Modernizing a Vintage Car

This car is 50 years old. But what if it wasn't? Two years ago, I brought this classic car back to life after it had sat for 20 years. After a lot of work, it's back on the road, but I found myself thinking, that's not enough. Vintage cars are super cool, but modern cars are generally more efficient, powerful, and reliable than cars of yesterday. I want those things for this. One thing I could do is swap out the old engine for a modern one, getting the best of both worlds, but actually no, I can't. So what if instead, there's a way to bring this old engine into the modern era, and maybe make even more power in the process. Today, I'm going to show you what I spent the past 8 months of my life working on. Starting with, what exactly is wrong with the car in the So let me get you up to speed. This is my Datsun 280z from 1978. And it's uh, kind of a piece of garbage. A lot of that is due to time being unkind to it, but also a lot is due to just how it is. Now the car does run, and the fuel injection system on it technically works, but that's about where any positives stop. According to the original owner's logbook, this thing gets around 15-17 miles per gallon. That's not great, though old marketing materials say it should produce around 170 horsepower at the engine. Power is always lost as it moves through the drivetrain, so there's probably a lot less than that actually hitting the road. But how much less? Not far from me is a performance shop called Stupid Fast, and I brought them a car that's stupid slow. We did a few runs on the dynamometer here, and the best we could get was 125 horsepower. For reference, that is less than a 2012 Chevrolet Cruze Eco, and that car is a piece of ----. I would know, I own one. So here it is, the Nissan L28 engine. Now before I go fully tearing apart this engine, I have two constraints for this project. One, any modification I make has to be fully reversible, that way I can go back to the old system if anything I do goes poorly, and two, no modifying anything inside the engine itself. Camshaft, pistons, valves, all stay original. I think the name of the game for improving this engine is going to be optimizing combustion. Engines need three things to run, air, fuel, and spark. In a four-stroke engine, air and fuel get sucked in, squeezed down, then ignited by the spark, only for the waste gases to be expelled or the cycle starts all over again. All this to turn linear motion into usable rotational energy. Here's the thing, air, fuel, and spark are all controlled via replaceable external components. Oh my god, the engine's blue. My thinking is, if we upgrade the old parts to better modern parts, get a new computer to use those parts, and add a whole bunch of sensors so the computer knows how best to use them, we should have a vintage engine that runs far more efficiently than it ever did in the 1970s, hopefully resulting in more power and better fuel economy. Now luckily, what makes the 280z so bad is also what should make this upgrade easy, and that is the fact that this car already has one of the earliest examples of computer-controlled fuel injection, the L-Jettronic system. Most cars of this age had carburetors to deliver fuel to the engine, which worked by black magic, but the 280z is basically the same setup as cars today, where a pump supplies pressurized fuel to fuel injectors that are activated by a computer. But a lot messier, and with a computer that's only slightly more advanced than what's inside a Tamagotchi, which is an actual Japanese computer. This may say made in Japan, but this whole system was actually developed by the German dishwasher company Bosch, making this the second worst collaboration between the two countries. Okay, I think the first thing I want to tackle is the air side of the equation. A big issue with the original system is that the computer only controlled fuel, meaning air was entirely analog, meaning 90% of what made up the system was just a lot of tubes. But why? When you press the gas pedal in a car, what you're actually doing is opening this valve, which, when installed, is basically the main doorway for air into your engine. More air, more explosion, more speed. But engines don't just need air whenever you choose to give it, they need air all the time, and for different reasons. So Bosch came up with 1, 2, 3, 4, 5 different components that all have the exact same job of letting more air into the engine. Like this guy which activates when you turn on the air conditioning, or this guy which activates when the engine is too cold. I don't want 5 parts with a million tubes, I want one part with zero tubes. This is an electronic throttle body. It's pretty much the standard for cars today, and is also made by Bosch. Instead of manually actuating the butterfly valve, this guy is controlled entirely from a motor inside the housing. That means if I connect it to a computer, along with a new digital pedal, we can control it, but so can the computer itself to always get the ideal engine speed in any situation. So I went to a junkyard, and picked up a pedal that has a position sensor built into it, and swapped out the old