April 2021 - Edition 24, Engineering Compromise


I was a little tardy on the previous two newsletters, so this would seem to be an attempt to correct that wrong by being on the earlier side. It’s not. It’s merely an opportunity for me to get something off of my chest based on something I just read.

Demetri's Corner

To quickly relate the happenings in my world, I’m still trying to build a fabrication table as a start for my eventual prototyping business. Details are mostly covered on the YouTube channel here, but the long delay in sticking a few pieces of steel together is mostly not my fault as I’m still waiting for a saw that never got delivered. Really annoying. I’ll probably fill my time making other random household items with my welder as I really need a LOT of practice.

So on to the reason I was motivated to write a newsletter when my end of the month deadline is still barely visible. Just to get it all out in the open, I regularly monitor articles in Jalopnik as I think their writing and topics covered are a refreshingly different approach to most car-based periodicals. I especially enjoy reading things written by Jason Torchinsky (a real name? I guess I can’t be one to judge) because his articles are a distillation of this interesting and different approach towards car editorials.

Just recently I was perusing the latest Jalopnik offerings, and saw an article about making cars more like personal computers (“Consumers Should Demand EVs Built Like PCs Even if Carmakers Don’t Want It). Mr. Torchinsky’s angle is that it would put the power to the consumers, but perhaps at the expense of some of the “fanciness” we’re used to getting in highly integrated cars. I don’t think I’m a good enough writer to steal thunder from Jason Torchinsky, but perhaps I can tap the lightening that caused the thunder and look at this from an engineer’s perspective.

Today's Subject - Engineering Compromise

If you haven’t noticed, there’s a theme to my writings on engineering. That theme is that there is no such thing as perfect and every solution is a compromise. It’s part of why engineering can be frustrating, or exceedingly beautiful (eye of the beholder I guess). I’ve covered in detail the concept of how engineering is constrained mostly by limited resources and how the judgement of the engineer defines the final product. In all of those cases I’ve been speaking as an engineer that works in a field were capability can never be in question – nuclear power plants, medical devices, firefighting, etc. All of these types of designs have one thing in common. The end result of your design has been determined for you, even before you start putting pen to paper (or mouse to screen?).

But consumer equipment is a whole different ballgame. Apple didn’t put out a spec sheet on the iPod to get interest and then build the consumer version to match customer expectations. Instead, they created a product and gave it as much capability as they could within pre-defined constraints of time, technology, and cost and then told the consumers what they were getting. This is probably the first extreme case of letting the tail wag the dog, but it set a precedent on how most consumer electronics are made. How does this relate to the car industry?

The Current Car Industry Model

Cars are a strange mix of the two types of engineering we discussed. The basic car itself – the chassis, drivetrain, safety equipment – is very much defined before an engineer gets started. This looks a lot like “classic” engineering where the minimum equipment requirements are set and trade-offs are really in the cost, and quality of the result. When we started adding options – infotainment systems, moonroofs, umbrellas – this looks a lot more like the iPod model. You don’t really ask for the car to say “Hello” when you get in, but the manufacturer found a way to squeeze in the capability and thought it would make the product more desirable.

What you’ll note is that for a complex mechanical system like a car powered by an internal combustion engine, it would be very difficult to divorce the baseline capability from the form of the car itself. This lends itself to the traditional engineering model where we define the capability up front. With the advent of the modular platforms and EVs in particular, this model is seeing some strain.

Cars as a System of iPods

If we were to take the extreme consumer approach to cars, we would probably just offer cars with minimal options and as a consumer you get what you get, with little possibility of customization. This is very much a Tesla type approach, where you need to go to a Tesla service representative to get new tires. So even though we’re making the case for a more “consumerized” car, the concept of the iPod as a car is likely not one that everyone is going to like (though some Teslastans are perfectly happy being told what they’re going to like).

What we really need is a bunch of iPods hooked together with pre-defined interfaces. This might just be the way to keep car companies engaged. I’m fine with one car company making a battery that has proprietary technology – as long as it still plugs into my Yugo frame. This allows the car companies (and battery, sensor, brake, motor..) to provide that “special experience” and market their products while opening their potential client base to a lot more than people capable and willing to drop the entire price of a car at one go.

What’s the Tradeoff?

Jason Torchinsky covered this in his article, but here’s my attempt at a (poor) summary. The less integrated the car design becomes, the less efficient the design. The loss of design efficiency is felt in the structure, power, weight, cost, etc. For a complex internal combustion car with integrated fluid-mechanical systems regulated by a computer with a myriad of sensors, the “system of iPods” approach would be untenable. We’d all be driving something that looks like it came out of Mad Max that was way too heavy, unreliable, and unsafe.

