Manifold Designs
Turbo Mini Forums

Hello folks, i've been following Paul and Rod with their testing in to the 5 port A series injection, which has made me start thinking of alternative solutions in terms of manifold designs.

Now, i'm not an expert at all, so most of what I say will probably not work in practice But I thought I would share with you my ideas anyway.

Ok, the first thing I would like to discuss is using seperators (baffles) within the inlet channels. This could be made from mild or stainless steel as part of the inlet manifold, which has a single full diameter plate on each inlet channel right down the centre, so to seperate the outer and inner cylinder air flow channels. You could then mount some slimline injectors side by side on both sides of the baffles to prevent any port robbing issues (in theory!). This is probably the least feasible of me ideas as the baffle plates would have to be custom made for each inlet, depending on weather the head has been ported or not, and the different heads that are available with the A series as well. I can do some paint sketches if it will help to present this idea.

My second idea is to simply use pencil spray type injectors, and angle them correctly from the manifold so that they are facing the valve heads of each inlet port, so to speak. The pencil type spray would then hopefully only enter the correct port as vacuum is present on either cylinder.

Both these ideas would not need the siamese type code as they would work in fully sequential mode, timed off of TDC.

I look forward to hearing the experts, and anyone else's responses on these ideas (please dont laugh though )

if you could get baffles in to the head all the way down to the valve, then in theory you will have a 7 port head, and injection is easy. Although the port size will be quite small.

Team www.sheepspeed.com Racing

turbo 16v k-series 11.9@118.9 :)

Denis O'Brien.

Hi,
Good idea, but I think there would be serious air flow limitations :(
Cheers,
Gavin :)

VEMs Authorised Installer / Re-seller. K head kits now available!

WB/EGT gauges. Click here for customers write-up

Visit www.doyouneedabrain.co.uk

My Mini build diary

The port separator has been done but it was at the expense of a significant loss in flow due to the resulting small ports (even extended to the maximum the head will take).

And the second idea won't work unless your injectors are big enough to send the fuel during the intake valve opening and from Paul's testing you'd need to have 2 injectors of about 1500cc/min to fuel a 200HP engine. With regular sequential injection (assuming you can time it correctly which may be difficult with a generic sequential ECU) you'd be able to use between 15 and 20% duty cycle from each injector instead of double that with the siamese code.

Even if you could manage to make the engine idle with those huge injectors (which is very unlikely), you'd still have to use 2 WBO2 sensors to set the timing correctly and check the AFR. So the only reason to go with such a setup would be if you really don't want to use a MegaSquirt ECU. And you'd still need to trim the fuel on the inner and outer cylinders because of wall wetting (which would still likely happen despite careful injector positioning).

But that's just my opinion. If you want to try either go ahead. However, since these are somewhat obvious and have been tried before and we don't see that as the solution, I think that you'd be better joining on what seems to be heading in a very good direction with Paul and Rod. Of course, there is still plenty of experimenting to be done in terms of manifold design (both intake and exhaust) but I think that the basic idea of the dedicated code has been validated.

Jean

http://www.jbperf.com/

Hi Jean, thanks for your input. With regards to the second idea, I do mean using 4 injectors instead of 2, on both designs, so I would have thought the load would not be as taxing on the injectors in this scenario? As apposed to having 2 very large injectors, which as you say would have great difficulty giving smooth idle.

It would have to be carefully designed as with the limited space around the port mouths and conflicts with exhaust ports, but I reckon it could be done.

Edited by Carl S on 11th Jun, 2009.

The 4 injectors will not be able to use more than the 15-20% duty cycle I quoted. Think about it: each injector can only inject when the intake valve is open for one cylinder. That's about 25% of the cycle and you have to remove some portion due to overlap and other factors. Even if you say that the factors which led Paul to assume a 15-20% duty cycle are limitations due to using only 2 injectors, it's still only a quarter of the cycle that can be used.

So even if you use 4 injectors, they still need to be 4 times as large as they would be for a "normal" engine. You don't gain anything over using 2 injectors. And with the siamese code, you may want to have 4 injectors but they will be used in a staged manner. If you want to use staged injection with normal sequential injection, you'd need 8 injectors (4 per port!!!) to get the same result.

Believe me, I have been thinking and reading about what can be done to do port fuel injection on the Mini engine for over 10 years and the siamese code has (or will soon have) all of the options that seem like a viable solution. I'm not saying that I know all that there is to know about it or that someone couldn't come up with something I haven't thought of or seen but splitting the ports or normal sequential injection are not viable solutions in my opinion.

Jean

http://www.jbperf.com/

The making a 7 port out of a 5 port idea has been done many times before. i believe that the manx 7 port head was just this.

As for flow, well on a full bore 1300+ turbo the ports probably are too small, but on a small bore the issue is not such an problem.

http://www.twin-turbo.co.uk
http://www.hillclimbandsprint.co.uk/default.asp

A man without a project is like a like a woman without a shopping list.

I think that we already have a solution to fuel inject a turbocharged 1275 to 200+hp and hope to prove that by Christmas.

