Speeduino 2T EFI Project

[TZ350]

Looks great, what does the MAP input do in the code? In Microsquirt there is a MAPxRPM table, the value in the table is a multiplier for the value in the TPSxRPM table. Or it can switch from one TPSxRPM table to another based on a threshold MAP value.

Hi Nath88.

Speeduino has two standard tables, VE and Alpha-n both can be multiplied by MAP and as the firmware code is open source you could have any scenario you like if you are prepared to re-program the firmware. Tuning Studio (Mega Squirt) is the application for Tuning the Speeduino firmware and Tuning Studio also handles all the data logging. Currently the software continuously looks for a maximum and minimum each cycle and works out the difference and translates that into a MAP value. The next move is to instantaneously measure the minimum at 10deg ATDC and Maximum at 160 deg ATDC only, which are the correct two places to look according to EngMod2T simulations.

All the 2S MAP Max Min stuff is add on code that I have done on my copy of Speeduino. The next move is to polish it up so its usable.

[husaberg]

You don't work for KTM by any chance?

I think thats Austria not Australia. Similar spelling but far worse smelling
As far as i know Flets still their go to guy for Southern Hemisphere R&D

[EfiCanuck]

TZ - thanks for doing all this work. I am a few steps behind with my 2 stroke EFI conversion, and this will help a lot.

Here are a few data points. Dont think they will help much, but are interesting.

- The OSSA trails bike - a 280 Keihin 2 stroke EFI - uses a map sensor in the air box. Very smooth and controllable bottom end. Perhaps since the TB is small (28mm) they can get enough of a reading to make it useful. Or perhaps they are reading it at specific crank angles like you are.

- The Vertigo trial bike. Uses a Athena GET ECU. Very smooth and controllable bottom end. They have an integrated IAT and MAP sensor on the airbox as well. No idea how they sample the sensor.

- The KTM 250/300 TPI MAP sensor is attached to the crankcase.

- I have a GET ECU on my KTM 250 TPI dirt bike and had a little chat with an Athena tech about how the crankcase pressure sensor is used. He indicated that the sensor on the crankcase is only used between 20%-30% TPS and 3.000-4.250 RPM. It was also implied that there was a "special" algorithm. Perhaps similar to yours - selecting values at specific crank angles. Of course this is for a enduro dirt bike, a street bike or race bike would be different.

[TZ350]

.

it´s my two stroke bike with speeduino https://youtu.be/zk55HR6bgO0

Great fuel injection project, I like it..... .... looks like more and more people are trying their hand at 2S EFI.

[husaberg]

.



Great fuel injection project, I like it..... .... looks like more and more people are trying their hand at 2S EFI.

He likely does even remember posting that or even joining KB
looks more like what we would call a Yamaha RS125.
I have seen the Rd125LC singles but never an aircooled one.

EDIT sorry its an RD135

Made mostly for the Brasilian market

looks like the cases are actually designed for a water cooled top end. looks like 5 speed box

Asia had the RZX135.

Yamaha RX-Z 135 was a two-stroke naked bike manufactured by Yamaha Motor Corporation. Debuted in April 1985, the RX-Z it was very popular in Malaysia and Singapore and has been sold for more than two decades before the production was ended in 2011. [2] In 2004, the model was revised cosmetically and a catalytic converter was installed.
The Yamaha RX-Z was originally equipped with a five-speed transmission with solid front disc brake. The model shared the same platform and chassis as its predecessor, the water-cooled Yamaha RD125. Shortly after the launch, the RX-Z became popular among young motorcyclists especially in Malaysia. A few years later, the engine was upgraded with the installation of a six-speed transmission, together with newer instrument panel and handlebar switches, as well as a cross-drilled front disc brake rotor.

Since then, the design of the RX-Z remained unchanged until 2004 when the model was updated, with the rear lights borrowed from the Yamaha Y125Z. The new RX-Z was equipped with a catalytic converter, sacrificing 1 PS of maximum power (original maximum power: 21 PS). However, the maximum torque remained unchanged but the low-end torque was improved compared with the early models. Some owners of the earlier RX-Z motorcycles may have problems during take-off because the engine tends to stall when an inexperienced rider tries to take off in the first gear. However, the problem was resolved in the new model.

