On Thursday I've realized that I should demo this at a Lemons race next weekend, and yesterday, 48 hours later, I've started engine #3 - that's the race can.
Not yet in shape for 14 hours of racing, but getting closer :) Oh, another thing worth mentioning - In this uber-minimalistic version the ECU board + brain board are really pushing the $100 threshold.
https://www.youtube.com/watch?v=t2y3h8rGKXQ
http://www.autoweek.com/article/20140424/CARNEWS02/140429923
And engine #4 is a Miata: https://www.youtube.com/watch?v=5wkH4HVtc50
With four test engines, now I really need more developers to get involved to get this useable.
Since you are open code, feel free to ask questions here vs. your forum.
Your next goal is to make a speed-density set up. So you have speed, MAP, and air temp. Realistically, you should do a combined actual intake air and engine temp- that's what really happens between the manifold getting hot and backflow from crossover.
S-D vs. MAF- there's a good A v. B debate there- both have big advantages and disadvantages to the other. Up until 2007, all of the development work I did was on MAF, but it needed a calculated SD correction for transients. Then we moved to making turbos, which changed the code to be S-D, which does work great. That's a pretty easy calculation, BTW- can be described as a simple quadratic- AX^2 + BX + C, which you can have different values for different engine speeds.
You'll want transient fuel compensation, somehow.
Oh, and I briefly looked at your forum. Yes, O2 sensors are fast- really fast. But that's not the reaction time that's a pain- there's a big time lag between the last fuel calculation to where it injects to where it burns to where it's exhausted to finally being measured. The O2 sensor is far better used as feedback correction.
And since you are just getting going- base O2 with heater is pretty easy to put in, but you should already integrate WB sensors. They are cheap enough to use on new cars + for good tuning, they are pretty much required. (and to extend that into a really effective but simple emissions controls is easy) Closed loop ALL the time when it's running, just need a good control system that reacts to the sensor.
Thinking about it a little more last night- the solution that the MS uses is probably the easiest. Use the ideal gas equation for a base amount of air based on the MAP + temperature, and then use a correction for that- they call it volumetric efficiency. Very elegant, and easy.
Combine that with a flow based injector calibration, and you have a strategy that is portable across engines. And is VERY easy to program.
That method of calculation is in all of the text books, so it's not proprietary to MS what so ever.
The key math part is blending between nodes. Do that well, and you can lower the number of nodes you need- engines are smooth and continuous, like math.
I know it's a little late, and likely redundant, but have you talked to the folks over at freeEMS? http://freeems.org/
I've been following them for a few years. You may be able to use a lot of their code, tuning software, etc.
burdickjp wrote: I know it's a little late, and likely redundant, but have you talked to the folks over at freeEMS? http://freeems.org/ I've been following them for a few years. You may be able to use a lot of their code, tuning software, etc.
Just spending 10 min on that site- they need someone driving that ship. It's great that there's a group of people who support each other in terms of making their own EFI set up. But if you want others to use the ideas, well....
For instance, I know a thing or two about the strategy and calibration that will make an engine run. Outside of beliving that electrons move, I barely know anything about electronics- so having a menu that I can buy the stuff from suppliers and assemble is required. I could not find that (found 5 different projects that people are working on, though). And while I know how I want the software to run, I don't know how to code it, and I don't know the base language that is used to do that.
basically, since there's no real direction to make the hardware or really do the software, then I can't really add input to the projects. Which I'd like to do.
I tried with the russefi and even saw there that they were looking at things that really are not important right now. While you are tying to get a progam that is easily portable between engines, how the O2 heater operates really insn't important.
alfadriver wrote: basically, since there's no real direction to make the hardware or really do the software, then I can't really add input to the projects. Which I'd like to do. I tried with the russefi and even saw there that they were looking at things that really are not important right now. While you are tying to get a progam that is easily portable between engines, how the O2 heater operates really insn't important.
