Hello!

[Sudo]

Hi, I just found this project while researching to develop my own ECU. While a DIY ECU is a very long and difficult task, I believe I have the electronics (hardware and firmware) skills needed to develop my own. What I lack is deep knowledge of EFI and I am still learning. So far, I really like the approach of this project. More specifically, I like the idea that everything is modular, including MCU and firmware. This will be a better starting point for me and I hope I can contribute to this project and offer my some of expertise and opinions and be allowed to disagree and not be banned.

[AndreyB]

Welcome Sudo!

I think long is the key word here - this is all not rocket science, just a lot of work.

At this point we hope we have a good-enough-for-now Frankenstein and a lot of open tickets in the tracker

A good starting point for software side would be this page.

[kb1gtt]

Hello and welcome along. Good to hear you have skills and a desire to help. Can you share bits of your project? I like to see what I can learn from others. I might be able to offer some suggestions on how to make it better.

About being banned, I've been there before. I would generally say that around here you should be pleasant and should disagree by either offering content, or by asking questions. For example, you could say bit of code xyz is wrong, so I made better code abc found here. That's a good way to get a disagreement taken seriously around here. Or if you say something like I see this does blah, but wouldn't it be better to do yadi yadi ya. That would also be a better way to get heard. When getting into a confrontation, please try to remember we have a limited time schedule, and try to keep thing constructive instead of destructive.

Also I'm not Admin, so take what I say with a grain of salt.

Probably the biggest thing, is that it's OK to not know stuff. Feel free to ask questions and we'll try to offer answers. Were all learning here, I'm sure I can learn from you and you can learn from me.

That said, welcome along and I look forward to seeing in the forums.

[Sudo]

Thanks. I agree with you kb1gtt. My project is still in the beginning stage of just gathering parts. Basically, I want to convert my Honda CB600 into fuel injection. I am somewhat stuck on an idea while brain storming. Since the OEM setup only have a crank position sensor and wasted spark. I am still trying to figure out a non invasive way to indicate which half of the cycle to perform sequential firing. Mechanically, I am not very strong at fabrication so I am searching for a software/logical way to determine which cylinder to fire. Another option is to go with bank injection. But I have am worried that it doesn't justify going fuel injection at all since the OEM carbs are already individual carbs and that bank injection could be a step backward? I guess if all else fails, I may have to fabricate a cam trigger.

[kb1gtt]

Here are some links that might be of interest http://code.google.com/p/daecu/wiki/Injector_driver_theory http://code.google.com/p/daecu/wiki/AN_and_DIGI_Protection

I really need to make them better. However they are a start.

Do you have pictures and / or a schematic of the existing system?

[abecedarian]

I believe the CB600 engine uses a 12-3 trigger wheel (12 teeth / 3 missing) inside the engine case with a single VR type sensor.

A thought for generating a pseudo-cam signal might be... since you'll most likely end up with independent throttle bodies, you might be able to pick up pulses on a MAP sensor attached to a TB runner. Theory being that whether at idle or WOT, you should get a 'flick' as the intake valve opens.

[Sudo]

MAP sensor is an excellent idea. I really like it! Is this commonly done with success?

I should scope out how the map sensor output look in detail and see what I can come up with. But just brainstorming, I suppose I can AC couple an analog map sensor signal and look for the transition.

[abecedarian]

I don't think it's been done, and I might be hesitant to try it- wonder what a backfire might do...?

Something that might work, if you have room in the head, would be to have something detect rocker arm motion, or maybe a VR sensor similar to what's on the crankshaft to detect a passing cam lobe.

[kb1gtt]

Most MAP sensors have a 1mS delay, and have a limited bandwidth, so at high rev's you won't be able to tell what's going on relative to the engine angle.

[abecedarian]

The MPX4250AP has a 1ms response time, so at 10K RPM, that's nearly 170 possible samples per second. So although I wouldn't recommend it, with some thought it might be possible.

