Another Canuck

[sparky]

Hi all,

Another Canadian here - from the west (left) coast. I am a professional enterprise developer with lots of fabrication, electronics (analog and embedded) and automotive skills. Offroad motorcycle enthusiast as well. So naturally a DIY EFI for a dirt bike would be just the thing to really learn about EFI/EMS.

I have been thinking about it for about 4 years. Although I have 23 other projects on the go (seriously), I would like to start the thinking process now, even though I might not get to implementation for quite a while. I think I can fit it in between 7:15 and 7:45PM each night

I was planning microsquirt, but was struggling with the closed nature of it. That left FreEMS and RusEFI. I like how you guys have progressed and thought I would join the club, even if it is just asking questions for now.

Once I get a good hold on EMS understanding, I would like to try some firmware development for offroad motorcycle specific features (a version of traction control that does not require a wheel speed sensor), controlling deceleration pop on 2 strokes, two stroke anti reverse, limp mode, bluetooth for trailside tuning, etc. We'll see how it goes.

thanks!

[AndreyB]

Welcome to the forum!

[kb1gtt]

Yes welcome along.

I've been tempted to do a Motorcycle spin. I was thinking a 2 cyl. I have a single cyl 79 Jawa Californian. Some day I want to EFI that. If you want to take a stab a motorcycle specific spin, let me know. I'd be happy to dedicate up to 10 minutes a week at spinning something smaller. I've also been tempted to make a modular design with some kind of communications bus connecting the modules.

[sparky]

Thanks guys, I am just running through various sensor & power options for the bike and will post up some pics/questions soon.

[sparky]

Hi gang, I was planning to do a casual EFI project using an older Honda 4 stroke, but plans have changed. I own a 2018 KTM 250 transfer port injection EFI bike, and I want to heavily modify it for technical riding. This is a 2 stroke bike. There are no mapping and tuning hacks for the current Synerject M3C ECU. One small company is building a tunable commercial ECU for the bike, but it is being plagued by delays. Another company has made a tunable ECU for the bike, but is extremely expensive and lacks a lot of features. So I am investigating rolling my own with RusEFI.

Here is the current configuration:
CPS - inductive with toothed wheel
IAT - thermistor
CLT - thermistor
TPS - current is hall effect, but I will probably go to a new throttle body that has resistive TPS
CCP - this is the "crankcase pressure sensor". Piezo. It is used for pressure sensing.
There is no O2 sensor. This is an open system with fixed map because it is an offroad two stroke.
Ignition - it is inductive (IDI) ignition, with the driver onboard the ECU. The coil is outboard.


So, I have lots of questions, but I will just post a few at a time, starting with the biggies:

1. I would still like to use the crankcase pressure sensor for a speed density system - is this possible and useful? Any other suggestions?

2. Can I tune this 2 stroke with closed system and an O2 sensor? If not, any suggestions for tuning an open system 2 stroke?

3. the ignition drive for the current IDI system is onboard the ECU. Any recommendations for a small outboard CDI system? Multispark would be nice.

4. specific to off road motorcycle ECUs are a couple of features I would like to try. One is "pipe bang" elimination, which basically shuts off fuel when the engine is decelerating and throttle is closed. The other is electronic traction control. This does *not* use a wheel speed sensors, but is sensed via engine load, RPM and TPS. Would features like this require programming, or is there any existing code I can leverage for features like this?

thanks for the help

[AndreyB]

4. specific to off road motorcycle ECUs are a couple of features I would like to try. One is "pipe bang" elimination, which basically shuts off fuel when the engine is decelerating and throttle is closed. The other is electronic traction control. This does *not* use a wheel speed sensors, but is sensed via engine load, RPM and TPS. Would features like this require programming, or is there any existing code I can leverage for features like this?

that would require extra programming for sure

[kb1gtt]

1: is possible but there may be some issues. Getting a good map value may need some programming. The short distance will likely make a noisy signal. The existing map filters may work as is but there is some risk as the short intake may cause some issues that we are not currently seeing.

