Project Information
Turbo Mini Forums

Hello there, I have recently decided to make my mini a little bit more interesting.

I would like to remove the carburettor system and I would like to replace the system with injection, also the ignition will be improved, which means the distributor will need to be blanked off. The pick up for the ignition will be driven off the crank (36-1 trigger type). Both the blanking plate and the trigger system will be designed on solidworks. Additionall as part of the project a manifold is going to be created witihin solid works, I was just curious if anyone had any good information about what the manifolds requirement is, maybe even some dimensions if possible (had no luck searching myself). The injectors position will be experimented with once the model has been produced, however the position in the inlet tract will be as close to the inlet valve as possible. I assume the injector should be positioned as centralised between the two inlet valves to allow for equal distribution for each cylinder, I am going to continue to use my current five port head. The injector sizing has been calculated to be around 220cc per cylinder, which means that each inlet tract will house a 440cc injector. The only way to accurately measure the amount of fuel entering the engine is to ensure the injector coincides with an open inlet valve, so the engine timing also needs to be concluded, i have read about more aggresive cams reducing the effectiveness of this kind of project due to the injection window being decreased further, giving less time to accurately inject.

Other things which I have calculated so far:
The fuel pump will need to deliver 52 litres per hour, a few assumptions were generated to provide these results, VE 110%, Max rpm 6800, AFR 11.5, Fuel density (STP) 0.74, Air density (STP) 0.0012 and the safety margin for the fuel pump (25%) so it does not run at 100% duty cycle.

I am still curious about the anticipated pressure the fuel rail will need to be regulated to, also the volume of the fuel rail. If the volume is incorrect I assume the injection pulses will decrease the overall pressure and the following pulses actually inject less fuel than desired (if the length of the pulse does not change).

If anyone can share some light on any of the areas meantioned before I start the work I will be very thankful.

James Aston

Have a search for posts by Paul S and Rod S. They have put together a guide to injection, uploaded manifold designs and much more.
Are you going turbo too?

1275 N/A Sprite, 998 T2 Turbo Mayfair
1275 EFi Turbo

Cheers, there is such a large archive I have been struggling to find the right thread.

Currently using a t2 turbo, this will remain the same. Its for a university project, just want to understand manifold geometry and the internal flow due to the pulsed nature of the flow. Done alot of reading on the pressure waves generated. I would like to incorporate some analysis of the flow of both manifolds.

If I have any further queries i'll just repost on this thread.

Which ECU are you planning on using? And you might want to reconsider using the distributor as a cam phase sensor. You don't need to use the full housing and you'll see examples in other threads but you will need some sort of a cam speed sensor to have the complete engine position and inject on an open valve. Semi-sequential injection has been shown to not be even close to optimal.

Jean

Edited by jbelanger on 21st Mar, 2013.

http://www.jbperf.com/

Just a very quick answer at the moment....

If you say you want the injection timing to co-incide with inlet valve openning, that is what we would call the siamese code on an MS2 or any other code that can vary injection timing between the inner and outer cylinders.

That means we attempt to get the fuel through the open valves so we can vary the amount of fuel between inner and outer cylinders and hence overcome the normal "charge robbing" of the A series.

But that means a lot of fuel has to go in in a very short space of time. Normal engines will be injecting fuel a long time before the inlet valves open.

The Rover MPI used 440cc injectors (maybe 480cc, I can't remember for sure) but that was normally aspirated. If you have a T2 turbo you are going to need much larger injectors (or pairs of injectors) to significantly increase that capacity if you want to time the injection on open valves (I currently use 1000cc).

And if you want to time injection, you will need to build a phase (cam) sensor into your system as the system will need to be fully sequential (easiest way is to modify a dizzy body).

For manifold geometry, search against Paul S, I tend to stick to just the electronics/ECU.

Sorry it's a rushed relpy, trying to make use of the first dry, snow free day we've had for ages....

EDIT - Jean beat me to the phase sensor bit, I type very slowly......

Edited by Rod S on 21st Mar, 2013.

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

it meant t3 not t2 :$

it meant t3 not t2 :$

Thank you for the prompt replys

Not entirely sure on the ECU yet to be honest, probably megasquirt and megajolt for the ignition. i know they are not the best but for what i am using them for they should be fine.
What sort of fuel rail pressure should i expect to use with 1000cc injectors then (3 bar)?
I will be using lambda sensors on two of the three exhaust branches. this should enable me to gain a finer control of AFR readings.