I've noticed the lack of direction as well. That seems to be a problem with many open-source and free projects.
I really want to see at least ONE of these projects succeed. I love my megasquirt, but we need a clean slate design.
In reply to burdickjp:
Well, everone's idea is the best one.
Gentelment, you are making a really good point RE: the documentation, or the poor quality of it. Please help me with that a bit. I guess it all begins with knowing the target audience :) For example, how many of us gear heads would be ready to solder surface-mount components? How many of us would be ready to solder 0805 resistors - 0805 stands for 0.079 in × 0.049 in, that's 2.0 mm × 1.25 mm. Would an average person choose to solder himself or buy the cheapest assembled board?
As for the lack of direction - unfortunately to some extent it is explained by the huge amount of work to be done. I am constantly torn apart by working on the hardware OR working on the software, both need attention.
Anyway, just a status update: We've tested our power supply & main chip schematics and we will now work on a full self-contained board for an older OEM ECU case.
Still making progress on the board for an old OEM ECU box:
So what are the chances of you making something that could support the OBD2 interface? I assume there is a pretty good reason MS & other stand-alones don't do this.
Define what exactly do you mean when you say "support OBD2 interface"? Obviously yes, we can use OBD2 over CAN for error codes, is that what you are looking for? This is in the pipeline eventually - we already have a CAN chip even on the current board.
russian wrote: Define what exactly do you mean when you say "support OBD2 interface"? Obviously yes, we can use OBD2 over CAN for error codes, is that what you are looking for? This is in the pipeline eventually - we already have a CAN chip even on the current board.
I mean... could cars with this ECU pass inspection in states that emissions test/inspect by plugging into the OBD2 port?
We are planning to report error codes via OBD-II - that's a useful feature for our race cars.
Federal government is against non-OEM engine control on public roads.
ProDarwin wrote: I mean... could cars with this ECU pass inspection in states that emissions test/inspect by plugging into the OBD2 port?
No ECU manufacturer wants to get hit with the same sort of judgement as what happened to Casper Electronics' O2 simulators. Note that the $80,000 fine the link mentions was reduced because Casper couldn't pay more, and that was on top of having to buy back as many of the simulators as they could.
ProDarwin wrote:russian wrote: Define what exactly do you mean when you say "support OBD2 interface"? Obviously yes, we can use OBD2 over CAN for error codes, is that what you are looking for? This is in the pipeline eventually - we already have a CAN chip even on the current board.I mean... could cars with this ECU pass inspection in states that emissions test/inspect by plugging into the OBD2 port?
The good reason nobody does that is that it's against the law.
The entire point of having an OBD interface is to show that a car is in compliance with air quality laws, so going around the interface would be very ilegal.
edit- which is what MMS just posted. 
on a side note- don't pretend that either this site or the russian site opreates in a vacuum, or that the epa or carb don't look into these things.
I understand that it may be extremely unlikely to happen. I'm not suggesting the ECU fool the emissions test (like an O2 simulator "defeat device"). I'm asking if it can actually report all the necessary things. Many many many modified cars can benefit from the ability to tune better yet still pass emissions.
Question: Is it illegal to run Hondata on an OBD2 S2000? HP Tuners on a GM ECU, etc?
Not sure what you mean about the epa/carb stuff? Are they going to come after me? They are welcome to, all of my car have all emissions equipment intact and pass emissions with zero problems - it would just be nice to have tuning options for some OBD2 cars that have never had the ECU "hacked".
ProDarwin wrote: I understand that it may be extremely unlikely to happen. I'm not suggesting the ECU fool the emissions test (like an O2 simulator "defeat device"). I'm asking if it can actually report all the necessary things. Many many many modified cars can benefit from the ability to tune better yet still pass emissions. Question: Is it illegal to run Hondata on an OBD2 S2000? HP Tuners on a GM ECU, etc? Not sure what you mean about the epa/carb stuff? Are they going to come after me? They are welcome to, all of my car have all emissions equipment intact and pass emissions with zero problems - it would just be nice to have tuning options for some OBD2 cars that have never had the ECU "hacked".
