injected N/A engine
Turbo Mini Forums

Hi, I am new to this forum.
I have a mini engined sprint/hillclimb car and would like to fuel inject it.

I have heard, and thinking about it, believe it is difficult to inject a siamesed ported A series.

For maximum performance I would rather use a twin port setup.
My question is, has anybody done this successfully and if so how (which ECU, injectors type, position etc).

My idea is to use a twin injector per port setup driven sequentially just as a conventional port throttled 4 cylinder and you can therefore phase the injection within the cycle correctly.

Many thanks

tubo dave's written a couple of really good articles on injecting the a series frequently refered to on this website.... do a search and you may find some of your questions will be answered.....

hope thats helps!

Pete

And on the 7th day........... God created turbochargers!

To add to what Peter is saying, you may find references to people claiming to have successfully done port injection with a standard ECU. However, if you look more closely you'll find that their setup is either one where you effectively have a wet manifold setup due to injector placement and manifold configuration or they have a severely limited performance results where you can't go from good idle to max top end without changing tuning parameters.

The problem is that even with sequential injection, mainstream ECUs are not setup to have injection timing which can vary with RPM and load and, since you have very small injection windows, you will need 4 massive injectors to make reasonable power. With regular sequential injection, if you can time injection correctly you have about 20-25% duty cycle available for injection while with the ECU we're developing, this can be doubled. So you would need injectors 4 times as big as what you would choose for a "normal" engine which will make idle and low load almost impossible to tune.

Even with the 40% IDC we can use, we still need injectors twice as big which could still cause problems. However, with 2 injectors per port, we can use one small and one big and stage them to cover the entire range efficiently. That's something you wouldn't be able to do with the setup you propose unless you would use 4 injectors per port which you would have fun fitting there but would be interesting to see.

I'm hoping the development will be finished this summer for a good working initial version. Additional options and features will be added later depending on testing and needs.

Jean

http://www.jbperf.com/

Thanks for the replies. It all sounds very interesting.
I have used ECU's in the past where the injection phase is mapped for speed and load but it was a long time ago and I can't remember which one. It could have been a version of Zytek or Bosch.
When you say WET what do you mean? Is it that the injectors are placed a long way from the valve? If so why is this a problem as you would get very good mixing and good charge cooling.
The obvious problem is the fuel hanging around in the valve closed period and all flowing into the first cylinder to open.
Am I being stupid (probably) but why is this so different from twin SU's

Easy way to think why it's different:
With Carbs, the engine takes in the fuel when it wants it.
With injection, you have to figure out exactly when it wants it at all rpm sites, and get it in there at the right time.

Another thought.
If the injectors were small and the duty cycle >80% at max power this must be similar to a carb with homogeneous mixture when mapped accordingly.
This is only part of the story since we need to map the full speed/load range.
The other thing surely is to make a cylinder fuelling correction but unless this was mapped against speed and load it would be pretty much useless.
Sorry for going on a bit I should read all the other posts before about this subject as I bet it has all been said before.

Start with reading this:

http://www.turbominis.co.uk/forums/index.php?p=vt&tid=6905

There's lots more in this section.

You need to fully understand the serious port robbing issues involved with injection the A Series engine.

Read and then you will understand why if anyone will get it right, these guys 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."

You definately need to read - especially if you're talking about 80% DC...

If you're talking about single point injection then you can use 80% duty cycle. You will then have an electronic equivalent of a single carb.

In that case, you no longer need to worry about injection timing but you have to inject 4 times per engine cycle (twice per rev) to have a suitably homogeneous mixture. This will reduce the effective duty cycle due to the number of opening-closing cycles but can be reduced by using 2 injectors in alternating fashion. This way you only open the injectors once per rev and their opening times can overlap.

This should be adequate but will have more wall wetting than port injection leading to less precise fuel metering for each cylinder due to port robbing. This is a simpler and cheaper solution that could be done with a standard Megasquirt and, since your engine is N/A, could be done with 2 reasonably sized injectors. You would basically have a better (IMO) and tuneable SPi.

Jean

http://www.jbperf.com/

Thanks for the link.
I do appreciate the fuel robbing problem as this was my first thought. I guess that twin large injectors per port are the way to go, driven sequentially with a 3d phased map facility.
With the injectors placed just after the throttle, (i.e maximum distance fro the valve in my instance) I would think this would give the fuel more time to evaporate and desensitize local high pressure fluctuations which could cause maldistribution. The instantaneus pressure near the throttle at WOT must be near atmospheric all the time whereas near the valve it will be all over the place and different on each cylinder (which is why, presumably scatter cams work)