Baffling the inlet port will severly limit air flow and power output.

Pointing the injectors directly at the valve is impractical in my opinion.

A well designed manifold with four injectors will do the job with the latest MS code.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Maybe then the first idea could work well for small bore applications?

Just looking at ways to bring it to the masses really, as apposed to achieving the most fuel efficient engines for coping with extreme power. I understand that Paul and Rod are putting in a fantastic effort to prove the best way of injecting the A series is with the Siamesed code, but I think this requires too much technical knowledge to get working correctly for each individuals engine.

If I could develop a standardised design along with a ECU kit, then I think there could be a massive potential market. The only question I have is: Would it be any better than the Rover SPi and MPi solutions?

If the answer is no, then I guess i'll have to join in with Rod and Paul :)

Developing an understanding of the siamese code and working setups is challenging and is not for the faint hearted.

Particularly the injection timing and VE differences between the innrer and outer cylinders.

However, the setup is no harder that normal VE tables or ignition timing. The problem is knowing what settings to use.

Hopefully soon I will have two engines to experiment with and Rod will also have his on the road. We can then build a set of standard setups that will be a good starting point.

However, with any injection/ignition ECU you will need to either rolling road it or invest time and money in instrumentation.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

With divided ports....how does the port area actually compare to the area that the intake valve allows to pass ?


More boost lol

9.85 @ 145mph
202mph standing mile
speed didn't kill me, but taxation probably will

My thoughts, pretty much as above...

I doubt you could achieve perfect seperation with a plate/baffle because the port profile changes down its length so getting a plate the right shape that would actually go in may be a challenge - although I suppose you could modify the port shape as well to achieve this....

The loss of CSA would be significant, taking the plate thickness into account, CSA would be less than half it was before. Even if this still isn't too bad compared to a valve curtain, it's for a MUCH longer length. But small bore, maybe not such an issue.

On the positive side, if you could achieve near perfect seperation, you could run "normal" injector duties and sensible sized injectors.

Having said that, Paul has already show you can have a smooth idle with fairly large injectors (when he doubled his up) and I achieved virtually the same last night when I got the timings about right on the siamese code (and I haven't even tuned the PWM parameters yet).

I think attempting to "aim" injectors at specific valves is a non-starter - they are just too far away......

With the siamese code, as Paul says, once there is more than one on the road, collectively we will be able to come up with some basic standards.

Schrödinger's cat - so which one am I ???

If a standardized setup is developed then tuning the injection timing would only have to be done once and could be re-used from then on. So the tuning for a new installation would only consist in tuning the mixture and ignition timing as for any injection setup. Of course that assumes the setup is really standard and that the injectors are large enough that there is some margin in the injection window for the inevitable differences in VE and air/fuel speed.

I'm biased on this but it seems to me that planning development with a known compromised setup is kind of pointless when there is an alternative right now. Of course more work needs to be done to really understand all the dynamic aspects but the potential is there and will only improve as more people get involved.

Jean

http://www.jbperf.com/

I'd imagine injection timing would change depending on airflow, and more important...cam timing.


Different cams will open/close at different times, and hence require different setups.

9.85 @ 145mph
202mph standing mile
speed didn't kill me, but taxation probably will

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

http://www.twin-turbo.co.uk
http://www.hillclimbandsprint.co.uk/default.asp

A man without a project is like a like a woman without a shopping list.

Of course, no matter what you do, working with a 5-port head will be a compromise.

It just seems to me that working without adding more compromises is a better thing to do. The software solution makes the most of what is there at the cost of more tuning time. Splitting the intake port adds more restriction to allow people to use a standard ECU and maybe save some tuning time. To me the best solution is obvious and I'd rather have the best efficiency but again, I'm biased.

Jean

http://www.jbperf.com/

Jean, I noticed you mentioned earlier about only around 20% of the time being available to inject a perticular cylinder with fuel at any one time, with the 20% roughly representing how much of 720 cycle the inlet valve is spent open. With a pencil stream type injector, are you not able to pre-prime each valve port with fuel before the valve opens, increasing the percentage of the cycle available? This is of course still assuming you can 'aim' the injectors at their retrospective ports inside the head.

Most of the fuel injected when the intake valve is close on the outer cylinders will be going to the inner cylinders but the opposite is not true. Due to the turbulence, any fuel that is in the air in the port will be ingested by the cylinder once the intake valve opens. And that will favor the inner cylinder 3:1 due to cylinder timing.

Also, the fuel injected will either be on the intake wall or in the air. The fuel on the walls will evaporate at a more or less constant rate and again this will favor the inner cylinders for the same reason mentioned above.

So you can't pre-prime each valve because the ports are not separate and most, if not all, of the fuel injected when the outers intake valve is closed will go to the inners. That's the charge robbing problem.

Jean

http://www.jbperf.com/

OK, that makes sense. Thanks for taking the time to explain.

I will continue to think about designs or revisions to current ideas and see if I can bring any new ideas to the table.