computer for one from this millennium. This is the EMU Black from a company called ECU Master, and it's a universal aftermarket computer specifically for controlling engines. It's a really powerful piece of hardware, and one of the more affordable options too considering what all I can do. Or at least I think. I didn't exactly pay for this. And that's because ECU Master USA is in Texas, where I live. Meaning it was really easy for me to find their offices, drive there, and very politely ask that they give me one free of charge. So after a lot of wiring, reading the manual, and some calibration... What the ****? It wasn't long before I had this throttle body fully operational. The fact that this one part replaces everything that used to be here is so cool. Or at least cool to me, because I do know that a lot of the older automotive enthusiasts don't like electronic throttles because they are computer controlled, where you lose the direct connection between the driver and the car. But listen, there are things you can do to get that direct connection back if you really care that much. So the question now is how do we get this onto the engine and the pedal onto the car? Well, the pedal is easy since the way the original one mounts is pretty much identical to the new one. There's just two holes instead of three. So I drew up this little adapter and sent it over to my 10 year old home-built 3D printer that I am very surprised still works, and got this part that is going to go right here. Now it didn't exactly work the first try. Feels weird. Feels like it's not in the right spot. That's because it's not in the right spot. Turns out the old pedal sticks out a lot farther than the new one, so I'll have to really offset the base to get them in a similar position. It took a couple more tries to get the ideal thickness and placement, but eventually I got one exactly where I wanted it. The throttle body isn't so straightforward. I chose this specific one because it has the exact same bolt spacing as the original, which sounds pretty nice, but listen closer. Right now, the bolts that held on the original throttle body just do not fit through the new one's mounting holes. The outlet bore is about 10 millimeters larger than the manifold's inlet, which is also this weird oblong shape that wouldn't be completely covered in with leak air like crazy, all while the side of the gear housing would be running right into the exhaust heat shielding. Yes, these are Discovery Channel Shark Week branded knockoff crocs. As far as I'm aware, no one has put one of these on a 280z before, so a readily available solution does not exist. So what do we do? Well, for starters, we can use smaller bolts that do fit the throttle body, they just don't fit the manifold. At least, not without installing some of these threaded adapters. Now, if you're wondering what threaded adapters are, they are adapters for threads. As for everything else, it seems like all we need to make this work is a spacer to offset the throttle body, a hole that's tapered for a smooth transition, with a little protrusion to cover the opening up top. So with some careful measurements and some time in Autodesk Fusion, it wasn't long before I had something that just might do the trick. And I'm happy to say that the very first test print fit perfectly in the trash can, but this next version was actually pretty good. Though right now, this isn't air tight. To seal it, I could use a gasket, but I hate gaskets and I just broke this gasket, so what about o-rings? They're cheap, can be formed to any shape, and won't tear apart even after a hundred disassemblies. All we need is a groove for them to live in. So I found 200 pages of extensive documentation specifically on how to design grooves for any type of o-ring seal that you can possibly imagine, immediately got overwhelmed, and ended up just eyeballing it, which worked out pretty well. The only issue now is that it's made of cheap PLA plastic that will literally melt in a hot engine bay. Now every car has plastic in the engine bay, including actual engine components, so I could just reprint this in a more suitable material, or could use the sponsor of this video, JLCCNC. JLCCNC is a custom parts manufacturing company specializing in aluminum, copper, brass, and seven different kinds of plastic. That's almost eight different kinds of plastic. All you do is upload your 3D model, choose the material and finish you want, and you're given a quote literally within seconds, and can have that part out for delivery in as soon as three business days. The pricing is also super competitive, whether you're a hobbyist who needs one part or a company that needs 1,000. Listen, I was going to use JLC regardless of whether or not they sponsored this video because I think they really are the best option, and I still have a lot more parts to make. So if you want to see the quality for yourself, you can click the link in this video description for $70 off your first order for whatever big project you're working on, or maybe you just want a little cover plate with a really cool engraving that's definitely not a penis. That's an option too. Thank you JLCCNC