Finally drove MFI
 

Some of these posts have been really interesting and informative- thank you for those!!

I’m in the process of removing the CIS from my 2.7 because even after a year of working on it, tuning all of the various elements and making countless repairs, it still isn’t great and frankly leaves a lot to be desired. Plus the expense in both money and time off the road is getting tiring.

I looked at carbs but felt that was trading one archaic and finicky technology for another. In addition the fuel economy stinks, they’re not great in 4-season environments and a set of really good, clean carbs cost more than I thought. They look and sound super cool but the functional trade offs were too many.

I looked at MFI and would have loved to go that route but again the cost of entry was surprisingly high and all of the offerings I saw were cores. Plus, having followed Jonny’s thread I wasn’t looking for that level of work tuning the MFI. It’s some of the coolest tech (born out of pilot’s dogfights!!) but again, the compromises were too high.

I ultimately went with one of Al’s X-Faktory ITB/EFI kits as I knew everything would be new and operable out of the box, he provides tons of support and the ability to tune it easily was appealing. I felt it offered the best of all worlds- throttle response of MFI (or at least close to it), sound and looks of carbs (or at least close to it), and fuel economy of CIS (whatever that’s worth)

I do have mixed feelings about EFI in a classic car but the reality is that I want to drive my car a lot and not only wrench on it and ITB/EFI seemed to combine the best of all the other offerings with minimal and acceptable compromise.

That said, now I want an MFI setup [emoji12]

I've made a career of working on computer controlled process systems and instrumentation - EFI always feels a bit like work to me, so one of the attractions of MFI was that at least some of what I need to do to get it to work involves using my hands and tools, not the laptop, and trying to get it to connect via a serial connection over a USB cable and hoping the software, firmware, blah blah blah all match.

If someone wants to pay me my full hourly rate to do that, I gladly will, but it's not something I do for fun if I can avoid it.

Yes I understand that I'm using datalogging and a laptop to do the tuning.....

Showdown, you made the right choice. In a couple of respects. Starting from ground zero, with nothing on hand and facing the "full retail value" of the three systems in contention - EFI/ITB, carbs, of MFI, I would have made the same choice. As I mentioned earlier, I "inherited" several full MFI systems back in the days when mechanics were still removing these "troublesome" systems from customers' cars. I got it all for free... And now, 20 years later, I cannot believe how much some folks seem to think this stuff is worth.

As discussed earlier, my use of the Electromotive crank fired ignition system very much falls into this same little quandary. Like I said, though, the sheer cost and lack of availability of "period correct" twin plug dizzies at the time I built my motor precluded their use. Same situation you are now seeing with MFI systems. Same justifications for me as well - this car is driven. A lot. It's been darn near 15 years since I finished my build, and it's going to roll over 100,000 miles very soon. Trouble free miles on both systems (the MFI and Electromotive), by the way.

That, and you chose Al, who I consider a good friend. Talk about irony, though - I just dropped off an MFI pump with him that he is putting on a mutual friend's 2.5 short stroke. Soon as he gets it on the motor and gets it running on his stand, the three of us will be reinstalling the motor in our buddy's car. Even our guru of EFI/ITB has now been caught playing with MFI... ;)

And, yeah, Jonny - laptops and datalogging. To tune 50 year old mechanical injection systems. The irony is not lost on me... But, we both know that what we have done would be virtually impossible without it. Not that it made it "easy", but it did make it "doable". I'll take it...

Bottom line is that anything is better than CIS [emoji12]

It is really cool seeing people take 50 year old equipment and using modern technology make it do things that were impossible back then.

Honestly, this is why these cars are so amazing.

oh the shame!........Jeff has revealed that even I am drawn to the dark seamy underbelly of the Porsche MFI world......aaaaarrrrgggghhhh! help me!

This leaves me wondering: if higher injection pressures result in more power, why does EFI use lower pressures? Is it due to a limitation of the electrically-operated injectors?

Yup, Showdown made a good choice. But I will say this: I've been driving CIS cars continuously since the 1970s (Rabbits) and when they are properly set up and working as designed, they work very well. Throttle response blows, of course. But all of my various Rabbits and 911SCs started on the first turn and instantly ran as though fully warmed up.

