82 GTBi Restoration/EFI Conversion

You might be interested in the EFI conversion I did on my Boxer ('83 BBi):
https://www.ferrarichat.com/forum/threads/efi-boxer.687781/

Lots of similar challenges to tackle; which often led to a combination of cut/bent parts from SendCutSend, as well as 3D printing.

I initially went down a similar path of putting EV14 injectors in custom bosses. The manifolds are different, but a similar idea. I considered a top-hat design like you have, but eventually went with a rail. I believe there's another 308 EFI conversion in the works using a top-hat to AN fitting:
https://www.efihardware.com/products/3378/injector-adapter-4an-with-stainless-steel-retaining-clamp

I also 3D printed a distributor cap that I used as a cam sensor. Being a flat 12, I have fewer ECU options. I went with a Motec 150.

Once the initial EFI conversion was finished, I converted the dual throttle body / long runner manifolds into a ITBs; which required custom stubby intake manifolds. I designed them in Fusion 360 based on the BBLM manifolds (weber carbs cut in half), printed prototypes on my Bambu X1C, and then had them printed in aluminum via CraftCloud (much cheaper than Xometry or Protolabs).

Good luck. It's a lot of work, but I have zero regrets having done the conversion.

 

What is the chassis number?

 

Good stuff, really glad you're getting it on the road again. Any other internal mods?

 

Hasn't run since 86?

Wonder why someone would park a 4 year old Ferrari and no one touched it again? Any history on that? Just curious.

 

I've never .messed the ecus you're looking at but they both look fine spec wise. I know I've read several here use link and are happy. I use enginelab and love it but it's not for everyone

I think the tua-x think is probably over rated. OEMs really need good mixture control to meet emissions targets... the rest of us can err on the rich side and never give it another thought

I did mine with tefzel... i break stuff and that is more forgiving

 

It's always neat to see other designs and how they tackle EFI on the 2V. Very cool.

The 2V has always been a real pita with injector conversion, the threaded boss, the angled ports. Have to use custom bent fuel lines for each injector. Then tie it all into a manifold to supply fuel.
The QV simplified so much that it's a slam dunk in comparison.

Project looks very good!

Gotta love the Advent of 3D printing and mfg to speed things up.

My preference is Link ecu's. Simple, easy and loads of tweaking with an intuitive to me gui.

 

My only fear with anything programmable retrofitted to these things is just that. It's a program and finding someone that can help with it if/when it goes south.

Stock they are simple and dependable. I have a Scirocco with Electromotive driven ignition and fuel injection and there is only ONE person, in his 80's, here in town that can talk to it with a laptop and he has had to reset it twice in the past 10 years. For unknown reasons (he says maybe voltage spikes, I dunno) the damn thing has just gone back to some baseline program. It runs great when it works and it is working fine now but if it fails again and he is dead, what do I do? Electromotive is gone. I have literally lost sleep worrying over this.

I would love to play around like this with my QV and Alfa Spider (82 with Bosch) because of the performance increases I have seen with my VW but the problems I have had spook the hell out of me.

 

Get with him and learn how to reset the rig yourself!

 

I kind of like the thought of the ecu dying....it means its way past time for a better one

Seriously, kind of the opposite of worrying about part is NLA...wait, you mean I NEED to upgrade?...how sad

 

I bought a cheap laptop at best buys and downloaded the program so I'm good there. I called him and he broke his hip and he cannot get around very well at the moment.

But, yes, that is exactly my intention.

 

Short answer is that I haven't seen any issues at all using the cam for the sync/ref sensors. I was a little concerned at first, but it's very easy to see the signal and how little jitter there is between what the ECU expects, and what it actually sees. A proper crank sensor is technically better, but using the cam is so much more accurate than the mechanical (rusty spring) advance that it replaced. As far as I'm concerned, I have much bigger fish to fry than trying to shave a millisecond off the timing calculation.

Regarding wiring, I would go for the Tefzel if you can swing it – along with Raychem DR25 heat shrink. That said, it's an expensive commitment to purchase proper cutting, stripping, and crimping tools for the various terminal types, so I wouldn't blame you for just going with a pre-terminated harness.

Let me know if there's anything you'd want to know about how I did mine. Good luck!

 

Just my 2c on the obsolescence of EFI hardware, but I felt that CIS was even worse in this regard. I personally find it easier to find a "tuner" (or learn it yourself) than it is to find a qualified CIS technician (or learn it yourself). CIS was a relatively short lived and niche phase compared to EFI; which has basically been around for 30+ years at this point and is ubiquitous.

