Porsche 12 Cylinder Fuel Injection

[dieselpilot]

None of those disposable drones are running EFI, carb is adequate. For an off the shelf EFI I was thinking megasquirt, microsquirt, speeduino, rusefi. TPS, RPM, head temp, and intake temp, will be enough. Some experimentation or measurements will be needed for a VE map. The biggest problem is off the shelf injectors are too big.

I knew nothing of the Lucas system. I now understand the shuttle metering valve is part of the distribution valve. Here I was, thinking I had dreamed up a rotary distribution valve. The metering distribution valve is fed fuel at over 100psi and the injectors open at 50psi. The single shuttle simplified example is for a two cylinder engine. There are a few basic designs it seems. Some run at 1/4 crank speed.

There's a 12 cylinder system at the bottom here. Two single shuttles in the rotor, and 12 outlets around the sleeve. Anyone have a metering unit for a Maserati or Ferrari laying around?

https://www.powerprops.com/race-systemen-te-koop/

[steamer]

Nozzles came.today   I'll open one up a.d look it.

Dave....about to deal with 2 feet of snow...

[petertha]

Probably yet another miniaturization challenge (or compromise). But how did the mechanically driven injectors back in the day handle 'accelerator pump' mode? ie. throttle is blipped relatively quickly, so instantaneous increased air flow through carb throat, so ATF mixture is temporarily lean. I'm sure they could measure manifold vacuum change, but how would a pressure signal alter mechanically actuated injection plungers to call on more fuel squirt during that demand period? I'm sure Dave wants to rev his model like the big boy! LOL

[steamer]

Wow Just read up on the latest threads!   It's amazing what has been done, and what technology has been doing!    Some of the ESC controlled units sound like old home days!   ( spent 6 years deep into RC car racing.....used and abused a lot of ESCs)

I want to investigate both mechanical and electronic ignition and see where they are and where they are going.   So Where to begin?

Well I have a great challenging project on the go!  It needs to have the block finished next and I dare say about another 60 percent of the engine to construct, debug, and hopefully test.   but I need to be realistic on my milestones.  Or I'll never get it done

So here's my plan

Im going to continue to have these dicussions on FI here, mechanical, and EFI, and continue to monitor what's going on with EFI as it's an evolving niche technology at the moment.

I'm going to continue the engine build all the way to a first run test.  For that I will probably build a simple dual carb 6 x 6 manifold to test the engine on, get the basics operational, and running on methanol and glow plugs.....this has been the plan thus far

A running engine even without the prototypical manifolds is a great milestone to direct further effort.

From there I think I need to either. Make a 2 carb mock 12 stack injection unit that at least looks right, or roll up my sleeves and jump on some version of the fuel injection system.   I don't think doing the mechanical one right now is the right thing to do.  I think my skills need to improve, and some engineering needs to happen first on that.   That said, it will be a good point to start as building a mech FI system will be as big a project as the engine is!

If I can get to that part, I'll get into a pair of distributors, some new FI nozzles for race gas and then we'll see...

That is as far as my crystal ball looks so far....or it could just be the oncoming train!

"Aint scared!!"

[jcge]

But how did the mechanically driven injectors back in the day handle 'accelerator pump' mode? ie. throttle is blipped relatively quickly, so instantaneous increased air flow through carb throat, so ATF mixture is temporarily lean. I'm sure they could measure manifold vacuum change, but how would a pressure signal alter mechanically actuated injection plungers to call on more fuel squirt during that demand period?

For normally aspirated racing purposes, Porsche removed the thermostat, enrichment solenoid (choke) and shutoff solenoid. That simply left the flywheel governor (engine speed), throttle position and barometric pressure as the controlling variables. The space cam (raumnocken) providing the "map".

Presumably there was some small lag in manifold air pressure on increasing throttle position which effectively negated the need for an "accelerator pump"

Control diagram for the EPRLA model regulator for reference.

Regards
John

[jcge]

Dave - the induction and injection arrangements and their possibilities is a rabbit hole for sure!

Sounds like you have a good (actually great) plan to progress your model engine.