for sponsoring this video, and thank you for clicking the link if you want to get custom parts made for as little as $1. So that's air completely upgraded, and now we're on to fuel and spark. And we're going to do spark first. This engine has six cylinders, and that means six spark plugs. In the original ignition system, there was one big coil that generated a huge electrical charge with a distributor to direct that charge to the spark plugs one at a time and in the correct order. That distributor went right here, and as the engine spun, so did it. The problem with that was that if the coil goes bad, which they do, the entire engine shuts down. Modern cars ditch the distributor altogether and instead have multiple tiny coils that the computer controls individually. That way, if one fails, you still have the rest of them to get you down the road into a part store, which I'll probably be doing soon since the coils I purchased for this are used ones from a Nissan Pathfinder with a hundred thousand miles on them. But I found out that if you just clean them, you can convince yourself, no, these are new. Though right now, they don't fit over the 280z spark plugs, at least not without modifying them. And in this case, I am not the first person to want R35 coils on the L28 engine, and there's already a guy online that sells adapters to make this work, making this the easiest upgrade for the entire project. As long as we ignore one tiny little issue. These actually won't work at all if the computer has no idea when to fire them, which will certainly be the case seeing as how no such device is in place to measure how fast the engine is spinning. Oh, don't worry. There is a solution. Am I supposed to know what that is? This is a hall effect sensor, and it's going to tell us exactly when to fire the spark plugs. This device detects any time a ferromagnetic material gets really close to it. So if we were to wire this into the EMU black, make something like a toothed wheel out of steel, and spin it right next to the sensor, the computer will see a pulse of electricity every time a tooth passes by. And because there's 36 teeth, each pulse equates to 10 degrees of rotation. If we tell the software that, it can calculate exactly how fast this is spinning in real time. So my idea is to scale this setup down and stick it right where the distributor used to be. That way, the computer will know how fast the distributor would be spinning if it were still there. And that seems doable, but speed alone isn't actually enough data. If I told you that someone got in their car, drove 100 miles per hour east for two hours, and then asked you what city they were in now, you would have no idea unless you also knew their starting point, which in this case was Miami, and they actually drove into the ocean. Basically, we need a reference marker for the computer to see so we can say, hey, this is when to fire the first spark plug. Technically, it's a bit more complicated than that, but also technically, I'm too lazy to explain it. Now unfortunately, this hall sensor is not able to see paint, so what we can do is actually remove that tooth. So long as we tell the computer that we did that, it's smart enough to recognize what is now a gap in the pattern and know that a tooth should be there, but it'll also be able to use that gap as the locating marker, which we can see is working perfectly because the RPMs aren't dropping, and this changes to synchronized. Okay, so the proof of concept seems to be working exactly as I want it to. Now we just need to scale this whole thing down and build an enclosure for this to mount to the engine. So fast forward two and a half weeks, and I have a design that I think should do just that. The final part will look something like this. There are three very important design considerations. One, I don't want to use this hall sensor for the final product, so I found this one that's super compact, has a built-in connector, and you're never going to guess who makes it. Two, to get the trigger wheel small enough to fit where I want it, I had to scale down to 24 teeth, so that's 15 degrees instead of 10, which is technically lower resolution, but it should still work. And three, the drive shaft that spun the distributor is made of case-hardened steel that would chew up aluminum if it came in contact with it, so the shaft that will be attached to the trigger wheel also needs to be case-hardened steel. That is not a material JLC offers, so my original plan was to buy a second distributor, tear it apart, and cut short the shaft from that. But then I found out that Nissan made this a very specific size that does not fit through any standardized bearing. Did I just destroy this for no reason? Yep. So I'm just gonna 3D print what I need in the meantime and let this be a problem for future Ronnie. I'm sure he won't mind. Okay, let me present the final 3D printed prototype. The shaft goes through two main bearings and gets sandwiched in place by screwing down the trigger wheel. Just like the distributor, it gets popped in place and spins around with the engine. And the hall sensor goes into this lid piece. Now obviously the