could i use the same arduino And your code for the bike-motorcicle efi system to switch to , but my bike already has the ecu and is carburated is a honda cb600f from year 2004?

I’m 41, and the first vehicle I ever drove on the street was a 1931 Model A, at 14. 👍

Thank you! I really appreciate it! I hope it performs nicely for you!

i got a question on the board there's 2 set of transistors T3-T8 not label on board can you reply with the right orientation

Thank you! Although I've never worked with turbocharged engines, one tip I would suggest is to ensure your MAP sensor (if you are using one) is compatible with both boost (positive pressure) and vacuum (negative pressure). Unlike naturally-aspirated engines, turbocharged and supercharged engines can develop manifold pressures higher than atmospheric when boosting. Also, be sure the ECU is able to correctly interpret the MAP signal as positive and negative, respectively. Good luck with your project!

I have a question, how to get the map sensor constant, basically to be able to implement any map sensor

I still have the fuel return system in place. Currently I draw fuel from the top of the tank using an immersed fuel sock, and I return the excess from the regulator through the bottom of the tank. This configuration is good because it prevents whirlpool formation at the bottom of the tank. Right now I have the regulator at the front of the car, so I do have a fairly long run of fuel line to the injectors at the rear. There is some load-dependent pressure drop over the line, but it can be fairly easily compensated out in the AFR table.

Great observation! In fact, there actually is a passage in the Zenith throttle body for vacuum take-off, and I have my MAP port aligned to it (see www.ebay.com/itm/154214676001 ) I actually do have a rate-of-change-of-MAP in the program code! However, I found that the acceleration was smooth enough without the extra fuel, and adding any additional fuel on MAP pressure increase tended to cause over-enrichment. I suspect some fuel gets condensed on the bottom of the manifold while cruising/idling, and this likely gets pulled in all at once when the throttle is opened quickly, thus negating the need for extra accelerator pump fuel.

Good point about damaging the catalytic converter, it makes sense that the unburnt fuel would get oxidized in the converter and would heat it up rapidly (if I had one installed). In this case I was a bit paranoid about entering a lean burn condition, and I figured the rev limiter would only be used in short, infrequent events. My main concern with cutting both spark and fuel was that if the fuel mixture became lean enough to detonate, it might ignite without a spark, causing detonation at the worst possible time (at redline). The Bosch Automotive Handbook sounds like a great resource, I will definitely check it out!

The existing system very well might work with a 6-cylinder engine as-is. Really the only limiting factor is the ability for the spark timing to accurately follow the distributor pulses. There's a "pulsesPerRevolution" term in the code that can be adjusted based on how many distributor pulses are expected for each engine revolution, and this can be arbitrarily configured. For a 6-cylinder engine, I believe this value would need to be set at 3.

@@DielectricVideos Nevermind, should have read the comments first. Thank you sir.

Yes, it would definitely be possible to implement the system using SMD components! For this one, I used THT parts to make the open-source boards as easy as possible to assemble. It could be made a lot more compact with SMD parts, and I would even opt to put the ATmega328p and associated support electronics directly on the board if I took this approach 👍

I'm currently runnning 89 octane pump gas, but with the larger injectors it should be able to run E51 or E85 with an adjustment to the fuel_comp term. Yes, the two things I did that really improved the lag were to add the AFR table, and also to increase the injector frequency to about 30 Hz for. Both provided a major advantage in reducing the lag, even without the need for the additional rate-of-change-of-MAP.

I saw a 33uh 4.4A 91Mohm inductor listed on your BOM, but didn't see it populated on the PCB. Is that component still utilized? Thanks!

@@mikedavis5784 Good catch, that was from an older version of the design where the components of the XL6009 unit were soldered directly to the board. I've updated the BOM to remove this, as it is no longer necessary.

My current engine is a 1776cc, but I have successfully run a very similar system on an old 1192cc VW engine as well! I think a single injector should be adequate for your Mini engine, provided it is a fairly high-flow injector like the one I have listed in the Bill of Materials. Sounds like a really cool project, I hope to hear how it goes!

