Finally, some pictures of my kit car project

In one of my projects I learned the ins and outs of the CAN protocol for a particular type of truck. It was very fun to be able to write code to read data and even control things over the CAN bus. I had a USB CAN dongle, so I could connect my Linux laptop up via USB to this truck and control the steering, brakes, emergency brakes, etc. It was really fun.

My subaru has some settings that can only be set over CAN. Specifically, the option to unlock all the doors when I unlock the driver’s door. I have wanted to get into it and find the bit to flip, but I bought the OBDII/BT dongle about 3 years ago and I haven’t touched it since…

1 Like

That sounds really cool. I hadn’t even thought of any of the other stuff. I just want to turn some codes off, lol. I’ve been reading about engine management for the last couple days and HOW sensors work vs what they do (which I pretty well understand) and exactly how the Subaru makes logic decisions and calculations. I figure it’s a month or so at least before I start messing with any tables so I need to read now.

I’m pretty sure CAN wasn’t introduced to these until later than my model year, but that STILL sounds really cool.

1 Like

CAN bus is kind of cool voodoo. Most of my tinker stuff is before its time, but my daily Ford is all on the bus. It makes some things very easy and others… not so much. On the whole, mostly easier. It certainly gets rid of a lot of wire.

My first exposure to it was reading an article on how many kg of copper it removed from one car (I think in regards a Lexus IS, as I recall.) It was an impressive number.

If the drivetrain is pre CAN, I’d give serious thought to stand-alone engine management. Given that you’re here, I’m guessing you’re a DIY kind of guy, a Megasquirt engine management system ought to have some appeal to it. It’s not exactly simple to get set up, but I seem to recall a base WRX management program available. (I was interested in the Toyota I6 engine tables, of course.) That will also give you lots of room to play with fuel delivery, ignition timing and load tables, with built-in wideband O2 feedback for everything up to crazy boost levels.

My problem with that… The ceramic ball bearing CHRA turbos, and 1000cc/min fuel injectors start looking really sweet when you know you can still tune them for daily street driving…

Speaking of wire…


Sure could do with a lot less, lol.

I’m really hopeful that I can do enough with romraider and/or ecuflash. Lots of work has been done to make the stock ecu flushable by clowns like me, and it seems plenty capable so far, if a little esoteric.

I considered megasquirt, and it’s still on the table if messing with the stock ecu fails. Second goal is to come in under budget, though (first goal is to finish). Got to prove to wifey that when I give a number, THAT’S the number.
Because I want to build a couple more cars, lol.

I may end up tearing this one apart and going nuts at some point, or selling it to build the version they’re working on with a small block. But the Type 65 looks SOOOOOOOOOO sweet. And of course the midlana looks like tons of fun to build (that’s another great candidate for megasquirt).

Before all that, I’m gonna have to build wifey an old style pickup truck I think, because she loves hallmark and you all know THAT deal.

So many projects, so little time.

1 Like

One of my previous incarnations was working with a company that provided integrated solutions for public transportation. I’m not sure exactly what they used, but they may have been using CAN bus as a low-level transport since the electrical environment of a city bus isn’t exactly quiet… Lots of embedded hardware (think fareboxes running two SoCs, one pseudo-proprietary RTO linux distro, and one Windows CE, plus driver displays with GPS and two-way comms to dispatch, etc., etc.)

Oh that is a lot of wires. That looks pretty fun, and stressful at the same time. I like organizing wires, and shortening them, but extending them drives me nuts!

It’s just like extending wood. Grab each end, and pull! :hernia: :andre_the_giant: :hulk:

Hmph… :weaksauce_emojis: :wink:

3 Likes

Yeah. It was bonkers, and with all the time I spent building and playing with my V1 machines and resulting printer, it ended up taking me a little over a year from cutting the first wire to the first engine re-start, lol.

My goal was to retain as much OEM functionality as I could, so I only removed things that I can’t use (airbags) or would rather upgrade (stereo).

I took some advice I was given to leave the wires as long as I could to minimize the number of splices and, by extension, possible failure points. But if I were to do it again (which I likely will) I’d just cut the connectors off, mount them where they belong, and run the wire as needed. I spent A LOT of time managing the wire coils and tangles.

1 Like

They certainly can! I was a contract worker at BMW in the early 2000s when they introduced a new 7 Series, at the time I was entering field reports in their data base. It had a CAN system and looking at a ‘ghost’ drawing of the car showing component locations you’d see 50+ control units of various functions. It seemed maybe they just all didn’t get along as most if not all dealers wound up with a service bay dedicated to reprogramming the entire car which could take several hours. GREAT! when it all works and a real PITA when it doesn’t.