Still, I think CIS was designed by Robert Bosch and Rube Goldberg.

I'll probably never have one, but I do want an MFI car. "Throttle response" makes many people wrinkle their nose. What is that, they wonder. Well, with MFI.....the feeling of cruising the engine to 4500 RPM and then stabbing the throttle is just...weird. And you know what? 911 driving is all about feel. We like the light weight. We like the lack of power steering. We drive them because of the "feel" and feedback. Guys sell their way-faster late model 911s and buy early cars because, well, they are just SO much more fun to drive. Electromotive does not inhibit this and MFI is actually an important part of that 911 experience, I think. That feeling of having your body and your mind connected directly to the machine and the road surface. Hard to explain, but I don't have to explain it to you guys.

Don't worry Al, there is help... "My name is Jeff, and I'm an MFI-aholic..." :D

MFI utilizes a mechanically driven camshaft (driven by a toothed belt off of the forward end of the left camshaft) to drive plungers inside of little cylinders, which is how it achieves such high pressures. Picture a two cycle engine - the "exhaust" port is where the fuel enters the six individual little cylinders in the MFI pump, then the plungers (or "pistons"), when driven up by the camshaft, force that fuel out through what would be the "spark plug" hole in our two cycle engine. The electric fuel pump merely supplies fuel to these cylinders in a constantly recirculating (back to the gas tank) flow.

EFI supplies fuel to the injectors directly off of the electric fuel pump at whatever pressure it can muster. EFI injectors are really just little servo or electromagnetically driven valves that open at a specified time for a specified duration to allow the pressurized fuel behind them (usually in what is known as a "common rail", or hollow tube full of fuel) through. There is no secondary "pump" of any kind in an EFI system to increase fuel pressure, so it can only run at whatever pressure the electric fuel pump runs.

Thank you Jeff, this is very informative. My first 911 back in 1975 was a '72 S, with the MFI. I loved the car, but sold it later on to "upgrade" to an SC. (can you imagine that now??). Since then my 911s have had Porsche EFI, or Carbs. It has been too long for me to recall my MFI experiences from the '70s. However, my wife's '72 2.7 engine has PMOs, and I've been considering putting MFI back on the car. Since the heads were changed to non MFI heads when the engine was done, I assume I need to get new heads since the MFI injects directly?

I use adapters in the heads. No need for both unless you have serious power levels where multiple fuel streams required. Without some help from a few turbos this is almost not going to happen.

I have had MFI cars, EFI cars with ITBs and now run EFI injectors into MFI ports combined with MFI TBs and stacks. This is about as close as you can get for direct comparison. Here are my thoughts.

MFI is immediate. The direct connection between the engine speed (cam speed) and injection pump is evident. The load changes with mechanical movement of the stylus is immediate. The high injection pressure and resulting fuel atomization combined with directed fuel spray into the chamber is a huge plus, especially at low rpms where the change in torque based on throttle position is greatest. If (big if) MFI is setup correctly which Jeff and Jonny have spent countless hours doing it is really hard to beat. that being said, MFI is a challenge to tune to this level of exactness. If you go outside the engine build norm of a space cam then your work is cutout for you.

EFI on ITBs is good. Really good. But what people say about delay is minute but still there. The computer is reactive. It reads a point in time and compares to its previous point in time and then adjusts on the next event. Yes, the clock cycles on a CPU are fast but the calculation loop does not run at CPU clock frequency. it runs slower and if you know what you are looking for it is noticeable. AEM and MoTeC (previously mentioned) have some of the highest horsepower processors to minimize this effect but it is not zero.

the EFI advantage is clear in that you can quickly map an engine to provide fuel and ignition timing optimized at EVERY load/rpm condition. Give the engine what it wants.

I do notice that the combination of the two MFI mechanicals with EFI control is a happy mix. modern injectors have exceptionally fine atomization and the spray pattern when mounted in the head and aligned with the intake valve are quite efficient. Sorta the best of both worlds if I can make that direct, real world experience relationship.

In my case, I built a 3.0 very similar to Jeff's with different cams. Stock MFI was not on my list for several reasons but the aesthetics of a vintage injection hardware very much is. My thoughts are that the EFI allowed my to optimize my particular engine spec on an MFI base platform. I think I made it better. and yes, its pretty immediate when the right foot goes down. If I would have had ALL the necessary parts to complete the MFI the decision might have been different.