@@DielectricVideos well, companies in the UK are expensive! Around £60 for a single board! That's close to $80. I looked at the company you used and I can get it made, and shipped to the UK for £5.38. That's around $7. I guess the brain of my mini will be interlectually and physically American!

The NPN transistors are T3, T5, and T7, and the PNP transistors are T4, T6, and T8. Thank you!

With some slight adjustment to the code, you can definitely use this board to test fuel injectors. The output duty cycle can be set to any value 0-100%, and the board is both Arduino-compatible and has the MOSFET to drive the injector (s)

Try adding a small resistance in series with the condenser, for example, 4.7 ohms. This will prevent an overcurrent condition when the MOSFEt first switches on. Also be sure you are using a high-resistance coil. If you want to use a low-resistance HEI coil, you need a pulse generator to limit the on-time of the coil.

@@DielectricVideos thanks for that will definitely try your suggestions.

I think that could work well, as it would basically behave like a 'hit and miss' engine while coasting down. I'd also suggest using accelerator pedal position as the primary fuel delivery input, then trim the fuel as needed using RPM and MAP values. The simplest system would be 1:1 TPS to duty cycle, but in reality there will likely be adjuments from RPM and MAP pressure to avoid overly rich or lean burn given the conditions.

@@DielectricVideos Thank you for your reply. I've had a go at it (accelerator pedal pot as input) and created a fork. I'm just a novice so I'm still working on the code to allow zero injector_ dc until idle catch. It's all fun stuff to me.

You can adjust the engine parameters in lines 42-60, and the AFR table can be updated on lines 93-102 on the Arduino program on GitHub. Hopefully this helps!

Using a TPS and engine RPM (Alpha-N control) requires empirical measurement of the airflow at multiple throttle positions and engine RPMs in order to build out a reference table for airflow calculation. This requires specialized airflow measurement tools and most likely dyno time. The obtained table will also be highly engine-specific, and will need to be changed with any modification to the intake or even operating altitude. While I wouldn't recommend alpha-N control for this system, you can learn more about it and potentially implement it on your car based on this link: tunertools.com/pages/load-control-and-calculation

I would suggest not using this system in an aircraft application, as such a use case requires a higher standard of reliability than a road car, and consumer-grade electronics such as the ones used here are not suitable for the application.

I initially used interrupts, but found they were prone to false triggering. There is a huge amount of spurious noise everywhere in the ECU application, so I found polling to be a lot more reliable. With polling, the likelihood of the sample being measured at the exact time of a transient is much lower (though it does occasionally happen).

@@DielectricVideos Interesting. Good to know, thanks.

If you have a 2-sensor digital ignition timing signal available, such as that provided by the optical board I designed, it is possible to use this system with any 4-cylinder gas engine. The board only supports one injector control channel, so you would need a single fuel injector or bank of concurrently operated fuel injectors. There is also an auxiliary output that you could use for the HHO delivery valve, but you would need to write custom firmware to control the metering as you need it.

Also, with this 1-coil board, a single traditional points system can also be used. The auxiliary output is also available on this board.

Ok thank

This board has a 150-ohm internal pullup resistor (R8) to +12V on the points input, but the input logic should work with a 5V square wave. If you want to remove this pullup resistor to reduce the current sinking requirements of your distributor circuit, you can remove it.

R_Liter_Pascal_Kelvin is the ideal gas constant used in PV=nRT. You should adjust vacuum_ADC_value and Pascals_Per_ADC_Unit to set your MAP sensor parameters. Arduino ADC inputs have 10 bits of resolution, and return values from 0-1023 representing voltages from 0V to V_REF (which is 5V by default, and can be changed). vacuum_ADC_value is the ADC value returned when the MAP sensor is at hard vacuum (0 PSIA), otherwise known as the offset. Pascals_Per_ADC_Unit is the number of Pascals of pressure represented by every ADC number unit. You can measure this empirically or derive it. For a 4-bar (400000 Pa) absolute pressure sensor, assuming an output range of the full ADC range ( 0-5V), the nominal value for Pascals_Per_ADC_Unit should be 400000/1023 = 391, but the exact value may need to be trimmed based on the accuracy of the sensor.

