend. 

I was puzzling out how I should enjoy a long weekend. I could call up my friends and see what they’re up to. I could go enjoy the great American outdoors. Or… I could build a floor on my floor.

It didn’t occur to me how much my subconscious was in control as I took out my trusty laser distance finder and started measuring the space I would soon carve out for my new workshop.

I know that in San Francisco all things are fleeting. The grand start-up today could be a whisper tomorrow. So I decided to design the work space so that it could one day move into a garage. Maybe it could move into a larger room. Mostly likely it could easily move into a storage container for safe keeping while I slept, unemployed, in my car; waiting for the glacial hiring processes in the area to run their course.

You might be wondering what plans by a mind such as mine look like. What does raw unadulterated genius look like on paper? Maybe the answer is out there, but my scribbles looked like this:

Even if it’s a bad plan, you should still have one.

I had a wish list:

  1. It needs a surface that an office chair could roll on.
  2. It needs surface that was cleanable and repairable.
  3. It shouldn’t look like total garbage (on Instagram, in person is fine).
  4. It should be rock solid (no one likes a wobbly work desk)
  5. It should be easy to disassemble.
  6. It needs a small standing work area for fine/assembly work.

It was enough to run to the Home Despot and buy some materials.

Next I laid out my prizes in their future location. For the first time the floor floor was taking definite shape. The madness within was bleeding into reality and I was too enthralled and unaware to question or stop it.

Hey… The measurements for the floor floor even lined up!

Since I have zero space and live in an overpriced shoe box with a window, I had but one choice. Paint the parts on the floor floor indoors. It was a nearly perfect plan, except that I had to sleep with all my windows open to avoid death from the fumes. I also went ahead and painted the expensive metal brackets I would use to hold the wood together.

Despite my best efforts, it all went well. I didn’t even permanently stain my apartment’s carpet. My deposit was safe.

Assembly came next. This part was a decent amount of fun, which surprised me. I’ve had experience working on projects with myself and I usually have to schedule in some self-loathing at this stage for some tedious mistake.

Next came the most rewarding part. The completion of the floor.

“The math checks out, trust me.” – A guy who didn’t do any math.

I used two strips of extra wide gorilla tape to hold the boards together, it’s been about three months and this is going strong. Since the joins will pretty much only ever see shear forces I don’t expect this to fail. Also note the 3D printed spacers on the edge that touches the wall. This has multiple purposes. One is simply to keep the wood from scuffing the baseboard. The other is one of ventilation. The only way I can see this whole project going sideways is to trap moisture underneath the floor floor and have to pay for replacing a moldy square of disappointment. This gives the floor underneath a better chance of survival.

Then came a really enjoyable hour of tiling. I had no idea using peel and stick tiles would be a relaxing activity, but they were. This not only covered all my mistakes, but it makes for a really nice desk surface. Plus if I ever burn the top with my soldering iron or otherwise ruin it. I can just peel up those tiles and place new. Great!

Looking fancy!

Of course I wasn’t done yet. I had more problems to solve. Seating was taken care of by getting an absolute steal of a deal on a Herman Miller chair. Storage was taken care of with a few trips to amazon and some surrounding box stores. I also had lighting to think of. I’m definitely a day creature and require surgical lighting to work under. I guess vampires are out of fashion anyway. I eventually solved it with some very bright LED bulbs from the Despot of Homes. It’s even controllable through my internet of home surveillance assistant.

So, the final result?

Aww yea, look at that quality work space.

Of course, there’s still a few items on my todo list:

  1. Organization: I have too many things. I need to throw some away. I’m also working on an over engineered custom inventory system, but that’s a story for another day.
  2. Wires. Too many dang wires. I don’t like how messy it is when you sit in my “living room”. I also want to move my laptop under the desk when it’s docked. I’m working on a nifty swing out table that lives under my desk for that. Watch out for it in a future episode of me solving the problems I made for myself.
  3. Robot vacuum can’t work here. I have an adorable robotic assistant named Cliston which refuses to learn the trick of climbing up a 3/4″ gap. I’m thinking of 3D printing him a ramp which I can place along the perimeter.

Any way, here’s a tour of some of the features:

“How much space do I have left in my apartment for anything else you ask?” That’s a great question!

Thanks for reading!

