Mechanical Fuel Injection, Because I Hate Electrons

alfadriver rocked spica on his turbo'd challenge alfa. i hope he chimes in soon.

Hilborn and Bosch had two very different approaches to pressure control. Hilborn systems used a pump driven with a timing belt and a pressure control valve on the throttle that would regulate the proportion of fuel returned to the tank vs through the injectors. I don't see that approach working with turbos.

Bosch used an electric pump with a device that resembled a flapper air flow meter but produced a pressure regulation output instead of a voltage output. That's the sort of system that would be best to copy for a turbo project - though engineering an easily tuned pressure controller won't be easy.

tuna55 wrote: I am working with VCH on this idea, and I think some of you need some framework: There are valid and good reasons for this choice, other than just hating electrons, and the carb-in-a-box isn't preferable for other reasons. This is not a street machine and a rulebook is in play.

I'll try to bite my tongue, but I'm at a loss as to what rule (short of "no electrons that don't drive spark plugs") would make MFI a better choice than EFI; I suspect that's the source of the difficulty in getting an answer to the actual question.

So, let's stay within the framework of: How to best develop a MFI system for a high boost pressure four cylinder motor. I want to know, for instance, how a MFI system would reference boost.

That was one of my first thoughts. I figured with a supercharger you might be able to get a decent throttle position map to work with MFI, but as soon as you get a turbo causing any given throttle position/rpm combination to have wildly varying fuel needs...

At one point I was talking to a guy who used to run a 2002 with the Kugelfischer injection (no turbo), and was going to help him build a system to adjust the KF's trim adjustment based on feedback from a wideband O2 sensor driving a servo. That was just to keep the thing in spec on his N/A engine with a big cam. (EDIT: and of course it involved an electron-based workaround to get feedback to the mechanobrain...)

I will readily admit that I long ago quit paying attention to MFI, and that if he had someone who really knew what he was doing with the KF he might not have needed to start mucking about with dynamic trim adjustments...

I assume that if you had a suitably trimmable MFI system, a diaphragm referencing manifold pressure could adjust it, but getting a sane relationship sounds.... non-trivial.

FWIW, the SPICA on the Alfas is a diesel based system- it was used on other stuff that SPICA supplied to.

It has inputs of engine speed, throttle position, and engine temp. Most of them also have a BP input, and all also have a crank fuel system. It's a great system for the Alfas- I much prefer them over webers.

Bosch was mentioned, but Kugelfischer has not been- Per is probably familar with that, since it's used on BMW's. The SPICA, Bosch, and KF will all be cheaper to get than Lucas or Kinsler, since they are production based, and not as widely used in either historic or current racing set ups.

I would shy away from making your own. That's a massive amount of engineering for mechanical body movement. Unless you want that kind of project- but it will probably be as much work as designing and making your own engine from scratch using random junk hardware. If you think you could translate Megasquirt's basic code into mechanical set ups- go for it.

In reply to alfadriver:

I did mention Kugelfischer (see above), however maintenance and upkeep is an issue.

tuna55 wrote: I want to know, for instance, how a MFI system would reference boost.

I'm going to throw out a guess it could work similar to a vacuum advance on a distributor. If we assume the mechanical pump adjusts the amount of fuel delivered as a function of rpm speed on a curve, then a pressure module could advance that curve and thus inject more fuel with more manifold pressure. How to physically make this work? Good luck.

I'm just glad the vintage car I want to build runs D-Jet - Bosch's first stab at EFI - which can be converted to MS fairly easily.

Kugelfischer seems ideal for what we're looking at- diesel-based, used in period road racing. From what I understand, though, part availability is a problem.

The diaphragm- based manifold pressure system seems like the best solution to me. Chevrolet used it on the MFI they developed for the SBC back in the 50's, and I believe the K-jetronic used a similar reference. Basically, all a carburetor does is reference vacuum, and it does pretty well just with that. If we throw a couple of more inputs in we might come up with a pretty tunable setup, even for positive manifold pressure.

I'm going to keep my eyes out for some SPICA gear, potentially, too.

By the way, sorry for my slightly-inflammatory, tongue-in-cheek post title. As Tuna mentioned, there are good reasons we're doing this THIS way, rather than EFI.

I do hate electrons, though. Tool of the devil, they are. ;-)

tuna55 wrote: I am working with VCH on this idea, and I think some of you need some framework: There are valid and good reasons for this choice, other than just hating electrons, and the carb-in-a-box isn't preferable for other reasons. This is not a street machine and a rulebook is in play. So, let's stay within the framework of: How to best develop a MFI system for a high boost pressure four cylinder motor. I want to know, for instance, how a MFI system would reference boost.

So whatcha buildin' Oh man I hope its a chainsaw. /best read in Homer Simpsons voice.

phaze1todd wrote: In reply to alfadriver: I did mention Kugelfischer (see above), however maintenance and upkeep is an issue.

My computer crashed just as i hit add post, so there are a few posts between where it should have been and where it ended up... Oh, well.

It's a good system, none the less.

AngryCorvair wrote: alfadriver rocked spica on his turbo'd challenge alfa. i hope he chimes in soon.

Actually, we didn't. It was electromotive powered. i'm ok with 'trons.

But I do know that there is a turbo kit just for SPICA- a buddy of mine even has some of the hardware. The way they did it was to replace the barometric reference thing with one with more range, and then run it off of boost pressure.

