I think a simple coil upgrade should work for NA and light to medium turbo/SC applications (12psi or less). If your going to go BIG turbo, then a more robust system may be in order. Per RG, the stock 2nd gen coil packs are really stout, so maybe we could adapt those, since finding them should be relativelky cheap and easy.

Once we get the NA setup worked out, then it should be easy for others to figure out how much more juice they need. The good news is that what we need to do to test the idea really first just involves getting the engine to start and idle on the MS, we can do that without upgrading the coils. If we can get that far along, we are on the serious downhile slope.

 

Thanks for clearing that up. I have no desire to run any more PSI than 10-11... I was just going to size a turbo that doesn't get on boil till at least 5000 rpm. I'm just fine with it not being torquey down low...

Ok... I'll just hold my horses till we figure out the NA side. You two are way more knowledgable than I am when it comes to this type of stuff...

 

The above long post of mine from last night was really just a brainstorm. New info has surfaced today that actually makes some of the above wrong and further explains other aspects of it. I'll get into details later but for right now I will say that coding may actually be much simpler than I made it sound above.

 

Hey I'm a low level programming guy that still programs in assembly language. My day job is in computer games and I have programmed tons of vector unit code for the Playstation2, assembly on the PC, GameCube, XBOX, etc... Believe me, the only way we can get performance for some of the subsystems is to use assembly. If you need any help, then let me know.

 

Thanks for the offer, we might need that. I've been going through the code over the last few days getting up to speed with it. While I'm not a programmer, I've done some C++, VB and HTML work so I kinda know my way around. The good news is that the code is rather simple in terms of what you prolly deal with. RG and I are working out the way it needs to read the wheel in "Pseudo Code" first, then I'm going to look at the current options and see if I can write the code based on the examples in front of me.

If you visit the website, there is the MS firmware base code that is just the Fuel injection none spark version which is here:

http://www.bgsoflex.com/megasquirt.asm

Now, to activate the Squirt and spark option, you have to download the following code as well (Extra024s9.zip), and have it update the firmware (there is an asm file in that zip, that is where I'm focusing my attention on):

http://megasquirt.sourceforge.net/extra/files/

This is just the firmware for the box, not the megatune software GUI interface you actually use on your laptop to control the box. I'm still working out how the Megatune ties into the firmware becuase even if I get the firmware code correct, I still need to be able to activate and control the option on megatune.

If you have some time I would download megatune version mt225b514-024s9_setup.exe from here:
http://www.not2fast.com/megasquirt/mt/prerelease/

this is the version which is upated to handle the spark options.

 

Although I never got back into the details of how the crank trigger actually works in terms of how code reads it, this will still be an ongoing thing to figure out and program. for the time being, Alan has purchased a 36-2 crank trigger wheel. We are going to attempt to integrate this wheel as well as another sensor onto the car in ADDITION to the stock wheel. The stock ecu will still read the standard wheel and everything will work properly and the Megasquirt will read the other wheel. This will allow us to make progress and show that the engine can be run off of a different system. It will also gives us the ability to develop the tuning that we want to. Later on as the code gets written, we will just need to redo 2 wires, a setting change in the software, and remove the new wheel and sensor. This would be very easy. It's a way to keep making progress even though we are stuck for the moment on one thing. The wheel is not here yet nor are the harnesses so we still won't have any progress from a functionality standpoint for some time yet.

 

Does that mean you have found a sorce for the plugs?
Where are you mounting the new wheel?
I assume there must be some space between the two wheels?

 

Yes we have found a source for the harnesses. They only sell them in batches of 1000 though! With 5 unique plugs this is alot of money and plugs lying around. Not going to do that. Fortunately we have found a way to acquire a couple of each to at least get us through the prototype stage.

The 36-2 wheel arrived today!

 

Wow, this is old school code. Just like the old days

Definetly go down the pseudo code route first, because then you can easily hand off the algoithms to somebody to code in assembler if you don't get confortable with it.

I'll read through some more from the links you have sent.

 

As an update I thought I'd mention I ordered the wiring harness. Hopefully get that in a week or so. RG and I are going to try to mount the wheel this weekend if possible so were ready to go for when RG gets the additional MS parts and I get the harness in hand.

 

Our new wiring harness is on it's way!

 

The wiring harness arives tomorrow, so I can throw it on the car, run it through the stock ECU and make sure it works before we begin the process this weekend to tap and label it.

Its possible that in the next two weeks we can fire the engine up with the MS controling the fuel only, just to see if it will work and the other stock systems, DBW, radio, A/C power steering etc... still function.

Assuming that works, then I have to wait for RG to finish building out his MS for the Squirt n Spark feature and we have to rig the decoder wheel to the front of my main pully.

But hey forward progress.

 

An Update:

I got the wiring harness in today, I tossed it on the car to test it and make sure all the connections were correct, I’ve done some driving around and everything seems to work fine, I’ll keep it on for the weekend to be on the safe side to be sure that all the connections are working.

As it was raining earlier, I also began the process of labeling the harness wires so we’ll know what to tap, I think I’ve got all the wires labeled but I’m going to have RG double check my work to be on the safe side.

I plan on creating the other half of the harness that connects to the Megasquirt this weekend, it uses a DB37 connector (think parallel printer cable), so It should be easy to get the parts.