In Malaysia, the RX-Z was one of the motorcycle models often associated with Mat Rempit street racers. [3][4] As a result, many Malay movies including Remp-It use the RX-Z in their movies.
The production of the RX-Z ended in 2011 after being sold for more than two decades
https://en.wikipedia.org/wiki/Yamaha_RX-Z

So what are are is a RD125LC with an air cooled top end and a 5 speedbox that was later changed to 6 again.


Mat Rempit
is a Malaysian term for "an individual who participates in immoral activities and public disturbance with a motorcycle as their main transport", usually involving underbone motorcycles,[1] colloquially known as Kapcai, or scooters. Mat Rempits are not involved in street racing but rather go against each other for cheap thrills and to rebel against authorities while some of them perform stunts for fun, such as the wheelie, superman (lying flat on the seat), wikang, and scorpion (standing on the seat with one leg during a wheelie).
The word "Rempit" comes from "ramp-(rev)-it" (ramp the throttle). According to Kamus Dewan, the definition of "Rempit" is "the act of whipping with a cane".[2] An alternative source of the word is derived from the noise made by a 2-stroke motorcycle. "Mat" is a Malay slang term used to address or to refer to a male person who is usually of Malay descent but is generally used derogatorily.
Organization and traits

Mat Rempits usually travel in groups and create disturbance in bustling city centres on weekend nights. Mat Rempits are usually associated with gangsterism, gang robbery, street fighting, assault vandalism, theft and bullying. Most motorcycles used by the Mat Rempits do not meet standard specifications, or have been modified extensively for greater speed, or just to make the exhaust noisier.
Additionally, most Mat Rempits do not have valid motorcycle licenses, do not pay road taxes, and ride stolen motorcycles.A growing number of housing estates have also been turned into racing tracks.[8] It is estimated that there are about 200,000 Mat Rempits in Malaysia

https://en.wikipedia.org/wiki/Mat_Rempit

Marlon Brando without the Triumph but 200,000 of them.

[TZ350]

I haven’t successfully injected a two stroke yet but I have done an EFI for Mazda rotary which has many of the same exhaust tuning sensitivities as a two stroke (big intake and exhaust port timing overlap).

I think you are on the right track going with speed density. Alpha N was a PITA to get tuned right, almost impossible. I used a MAP sensor fed by an accumulator in turn connected to manifold with small gauge hose or with a small orifice restriction.

This formed a filter to mechanically average out the rapidly fluctuating manifold pressure. Only down side to this approach is throttle response may be less than instant. That was not a problem in my application (airplane) where all throttle changes are relatively slow.

I have tried the accumulator idea and find that with a 2S all you get is an average pressure. And the interesting thing about that is that at idle the average is near ambient and at WOT the average is a little lower.

I have written the code for a Arduino Nano to take a high pressure measurement at 135 deg ATDC and a low pressure reading at 315 deg ATDC. The code is triggered by a Timer Interrupt and an interrupt attached to the ignition trigger calculates the period. The end result after a bit of math manipulation is a MAP value that looks like a traditional 4S MAP. I am in the process of de bugging the code and hope to get to try it all out this weekend.

Small amplifier board on the left to amplify the ignition pulse simulated by my computer. Arduino Nano in the middle on a break out board with screw terminals. 12 bit DAC on the right near the calculator.

.

[OopsClunkThud]

I'm sure you know this but keeping the interrupt code as small and quick as possible is important. Just enough to capture the data. Do everything else in the normal loop.

[Vannik]

I have been following what you are doing here and on the Speeduino forum and are beginning to think you need to sample at a much higher rate, like two times per degree at max rpm and have a very fast processor to do some signal conditioning before using it. We often find the signal to noise ratio with pressure signals very bad where the noise has a 3 or more times bigger amplitude than the pressure amplitude. But not my area of expertice!