The Unlikely Long Life of My AKG-Y50s

2 Years and 10 Months ago I was gifted a pair of AKG Y50s by my father for Christmas. I love these headphones. I’ve worn them daily, sometimes for hours at a time. I can say for certain:They were not meant to last this long. Yet, thanks to a lifetime of fixing things, they’re still going strong: To this date I have:

  1. Bought a carrying case for the headphones
  2. Replaced the cable
  3. Replaced the cable
  4. Replaced the cable
  5. Repaired the right ear cup (it fell off).
  6. Repaired the left ear cup (it fell off).
  7. Repaired the right ear cup again (the screw loosened and it fell off).
  8. Replaced both ear pads. 

The first thing I noted about half a year into the ownership of these headphones is that the nice neoprene carrying case that came with them was completely incapable of protecting them properly. So I bought a hard shell headphone box from Amazon, which, as a bonus, had a perfect slot to carry my graphing calculator. What can I say? I have specific needs.

Next, the cable that it came with was no longer working. So I contacted AKG. My first shock was that, for a 100 dollar pair of headphones, they wanted 30 dollars + another 12.00 for shipping to send me a replacement cable. This was unacceptable:

So I did what any self-respecting modern person would do. I searched Amazon for a replacement cable. They had one! and for $12.00, shipping included. The off-brand cable worked just as well as the Y50’s original, and no one could tell the difference in the microphone quality.Then,half a year later I replaced the cable again. This third one has continued to work.

Then I had a new problem. The ear cup fell off! It would just forlornly dangle from its cord. I tried to fix it a few times by carefully rerouting the wire the way AKG intended and re-tightening the screw. However, despite loctite and other measures, it would inevitably work its way loose. It ends up the threads themselves wore from regular use. So I broke out a drill and a soldering iron. I routed the wire through the top of the ear cup and put a new self-tapping screw in the side which held much better.

This worked for a while and then the left ear cup had the same issue. Fortunately my fix/redesign has held up. So far the screw has only come loose once, but I was able to simply tighten it and it returned to normal operation. I suppose there is a chance I messed with the tuning of the headphones this way, but as far as my untrained ear goes, they still sound like a mid-range-decently-good-but-not-amazing-you-won’t-be-disappointed set of headphones. 

Most recently, after nearly three years the ear pads had started to wear down and the foam was poking through. Much to my surprise I could get another set on eBay for 13 dollars. They even had the silk printing on the inside of the cups for left and right!


  1. Brand Name and Popularity: If I had purchased a less popular set of headphones from a different brand I might not have had access to the same level of repair parts. Simply typing in Y50 got me access to all the replacement cables and ear pads I could want.
  2. Protect your items: Honestly, I should have had a hard case for these from the beginning. I think there’s a chance that I could have at least made it to year two without the cups falling off had I done that.
  3. If it’s already broke, how bad can you fuck it up? Last is just a general rule of repair. Take your time, think through it, but if it’s between a new one and fixing it: Don’t be afraid to bust out the soldering iron and drill. 

All in all these were a great pair of headphones. If it wasn’t for the ear cups falling off, I’d rate this an extremely durable pair of headphones. I consider ear pads and cables to be reasonable wear items; though I was disappointed at the high cost of ordering replacement wear items from AKG. I think this is a failure on their part.  

The thing that bothers me about the ear cups falling off is that it really seemed to be something that could show up in testing. Everything else worked great and is still working within a reasonable lifespan. The headband adjusts. The joints tilt and move like they’re supposed to. It’s just this one joint that failed, and it failed the same way multiple times in a row. I’d have felt really betrayed if I’d had to order a new set of headphones a year into ownership. We shouldn’t be making trash to fill our planet with garbage.

I think it should, at some point, become the ethical and moral obligation of engineers and hardware companies to not produce trash. I also think it’s our responsibility as consumers of objects which consume vast supply chains, enterprises, and companies to try and maintain and repair these miracles for as long as we can. 