For EVs, the calculus changes. For the most part the portions of car design that necessitated the “traditional approach” no longer require tight integration. They can be defined as a system of interfaces. To be fair, this is true for a standard car as well, but those interfaces tend to be peskily hard to align items like driveshafts, fluid systems, and stressed casings. For EVs, if properly defined, these interfaces could be cable and connectors, coolant hoses, and bolt locations.

Defining the Interfaces

The real challenge in this approach is determining where one component ends and the other begins and how they are connected. A lot of teeth gnashing would be spent on this as it has two implications: it defines the size of a single system, and defines the amount of customization allowed for a solution. The smaller the single system and the less customization allowed, the more the market opens up for “small guys”. Most car makers are not cool with that.

The car makers would also have to determine the interfaces and make them standardized. This requires a lot of forethought because the interface has to “see the future” and ensure it will be essentially future proof. I bring to you as an example the USB interface as something that went very right a while back but is coming under scrutiny now with our higher power electronics.

To some extent this is already happening. Standardized head unit connections for car radios was an early attempt (though we’ve gone a little backwards there with the prevalence of movie theaters in the front seats pretending to be control panels). The integration of our phone with some standard data interface is also becoming pretty standard, which is truly a step in the right direction. There are some things that should have been standard a long time ago (why are there so many different bulbs for side indicators?). For purposes of this conversation, the real question is what is unique to EVs that make this easier to implement and more desirable? The Jalopnik article goes into more detail about all of the systems that could be modularized, but I’m going to focus on the two heaviest hitters that are unique to EVs.

The easiest and most revolutionary change is the battery. Whether you mount it in the trunk, floor, roof, or the passenger compartment, you just need wires to hook it up to the drivetrain. Sure, a lot of fancy data gathering and sensors could be involved, but they could all be within the battery package. You don’t actually lose a lot of packaging efficiency with breaking the battery up into bite-sized pieces either, so you could make the base unit pretty small so you could have as few or as many as your application desires/needs. There is the issue of cooling which is a complication, but we’ll discuss that later.

The second most obvious answer is the drivetrain itself. There are two very different approaches here – hub motors and shaft motors. The former certainly fits the bill in terms of packaging efficiency and compatibility, but suffers a lot technically, so we’ll assume that nobody is really going to want their main car to be compared to an electric scooter. For shaft driven motors, the integration becomes more complicated. At a minimum the interfaces would be bolting locations, power feeds, and controls. Current EVs have motors integrated with the suspension, which complicates the interface significantly. What you really need is a motor well of sorts where you drop in the motor that mates to a common coupling. The chassis would take care of the suspension and power transmission. Then the motor would just need software so it knows what type of drivetrain it’s powering. Ironically I thought about this a looong time ago (when I was seven). I dreamed of an electric supercar that in its trunk had a wall of motor input shafts (up to six). The motors would only be installed, or even clutched in by the control system as needed for acceleration or regeneration. The concept has aged well…

I do need to note that some work has already occurred in this space – in reverse. There are companies making “crate emotors” that are direct swaps with standard enthusiast motors like Chevy small blocks. These package an electric motor and electronics into a V-8 shaped box with RPM and torque outputs that closely resemble the motors they replace. I don’t advocate this approach in general as you are literally reshaping a round motor to fit in a squarish hole, but it does prove the concept.

I think where we hit a wall is with the rolling chassis. I don’t see too many opportunities to make this plug and play without significant losses in capability. That doesn’t mean it can’t be customized and enhanced. This is pretty much standard already with body kits, aftermarket seats, upgraded suspension parts, etc. Considered as one module, it’s a pretty massive part of the car and seems to undo the very purpose of modularization. But wait! You would think that the rolling part of the car, which is its primary function, would be the most costly. But generally these are now stamped in a few processes and assembled by robots. The expense is the batteries, motors, and electronics, so even though the bulk of the car by weight is still essentially proprietary, the really expensive bits (and those that frequently need replacement) are prime for modularization.

How Does this Look?

Imagine this idyllic vision of car modularization comes to pass. What is your car buying experience? Certainly, car companies will continue to offer pre-made packages you can drive off the lot. Those would look  a lot like they do today. Let’s look at the a la carte version.

You’re low on cash but need wheels because you live in America. You go to the Jeep dealership and get a rolling chassis stripped to its bare minimum. Then you have it towed to Nissan where they fit a Leaf motor and battery system. You’re on the road already for the price of a cheap car today, but the superior road hogging capability of a Jeep! Granted it probably won’t get up to highway speeds or go more than 50 miles, but if you just need something to get you to work in town it would do.

But what if you hit the jackpot in a few years? You take your Leaf Renegade to the Tesla dealership and have the roadster motor and electronics fitted. Then you max out the existing battery bays with high performance batteries. For kicks you get an aftermarket battery roof rack installed and put some more juice there (perhaps a tipsy solution). You’d need to upgrade some other support equipment (maybe wiring harnesses in the chassis, cooling systems, etc). You just built yourself a Baja racer! Nice!