final thing will be fully enclosed, I just thought it'd be cool to have a cutaway view to see how it works. And let me tell you, it does not. Yeah. It just kept losing synchronization. Looking a little bit closer, I realized the trigger wheel is wobbling around. There's an ideal distance for how far the teeth should be from the sensor to get an accurate reading, and it looks like it's moving both too far and too close. At first I couldn't figure out why, but then I saw the reason was very stupid and was staring me in the face. It's me. The way I designed this requires there to be a wave washer, which is basically a spring, to apply a preload on a thrust bearing. Without the washer, this thing rattles around, and without the thrust bearing, these parts just scrape together. So they both have to be there. However, I completely overlooked the fact that having the trigger wheel directly on a spring makes it so the trigger wheel is directly on a spring. Yeah, give me a day or two to figure out this one. Okay, I got it. For some reason, I was dead set on the idea that the base of this shaft needs to be an exact copy of the distributor shaft, but I realized it doesn't. As long as the hole in the slot stay the same size, I can make this however tall I want. So in that case, here is the final, final prototype. This time, bearings get locked in place with a snap ring, and the taller shoulder of the new shaft butts right up against the inner race of the lower bearing. So no thrust bearing needed, therefore no wave washer. This thing is rock solid. But just to be sure, I tested and tested until I was positive this wasn't going to fail. The only thing now, it's uh, it's plastic. Oh hell yeah. JLCCNC is once again the solution for the aluminum pieces, but I still need this to be made of case-hardened steel. But I realize this doesn't have to be a for me at all, if I instead make it a problem for someone else entirely. If you weren't aware, there's a man in Switzerland called My Mechanics who's restoring a Nissan 240Z to a condition better than it ever was when it left the factory, who happens to be quite the skilled machinist. Wow. Wow. And also a good friend. He's got a video on his second channel about everything it took to make this part, so go check that out. After you finish my video. Okay, so final assembly. Once the oil seal goes in and the bearings get locked in place, this thing is ready to be installed. But just to be sure, I tested one more time on the bench. This, this is good. And now it gets installed for good. I've never created something as complex as this before. Machined pieces, moving parts, and electronics. To be honest, I am insanely proud of myself. I just think it's cool that- Technically you outsourced literally 100% of the manufacturing, so it's not like you actually made this. And also you're not even the first person to have- Shut the f- So with this in place, now that is spark fully upgraded. And technically it's also fuel. By far the biggest issue with the original LJetronic system was that the computer, whose only job was to fire the injectors, had no idea when to fire the injectors. So it instead just fired all of them at the same time, all the time. That is one, not a joke. Two, extremely wasteful. And three, no longer the case. And that's because the distributor spins at the exact same speed as the camshaft, which is what controls the opening and closing of the intake and exhaust valves. So if we know the position of the distributor, we also know which cylinder is in what stage of combustion when. So for all intents and purposes, this is a camshaft angle sensor. Meaning the EMU Black now also has enough information to fire the injectors sequentially, which should save a lot of fuel from being wasted. Nevertheless, I hate everything about this. So I got modern EV14 style injectors to replace the old ones and a much simpler fuel rail, freeing the engine of all convoluted tubes and finalizing the upgrade of air, fuel, and spark. Wait. Can you sense that? Probably not. We don't have enough sensors. To measure airflow into the engine, there was originally this massive airflow meter, but this computer has a built-in pressure sensor instead that we can simply connect to one of the vacuum ports already on the manifold. So this goes away and in its place will be this simple aluminum tube. But I went a step further and had a threaded bung welded onto it. That way we can also measure the temperature of the air coming in. As for the temperature of the engine itself, I had to make some slight modifications to get this new sensor to fit where the old one went, but the biggest upgrade by far will be the installation of the lambda sensor, also called the oxygen sensor. There's an ideal ratio of air and fuel for gasoline combustion and this guy will tell us if there's too much air or too much fuel. The best place for it will be in the exhaust pipe somewhere around here, but I don't know if that'll hit the transmission or the body of the car, so I installed the transmission and the body of the car. And it turns out there's tons of clearance, so I drilled a hole, attached the O2 sensor adapter, and threaded this