@@DielectricVideos Thank you so much for taking the time to explain all that for me. I really appreciate it a lot.

All the files (software as well as PCB) for both the single-coil and 4-coil ECM designs can be found here: github.com/JohnPattersonConsulting/DuneBuggyEFI If you'd like to reach out for more information, you can also email john@johnpattersonconsulting.com. Thanks!

If the engine has a points-type system on the magneto, then it should be possible to use this system. The board has an internal pullup resistor, and it can be connected to any "pull-to-ground" switch for ignition timing and RPM sensing. The only issue you may encounter would be due to the small delay between the points opening and the coil firing, as the coil takes a small amount of time (1 millisecond) to charge and fire after it sees the points state change. If the engine has fixed points timing, the ignition may fire slightly late as compared to the original magneto.

@@DielectricVideosthank you for the quick reply! Unfortunately the engine has a magneto ignition system without any points, they sadly got rid of the points on the modern engines. Do you think it would be possible to install a sensor that grounds the input to the ECM somehow else? Or does that not work?

It should work, assuming you can get a pulse-train ignition signal from the magneto or other ignition signal source. You'd have to adjust the value of pulsesPerRevolution to match the magneto or ignition pickup output, which I believe is 1 pulse per revolution on most single-cylinder engines. Other than that, you would also need to adjust the displacement value to match your engine.

Yep, although it's shimmed to the max. The belt on there came with the engine and had been sitting for many years, so I think it either expanded, or was the wrong size to begin with. I have a replacement one that's a closer fit, so I'll probably put it in soon 👍

I've made quite a bit of progress on the 4-coil system since filming this video, and I actually have the new ECU mounted in a more solid enclosure inside the cabin/body now! That being said, the enclosure in the video did hold up to about 2000 miles of roadservice, including quite a bit of idling in 90+ degree weather last fall, so it didn't turn out to be much of an issue in any case 👍

The main advantage of speed-density over alpha-n is that there is no need to map/tune throttle position vs. airflow over a wide RPM range. MAP sensors can also be installed using direct connection to the manifold, so there is no need for a throttle body with throttle position sensor. There is still some tuning to perform even for speed-density, because volumetric efficiency is not 100% on any engine.

Here's where I posted the code that I use on my car, assuming you want to do both ignition control and injection control: github.com/JohnPattersonConsulting/DuneBuggyEFI/tree/main/EFI%2BIgnition/Arduino%20Code If you just want to do injection control, you can use this code: github.com/JohnPattersonConsulting/DuneBuggyEFI/blob/main/MAP_EFI_V2.1.ino For your Suzuki Volty, there are a few adjustments to the program constants that will need to be made. In particular, line 24 [line 26 in the injection-only code] (pulsesPerRevolution) will need to be changed to reflect the number of ignition pulses your distributor points give per engine revolution, line 48 [line 27 in the injection-only code] (displacement_Liters) will need to be changed to the displacement of your engine, and line 50 [line 28 in the injection-only code](injector_MaxGramsPerMinute) will need to be updated to the total capacity of your injector at the operating fuel pressure, in grams per minute. There are a few other constants to adjust for things like idle duty cycle, ignition timing advance, and AFR table values once you are ready to fine-tune the system.

We did these all adjustments in the code that you given. But the engine didn’t start. We are using a denso MAP sensor (079800-5690). Is there any changes have to do Due to that MAP sensor ??

@@heshanthalpawila561 First be sure the engine starts when supplied manually with fuel from a spray can or other source. If it does, next check that the activity LED changes form flashing (stalled) to steady (engine running). If the LED is steady and there is still no fuel being supplied, check injector wiring. Because there is a minimum idle injector duty cycle, as long as the ECU detects the engine turning, some amount of fuel should be delivered regardless of the MAP sensor state.

