This is my Jaguar... a 1977 XJ12L
Alternator
Part 1: The Idea
Most of the recurring problems with the car centered around the fuel, ignition and cooling systems. A single coil just wasn't producing enough spark to reliably run the V12 engine to its 6500 rpm redline. Any minor variation from perfectly optimal conditions was guaranteed to cause trouble. On top of that, fuel economy was dismal. The original Bosch D-jetronic system was probably a marvel of modern technology in 1977, but it wasn't cutting it in 2001.
From these circumstances an idea was born: Why not replace the Jaguar/Bosch system with a completely new EFI system. A search of the web quickly revealed why "not": Basic after market systems were $1500 and up, and most did nothing to help the weak Jaguar ignition system.
I continued to drive the car as it was until June 2001, when I had another idea: why not use a modern EFI system off another, junked car. It seemed like a good idea, and I started researching various systems.
It quickly became clear that there were only two realistic choices: Ford or GM. I not a particular fan of either company, but compared to Lucas, Ford and GM are golden. On the web I discovered that the entire GM fuel and ignition system has been almost completely reverse-engineered, primarily by a group of hobbyists such as myself at www.diy-efi.org. Another advantage of using GM parts is that GM cars seem to die young (hopefully not due to EFI problems!). The junkyards in California are literally overflowing with late 80's GM cars.
So, armed with that seemingly simple idea and some printouts from the web, I headed for the nearest junkyard.
Part 2: The Parts
After a day of scouring the junkyards, I settled on the following configuration for my GM-Jag EFI system:
2) Two sets of DIS coil packs for the same V6 engine
3) 12 GM fuel injectors off the same.
4) All the related MAP, CTS, MAT, etc. sensors and connectors that go with it.
I made it out of the junkyard for under $200. GM ECMs sell for just $36 at Ecology Auto Wrecking in San Diego!
Part 3: The Installation
The first task was to set up all the pieces on the bench (or, in this case, the garage floor) and see if they worked.
OK, go to "plan B": Drill 7 holes around the circumference of the front pulley, and mount the two crank position sensors 60 degrees apart on the front cover. Ah-ha! Sparks jumping and injectors ticking were the best sounds I had heard all week.
Now the real work begins: Installing all the bits and pieces to look "factory" on the Jaguar engine. This link (coming eventually, sorry) gives the pin-out color codes for the '727 ecm. This information seems to be in short supply on the web, and I ended up figuring most of it out myself from various sources. Here is the wiring harness that I created to interface the various sensors with the ECMs.
With the battery, ECMs and sensors in place, it was time to start the engine. Double check all the connections, hold breath, turn key and...LOTS of black smoke. But it ran!
I spent the next two days trying to figure out why the engine was running so rich. I soon realized that my plan to use the original Bosch injectors and fuel rail was not going to work. The Bosch injectors were flowing about 2 times as much as the GM injectors. It was like having a fire hose mounted on the intake manifold. No wonder Jaguar had to use so many stupid tricks to pass emissions at idle.
Here is what the engine compartment looked like after I finished fabricating my own fuel rail and modifying the jaguar manifolds to accept the GM injectors:
Unfortunately, most GM IACs are designed to mount inside the throttle housing, a problem for people like myself who are using a completely different setup. But while staring at a 3.1l V6 in the junkyard, I discovered that the IAC on this particular motor mounts to the top of the throttle body by 4 screws. A little cutting here and filing there, and I had a very nice IAC installation on the A bank air cleaner housing.
Now things were coming along smoothly. The engine started ran, and even drove pretty well. About this time I also FINALLY managed to establish ALDL (serial) communications with the ECMs after about 6 attempts at building and re-building my interface board, rewriting the communication software, and otherwise pulling my hair out for an entire weekend. After all that trouble, it turns out the the problem was simple: '727 ecm's out of Oldsmobiles use ALDL address $90, NOT $94 or $92 as commonly reported on the web.
With communications working it was a wonderful feeling to be able to "peek inside" the running ECM. It is truly amazing the data that is available on the serial port of a GM ECM. Sensor values, fuel ratio, spark advance, injector pulse width - it is all readily available.