on for good, finishing the final step of my three-step plan. So step four of my three-step plan is plumbing up the fuel system, which was easy enough, just some hoses and a fuel pressure regulator, as well as connecting all of the electronic parts to the computer. Now I love wiring, in the same way I love getting stabbed in the face 47 times. None of the original engine harness is usable, so I had to make my own completely from scratch, along with a custom power distribution system for all the new components. And because I want this key to turn on the computer just like before, that meant integrating part of that system into the factory body harness, where the only available diagram is this black and white picture. So two and a half weeks of researching, modifying, routing, sleeving, and so much heat shrinking later, we have a harness that actually looks pretty good. I also made this metal bracket, that way the EMU Black can live in the same spot as the old computer. But seeing as how this was my first time ever making an engine wiring harness, I had no idea if this would actually work. Oh, the gas pedal. However, as soon as I turned the key, it did. Drive by wire, operational. Temperature and pressure, accurate. And fuel pump, pumping. But I uh, accidentally drowned a spider in the gasoline when testing this. Oh no, you can see the bubbles of his last breath. But what about the camshaft angle sensor? We know it works, but it needs to be calibrated. So after I put oil back in the engine and transmission, I told the computer to fire the cylinder one spark plug when this line on the main pulley is at the zero degree marker, which is the point where the piston is at the very top of the stroke. And we can verify that with this timing light that only flashes when the spark plug fires. In case you missed it, should be here, but it's actually here. So after adjusting that offset in software and trying again, we got it right on the money. The computer now knows exactly where the engine is, which means as soon as I calibrate the fuel pressure and reactivate the injectors, there's no reason this engine shouldn't start right now. I'm going to start the car. God damn it. Except of course for maybe a failing starter motor. See if that works. That was close. But unfortunately, close is all it ever got. No matter how many times I kept trying, it never started. Could it be that I forgot to install the alternator belt and the battery is draining too low for the computer to work properly? No. Could it be that I forgot to plug in this vacuum line that tees into the pressure sensor causing there to be inaccurate readings? No. Could it be that I accidentally disabled the throttle body causing no air to actually enter the engine? Also, no. So what the hell is going on? Well, it turns out that you can't actually just plug stuff in and hope for the best. The computer needs to know everything about the engine it's controlling. I told it how big this is, I told it how many cylinders this has, and I told it what order to fire those cylinders, but I did not tell it how much this thing sucks air in. It turns out all engines get better at sucking in air the faster they spin until they start getting a little bit worse. So the computer needs what is called a volumetric efficiency table that should look something like this, but mine does not. So the computer thinks this engine is really bad at sucking in air, so it's giving it really little fuel to match what little air it thinks there is. I could not find a table for this engine online, but anything is better than nothing, so I loaded one in for a completely different car just to see what would happen. Let's see if this goes. Okay, so it works. Yeah, it runs, but it won't for long if I don't install the cooling system. So after filling this up without spilling a drop, oh my god, I also threw on the rest of the exhaust, did some research about making my own VE table, and before I knew it, I was on the road for the first time with a new system. I mean it feels like it's driving better than it did before, which is very possibly placebo. That's pretty good, but it can be better, and the only way to do that is to take it to someone who actually knows what they're doing. So we're back at Stupid Fest eight months after I first dyno'd the car. Jason told me I did a pretty good job on everything, and I felt really good, until the car failed to start, and I had to bring out the hammer. I had zero expectations for how this was going to go, but after a lot of adjustments and many test runs, we made 142 horsepower, and that is more than a 2012 Chevy Cruze Eco. It's crazy to me that everything on the inside is still 1978 tech, and only changing the external parts to burn fuel more efficiently results in a 13% power increase. And speaking of burning fuel more efficiently, I drove this car everywhere. On the highway, at the drive-thru, grocery shopping, and road trips with friends, and I was able to average 19.9 miles per gallon, which is better than the 17 I was getting. So if my math is right, with the amount of money I'll be saving on gas, this entire project will have paid for itself in only 63 years.