Possibly. The key to operating a diesel engine is the correct timing of the injectors, as the injection timing is what initiates the combustion stroke. If this board were used to drive diesel injectors, it would be necessary to synchronize the injector with a known crank position source. Unlike a gas engine, there is no need to perform AFR calculations, as the diesel engine operates with unrestricted airflow in a lean-burn condition. The injector on-time would be synonymous with throttle position, and as a result, could be obtained from a pedal encoder or potentiometer. If your intention is to use diesel solenoid injectors on a gas engine, then it will most likely not work. Diesel solenoid injectors typically operate at extremely high pressures (10,000s of PSI), and they are also lubricated by the diesel itself. As a result, they would not work well with a low-pressure gasoline system.

@@DielectricVideos Interesting! do you think a signal could be taken from the existing diesel injector driver circuit in place of the distributor for timing, and then be manipulated to extend/shorten output injector pulse duration according to throttle position?

@@scottneels2628 Typically diesel engines operate in lean combustion with no air throttling, and instead throttle only by fuel delivery. For a diesel cycle, it may be better to simply pick up accelerator pedal position using a potentiometer or "pedal by wire", then convert it to an injector PWM signal. You might be able to do this with just a 555 chip and a MOSFET.

The TPS would be helpful for faster response, though either MAF (mass airflow) or speed-density (MAP/RPM) would still be needed to get the fuel metering correct. Since it's not such a fast system, the 328p is adequate for now, but a faster IC would be useful for engines that rev higher or have more cylinders. The 4-coil system I'm working on actually uses 2 Nanos to split the load on fuel metering and spark timing, but I'm also looking into running an ESP32 as the MCU so that I can run a small web server on it to do realtime engine monitoring!

I would use a "Blue Pill" development board based on STMicroelectronics’ STM32F103C9T6 microcontroller that has an ARM Cortex-M3 32bit core that runs at 72MHz max. Software libraries are available that allow users to program the chip using the Arduino IDE.

They're currently not for sale, but you can view and use all the files needed to build a system here: github.com/JohnPattersonConsulting/DuneBuggyEFI

Based on the price of the components today, I estimate the total component cost to be about $30.77 and the PCB cost to be about $7.10, for a combined ECM cost of about $37.87. It takes about an hour to assemble one, so consider the value of your time as well. Note that this cost does not include the wiring, coilpack, fuel injectors, fuel line, pump, filter, etc. and is only reflective of the ECM cost. The shipping costs are also not included in this calculation, and prices of components may vary from day to day.

Selection of map sensors Pressure range and signal voltage range thank you very much

Yes, 35um (1oz) copper is what I used 👍

@@DielectricVideos My motorcycle is a boxer engine with 180° trigger. I want to modify your Optical_Spark_V2.0. What software did you use to edit? protel 99se autodesk eagle Altium Designer Summer 09

@@zhanlingsong3495 The PCB files are all Autodesk Eagle files. I believe they were created in version 9.2.2.

@@DielectricVideos Thank you again

@@DielectricVideos If I retrofit your work to my 750cc boxer engine, can I use the FAI injectors?

I was actually thinking of designing a new version of the system in the future to use an ESP32 board with WiFi. I would run a small web server on one of the ESP32's cores, the use the other core for engine management. As a result, I could show key engine parameters on an in-car tablet connected to an in-car network. It's just an idea, but I think it would be really cool to try at some point!

It may be interesting to modify the arf table in a graphical environment, I'm still a bit intrigued with an injection synchronized with the engine, but with what I've been researching, it wouldn't be easy

@@DielectricVideos That would be seriously cool!

I set the centrifugal advance constant in the program to 0 degrees/RPM, so the advance in the distributor is perfectly OK to have still. It would also be OK to lock it out and set a nonzero value in the program 👍