[TZ350]

I'm sure you know this but keeping the interrupt code as small and quick as possible is important. Just enough to capture the data. Do everything else in the normal loop.

I am just finding my way here and appreciated the reminder about the code.

[TZ350]

.... beginning to think you need to sample at a much higher rate, like two times per degree at max rpm and have a very fast processor to do some signal conditioning before using it. We often find the signal to noise ratio with pressure signals very bad where the noise has a 3 or more times bigger amplitude than the pressure amplitude.

As I am not looking to track a complete cycle now but only looking to have two pressure measurement values taken at two distinct fixed points I was thinking of only two samples but if I can manage it I will make a series of readings of the MAP sensor clustered around each point and smooth them. Thanks for the pointer about noise.

I am using a CTC Timer Interrupt, being a newbe to coding, the challenge will be to reset the counter quickly enough.

It was a simple idea. Take a high and low crankcase pressure reading and use the difference between them as an indication of airflow. Easy to say but it sure has been harder to implement.
.

[Vannik]

I understand what you are trying to do but would have started with a full high performance super fast system and then worked at simplifying (less expensive) it rather than trying from the less complex side, adding more and more complexity and maybe missing something. But all in all a very worthwhile and noteworthy effort.

[TZ350]

I understand what you are trying to do but would have started with a full high performance super fast system ...

Problem is, I have come from a place of complete computing/programming ignorance and would not know a high performance super fast system if it hit me on the head........ ... I have to work with the resources I can and in the end it may not be enough.

It has been suggested to me that a mechanical system might work for me better. Something that opens a port to a small chamber at the right time for the MAP sensor to read the crankcase pressure. Easy to do with a hole in the skirt of the piston and a gallery through the cylinder to a small chamber. The chamber could hold sufficient volume/pressure to allow the MAP sensor to settle. It is a tempting idea.

I am getting the code sorted, but it is like polishing a Turd, just takes time.

[husaberg]

Problem is, I have come from a place of complete computing/programming ignorance and would not know a high performance super fast system if it hit me on the head........ ... I have to work with the resources I can and in the end it may not be enough.

It has been suggested to me that a mechanical system might work for me better. Something that opens a port to a small chamber at the right time for the MAP sensor to read the crankcase pressure. Easy to do with a hole in the skirt of the piston and a gallery through the cylinder to a small chamber. The chamber could hold sufficient volume/pressure to allow the MAP sensor to settle. It is a tempting idea.

I am getting the code sorted, but it is like polishing a Turd, just takes time.

I came across this Rob
https://www.megunolink.com/articles/...-measurements/
but you know i like the remote chamber KISS reading

[TZ350]

I came across this Rob
https://www.megunolink.com/articles/...-measurements/
but you know i like the remote chamber KISS reading

Interesting link Husa, thanks.

I have been using a traditional exponential smoothing function. Exponential uses very little memory and processing time unlike other methods that require arrays. The remote chamber thing looks like a good fallback position if I really have to go there. But the Arduino Nano will be "KISS" once it is sorted as copies could easily be added to a Link, Microsquirt, Megasquirt, or any other EFI system that takes a MAP input.

[TZ350]

.
Making progress with the Arduino Nano MAP co processor.



The code runs one event interrupt for sync of a single lobe trigger 20deg BTDC and timer interrupt 1 with a 64 prescaler. Timer interrupt 1 triggers twice a crank revolution, once at 135deg ATDC and again at 315deg ATDC. So the compare match register changes twice a crank revolution and is slightly asymmetrical.



It all works in theory on the test bench but I am a little concerned that 560 microseconds at 13,000 rpm will not be enough time for timer 1's code to make two low pressure analog reads and do a bit of math before the sync interrupt gets triggered 45deg later.

Anyway, it's time to tidy the code up, annotate it properly and remove the de bugging traces and see what happens when its run in real life on the engine.



The trace in the box on the right shows the numbers are in the right place at a simulated 13,000 RPM and the numbers are the right size.
.