Regards
John

[Vixen]

Probably yet another miniaturization challenge (or compromise). But how did the mechanically driven injectors back in the day handle 'accelerator pump' mode? ie. throttle is blipped relatively quickly, so instantaneous increased air flow through carb throat, so ATF mixture is temporarily lean. I'm sure they could measure manifold vacuum change, but how would a pressure signal alter mechanically actuated injection plungers to call on more fuel squirt during that demand period? I'm sure Dave wants to rev his model like the big boy! LOL

Peter,

You have hit the petrol injection nail square on the head.

A Compression Ignition (CI) engine's speed and power is controlled only by varying the amount of fuel delivered. A Petrol Injection (PI) we can also read Methanol Injection, is controlled by the amount of AIR and FUEL mixture delivered through a carburettor or by injection. The air throttle and the fuel injection halves need to work in perfect harmony to deliver the correct amount of fuel in all throttle/ speed/ load conditions.

The Schilling injection drawing was completely devoid of any air flow/fuel flow control. In fact, there was no consideration at all given to the air side.

You are quite right about the 'accelerator pump mode', but that's only part of the equation. Perhaps just as important is what happens when you snap the throttle shut. The throttle cuts the air supply and the controller must in response cut the fuel. otherwise the engine fills with unburnt fuel. The fuel air mixture needs to be tightly controlled at all points between. That is what make PI so much more of a challenge than CI.

You asked about how they did it 'back in the day'? Well .....BMW developed this 'Kommandogerat' for the PI on their huge BMW 801.



It's the mechanical equivalent of the fuel/ air map within an electronic EFI.

Read Calum Douglas's book 'The Secret Horsepower Race'. it tells all about fighter engine developments during WW2 and deals extensively with German mechanical fuel injection development.

Mike

[Vixen]

Dave,

I am sure the step by step approach is absolutely the correct one.  Start with carburetor(s) on a non-scale manifold then migrate to single point EFI until you are in a position to go multi-point to refine the looks.

Remember Keith had a single injector feeding all cylinders on his incredible V10. Single point injection using Megasquirt  EFI development kit (or was it a Microsquirt) from the same people. He built the engine and developed his own EFI.    Keith has an excellent build log here on the MEM forum.
Who says it cannot be done?

Cheers

Mike

[steamer]

Dave,

I am sure the step by step approach is absolutely the correct one.  Start with carburetor(s) on a non-scale manifold then migrate to single point EFI until you are in a position to go muli-point to refine the looks.

Remember Kieth had a single injector feeding all cylinders on his ineradicable V10. Single point injection using Megasquirt  EFI development kit (or was it a Microsquirt) from the same people. He built the engine and developed his own EFI.    Keith has an excellent build log here on the MEM forum.
Who says it cannot be done?

Cheers

Mike

Step by and incremental progress.    Yup!

I'm a firm believer of the 11th commandment

"Thou shalt not kid thyself!"

[steamer]

Lots of useful and interesting information     

As we are not looking at passing emissions tests or achieving maximum power the actual mixture control is not too important. It would be nice to have it accurate enough to use petrol and spark ignition. That would also give Dave the chance to enjoy building a couple of 12 cylinder distributors 

As has been said before Schillings used the wide air fuel ratio tolerance of methanol and glow plugs to simplify things.

If more than one injector is connected to a metering device ensuring equal flow to each injector is difficult. This is a problem for all the continuous flow systems. Bosch used some very precise machining on the K-Jetronic system and Schillings used individually adjustable jets and measured the cylinder temperatures with a finger to get some sort of balance.

Timed injection pulses are not actually necessary unless working with stratified charge combustion, but the individual metering systems required makes balancing the injection quantities somewhat easier.

The 6.4cc cylinders will work with a 2mm bore injection pump. My 12cc twin is running with such a system.

<a href="https://www.youtube.com/watch?v=FJ6EHk4KvJw" target="_blank">http://www.youtube.com/watch?v=FJ6EHk4KvJw</a>

Can a mechanical pump and injector system be made that is reasonably to scale? Guesstimating from the drawings the inlet trumpets look to be around 10mm diameter (half the cylinder bore) and the injection pump is around 25mm wide (a bit more than the bore diameter) and 40-50mm tall.