Ultimately, do what works best for your car and style.

I am still smiling since I just got back from a nice 3 hour drive.

When I built my 3.0, I used an '83 SC motor as the starting point. In addition to having no MFI injector ports, the heads had the CIS scallops on the one side of the intake ports. I had the scallops welded over, then opened up the "small port" (I believe 35mm as opposed to the earlier "large port" 39mm) intake ports to 38mm. With that little obstacle behind me, it was a simple matter to drill and tap for the MFI injectors. A good buddy who owns a local shop (Monty Jarvies at Redmond European) has a little drill jig that he made that bolts to the two intake port studs. Pretty simple little tool really, that makes drilling for the MFI injector location a piece of cake. I bet a lot of old time shops have a similar little tool. If you get to that point and cannot find anyone locally, I'm sure Monty would be happy to help you out. Or, probably better for you (since you are in San Pedro), I'll bet that Henry Schmidt at Supertec has drilled hundreds of heads for the MFI injectors. Pretty easy thing to do.

Thank you
 

What a clear summary of this comparison. Thank you for this. In the end it’s all about the smile! And not really about the bits that put it there.

P.S. ear to ear at 5 this morning.

Let me rephrase my question. Given that fuel atomization is better at higher pressures, why don't modern EFI systems run at higher pressures on the common rail? For example, current diesel engines are running at more than 20,000 psi (way higher than combustion chamber pressure after the compression stroke). I would have thought that it would be technically possible to run high pressures on the common rail of a gas EFI engine too. The fact that we don't see performance engines with indirect EFI running at high pressures makes me wonder if there is much performance in there.

That seems like a great question.

Isn't atomization a function of fuel line pressure and the orifices the fuel exits from? I just assumed modern fuel injectors can made to much tighter tolerances and/or patterns to get maximum atomization with less pressure.

I am probably missing something about this part of the discussion...

Modern EFI injectors don't have just one hole to squirt the fuel, they have several. 6 holes are common, obviating the need for very high fuel pressures.

Having the injector spraying at the intake valve does NOT produce the most power. Putting the injector as high up in the intake tract as possible does. If you don't believe me, look where Porsche put them on the 908 and 917 engines. Or the 935's. Or look where Formula 1 injectors are located.

Great question, but I'm afraid I don't have an answer. I can guess, I suppose, and if I were to guess, I would guess that part of it boils down to cost and complexity. 30-40 PSI as delivered by the electric pump is certainly "good enough" to get the job done with regards to fuel delivery into the throttle bodies from the common rail. I mean, for carburetors, atmospheric is "good enough", isn't it?

Additionally, I think any gains made in fuel atomization will be lost by the time the incoming charge has made it into the cylinder. With that earlier, higher introduction of the fuel spray, all it is going to do is splash and collect up against the far side of the throttle body or somewhere down the intake port (depending on where the injector is aimed).

I suppose EFI setups can be made to spray directly into the cylinders as well. Jamie has done it, utilizing the MFI injector position on his EFI setup. I think, however, having to drill and tap cylinder heads for injectors would significantly reduce the appeal for a lot of people, while raising the cost significantly. It's no longer just a "bolt-on" installation if we have to remove the heads so we can drill and tap into the intake ports.

So, yeah, I think the added cost and complexity of introducing some sort of a high pressure pump, of having to at least partially disassemble the engine to facilitate installation of the injectors, and that kind of thing might be why we don't see it, at least in aftermarket kits sold in the hot rod hobbyist industry.

What kinds of pressures does modern gasoline direct injection run at? This is very similar to how we introduce the fuel with MFI. And, coincidentally, we are seeing some fantastic gains in power and efficiency with these systems. I have to think it would be awfully expensive to adapt something like this to our engines, or to offer it to the aftermarket hobbyist hot-rodding world, however. Common rail, low pressure EFI/ITB is very cheap and easy by comparison. And it works quite well.

That's my guess anyway.