Part 4: Tuning
Tuning the GM fuel injection to behave nicely on its new engine was not an easy task. It was all about learning what the engine expects from the ECM, and what the ECM expects from the engine. The first one was easy: how much fuel, and how much spark. The second one is where I ran into a lot of trouble. Here are some examples, if you can avoid these, you will save yourself a lot of time:
Part 5: Update August 11, 2003: It's Done!
Well, 18+ months after I started this project, it is finally running. Not only running but it passed smog. I am on the 35th chip iteration of the second generation system. I have probably burned over 100 chips if I count back to the very beginning. I guess it is a bit over-simplified to say it is "done". Projects like this are never really done, they just reach incresing levels of refinement. At the moment, the Jag has reached the level of "pretty darn good".
Here is an update of the parts I use to make this thing tick:
Go back Home.
It looks pretty nice here, and in many ways it is the best car I have ever owned. Reliability is not one of them. To give you a sense of perspective, let me briefly list a few of the things that I have repaired/replaced/rebuilt on this car in the first 20,000 miles that I owned it:
Brake calipers
Suspension bushings
Radiator (3 times)
Air conditioning
Transmission (leaks)
Speedometer
Tachometer
Distributor
Fuel injectors
And that is just a partial list.
1) A pair of '727 ECMs off of late 80's GM 2.8l V6 cars. (one for each bank of my V12)
I selected this ECM because:
Now the real fun begins.
I had originally planned to use the bottom half of the distributor as a trigger source for the EFI system. Unfortunately, the small diameter of the reluctor wheel didn't provide enough room between the TDC and sync notch to get a clear signal to the ignition modules.
Hmm, it looks deceptively simple in the photo, but in reality it took hours to cut, strip, and solder the 30+ connectors and 80 or so wires. The best part of the installation was putting al the hardware bits into place. I had originally planned to move the battery to the trunk to make space for the ECMs under hood, like this:
But after I accidentally killed the original battery, I discovered this wasn't necessary. In a fit of frustration I went out to my Honda 600 and borrowed the much smaller group 51 battery out of it to power the car temporarily. Much to my delight, the Honda battery nestles nicely next to the ECMs, and saves about 10 lb. in the process. So far, there have been no problems turning over the big V12 with the smaller battery.
At this point I hadn't done much to make it look pretty, I just want it to RUN! Recheck all connections... fuel lines look ok... turn the key - and Hey! It works! And not too bad either, considering that it was running on a completely different engine than originally designed for. As the engine warmed up, the ECMs started to get a little confused because the engine wasn't responding to commands to bring down the idle speed. I was running without an Idle Air Control (IAC) valve, since the original Bosch D-jet design had no real idle control (not counting the lame thermo-valve that had long since crapped out on my car). The ECMs didnt seem to care for this situation, and besides, I thought it would be cool to have idle control.
Can you find it? It's located just to the left of the throttle body, below the horizontal throttle shaft. Like it was made to be there.
Once you have sorted out the various trouble coses, the rest is: drive, log data, analyze it, burn a new chip, repeat as needed. Here are the tools I used:
I started out tuning static idle in my garage and then worked my way up through city driving, highway cruise, hill climbing, and finally wide open throtttle. It took about 10 iterations before it really started to run smoothly, and about 25 before the BLMs were settled at 128 across the normal operating range (If you don't know what a BLM is, think of it as a measure of how well the fuel delivery is tuned to the engine, where 128 is perfect, less than 128 is too rich, and more than 128 is too lean). The hardest parts to tune are the transistions and the fringes of the operating range. In both of these cases, the conditions only last for a second or less, so it is hard to collect much data about what is going on.
If you read the whole thing from the top, you might have noticed a few things that have changed since the beginning. This project went through two complete generations; the second one was when I installed the the 6 liter engine and 5 speed. I had to redo the whole wiring harness anyway, so I decided to make a few other changes:
Overall, I'm thrilled with the performance of the efi system. It still has a few calibration bugs to iron out, but for the most part it is very pleasing to drive.