A couple of other notes and thoughts:

I’ve been pulling together the engine info I need to enter into the MS software such as displacement issues, injector size etc…I want to double check the injector sizing here that the primary’s are 290cc and the primary 2 and secondary’s are 380cc.

Since we intend to attempt a firing up of the engine under MS using fuel only to test out how the ECU handles not being in full control over the engine in a week or so, I’m thinking to be on the safe side I had better premix some oil just in case the oil metering pump don’t play nice. Any thoughts on this from the group? I don’t want to much oil in there due to potential carbon build up issues, but I figure I can burn it off with some hard driving.

More fun to come shortly.

 

From previous experience having installed a Motec on a third gen RX-7 street car, I've found that the oil metering pump "failed" in semi to full open mode. In my particular case, the stock ECU was left to power the OMP while monitoring the map and tps sensors only. Under that condition, the check engine light came on but the motor still received oil, while the M4 unit continued to control fuel and ignition.

 

Alan dropped off the harness with me today. The first thing to do is to get the intake temperature sensor wired in, the coolant temperature sensor wired in, and the throttle position sensor wired in. Then the car can be turned on and I can see what Megasquirt sees as far as reading the engine. If this all works out fine and the factory ecu hasn't done anythign weird (it shouldn't) then I will wire in all of the fuel injectors and completely sever them from the factory ecu. At this point an attempt can be made at firing the car up with the stock ecu controlling ignition and the Megasquirt controlling fuel. It's kind of like a 3 legged race in grade school. It may work well this way or it may fall flat on it's face. We'll see what happens. We'll know later this next week though.

 

Depending on how the MegaSquirt is manufactured, you may also want to consider the possibility of existing pullup or pulldown resistance on the MS sensor inputs, and how that situation might give the stock ECU some false sensor readings after you splice into the factory signals. You can check this by testing for any voltage (0, 5, 8, 12, or similar) on the sensor input pins of the Megasquirt and then seeing if there is resistance to any of the sensor reference voltage outputs. If any voltage exists, then you may run the risk of modifying the sensor signal going to the stock ECU after you splice into the harness.

 

@Gin, that was of some concern to me earlier in the brainstorming process which was why I was asking if it would be better to just add the cheap GM sensors and leave the factory sensors alone... Simple physics tells us that signals change when the circuit is modified... we just don't know if the change is too large to throw it out of spec...

 

I really don't think there will be a problem with it. Even if there is a slight change, it will be a very small resistance change with the sensors. All this will do is make the ecu think the car is at a slightly different temperature. Mind you I'm not talking about several degrees. The range resistance wise is so great that it will only be in fractions of a degree. Not enough to matter. I can also calibrate MS to whatever I want to and will ahve to since the default is for GM sensors.

 

the canzoomer device also tapped the ECU and had no signals issues I was aware of.

 

Here are some pics of the harness in my car being tested for to mke sure everything works.

 

This is great stuff! I will definitly be getting myself a megasquirt if this works out! Keep the reports coming! Do you have any plans of installing this on a turbo'd 8 if Brillo's car works out?

 

I don't personally plan to use a turbo, if I do anything it will be a supercharger, beit the Richard Paul or Pettit version, but there is no reason why you couldn't build your own turbo with this. really, all you need is a manifold, and you could build your own custom setup.

Turbo's seem like to much of a pain for a street car. Its more parts, weight, HEAT, and installation. But to each his own, we just want to get a system out so people can expand the FI market a bit

 

Fred, I got the shop manual today so if there is something in there you need just ask. In reading this thing I see that the system looks at the TPS and compares it to the MAF. Within limits that is. I can't find what the progaming is though.

Back to me, in my system I only need to cut into the spark. Would I have to go through the whole harness mess like you have there? I only want to figure the added mass and put in that much fuel. The O2 sensors will not pick anything up because the AF will be ok. The only thing is if the MAF sends signals that don't match the TPS. Actually it is overflowing and maybe the thing will pull the throtlle back. How can we give it a limited voltage?

 

This generation of ECU's do MANY things that you might not expect. For example, the ECU will sometimes use an aggressive fuel cut strategy on deceleration from partial throttle to learn the deceleration enrichment profile. Also, I believe the ECU will use different strategies to time when to cycle which injectors. Even the knock sensor alters its DSP profile based on 'inferred' engine torque at various loadings. There are also fairly long windows in which the ECU will try to relearn parts of KAM, some of which may never be set until ambient temps get in a certain range (say about 40 deg F).

I strongly suspect that without full control of fuel and spark, any FI system is going to have issues, primarily around the closed-loop areas of the map. You can get a glimpse of what the ECU is doing by reading the Ford ECU systems operations summary. http://www.motorcraftservice.com/vdi...f/OBDSM505.pdf

The ECU is really better than we give it credit for, we just don't like the way its programmed. Why not just go to a Motec M600? It would take a lot of work BUT I think the Motec could provide ALL of the stock ECU functionality, including DBW throttle, DSC/TCS, secondary and aux port actuators, 6 injectors, 6 ignition, gear-position input, etc.

I really think that unless stock ECU reprogramming becomes possible for individuals, FI is going to be limited by the stock unit's attempt to compensate. Greddy owners have to reset their long-term fuel trims every few weeks.

 

Just for comparison sake, Motec at around $5000 or Megasquirt at around $200. If the Motec gets a $4800 discount we may think about it. The Motec is awesome but very expensive.