That's certainly a reasonable observation. There are, however, other factors in play with these engines. Camshaft timing and intake charge temperatures are a couple of them. Introducing the fuel higher up has a marked effect on that intake charge temperature, particularly in forced induction setups. Maybe a little less so, but it helps with naturally aspirated setups as well.

An interesting example of this affect was seen in NHRA Pro Stock drag racing. They got away from their decades old, traditional twin four barrel carbs sitting high up on tunnel ram manifolds and adopted fuel injection a few years ago. The teams had one hell of a time with this at first and couldn't figure out why the motors were making less power and suffering severe detonation. It turned out that introducing the fuel down at the juncture of the intake manifold and intake ports was resulting in much higher intake charge temperatures. That continuous fog of gasoline in that big tall tunnel ram was having a significant cooling affect, which was now absent.

There are many reasons for introducing the fuel in different places in the intake tract. It's not always as obvious as it may appear. Incidentally, a cursory look at some of the RSR's at the last Rennsport revealed many of them had ditched their high butterfly or slide valve setups, and had been fitted with the low throttle bodies and injectors in the intake ports. It really surprised me to see that.

To be contrary to Arne2, I think that you are on the wrong page.

Throttle response while not in gear is kind of a useless measure. What good is that except to maybe impress at a stop-light or make you feel good in the driveway? Also, what I would argue you are feeling in this situation is inertia. Motorcycle and F1 engines have very low inertia because of they way that they are designed and the low masses of their internal components. As a result you touch the throttle with no load on the engine and it spins up fast.

Throttle response implies the engine is under load. CIS engines (to pick a somewhat extreme example) react like this when you add throttle while under load.
1) Throttle opens when you push the throttle pedal down.
2) Vacuum builds upstream of the throttle. If you could watch this in slow motion, it would start at the throttle butterfly and work its way up the intake track, first to the plenum, and then from there back up to the airflow sensor plate. So the length and volume of the intake track matters. The greater the length and the larger the volume (compared to the cylinder volume), the longer that this takes. In reality it's not this simple since there are also harmonic effects to take into account. CIS engines generally have camshafts with little overlap in order to minimize the effects of the harmonics. A CIS engine with lots of overlap will be very difficult to tune at part throttle situation and have noticeable flat-spots in the rev range depending on the amount of overlap. Being able to use a cam with more overlap results in more HP when above ~3000 RPM. It's the harmonics again.
3) Now that the increase in vacuum has reached the airflow sensor, it needs to build to a level that will actually move the sensor plate against the spring, overcoming any stickiness and inertia in the mechanism and the resistance of the spring.
4) Finally, now that the sensor plate has moved, the fuel pressure at the injectors is increased and additional fuel is injected into the intakes to be drawn into the cylinders on the next cycle.

For comparison, when you tip in the throttle on an MFI car, the process looks like this:
1) Throttle opens when you push the throttle pedal down. At the same time the mechanical linkages move the MFI cam to the location that matches that RPM, throttle position and temperature. This immediately moves the MFI control rack which changes the volume of fuel being injected into the intake track at that moment. The injector open at that moment gets a partial increase in fuel, and the next injector gets the newly metered amount of fuel.

As you can see, there is a lot less time in an MFI between when you touch the throttle, and when the cylinders get the increase air and fuel in the correct proportions. This is what results in the significantly better throttle response.

How does MFI compare to modern EFI systems?
It depends. Compared to more-modern airflow sensing EFI systems, MFI has a faster throttle response because any airflow sensing system has to follow some version of the steps 1-4 that I listed above*

Compared to modern fully mapped EFI systems, MFI will have about the same throttle response. This is because they have merely replaced the mechanical map (imaged in the earlier post) with an electronic map.

* Most modern EFI systems are a combination of mapped and sensing. They will sense airflow at idle and lower RPM or load settings, and then transition to a fully mapped set-up at higher loads or RPM. To be honest, the airflow sensing function is really only useful at smaller airflow rates, in order to keep the mixture correct for good emissions. But the sensor just gets in the way at WOT. Airflow sensors are also rarely plates or flaps nowadays, but rather a hot-wire which barely impedes the flow of air.

There are also a couple of other factors that make MFI great compared to Carbs or more modern EFI systems.