I think, based on my current versions, that a poppet injector with an external nozzle diameter of 3mm using M4 x 0.4 threads for the body is practical. I am using 1mm bore 2mm od copper pipe for the injection pipes. We would have to look for a suitable fuel resistant hard plastic pipe of similar dimensions as well as a suitable way of fitting it to the unions. The smallest injector in the picture has an M1 thread. They all have 5mm outside diameter to fit my current engines.

The injection pump could probably be simplified from my diesel based versions as the required pressure is only 3-6 bar. 12 individual plungers with appropriate adjustments to balance the quantities would not be easy in the available space. The Lucas shuttle pump has possibilities. Two 6 cylinder pumps would be easier than one 12 cylinder unit.   A 2mm diameter shuttle is ok (cut from a pin gauge?). The rotating sleeve could be 6mm diameter with a lapped 2mm bore for the shuttle. This just has two ports. The fixed sleeve has the inlet and outlet ports with a lapped 6mm bore for the rotating sleeve.

All timed injection systems automatically adjust for engine speed so just the throttle position needs to be considered for the fuel volume control. A simple profiled cam would probably be sufficient (this is how my engine is set up) although an electronic system could be used. When used in competition engines the Lucas shuttle system was usually mechanically controlled by the throttle position as it was not possible to get a sensible manifold vacuum signal with individual inlets and high lift long period cam profiles.

You must help me with the force oh obiwon...   we be talking about these engines for sure

[dieselpilot]

The Lucas system simply didn't have acceleration enrichment and had to tuned rich enough or driven accordingly.

[Roger B]

Well as I have said that I think a mechanical injection system is possible I need to make some trials 

I have drawn up a smaller version of the 1mm bore poppet injector used in the last trial. It looks feasible, however I need to consider the much lower pressures used in the Lucas PI system. To achieve an injection pressure of around 3-3.5 bar the spring compression, with the springs that I have, is around 0.2mm. The M1thread has a pitch of 0.25 mm so this should be OK.

I will try and make a two cylinder shuttle pump, which can be scaled up to 6 cylinders. A twin shuttle 12 cylinder pump is beyond my current abilities and equipment. This will need some additional undersize reamers and laps the produce the 6mm rotor to sleeve fit. I have all I need for the 2mm shuttle. I think that the stubs on the shuttle stop and stroke adjuster will need to be ground, but I can use the technique I developed for the needle injectors.

As this will require a fuel pressure on around 7 bar I need to consider the safety of the trials. A small all metal fuel tank and metal to metal pipe connections are a must. The fuel pressure pump (gear pump) will need to just be enough to reach the required pressure and flow and the pressure control/relief valve will need to be leak free. I may consider a commercial item but it must be petrol/alkylate compatible.

Lots to think about 

[Roger B]

The Lucas system simply didn't have acceleration enrichment and had to tuned rich enough or driven accordingly.

Injection tends not to need an accelerator pump. When a carburettor engine is idling or running at low speed the fuel is fed via the slow running system. If the throttle is rapidly opened the manifold vacuum drops to zero and the fuel flow stops as there is no suction. Fuel needs to be supplied until the gas velocity increases enough to produce a vacuum on the main jet system. This is not a problem for a pumped injection system. There is an additional effect that as the manifold pressure increases (vacuum decreases) some of the fuel will condense out onto the manifold walls further reducing the mixture strength.

[petertha]

Ah! (head slap). The main 'leaning out' issue is probably because the carb venturi throat has lost much of it vacuum under sudden acceleration demand & thus ability to suck fuel through the jets during that transition. RPM based injectors would not have to deal with this, they are no wiser to the state of vacuum. Probably not as simple as that but probably explains a lot?

[Vixen]

A Compression Ignition (CI) engine uses the Diesel Cycle, its speed and torque are controlled only by varying the amount of fuel delivered. A Petrol (methanol) Injection (PI) uses the Otto Cycle where the its speed and torque are controlled by the amount of AIR and FUEL mixture delivered through a carburettor or by injection. The air throttle and the fuel injection halves need to work in perfect harmony to deliver the correct amount of fuel in all throttle/ speed/ load conditions. A carburettor does that automatically by air flow though the venturii past the fuel jet.

How are you proposing to get the air side to control the fuel side?
Mike