Well I found where the oil leak is coming from. I must of misaligned the rear main seal when installing it, even with the help of the 3D printed tool that I made.
My goal is to replace the seal and reinstall the transmission in the car this weekend. The leaves are about to change colors here in Michigan, making a great backdrop for some photos.
This interview with the inventor of the inerter is fascinating. Mechanical engineering has been mostly figured out, we just remix designs and inventions discovered hundreds of years ago. The inerter is an exception to this rule, a novel invention that still hasn’t broken out of niche use cases like Formula One suspension systems.
When asked if I’m a motorsports fan, my common response is that I’m more into the “motor” than the “sport”. I’d much rather spend my sundays learning about the engineering of the vehicles than watching the cars go in circles.
Last Thursday was supposed to be a huge milestone in the restoration of my 1995 Mercedes-Benz SL500. After a year-long process of removing, rebuilding, and reinstalling the transmission, the car was finally ready to drive. I took a glorious maiden voyage and snapped the picture at the beginning of this post. However, there are a few problems.
The first problem is that at higher throttle positions the transmission isn’t downshifting as expected. I think I can adjust the throttle linkage to solve this, but the brittle plastic adjuster screw is cracked. Otherwise, the transmission is working wonderfully: shift quality rivals modern automatic transmissions, all gears engage, and reverse now engages once the car is warmed up.
The second problem is a pretty significant oil leak from the bell housing between the transmission and the engine. It’s almost certainly due to a mistake I made when replacing the rear main seal of the engine - a job that wasn’t even necessary, but one that I did “while I was in there.” To fix this, I have to pull the transmission again.
The last problem is more annoying than serious. Before my first drive, I queued up my favorite yacht rock, but when I turned on the head unit, the speakers only output a high-pitched whine at full volume. Somehow, over the past year, the amplifier died even though it wasn’t even connected to power or being used. The same company that rebuilt my head unit and added bluetooth also repairs these amplifiers, but it isn’t cheap, so I might try to DIY this repair.
The adage of “there’s nothing more expensive than a cheap German car” is ringing true. I bought the car last February for $5,000 and I’ve spent an additional $9,000 on maintenance, taxes, and registration. The only time I have paid for labor on this vehicle was for an alignment. I don’t want to imagine what a shop would charge for the work I’ve done. Here is an up-to-date detailed list of the costs for this project.
There is an argument that I should have just spent double and bought a better example of this model and I would’ve ended up with a better car. But that ignores the less tangible benefits of this endeavor. First, I am preserving a car that might otherwise have been discarded or neglected. Second, I am learning a ton. I work professionally as an Automotive Engineer, and projects like these are essential to intimately understand how a car works and also helps me empathize with mechanics when designing systems and components.
If I ever want to quit renting and become a homeowner, I’ll need to rein in spending on this project. So until I buy a house, I’ll only do what’s absolutely necessary to drive this car without causing an environmental hazard. My goal is to have the transmission out by the end of this weekend and reinstalled by the middle of next month.
Not having all the information you need is never a satisfactory excuse for not starting the analysis.
Making the world’s fastest RC car is a lofty goal, and it would be embarrassing to build a car based on vibes that comes nowhere close to breaking the record. To drive confidence in this idea, I decided to make a program that would simulate a straight-line top speed run given basic vehicle parameters. This program is written in Python and utilizes a basic road loads model combined with a fan model based upon momentum theory.
Some caveats before we begin. Firstly, the adage of “all simulations are wrong, some are useful” definitely applies here. This simulator makes a ton of assumptions and relies upon limited information that is given for most off the shelf fans. As design is an iterative process, I can always go back and refine this simulator with any testing results I get later on. Secondly, the vehicle parameters input in the model are based upon using a pair of 90mm FMS EDF units, two Aerotech G80-T rocket motors, and a whole lot of guessing for everything else. As I get further into the design process I will refine these parameters to get a better idea of how the car will actually perform.
With a top speed of almost 500 kph (310 mph) this design beats my 400 kph (250 mph) target top speed comfortably without the use of rocket motors. Unfortunately, it takes about 3 kilometers (1.9 miles) to get there. Finding a road of this length will be a challenge, but even at a mile I’m comfortably above my goal. Let’s call this cushion a “safety factor”.
If my previous safety factor isn’t enough, this surely will be enough with a top speed of 635 kph (395 mph). Figuring out when to fire the rockets isn’t immediately obvious, so I also added a tool that will find the optimal time to light the rockets that maximizes the top speed of the vehicle. In reality, I will probably be space limited, so I will need to optimize around finding the firing time that maximizes the top speed within a set distance.