1) MFI has no restrictions in the intake track. In a carb you need to have venturis which by definition create a restriction and vacuum downstream in order to suck the fuel out of the bowls. If for example this restriction were to consist of 10% of the cross-sectional area of the intake track, then a carb will flow 10% less air than a comparable MFI system. Less air equals less HP. This is primarily why MFI'd engines generate about 10% more HP than an equivalent carb'd engine. Modern EFI systems will perform similar to an MFI system in this regard.
2) MFI injects the fuel into the intakes immediately upstream of the intake valves This means that fuel doesn't have to flow through the entire intake track. Remember that in an ideal mixture, there are 14.7 parts air to 1 part fuel. So fuel makes up about 6% of the volume in the intake. As I mentioned above, a carb intake track will flow less air than a comparable MFI intake track. This is because downstream of the carbs (usually a length of about 10"), the intake tracks have to carry a mixture of air and fuel. The atomized fuel takes up some of the volume of the intake track, so there is less room for the air given a specific cross sectional area. You can increase the cross section of the intake, but this will than decrease the velocity of the air-fuel mixture, which will negatively influence the swirl and tumble in the cylinders -- which can mean less HP. Some of the fuel will also stick to the walls of the intake track, especially when the air or the intake track are cold. Less fuel into the cylinder means less HP. The MFI injectors on the other hand can impart some inertia on the air as it flows past the intake valve and into the cylinder. Once again, modern EFI systems will perform similar to MFI systems in this regard.
3) MFI systems inject fuel into the intake track at much higher pressures than carbs, or until recently EFI systems. Until the last 10-15 years (and the advent of direct injection), most EFI systems injected fuel at pressures of 30 to 50 PSI. A properly functioning MFI system would inject fuel at pressures of about 120 PSI. Higher pressures result in better atomization of the fuel, which results in better combustion and better throttle response and more HP.

How about Direct Injection systems?
MFI was actually originally developed for direct injection of fuel into diesel engines in the 1920's or 30s. It was than adapted for widespread use in gasoline airplane engines (non-direct injection) in WWII because it could increase HP (see above) as well as be less sensitive to the G-forces being exerted on the engines. This was predominantly (but not exclusively) an application by Bosch on German airplane engines. Bosch, Kugelfischer (I believe licensing the technology from Bosch), Bendex and others continued to develop the technology.
Than post-WWII, Mercedes Benz and Bosch applied the MFI technology of directly injecting fuel into the cylinders of gasoline race engines -- once again for the reasons mentioned above. It worked, but introduced other issues -- which is a longer story. Porsche (Bosch) and BMW (Kugelfischer) than started to use the non-direct MFI injection technology in the late 1960's. With the invention of electronically controlled, high-pressure piezoelectric injection systems, we have finally gotten to electronically controlled, direct injection gasoline injection systems which finally achieve the full performance that Mercedes was trying to accomplish with the 300 SLC engines. Fuel is precisely delivered directly into the hot cylinder, where it can atomize faster (also cooling the intake charge), without taking up volume in the intake track. It can also be shut-off on trailing throttle to save fuel and reduce emissions.

DFI needs the higher pressure because it works in a MUCH smaller time window. To get the required fuel in the combustion chamber at the right time is much more timing critical than with PFI used by Porsche’s MFI setup or the standard ITB EFI setup.

You don’t need any more than 45 psi or so with modern fuel injectors to get the necessary fuel atomization. Also, the newer ITBs are the direct to head type and the fuel injector is aimed right at the back of the intake valve.

The idea that EFI is slower than MFI is crazy. Both systems have to light up the spark plugs at a very precise time in the cycle. But, the window to get the fuel into the intake port is not small. EFI has no problem getting fuel into the intake at the right time every time. EFI being able to vary the fuel injected and timing based on conditions is an advantage over MFI, not a disadvantage.

My 3.6L air-cooled race engine makes over 119 HP per liter using modern EFI/ITBs with and inductive ignition system.

I get the love for MFI. But let’s not turn it into a religion and make claims that cannot be substantiated.

Just to clarify -- "EFI" covers a wide range of systems, everything from D-Jetronic and L-Jetronic up to the latest systems. Most people are not putting the most recent direct injection on their air-cooled 911s.

I did also say...