Theoretically, any ECU can satisfy the OBD requirements. Realistically, there are many tests that have to be done to demonstrate that it does. Starting with actually meeting the original certified emmissions standards. Then moving onto ALL of the faults- from basic hardware faults to things that are operating, but make the engine run outside of the standards, and proving that the trigger is right at the emmissions fault limit. The OBD part of development is probably a little harder than meeting the emissions requirements.
As for the tuned ECU's- many of them are defeat devices. Not knowing the specifics of what Hondatune or HP do, I can't judge. But I know that some of the tuners of the Ecoboost do violate the law. Whether anyone goes after them is more of how bad the violation is- number of cars affected + how much worse the emissions are. (the link above to the settlement shows that there was a good model on what was affected)
My point about EPA and CARB is that people bring up ideas to get around their requirements, seemingly pretending that they don't look at this board. They do. EPA can't do anything to an individual, but they can go after people selling stuff. CARB has more leeway in going after individuals, since they have a car inspection system. One more thing- just because you have all the hardware there does not make it compliant if you are using an aftermarket ECU- there's many hours and many dollars spend developing the calibration to make it compliant. I doubt that any aftermarket ECU has access to even 10% of the testing equipment that is used for tuning. So you know....
russian wrote: Gentelment, you are making a really good point RE: the documentation, or the poor quality of it. Please help me with that a bit. I guess it all begins with knowing the target audience :) For example, how many of us gear heads would be ready to solder surface-mount components? How many of us would be ready to solder 0805 resistors - 0805 stands for 0.079 in × 0.049 in, that's 2.0 mm × 1.25 mm. Would an average person choose to solder himself or buy the cheapest assembled board?
No idea if I'm the average person, but my hands aren't steady enough and/or dextrous enough to solder surface mount components.
When I still had my Miata I was thinking about building a DIYPNP, but only after verifying that I didn't have to solder any SMD components.
In reply to BoxheadTim:
Problem is all the cool stuff is SMD-only now, that's what I was told by the EE guys. By the way I myself is not dextrous enough to solder the main chip, so a fully-assembled option would always be available.
On the other hand the software does not care about the way the board was soldered, so if anyone would design a through-hole version of the board such version can always co-exist.
In reply to ProDarwin:
As soon as you swap the ECU completely, why do you care if the car was OBD2 originally or not? The only legit use-case is if you want to keep using your favorite OBD scanner, isn't it?
russian wrote: In reply to ProDarwin: As soon as you swap the ECU completely, why do you care if the car was OBD2 originally or not? The only legit use-case is if you want to keep using your favorite OBD scanner, isn't it?
Well, I was hoping that satisfying emissions requirements was a legit use case as well.
In reply to ProDarwin:
Technically US is not the only jurisdiction on this planet so we have to see if anyone would step up and implement ISO 15765, I believe that's the protocol for OBD via CAN.
I am pretty sure you can have the boards populated and soldered overseas or at a local shop. Surface mount tech is remarkably simple, use a "screen printing" process (like for a t-shirt) to put the solder mask down, pick and place machine drops on parts, run through oven. Flip and repeat. Note for 2 sided boards you like to try and put all the heavy stuff on 2nd side. Given the simplicity of your board layout (ie its not 32 layers with 1200 parts with RF pathways, inductive pathways etc) it would probably be manually placed parts. I have a good friend who still works in this field, PM me if you need some contact details.
ProDarwin wrote: Well, I was hoping that satisfying emissions requirements was a legit use case as well.
As Alfadriver mentioned, building an ECU to satisfy emissions requirements is a tall order. I am aware of two aftermarket EFI systems that managed to get a CARB EO number, both as retrofits for cars well before OBD2, and both required massive testing budgets.
You'll need to log in to post.