2 Likes

The problem with extending wood using the brute force method, is it tends to get skinny in the middle. Not a big problem if you were going to cut it down anyway, but…

CAN is very good at dealing with noise. It uses two wires which are both at 2.5V when idle. When a message is sent, it pulls one down to ground and the other up to 5V for a 1 and then the first to 5V and the second to 0V for a 0.

That’s not the most interesting part though. All of the parts talk on the same 2 wires. As long as you can connect the next one, you can extend it really far. They can all try to tall at the same time, but the electronics make sure that the message with the most zeroes wins.

So important messages, like the engine throttle command get IDs like 3 (the transmission can send this message to slow down the engine during gear shifts, for example). Boring numbers like cabin temperature control get IDs like 103.

Each message is an ID, and then 8 bytes. Everything has to fit into those messages. And if you need something like the transmission to tell the engine something, then they have to agree on the same message ids and structure. The standard I worked on was SAE J1939. Which has a lot of common messages and space for some special ones.

I can easily imagine a new system having a ton of problems. If one of the 50+ things turns into a “babbling idiot” (technical term for they won’t shut up), then the whole network can get clogged. Imagine a door sensor for the trunk, and it goes haywire, thinking you’re opening the truck at 1kHz. Anything with a lower priority will get crushed and then the AC isn’t working anymore. I think most of these lessons have been learned. But those are nasty things to trace.

That is some cool insight into how it works. I would not have guessed that. I try to make everything work on it’s own pin. One day I will learn something other than an arduino.

I knew CAN was similar to the 1-wire stuff I use in projects, but thanks for the additional info. I might have to snag a reader and play with my truck some.

I don’t envy the wire mess in those photos. I finally got around to running a hardwired power to my dashcam and thought that one wire was a PITA. I can’t imagine having to deal with all of those.

I also played with ECUs quite a bit with my last Harley. The stock ECU was MAP based and had no O2 sensors so it was open loop. I replaced it with an alpha-n system that had a wide band sensor. I didn’t have a dyno so it took a few months of riding around with a laptop and adjusting numbers for me to tune it.

2 Likes

Yeah the wiring was…and still is…something of a mess. At some point, this car is coming all apart and going back together. I’ve gotten better at some things, acquired some tools since I started (like a cnc) and I think I can do better.
The wiring took forever, but really it was more like a puzzle for me because I cut it all apart. Every single wire. Put it all back together by reading the schematic.

As an update, I got my tactrix cable that lets me read and flash the ecu yesterday. Managed to disable the codes on the first go. Fired up and the only light left was airbags, so that’s a huge win. Now I have to tighten up my vacuum and emission hose mess and then back to mechanical stuff!

2 Likes

Ya had to bring up the topic of “kit cars”. The year was 1983, I had a pocket full of cash from working my young butt off, and the choice was down to an older 911 that needed some body work, or a father and son kit car project that ran out of gas. I had seen the kit car that spring at the Chicago auto show, and the “built” price was near $16,000 as I remember. Well, $7,000 later I was the proud owner of an “almost complete” Bremen Sebring, built on a vw chassis.
I spent most of my spare time the next year working on it, and my best friend and I took it out on a summer day for it’s maiden voyage. It might have gone better had the positive battery cable not disconnected from the battery somewhere on the trip. I pulled it back in the driveway, went to power up the top, and nothing. For the next hour we struggled to unscrew one of the windows to climb out. The next week I had the sunroof installed. Over the next 15 years, and another $3-5,000 later (that’s all that I will admit to), rebuild & bored out engine & carbs, upgraded heat, new vgo gauges & dash, rewired completey, I could go on, but it never made it out on the street again, I was always tweaking and upgrading it, and never had it ready to run to my satisfaction. Finally my wife was tired of it taking up half the garage (and two kids getting close to college and money was tight) So I was convinced to sell it on ebay. It received one bid of $3,000. Two young (under 25) kids came by with a tilt-bed and took it away looking quite happy. To this day, I wonder if they ever finished it and how my life would have gone had I bought the 911 instead :slight_smile:
I used to have polaroid’s of the car but time and heat have pretty much destroyed them. Here is the closest I could find to my cream-puff…


Pretty darn close to mine except for the added fog? lights in the front, right color, wheels & sunroof -

Maybe the kids did finish it after all!! :smiley:

2 Likes

Sweet, I just stumbled upon this thread and will be following along with interest. I have a '73 Porsche 911T that needs a motor (the '87 3.4L RUF carrera motor it had burned in a shop fire). My buddy who is a mechanic talked about subyizing it, that there were even kits available for that and he would help. I started doing a proper restore out of college, and it got put it aside when I had kids (mostly done except the interior). Now my son is getting to that age, and now there is a bit more pressure on me to get the project going (and hopefully hand it off to my son to finish). I know there’s probably more details than anyone could shake a stick at when it comes to pulling off a suby swap like this… but heck maybe this thread may offer some insight. All I know is when my buddy mentioned it could easily do 400hp and be streetable, I got all excited… car already pulled mid 13’s with the 270hp carrera motor lol.

edit: Reading up on some comments… my buddy mentioned he knows the ecu’s pretty well and is pretty confident remapping my motor as needed. Sounds like a steep learning curve to do it myself… that or a fat check to someone who does.