But I do suggest a VR near a cam lobe to see if that's possible. Otherwise, it might be possible to get something off a cam gear if there's clearance.

I don't know if it's possible to transfer later CB600 parts over, but that may be an option too since they were FI.

[kb1gtt]

It's debatable if you need a MAP at all. I know ion-sense can sense crank angle. Perhaps that would be one way to combine sensors, as you're certainly going to need a spark plug.

[AndreyB]

... but it would be much easier to figure out a plan old trigger wheel...

[kb1gtt]

I think he noted that mechanical and fab was an issue. I thought he may want to use more familiar tools. I agree a crank wheel and hall is a really good option. However getting some kind of wheel or sensor seems like it might be a problem for this fellow. If they can get a wheel and sensor, that's going to be one of the best approaches I'm sure.

[abecedarian]

Ion sense on a carbureted 599CC motorcycle? Seriously? A single spark plug would cost as much as the engine, chassis and new tires (tyres for those near EU), best I can tell.

I understand it's a worthwhile endeavor for those with more green in their wallet than soul on St. Patrick, but really?

If economy and ecology were all what's important, it'd be a new bike, not some retrofit thing. But then there's the environmental cost of manufactured plastics and mining ores for engine and frame construction, so maybe keeping the old one on the street is more practical?

Keep it real!

[puff]

what bike is that? could you post a picture?

[kb1gtt]

Harley Davidson has been using Ion sense in many of their OEM bikes for several years. I picked up a coil for a bike on eBay for something like $20, and it used a standard spark plug. I then set fire to the coil, as my connector at the coil couldn't maintain a proper voltage. Ooops. So it's toast now.

[abecedarian]

Harley Davidson has been using Ion sense in many of their OEM bikes for several years. I picked up a coil for a bike on eBay for something like $20, and it used a standard spark plug. I then set fire to the coil, as my connector at the coil couldn't maintain a proper voltage. Ooops. So it's toast now.

I stand corrected. But I am still willing to wager a fair penny that the control mechanisms cost quite the coin.
It's not just the 'coil' that does the work.

what bike is that? could you post a picture?

CB600 (599 in the US) are open / naked and detuned versions of the CBR600.

[kb1gtt]

http://www.aces.edu/~parmega/efi/temp/DIY-Ion-Sensing.pdf

I wonder have we completely hijacked the fellows "hello" thread? we haven't heard from them to see if this is up their alley or not.

[Sudo]

Ouch, 1ms delay sounds terrible. Thanks for the info, so I didn't have to find out the hard way. I think I will either fabricate a trigger wheel or perform bank injection.

I haven't heard of ION sensing, but I will read up on that. Since it is costly as mentioned, I may not want it. I am pretty cheap! And it sounds like it is cheaper to have a trigger wheel CNC milled.

This is my bike. It is still in 1 piece and runs good. But I think EFI will make it the perfect bike.

[Sudo]

Skimming through that ION sensing article, it seems that it requires a detonation initially before it can sense the angle. Perhaps one can use bank injection in the beginning to get it started, Then use ION sensing for precise angle and sequence. I am afraid this might be out of my range of skills. It sounds very interesting though and I will keep it in the back of my mind. I will probably pursue the true and tried method of adding cam trigger wheel.

[kb1gtt]

The quick and basic overview of ion-sense is that when you are not in a spark event, you generate around 80V to 300V on the spark plug. This will cause a certain amount of electrons to wonder across the gap. As you compressor, the number of electrons will change, and by measuring this electron leakage, you can determine when you are compressed to the point where you can fire a combustion event. A secondary benefit, is that if you are running hot and if you are on the edge of detonation, the hot temperatures will cause the air in the gap to start to ionize and you'll get a specific leakage tendency which can be measured and identified. This leakage because of ionization is why many people like ion-sense. It allows you to determine when you are about to detonate, before you detonate. A knock sensor is post detonation, detection.

I agree that a typical hall or VR sensor is generally the best way to go.