Anyway, if you’re handy, buy these headphones. They’re not bad. Though, since my Pixel 2 needs a dongle, I think my next pair will be wireless

Open Prony Part 1: Initial Design and Theory

I have a problem with buying motors for mass manufacture; it’s a problem of trust. Take, for example, the Solarbotics GM3; every motor manufacturer in China has a copy of this motor collecting dust in their catalog. Each and every one of them claims that their motor is stronger, faster, etc. I need to decide if I want to order thousands of these motors and I have no easy way to check their performance.
I’ve observed that this motor has played a form of the telephone game for mechanical engineering. A great example of this is the compliant clutch. The original version of this motor has a little two part clutch in it that can flex out of the way when the load is exceeded. Most of the Chinese copies of this motor still have a two part gear, they’ve just disabled the clutch entirely. My best guess is that at some point a manufacturer made a quick tooling change for a customer to disable the clutch. Then another manufacturer got one to copy, and copied the change without understanding its purpose.

The datasheet still claims that this feature works, of course.

Anyway, I started looking around online for a way to measure the motor’s output. It’s a small motor, but I figured someone has solved this problem. Nothing exactly right. So I got to looking a bit more and discovered a measurement tool called a Prony Brake.  This seemed interesting, but I couldn’t imagine a practical way to get it to work on such small motors. Then I stumbled across a variation called the rope brake dynamometer. This, I felt, was something I could work with.

Rope Brake

In a normal rope brake dynamometer you’ve got a rope around a wheel. On one end you have a weight and on the other a spring scale. As the motor changes speed you can see the force on the spring scale increase or decrease depending on which side of the rotation you have it on. Then it’s a formula away from having the power curve.
I wanted a way to increase the “mass” of the weight so I could stall the motor and see the full torque curve. I also wanted to be able to arbitrarily set the mass depending on the motor, to do this I came up with this configuration:
My plan is to wrap a small thread around a known spindle size. On each end of the thread I’ll have an extension spring. As I expand the springs, I’ll measure the force at each end. Then it’s a matter of subtracting the readings to get my torque. After that it’s only an angular velocity measurement away from getting the power output of the motor.

To measure the force produced I settled on some of these generic load cells from Sparkfun. I thought about ordering cheaper ones from eBay or Amazon, but I figure that if one is designing measuring equipment you might as well start with something that has a little QC behind it.  They also sell a breakout board for an interesting chip called HX711. This takes all the headache out of interpreting the output from the load cell and should generally make things a bit more repeatable.
On top of that I’ve got a teensy on order, some stepper drivers in a box somewhere, and a motor driver likewise. Hopefully I can cobble together the electronics and most of the mechanics from 3D printer discards.
With the electronics worked out I started to design the perfect mechanical assembly. It would be a nice precision construction; made out of metal. It would move beautifully, mechanical engineers would admire it for years. It would cost hundreds—- wait, maybe I should check if this whole idea actually works first. 
So I paused that design and forced myself to make a really crap version out of plastic first. Rather than being able to support any arbitrary motor, this one would be designed for exactly one motor. All I need is something to test. Which is where I am today.

So my next steps are simple in concept. I’ve ordered all the electronics. I’ve started my 3D print. I’ll assemble it all, wire it up, write some code, and see if it works. If I can generate some curves I’ll move on to some fancier design work.
My eventual goal is an open source design that can measure just about any small motor.
Next time on Part 2: Does it work?

Hackaday OC: Resin Casting Tutorial and Filament Thickness Sensors

The first one, which took an absolutely enormous amount of time, is a tutorial on resin casting. This will be part of an ongoing series I’m working on for quick prototyping using these methods. In this turtorial I duplicate some buttons to keep my favorite coat in service for another season.
The next piece is some research on filament thickness sensors, an interesting development in the 3d printing community.

Hackaday OC: 3d Printing Temperatures and Modeling in Fusion 360

I wrote two new articles for hackaday. The first is, “Keep Your Nozzle Hot and Your Prints Cool.” This is about my experince in tuning up the quality of my 3d prints by adding sufficient cooling. It really helped!
The next took me quite a while to write. I really like Fusion 360 as a mid-range (for now) alternative to Solidworks. This is a short tutorial on using some of its parametric cad features. Making Parametric Models in Fusion 360

Hackaday OC: What's in a Tool and Adventures in Resin Casting.

I’ve written two new articles for Hackaday. The first is “What’s in a Tool: A Case for Made in USA”. This article made front page of slashdot, has over 50,000 views, and nearly 500 combined comments. I’m very excited about the good reception. Read it on Hackaday.
The second is an article about my adventures in resin casting. Unfortunately it’s a bout a learning experience I had with different sorts of mold release for silicone molds. Also on Hackaday.