Initial siamese-port MSII code
Turbo Mini Forums

I thought it would a good idea to repost the information here to start the discussion.

I have gone through the MSII code and produced a quick and dirty initial version of a siamese port version using the MSII v2.33 code. It steals the ignition advance table to be used as an injection advance table. So the code will no longer support ignition but it has all the other features. It will require a single trigger per revolution which occurs about TDC of cylinder 2 (and 3) and the engine has to be configured as a 2 cylinder engine and injection advance set with respect to TDC of cylinders 2 and 3. Injection advance is for start of injection.

The reason behind doing it that way was to have a minimum of code change and keep most compatibility with the current tuning configuration file. Also, since there's a new version of the code coming shortly that will support wheel decoding, I didn't want to put too much effort into this initial version.

I have not been able to test the code since I have neither an MSII nor a Mini at the moment. So, if someone is interested in lending me an MSII, I could test the code since I have an MS v2.2 board, a stim, and an oscilloscope.

I don't know if I can attach files here but the people interested in looking at it or test it can get the source file, S19 file, and Megatune ini file at the link below.

http://www.msefi.com/viewtopic.php?t=14617

What the code does is pulse the injectors every engine revolution. That means that fuel is injected on a closed valve for the inside cylinders and on an open valve for the outside cylinders, as long as the timing in the advance table has been set correctly and the trigger has been set at the right place.

Since the inside cylinders open their intake valve first, they will get whatever fuel was injected in the port with the valve closed and then, the outside cylinders will get their fuel delivered when their intake valve is open. You don't need a cam sensor to do that, only a single crank trigger since on one revolution it's going to be inside cylinder on one port and outside cylinder on the other port and then on the next revolution it's the reverse. You can read more on this on Marcel Chichak's site. http://www.starchak.ca/

As for injector size, you'll probably need injectors about twice as big as you would need for a regular engine but that shouldn't be too much of an issue with MSII. Also, the code still keeps the dual table feature so you can use 2 injectors per port since all the injectors fire at the same time.

The way the advance table is used in the code is to set the beginning of injection at a certain number of degrees before the trigger. The engine is set as a 2-cylinder engine and the trigger should be set at about TDC of cylinder 2 (and 3). So for low speed and low load, you would set the advance to about 90 degrees. This would inject at 90 degrees ATDC of cylinder 1 in the first port (open valve) and 90 degrees BTDC of cylinder 3 (closed valve). On the next revolution, the injection will occur at 90 degrees BTDC of cylinder 2 on the first port and 90 degrees ATDC of cylinder 4, and so on. Then, as load and rpm increase, you would set the advance to something like 180 degrees resulting in injection at TDC for cylinders 1 and 4 and 180 degrees BTDC for cylinders 2 and 3. For even higher rpm and load, you will have to start even before to leave enough time for the injector to open and fuel to reach the cylinder and still finish with the valve open on cylinders 1 and 4. So you might have to set the advance to 270 degrees which will result in injection starting at 90 degrees BTDC for cylinders 1 and 4 and 270 degrees for cylinders 2 and 3.

I know this might not be clear at first and with only text but think about it and you'll see it makes sense. The tricky part will be tuning. Ideally, you'll need 2 (or 3) WBO2 sensor, 1 (or 2) for an outside cylinder and 1 for the inside cylinders. Then you'll have to tune the advance by watching the difference in AFR between the inside and outside cylinders and advance or retard accordingly. It's also possible that at high rpm and load you will not be able to get a correct AFR balance because you've reach the limit of the open valve window for the outside cylinder so you'll see it go leaner if you advance or retard. This means you'll have to go with either a bigger injector or an additional one with dual table.

If you have questions on my explanation, on the code, or on the logic please ask. If you question my implementation, have suggestions, or just want to shoot me down, go ahead. I am aware that what I have done is a very crude way of doing it but I'm waiting for the wheel decoding MSII code to implement a better solution that will also keep the ignition part of the code and give better accuracy of the injection timing. In the mean time, I think that this code is enough to start experimenting with the tuning aspects of this type of sequential injection. Unfortunately this can't done by me since at the moment, I don't have an MSII or a Mini. I will give all the support I can to anyone wanting to work with this.

