5th Fuel Injector ?? The Wideband says the A/F is moving around quite a bit and is also a little high, There is a dead spot below 2psi of boost. If I go to 6psi, the fuel pressure may get too high for comfort. My plan is to augment the FMU with a 5th injector added into the charge pipe (between the IC and TB) angled so it points down the pipe. I'll control it from the Arduino blower controller (ie the "SuperController"). The SuperController knows boost, rpm, fuel pressure and throttle position. It also knows time from boost start and shutdown. I think that is all it needs to pulse width modulate the 5th injector to smooth out the FMU. Should be able to handle steady as well as transient situations. I'm currenly using only 1 of the 2 processors available on the ESP32, so the 2nd processor can handle the injector easily. Just need to add a FET to drive the injector and write some code (ChatGPT is very helpful to a hack like me). My thoughts now are to rely on the FMU for most of the fuel, and use the 5th injector to get the A/F close to perfect. The downside of the 5th injector (I think), is it won't get the fueling as even between cylinders as the FMU can. Also may have some wall wetting.

I eventually plan to add some failsafe features, like lowering boost if A/F ratio gets high.

Anybody done something like this before? Suggestions?

Also, one of the radiator fans is blowing hot air directly on the air filter, so I need to deflect that air and get cooler air to it. Its been around 75 deg around here (Florida) - the highest IAT I have seen was abouit 90 - I guess the IC is doing a pretty good job.

 

The 5th injector approach can work, but it's honestly a band-aid on top of a band-aid at this point. You're building a pretty sophisticated control system to compensate for an FMU that's fundamentally the wrong tool for the job.

You're basically building a standalone ECU fuel controller at this point. Between your SuperController managing boost, injector duty, safety logic, and data logging... you're doing 80% of what a Haltech does, just in a more fragmented way.

 

Thanks for the insight, and I totally appreciate what you are saying. If my objective was to get the best result with the least effort and risk, that is how I would probably go. But for me, its the journey that matters, and doing my own thing. I realized from the start that going a different direction from everyone else would be a ton of work and a little risky. But then I wouldnt end up with a totally unique s2000, where I designed and built as much as I could.

 

The electric SC setup you got is pretty unique on it's own.
These engines are not cheap to rebuild, I would 100% recommend using a standalone ecu to control the engine in a safe way.
That way you can concentrate your efforts to improve the rest of the setup instead.

 

I have the design of my diy fuel system just about finished and I'm ready to build it. I think it will work well up to 6psi. It has multiple failsafes. But.... before I start, can the experts chime in on the best replacement ECU options. I really want to know 3 things: 1) what does it takes to get the unboosted driveability back to stock. What do I lose in normal use. I'd like to meet emissions and not lose anything if possible. Is there a perfect base tune, if not what is involved? 2) What is involved with the install? Changing sensors, injectors, wiring, cost, etc. 3) How much hastle getting it all working perfectly?

Please chime in!!

 

You mean perfect stock tune? Not base tune. Base tune means a suitable starting point for a tune, that is safe, but far from perfect.

 

Definitely no expert, but have taught myself and self tuned my AP1 successfully.

1) "Back to stock" can take some tinkering and therefore time with regards to things like starting and idling like stock. Perfecting part throttle can also take some time. Getting it close and getting it perfect are definitely two different investments in time. Emissions vary from State to State, so hard to answer this but some people tend to swap back to the OEM ECU and injectors to pass emissions. You need to understand how to apply basic tuning theory through the software your chosen ECU utilizes. People tend to have preferences in the software, but all are ultimately doing the same thing. I'm currently using an AEM Series 2 and find the software extremely user friendly. There's HP Academy which has a lot of classes available and some YouTube videos which can be helpful if you're interested in learning tuning theory

2) Some options are more plug and play than others and pricing varies widely. There's always useful sensors to wire in even with the plug and play options, you should at a minimum wire in your wideband o2

3) Depends on your definition of hassle and perfectly. It can take a time investment for perfection so you definitely need to expect that especially if you've not worked with a standalone ECU before. Dyno is always recommended, but it still will take some time road tuning to perfect everything.

Popular option right now is the Link, Haltech also has the Elite and new Nexus ECUs which are nice and have plug and play harness available. The kpro is going to operate most similarly to the OEM ECU, and has some advantages because of that like starting and idling. But it also has some limitations when compared to a full standalone ECU.

 

Thanks - that's a very helpful summary

 

I don't think I like the FMU, at least not a 10/1 pressure ratio. The fuel pressure gets too high and it seems a little flaky. I plan to try a charge pipe injector, and if that doesn't look good, I will probably give in and go new ECU. However, I plan to test the CPI thoroughly before putting it on the car. I'm an engineer, and I like lots of data. So here is a look at my first test:

The test case uses a Bosch EV14 compact injector. I used a 360cc injector with a cone spray, driving it with an Arduino + MOSFET that pulse width modulated the injector. I injected gasoline at 50psi into the air stream of a leaf blower. The picture shows the result. The spray pattern is much more tightly focused than I anticipated. It will shoot 10 feet through the air if I point it into free air. (I did this outside in the wide open, with a fire extiguisher near!) I also tried holding the charge pipe so the leaf blower blew the gas into my charge pipe. It pretty much blasts right across the charge pipe, hits the other side, and splashes liquid gas everywhere. When it hits the charge pipe, a lot of the mist turns into large drops of gasoline so that liquid gas runs out of the end of the charge pipe. I tried all angles between perpindicular to the flow, to shooting it almost straight down the pipe. When it shoots straight down the pipe, it remains a tight stream, until it hits the 90deg elbow, turns and then splashes out toward the throttle body. When it turned the corner of the 90, it hugged the outside and entered the TB as a narrow stream of spray + liquid. I tried spraying into the flow as well. The flow seemed to disrupt the spray, so when it came back the other way, the drops were larger. Perpindicular looked to be the best, as the shear from the air flow helped disperse the spray, But any way I tried it, quite a bit of the gas ends up liquid coming out of the 90deg bend to the throttle body. I never did get a mist that came anywhere near filling the charge pipe. Bosch makes EV14's with wider spray patterns, but I couldn't find any that were much wider. My conclusion is the EV14 will not work well enough.

I read somewhere that the tight spray pattern of the EV14 is typical of modern injectors - they aim directly at the intake valve and let the hot valve do the atomization. Older, pintle style injectors apparently generate a finer mist. I may try one. The problem is they are a lot larger than the EV14, and my space is limited. I understand that old throttle body injectors may also be better. I also am interested in trying something called a "Fuel-It" charge pipe injector. I contacted them for more data.

Any suggestions? (besides "change your ECU" I'm not quite there yet!).