This comparison tool will be very helpful in design to understand the trade-offs between different components and designs on the performance of the vehicle. This example shows that a 10 percent improvement in mass, frontal area, and drag coefficient will increase the top speed by a little over 5 percent.
Road loads are simply all of the forces acting upon a vehicle in motion. The main loads that need to be modeled are the propulsion system loads, aerodynamic loads, drivetrain loads, and gravitational loads. Road load models are the first step automotive engineers use to size powertrains, estimate energy usage (fuel economy, range, etc.), and to estimate basic performance attributes. I strongly believe that every automotive journalist should have to pass a test over the basic relationships of road loads and how they affect vehicle performance, I’m really tired of reading borderline misinformation in large car magazines.
Here are the road load relationships that I believe are the most critical to understand:
I’ve been down quite the rabbit hole when it comes to modeling fans. My initial confidence in embarking on this project stemmed from having a poor understanding of how fans work. Initially, I assumed that the quoted “static” thrust of the fans would be constant across all inlet velocities. This isn’t true and I ended up with napkin math telling me that I could easily break the sound barrier with a big enough fan. When I realized that air is stuff again, I was forced to reopen the text books from my least favorite classes.
I landed on using momentum theory and a (probably wrong) assumption that air power will be constant. I didn’t model fan rpm variation as a function of vehicle speed, transient behavior, and I assumed perfect ducting.
I’m a Mechanical Engineer by “training”, which while a pretty well rounded discipline it did not teach me how to do the vast majority of the programming required to do this project. So I used this project (and this website) as an opportunity to explore the new LLM based coding tools that are now available. For this project I used Visual Studio Code with the Github Copilot extensions to help develop this project. Without these tools, this simulator would be more of a command line based “script” than a “program” with a GUI. LLM caveats definitely apply here, you can’t blindly trust the outputs.
I’m well aware of the societal and environmental downsides of these technologies and I honestly don’t have a good excuse to justify why I use them in spite of these downsides. But I also don’t have a good excuse for eating meat or taking long hot showers.
I’ve spent probably way too long developing this simulator, but I learned a ton and I don’t feel like it was a complete waste of time. If anyone wants to run the program themselves, shoot me an email ([email protected]) and I’ll add you to the GitHub repo. I’m hesitant to make it completely public, because I want to make sure its accurate and a little more user friendly.
My next post will be about the component selection process for this car and also some sketches of what the car looks like in my head.
From Mercedes Benz media:
On the legendary high-speed test track in Nardò, Italy – a site where the world’s fastest cars have been tested for decades – the CONCEPT AMG GT XX broke a total of 25 long-distance records. Among them was the record for the greatest distance ever covered by an electric vehicle within 24 hours. The previous record was under 4,000 kilometers. The CONCEPT AMG GT XX has far surpassed this value: The technology platform traveled 5,479 kilometers in 24 hours – that is 1,518 kilometers more or around 38 percent further than the previous best.
Average speed of 228 kph or 141 mph for my fellow Americans, and that’s including charging stops!
In the early morning of 25 August 2025, the moment finally arrived: mission accomplished! The distance of 40,075 kilometres (24,901 miles) was covered in exactly 7 days, 13 hours, 24 minutes and 07 seconds – beating its own target of eight days by a wide margin.
Such an insane demonstration of engineering prowess and proves that the future is bright when it comes to battery electric vehicle technology. It’s also nice to see this coming from a western automaker, when the prevailing media narrative is that Chinese EV technology is insurmountably ahead.
Still waiting on some of the Oldsmobile Aerotech speed records to be beaten…
Every good design project needs a mood board, here’s what is currently inspiring me as I begin designing:
Last year, the lease on my wife’s car was ending, and I saw it as a great opportunity to get a new (to me) daily driver while she borrowed my 2017 VW Golf R during her final stretch of school. Instead of choosing a sensible car, I convinced myself that a $5,000, 30-year-old German convertible would be a perfectly reliable mode of transportation. That might have been true—if I had bought the right car. Instead, I “saved” $10,000 by picking one in the perfect color (its color scheme actually inspired this website), but it came with no maintenance history and obvious must-fix issues. My naive optimism led me to believe that the premium for a lower-mileage car with service records wasn’t worth the five-figure difference.
Despite its quirks, I have no regrets about buying this car, and I’m determined to preserve it for decades to come. The R129 SL represents a golden era of Mercedes-Benz, when engineering and craftsmanship were at the forefront. With an original MSRP of $90,000—equivalent to nearly $200,000 today—it stood as the brand’s flagship model. Even now, it offers impressive performance, timeless design, and a driving experience that feels genuinely elevated.
In this post, I’m going to document where I’m currently at in “restoring” this car and what I plan to do in the future.