1 Like

Man, I saw those when I first started looking. SOOOOO many of the older kit cars were junk, but THAT one just looks super sweet. No idea what the performance would be like, but it’s hard to imagine it would compete with modern tech. Still…so cool and I bet it would turn some heads

I’m certainly gaining a bit more respect for these motors, but it was a huge pita to build. For starters, the closest guy to me that could bore the block was 2 hour ride each way and he was so backed up I didn’t get my block back for 4 months. Nobody else would even touch it.

Then one of the headbolt threads let go while I was assembling it. The machinist apparently had some issues with the threads in 3 holes he didn’t bother to tell me about because he chased them out and bolted his torque plate on OK. Apparently putting a steel bolt with FINE THREADS in an aluminum head (including the water jacket) and running grounding voltage through it for 15 years tends to booger things up. I actually had to use my shovel as a breaker bar on disassembly for exactly 3 of the bolts, too, so I wasn’t surprised.

And none of the stuff that goes under the intake can go on later. Since I forgot stuff 3 or 4 times, that was a lot of assembly/disassembly.

But there are no mains, in the classic sense. The crank goes in the middle of two block halves, so those have to get bolted together first and ARE the mains. Which means the pistons go in AFTER the rods, (or else you can’t get to the rod caps) so you’re fishing wrist pins and clips through access holes in the block. On the other hand, no balance shafts, so that’s nice, lol.

Obviously not impossible, and at the end of the day it’s all nuts and bolts. Just nowhere near as easy as some other motors I’ve built or seen built (used to work at a machine shop in the early 2ks, and they taught me a bunch).

The other thing to watch out for is that it’s SOOOOOO hard to put one of the engines together for less than $5k. Or maybe I should say it’s so easy to spend $5k building one. I bought an import kit for a stock build and didn’t upgrade anything but headbolts (arp studs for the win, nothing less in a boosted engine for me). Then again, I expect my car to be about 40% lighter than the WRX, so I don’t need big power. Go check out any of the subaru shops, though, and it’s $3k for parts. No such thing as a rebuild kit -everything is a la cart so it’s easy for “small” upgrades to start adding up, especially after machine work. And speaking of that, some people will tell you that you absolutely HAVE to replace things like the bucket shims (they set the valve lash on these) if your newly machined head is too tight instead of just rearranging them for best fit and tipping the valves where necessary like people have done for…ever. Real easy for other people to spend your money there.

All THAT said, it’s still a pretty good motor, especially for a small displacement 4cyl, and responds very well to tuning likely because Subaru left a lot of power on the table in favor of emissions, economy, and reliability. My friend had his tuned at a shop (something like $600 for the accessport and another $300-400 for the dyno tuning) and ended up with 270 at the wheels by only fixing up the fuel and spark curves. Turbos are AWESOME for streetability, imo, too. They don’t often like to start making boost at grocery-getting RPMs, so that makes them pretty tame overall, and being able to set the boost by throttle position (which these ECUs do) means you can effectively limit your horsepower with your right foot. I’m not sure exactly how steep the learning curve is going to be, but it seems you either want to pay someone for dyno time who has tuned a bunch of these and knows where to start, or be prepared to spend months making small iterative changes.

New project…bell crank. Have to reverse the motion on both cables to tidy up the rear and be able to flip the shifter. The plan is to make the bracket or is steel and the other pieces from alunimimium. I’ll have to resize the primo to attempt this, though. As for the steel, I’m not sure how many times I want to try it. At least once, no doubt, but maybe if it doesn’t go well I’ll strap on a marker then cut it out by hand or send it off to the water jet guy. That’s what I built the first burly for anyway. Mock up parts and either mark them out or send them off.

I need to add some rods for linkage to the selector, then figure out the selector bracket, still.

Here is some video of somebody else’s solution with a different shifter that only needed one direction change, but the video is a lot better at showing the idea.

6 Likes

I wish you were my neighbor…I wanna play!

2 Likes

Man, you’d be a great neighbor to have! I wish I could tell you how many hours I spent coming up with ideas and rejecting them before i got this far. I bet whatever time you cut off my thinking would be worth however much beer and tacos you could consume while doing it.

2 Likes