The big problem with mechanical systems is that they have no adjustment for air temp- and air charge changes a lot with how cool the charge is, very non linear. It almost speaks of not having an intercooler to change the temps too much.

actually, carbed systems don't do intercoolers very well, either- you'll note how few race turbo systems used them, as well.

anyway- good luck with it, it's a very tough challenge. hope you can find a boosted application to mooch off of.

I dunno about KF or SPICA, but I'm pretty sure that Bosch's MFI pump can do temperature compensation. On my 911T there is a hose running from the intake to the pump, with some bimetallic disks that expand/contract a rod into the rube goldberg portion of the pump.

I know they used the same/similar pump for the 935s and 934s, but I can't think of how they did pressure compensation.

Edit: I forgot about the pressure compensator on top of the MFI pump. Looks like Porsche put a special diaphram and manifold on them to get boost compensation.

Is this an "in budget" project? I can't think of ANY way to get this under budget without designing something from scratch. A good used Porsche pump is in the 1K range.

tpwalsh gets bonus points for being the first person to reference Rube Goldberg. Surprised it took two pages of MFI discussion to get there

ransom wrote: tpwalsh gets bonus points for being the first person to reference Rube Goldberg. Surprised it took two pages of MFI discussion to get there

I would have posted earlier, but got swamped at work. Sorry for the late reply. . They ARE rubegoldberg devices. If you've every opened up the front half of one, you'll want to shut it immediately. There's levers and springs, and dohickies EVERYWHERE inside. The back half where the plungers and rack are that actually deliver the fuel at least looks straightforward

Image of said 935 pump:

Note that it is missing both the temperature comensator, as well as the overun solenoid. The upper round cylinder on the right is the temperature compensator, and the lower one is the overrun solenoid. The diaphragm on top of the pump is the atmospheric compensator.

In reply to tpwalsh:

Besided the P's- there were a number of BMW's that were turboed in the 70s, as well. Maybe there's more information in those systems.

In reply to alfadriver:

Those might be a better option anyway since he said "turbo 4 cylinder". I just happen to know the bosch/Porsche system MUCH more so than the others. The sequential MFI systems that are Bosch MFI/KF/SPICA all work about the same though.

Honestly now that I think about it, the CIS (K-Jetronic) is probably the best, easier system for a turbo car. Porsche used it until the 90's on their turbo cars, both pure racers and street cars. turbos and CIS seem like a match made in heaven. Draw air through the meter into the turbo that opens up the fuel distributor for more fuel. Easy Peasy. The plumbing is the hard part.

In reply to tpwalsh:

And there's a plethora of 4 cyl cars that have CIS on it. I'm sure it would be a lot easier to find one of those vs. a SPICA even. WAAAAY more VW's sold than Alfas.

What about using a stock mid 80's VW MFI and a "cold start" fuel injector wired (maybe five wires in total?) to a mech pressure switch? Not an full MFI, but very close, and you could install an intercooler.

Paul B

Given the number of guys I see with Porsche 924's (especially 924 turbos) that complain about getting their CIS systems working well, or at all, would be enough of a warning.

Some of this could due to old fuel system components and changes in gas formulation, but either way, it seems that quite a bit of the problems on the 924 message board is related to CIS issues.

If you go the CIS route, get some books and the proper CIS pressure tools and arm yourself appropriately.

The 924 Turbo that they ran at LeMans used a 928 fuel distributor and simply combined the 8 lines to 4 for the injectors. Though the stock 924 turbo solution will provide plenty of fuel to support 1+bar of boost.

With that said, some are starting to adapt the EFI from other VW's with similar sized motors instead of going MS. Fairly simple conversion and it just seems to work.

Just a few thoughts. Good luck on the project (whatever it is) and I hope you can share it with us as it sounds pretty interesting.

Lots of boosted BMW E 21's running around with cis. I would go that way, I understand it better.

nicksta43 wrote: Lots of boosted BMW E 21's running around with cis. I would go that way, I understand it better.

What system does the 2002tii Turbo use? While not exactly common for parts, it might still be a source for copying ideas.

Saab 900 Turbos from 79-84 had K jet. VWs had much better fuel lines, though. Common mod is to fit an MB dist.

Driveability. driveability, driveability .... Changing just about anything on an EFI system = a few keystrokes. Doing the same on a mechanical system .....

Begs for a dyno showdown ....
Test for HP vs Mileage vs Emissions vs Driveability.
4 hour limit for tuning
I'm betting on the electrons FTW

OK guys, I am very glad to see that each of you can pee very very far, very impressive. Now.

EFI is off the table. It just is.

I'd tell you all why, but it is one of those "don't want to get your hopes up" kind of things. It's to satisfy a rule book.

Figure 2 liter engine at 36-40 psi on race gas. Looking for something like 600 hp. Streetable on occasion would be a great goal. The ability to start with a key and idle for a few minutes without fouling plugs or overheating would be a huge benefit.

We have looked at these options put forth thus far and I see $$$$ -> science experiment, but nothing really in between. Right now the best we can figure is a diesel V8 pickup setup with two injectors per cylinder. It's not a challenge budget, but it's not "How can I spend 15K for a box of parts??" either.

I think we should set up some mechanical injectors and some electrical injectors and see how far each set can squirt. Now that would be a fuel injection pissing contest.