2: I would not try this. The oil in 2 Stoke engine pretty much makes this a no go with any technology. Aka oil on the sensor causes problems. You really shouldn't tune via O2 any how. O2 gets you close but dyno tuning is the real way to tune.

3: you need an external igniter. Perhaps something like a j701 could be used. I know the J701 and several others have been used. It's possible but would need an extra part.

[stefanst]

2. How about measuring exhaust gas temperature? That should give some decent clues about dangerous lean conditions.

[Simon@FutureProof]

It does but only really if you have a solid base line for engine to know what is a high or low temperature.

For example we have 2 engines on different dynos at the moment, one is hitting a max of about 800 degrees EGT and the other hits 1080 Degrees C regularly!

If its possible to base line the engine on a good tune and set boundaries from that then it could be useful or if you can compare EGT across a few cylinders you can get a good picture of how each cylinder is running.

RE-Pipe bang elimination, I would have thought that the overrun fuel cut would achieve that if you were aggressive enough with it, after all that is more or less what it is designed to do for a car, cut the fuel to prevent wasted fuel and exhaust banging. In later cars it does get used to help keep the cat ok with similar goals of eliminating damaging banging in the exhaust.

[sparky]

Thanks for the quick replies guys.

I notice two camps about using O2 sensor on two strokes for tuning. Once camp is: "no - it wrecks the sensor and the results are no good anyways because of the oil" - the other camp says "no problem with the right sensor for short term tuning use. Must mount it in the right place and position". Seems a lot of 2-stroke sled guys and kart guys use the o2 sensor for tuning. I dont see it much in the dirt bike world as it is mostly carbs and old style tuning methods.

Unfortunately I do not have access to a dyno, and the closest one is 6 hours away. I cannot imagine it being cheap either. I am not building a race bike for top HP. It is a technical bike - torque and linear power delivery is king.

So - is all the tuning for the DIY EFI projects done on a dyno? If not - how to they do it? Plug readings?

[kb1gtt]

Don't forget the butt dyno. That works surprisingly well.

Most folks suggest EGT's for detecting problems. Generally you tune at the dyno to determine peak HP, or what ever characteristic you are looking for. The dyno will typically have a sensor which you can be put up your pipe to ensure your AFR is in the safe zone. Once you know peak HP and safe AFR, you then know your typical temp. From that point, you then look for a variation of that temp.

If you move your EGT even by a small amount, it can drastically change your measured value. So looking at the actual value is often of less interest. It does tell you some things, but it's of a lower importance. I understand that generally you look for a variation from when it was on the dyno, as well you look for variations from cyl to cyl.

The O2 sensors need to be hot to operate. Internal heaters help heat them up, or keep them hot if you put them slightly wrong in the pipe. I have not seen a successful application for 2 cycle. However I also do not know anyone who has tried. My understanding is that the build up will skew the results potentially causing the ECU to think it's AFR is wrong when it really is correct. However I believe it causes it to run rich, so that's typically a safe condition. Not environmentally friendly, but safe.

As well you could try it for a tune, then turn it off. AKA a 50k mile sensor, might only get something like 10k miles because of the oil. However you can do a tune in less than 10k miles. Once you have your auto-tuned table, then you can simply run with that table.

[JBR21]

Hi guys,

I know it's been a while...But is there anything new in this project?

I want to convert my sx125 TBI to a standalone ST32 project that runs with rusefi FW.

What I think will be a big issue on this small bike is the mapsensor sampling.

The original ecu does this 2 times per cycle and takes the lowest value for calculation- but the values it reads are really small.

So there has to be a factor, that makes it "noticeable".

Do you guys think, that we can get that thing running??

Thanks guys!

[AndreyB]

Do you guys think, that we can get that thing running??

yes.