What are the Downsides

Clearly the most obvious downside is the lack of efficiency in space packing and weight that comes with modularization. The only major components we’re talking about modularizing are the batteries and motor, so lets talk about them separately.

The batteries are ripe for modularization. We already do this. As a matter of fact, the vaunted battery packs in Teslas are essentially glorified AA’s all soldered together with a cooling system. But that brings up the point of the cooling system. High performance electronics get hot, so you need to cool them down. Typically this is done with some liquid that is chilled by another system. Liquid cooling systems have become compact, so even though they are not a main technology, it could also be modularized. You could be running a VAG battery pack with a GM cooling system if we standardize what fluid is used and your cooling system has the appropriate capacity. But you may be putting a cooling system sized for a Leaf battery into a bay that can fit a Hummer cooling system so there is some space efficiency issue there. There’s also a trend right now to make the batteries structural elements in the car, but I’m less than enthused by this idea. Putting stress on a shell that contains nasty chemicals seems like a recipe for disaster. I would rather emphasize making the battery structure minimal and having them “float” in a carrier. It might be heavier, but I’m not convinced the difference is going to result in any performance or packaging issues a consumer would notice.

The motor is a little bit different. We’re talking about something that is potentially big and heavy, or small and light. The location where the motor is mounted coupled with the weight of the motor will change vehicle dynamics. Add to that the same packaging problem we talked about with the cooling system, but more significant. We may be looking at limitations so you can’t ever upgrade your Mini to the Cybertruck motor. The other problem is power transmission. We’re used to motors producing prodigious amounts of torque that must be held by the chassis, really defining the motor mounting points. I think this is completely solvable if the motor module output is always at very high RPM and low torque. We then decouple the complex mounting interface as the torque handling is all covered inside the rolling chassis. It also makes the motor casing lighter.

As a tangent, another novel approach is more like my seven year old dream and force motors to be small, like in no more than a current 50 hp motor. You might need four or five, but you could install only two for every day and then five for the race. And if one broke on the road, you pull it out and go along your merry way! They would be small enough that it’s probably something you could drop in without a service center. There’s a lot of flexibility in this system. Some chassis may only be able to fit two where some can fit six. And manufacturer’s could compete in motor power density, fitting 120 hp motor into that 50 hp space.

A less obvious consideration is that this could make the type of car we’re used to much more expensive in the long run. Without a doubt, a similarly outfitted car will be more expensive a la carte than if it was pre-packaged. Anyone that has gone to a sushi restaurant knows what I mean. So even though the major advantage is the ability to get on the road cheaply, the major disadvantage is the expense in having to get on the road “nicely”. This might be the hook we can use to convince the car industry that their dealerships are still viable. They can still be competitive with the modders by providing nicer packages for less than the individual components. Those with means (and lack of interest) will buy these package deals. The rest will be putting together a volksferravette-150.


I’m very excited about a future where I can put a e-Viper motor in my minivan and strap on some batteries for that long road trip. I don’t think it will be around when I buy my next car, but perhaps when it’s time for my children to buy their second or third carlike contraption. This is going to take a while. But I’m with Jason Torchinsky on this. Ultimately the power is in the hands of the consumer and we just need a few brave companies (of which he has suggestions) to take this seriously.

Dose of Aphorisms

I wrote this newsletter mostly to appease my burning desire to inject my two cents worth into something I felt I’ve been advocating for a long time – accepting that no solution is optimal, and we need to understand the tradeoffs as we make decisions for the future and be willing to compromise. Changing the automotive industry to serve the consumer rather than the shareholder is a worthwhile task, but approaching it without considering the compromises to be made by both sides is short sighted.

A failure to compromise on a solution is a failure to have a viable response to the problem.

Explanation of Fields in the SMARRT form submission

Reference Scenario Inputs:

Number of People Infected – How many potential members of the gathering are infectious. The simulation starts when they enter (time=0).

Type of Activity – Impacts the number of particles spread as aerosols per respiration. More strenuous activities result in more viroid particles being released.

Air Changes per Hour – This is the air exchange rate with fresh air for the volume of air being breathed by the gathering. If you use forced air exchange, you can calculate the number of air changes per hour for your specific situation.

Space Floor Area and Ceiling Height – These are used to calculate the total space volume.

Duration Infectious Person is Present – This is how long the infectious person stays in the space after their initial entry. For the reference scenario, this defines the end of the simulation.

Gathering Scenario Inputs:

See the reference scenario for all inputs up to Time of space entry.

Time of space entry and exit – These values represent when you enter and leave the space referenced to the infectious person. For example, if you show up fifteen minutes late, but stay an hour after the end of a one hour party, the Duration Infectious Person is Present is 60 minutes, the Time of Space Entry 15 minutes, and the Time of Space Exit 120 minutes