When I bought the car there were a few big ticket items that I knew needed addressing, and some more minor maintenance stuff that I did preventatively. I spent about 3 months last year from February until early May doing these tasks in the evenings and on the weekend. I’m going to do a follow-up post discussing the costs of everything that I’ve done.
The Engine Wiring Harnesses — In the early nineties Mercedes switched the jacketing material on some of their wiring harnesses to one that over time will plasticize and crumble. The common story for the reasoning for this material change is that they wanted something “bio-degradable,” but after working in the automotive industry for the past decade I’m skeptical of this story. I think it’s much more likely that the material change was driven by other factors and they were ignorant to the long term durability issues this change made. None of this changes the fact that two of the main engine wiring harnesses on this vehicle were crumbling and needed to be fixed. The more complex of the two called the “upper harness” can be purchased and the installation only takes an afternoon’s worth of work. The less complex of the two called the “lower harness” is not available anymore so I had to remake this harness from scratch. Luckily this lower harness is only a few connectors, but unluckily some of those connectors are very difficult to find and expensive. I remade this harness to as close to “motorsports” standards as I could using Raychem DR-25 heat shrink as the wire covering and MilSpec /32 wires.
The Convertible Top Hydraulics — The R129 was the first convertible vehicle on sale with a fully automatic roof system. This system has twelve hydraulic cylinders and one distribution block that contain seals that will eventually leak if not properly addressed. Last winter I removed all of these components from the vehicle and sent them to Top Hydraulics in Oregon to be rebuilt. While theoretically I could have bought the seals and rebuilt them myself, if I were to damage a single cylinder in that process it would have cost more than the entire cost of getting them rebuilt professionally.
Everything Else — Here is a non-exhaustive list of the other maintenance tasks I did last winter (many of them preventative)
Because there is nothing more expensive than a used German car there is still a laundry list of things that I should address before I do any cosmetic or performance upgrades on the car. I did spend last summer driving the car as my main transportation method which exposed some issues that still need to be fixed.
Merlin Mann, creator of many great shows, has also been writing and collecting a great many sayings that he calls “wisdoms.” Always a great read and related to the previous few posts.
Some highlights:
- Never organize anything you should discard.
- Sometimes, a person will confess something embarrassing that obviously makes them feel really dumb and vulnerable. That is never the time to say “I told you so,” and it is rarely the best moment to offer advice that they never asked for. Just shut the fuck up and listen.
- Buy the nicest screwdrivers you can afford.
- Whenever you meet someone new, ask them what they’re most excited about right now. Everyone interesting is excited about something right now, and they’d probably love to tell you about it.
- Three is two, two is one, and one is none.
- If you’re not sure who’s doing your emotional labor, it’s probably everyone you know.
- Try to become someone who’s fun to teach things to.
- Nearly everything that happens in the world doesn’t involve you. How often do you think about this?
- Sometimes we make things to teach, but often we make things to learn.
Another “Chicken Soup for the Soul”, but this is spacecraft edition.
A few highlights:
- 2 — To design a spacecraft right takes an infinite amount of effort. This is why it’s a good idea to design them to operate when some things are wrong .
- 4 — Your best design efforts will inevitably wind up being useless in the final design. Learn to live with the disappointment.
- 7 — At the start of any design effort, the person who most wants to be team leader is least likely to be capable of it.
- 11 — Sometimes, the fastest way to get to the end is to throw everything out and start over.
- 19 — The odds are greatly against you being immensely smarter than everyone else in the field. If your analysis says your terminal velocity is twice the speed of light, you may have invented warp drive, but the chances are a lot better that you’ve screwed up.
- 23 — The schedule you develop will seem like a complete work of fiction up until the time your customer fires you for not meeting it.
- 36 — Any run-of-the-mill engineer can design something which is elegant. A good engineer designs systems to be efficient. A great engineer designs them to be effective.
I’ve always been a big fan of “Chicken Soup for the Soul” style sayings and these rules are at the intersection of that interest and my obsession with automotive engineering. On my Formula SAE team in college we had these rules hung on the wall above our design computer.
Here are a few highlights from this list:
- 1 — The keywords are elegant simplicity. Each part doing three jobs at least.
- 6 — Murphy is a racing mechanic and works for Team Lotus. If it can be assembled the wrong way around he will
- 16 — Remember, all engineering materials are rubber like. They deflect to some degree under the slightest load and will expand to some degree in the presence of even a lighted match. Everything resonates, even brake pipes.
- 31 — Air in motion is the best heat insulator.
- 33 — You cannot change the basic laws of nature, no one has yet. Let Newton, Charles, Boyle, Bramah and Bernoulli rest in peace.