A good ECU can work with various sensor signals (new and old). Eg, my modern Motec M150 can just as easily read a signal from my 40 year old sensors as it can a modern digital sensor. When that ECU dies, I expect I'll be able to replace it with one even more capable.

When you factor in performance gains, fuel flexibility, and the ability to upgrade cams (CIS is very sensitive to intake reversion), then it's kind of no brainer, IMO.

Carbs are a different story.

 

Enginelab was the supplier to AEM before holley bought AEM for the other products and dropped the ECUs. At that point Jim (the owner of EngineLab) apparently decided to focus on other business. I know he was trying to sell into the general industrial controllers space. Anyway, he decided supporting individuals was taking too much of his time and stopped. They are still selling and I may reach out again, they were talking about releasing a model library to make it easy to people to build working models....but about 1/2 dozen people have found me to ask for help with 1 thing or another and if you chose this path I'd be happy to share my working model and any knowledge I have.

Tau-x - this is basically right and I do have a version of this strategy in my model but I haven't played with tuning it yet. MoTec uses a version as well, you could look at their accel enrichment function/tuning. The tuning is kind of the important piece....MegaSquirt added a true tua-x but I think have dropped it because their user base couldn't make it work....the base tunes weren't that good, the tuning MAPs not that large so lacking OEM level accuracy and ALL the other correction need to be spot on before the tau-x will work. The MoTec system was more designed to be tunable and I kind of followed that path with mine, at least I think I did. I did a write-up on my version for someone else, but have attached here for you. If cost is no object, the M1series motecs have a development package (they sell the HW and SW separate, $4k and $2.5k last I knew, then logging it extra, everything is extra) that give you their full library and lets you create your own functions. I had an M150 ECU, then realized how much I still needed to spend and sold it and bough the EngineLab that gave me basically all the same stuff, other than the library.

Trigger wheels....every ECU on the planet will support a 36-1 +1cam or 60-2 + 1 cam. Those have been the preferred patterns, since the 90s? at least 36 teeth is kind of where you need to be to hit the OEM spec of 0.1 degree ignition timing accuracy and get misfire detection working right (which I have not yet mastered). Lower tooth count patterns work just fine, my very first setup I used 4 (ECUs had JUST gone from 8 to 16bit) but at that low driving is fine, but you notice the ignition timing becomes retarded during rapid acceleration. Most of the really strange most patterns were created to rapid start, surveys says consumers see quick starting as a sign of quality and the older patters require up to 2 full engine revolution to synchronize but it could be much quicker so it starts randomly between like 1/4 and 2 1/4 revolutions where the new patterns will start constantly in like 1/4-1/2 revolution. I think there should be oil pressure so I prefer the older patterns

The 308 CIS flywheel has 2 pins on it that the ignition reads, the way the webbing is it just kind of works out that you can add 33 more pins to make a 36-1 and use either (not both) of the OEM pickups. That is I think the neatest way to do it. The easier way to mount to the front damper or prehaps easier yet the cam mounted but that will the timing to move a little with RPM but it should be pretty repeatable so most of that should come out in tuning leaving whatever error the lower tooth count causes, 1/2 max I guess which is just touching the you can maybe see in on a dyno line.

 

It's a sine wave. You can calibrate the ECU to trigger off any part of the wave (but usually the middle). My car has 12 teeth on the distributor reluctance sensor. I locked out the mechanical advance, but otherwise kept the wheel and sensor as-is. I then removed the original distributor "arm" that does the actual ignition distribution, and 3D printed a new arm with a magnet on the end. This acts as the sync signal (using a digital hall effect sensor). It essentially works the same way, but it's a digital step, rather than an analog wave.

There are some best practices when it comes to triggers. Here's an excerpt from a Hall Effect sensor guide:

Image Unavailable, Please Login

From my experience, the tick over is not "instant", but it's only taking a few seconds to fire up. Honestly, you'd hardly even think to critique it. The fact that it builds a little pressure is a kind of happy accident.

You can split the engine management, and even some modern cars do that. I'm not sure I've seen it done for less than 10 cylinders though. Even on a Boxer, which effectively behaves as two 6 cylinder engines, I still used just one ECU.

 

The oems do or did that for cost, 4&6 cyl ecus are high volume and cheap, 10-12 not so much.

The 308 electric ignition was 2 4cyl units which was cost but also speed, they have old slow 8 bit processors and using 2 doubles the time they get to process between spark events. The 10-12 engines really make the ecu processor work. The motec M1 series use a processor with 2 32 channel time coprrocessing units to off load all the time critical stuff....same processor used in old (pre spec ecu) F1 applications. It was a big problem 20 years ago is my point.