Will you be able to direct me to the information for the phase sensor, sounds like an interesting method of controlling the injection pulses.

Cheers

You don't need megajolt for ignition if you are using megasquirt (any version).

Megasquirt (MS2) is probably the only ECU that has code (written by Jean) specifically to run sequential timed injection on an A series engine (as well as ignition) specifically to target valve opennings.

Other ECUs can achieve the same by different methods but the MS2 (siamese code) is the only one written (and proven on this forum by meausuring individual exhaust AFRs on a turbo) specifically to do it.

Not sure what you mean by measuring AFRs on the downpipes, if you are using a turbo (you said T3 not T2 earlier) there either won't be any downpipes, or they will be too short, and anyway, the wideband LSUs don't work properly under positive pressure.

Where are you based ??? (fill in your user profile).

Edited by Rod S on 21st Mar, 2013.

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

Ahh ok, that clears a few things up.

I see what you mean about the limitation due to space, would be nice to have a method of control though.

Would be nice to start modelling as the dead line is now two months away, also need to do a 15,000 word report. Not looking forward to that!!

Glad to hear that the coding is available to be honest, I think i could set a base map however I think coding the injection and ignition would be a little too complicated for the time frame I have to work with.

I had considered variable diameter, however the electronics alone are a dissertation, the main reason I researched this was to try and improve the port velocity over the whole range.

The space limitation of widebands before the turbo on a short exhaust manifold isn't really the main issue.

The wideband LSUs (sensors) would be well outside their temperature rating pre-turbo (but would probably survive long enough for your project) but, much more importantly, the accuracy of their output is very pressure dependant. On a normally aspirated engine the sensors will see little variation in gas pressure (unless the exhaust system is really restrictive) but on a turbo installation, if they are before the turbine, could see up to 3bar variation between on boost and off boost because of the pressure needed to drive the turbine on boost. Those of us running multiple widebands pre-turbo have devised sampling chamber systems so the LSU is exposed to the exhaust gas pre-turbo through a small sample tube then vented to the exhaust system by a much larger tube connecting after the turbo, so never see any pressure extremes. Without doing this, we believe the readings are meaningless.

The phase sensors we use are either hall switches or opto switches built into cut down dizzy bodies apart from Paul who has used a hall switch pointing at an extended bolt on the back of the large timing chain wheel. Anything that measures cam rotation once per cam rev.

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

I have decided to re-design the plenum as well so that i am able to keep both the tracts seperate from the plenum to the engine. The plenums overall volume is yet to be confirmed, however i have read that i should use a 110% - 120% displacement of the engine. Making a 1.275 litre capacity 1530cc at the plenum.

http://www.geocities.ws/MotorCity/Factory/...1380turbo4.html

The volume for his plenum should be slightly larger due to his engine having a slightly higher overall volume.

Would you be able to show me a photo of the sampling chamber, that is quite an interesting concept. Never heard of that kind of control.

I am going to be using hall effect as I have more of an understanding of its duty. Its been a long time since I did trigger types at college.

I will get back to you in a few days, i'm a bit snowed under with various tasks at the minute.

Attachments:

178.39kb

The plenum being about 130% the engine capacity is a common value, can't remember where it came from, so 110-120 is in the right ball park.

Sample chambers weren't my idea, but they work...
They overcome the temperature and pressure effects of just connecting into the inner and outer exhaust runners before the turbo.

My version





If you search against paul and Graham you will find other ways of doing it but the principle is the same, small sample tube to the LSU from each exhuast runner before the turbo and a large pipe to lose the pressure down into the downpipe after the turbo.

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

Cheers again for the explanations, your really helping me get my head around some of the more complex tasks.

I am relatively happy with the plenum now, only problem been I need to go home so I can physically get some measurements so the dimensions for solid works are spot on.

I calculated the runner length to be 35 centimetres using David Vizards calculations. The runners diameter is calculated to be 39.2cm

If i am right the inlet port is 35mm so i will be able to incorporate a slight taper so that i am able to bump the port velocity up just because it enters the port in the cylinder head.

I also still need to think about the injector type, what kind of injection nozzle would be advisable?

Do any of you have a Metro Turbo Injector conversion that i would be able to come and look at, currently in Derby however would be prepared to travel, I do have my own car.