Regards,
Jean

http://www.jbperf.com/

Good Stuff Jean.

In response to the thread over on the MS forums.
"First of all, I think that you might be over simplifying things by saying that port velocity is equal to VE. You're not taking into account the exhaust contribution in there which will affect VE but not necessarily intake port velocity linearly. I may be wrong but that would need to be verified."

It's entirely possible I'm oversimplifying things and maybe I don't have the VE table entirely pegged right, but it was my understanding that VE table is exactly what compensates for all interactions of cyl filling, therefore any scavenging or ramcharging effects are accounted for. I'm not sure why this wouldn't affect intake port velocity? Certainly it wouldn't be linear, but that's why the VE table exists. If it were linear it would be a singular value.

-Tyler
DO NOT TOUCH MY HORNS OF DOOM!

Thinking more about it, you're probable right. However, since we don't have constant size injection pulses meaning we don't necessarily start and end at the same time during the intake duration and since the port velocity is certainly not constant throughout the intake duration, I'm not sure we can or want to simplify things too much. But until someone goes ahead and try it, we won't know for sure.

By the way, I didn't think you were trashing my work on the MS forum. It's just that I'm anxious to see someone try this code to see if it can be made to work.

Regards,
Jean

http://www.jbperf.com/

Agreed... the only way we'll know anything is if someone sets it up and tries it...
...in that vein, a guy I work w/ has an MS that he's not using right now, and he said I could play with it. My mini isn't in a drivable condition(tho the motor runs fine), so it's a perfect time to test it out. I need to get some hard parts tho. I need to build an intake manifold and such, but I haven't had time to srouce any parts yet...

-Tyler
DO NOT TOUCH MY HORNS OF DOOM!

Something to keep in mind - is are you looking at using every last degree of injection window - or just accepting that a simple advance will do the job - and just use injectors (maybe) 5% larger than they need to be in order to avoid the charge-stealling....

I would not size the injector to use all the injection window unless you take into account every last possible increment the fuel computation might give you. And even then, I don't see the point. You'd be much better giving yourself some margin (at least 5-10%) and make tuning advance that much simpler and safer.

And if you use staged injectors, I'd go with even more margin since you probably can afford to.

Regards,
Jean

http://www.jbperf.com/

I'm not sure I quite follow your question, dave.
I believe we will be required to use the entire injection window, at least at high RPM. Simple advance should be enough to get the fuel in the right place at the right time. Proper fuel injector sizing will still be needed to get the proper amount of fuel in a small enough window and still provide a decent idle. I did the math at one point, and I think I came up w/ around 330cc for my 998 motor, but I've had so many other numbers in my head, I'll have to redo those calcs. But they were based on marcel's calculation of getting all the fuel in something like 30% IDC.

-Tyler
DO NOT TOUCH MY HORNS OF DOOM!

My concern w/ staged injectors is I believe we will need to use the available 2 banks for our 2 primary injectors(one on each bank).
Unless we are planning on using the router board, but that doesn't seem completely ironed out yet

-Tyler
DO NOT TOUCH MY HORNS OF DOOM!

All I was saying - is if 'just' advance alone can get you say 90% there, then it'd be much easier to accept this, and just use a slightly larger injector and 'advance'; Rather than a slightly smaller injector and struggling with a VE AND advance table to get the same (basic) end result...

It's the testing that needs to be borne out on a working engine of course; and probably one tested under load aswell as free revving.

So is your L&S dyno working yet Marcel?

Edited by TurboDave16V on 4th Jan, 2006.

mrbell,

If we keep the current method of a single injection event per 360 degrees, we only need a single driver for both ports and the second driver can be used for staging. I guess this relates to your post in the other thread.

http://www.jbperf.com/