FreeFuel ECU patch - a flex fuel implementation to Evo ECU
Apr 1, 2015, 02:23 PM
#286
Newbie
Thread Starter
Nice to see that some people are using FreeFuel in the US and are happy with it.
I wonder if anybody in the US would be interested in adapter / converter designed for FreeFuel to convert sensor signal to ECU analog input. It's a fully digital adapter designed by me. Unlike analog solutions output voltage range is optimized to allow using safety features built into the FreeFuel patch and maximum ethanol content resolution. Output is also very stable is much faster to react to sensor signal change than analog designs. It is currently in use in some cars in Finland with good feedback. Installation is made as simple as possible: just pull one wire through firewall to engine bay, push pins to sensor connector frame and solder three wires to ECU.
Price would be around 100 USD shipped to US from Finland. Not sure if you have to pay taxes on top of that.
I have attached a spec sheet to this message.
I would appreciate if the discussion of the other FlexFuel solution, apparently called FiTronic, would be kept outside this thread. As far as I know it has nothing to do with this patch and it can be extremely confusing to mix the discussion about it here. Thank you.
I wonder if anybody in the US would be interested in adapter / converter designed for FreeFuel to convert sensor signal to ECU analog input. It's a fully digital adapter designed by me. Unlike analog solutions output voltage range is optimized to allow using safety features built into the FreeFuel patch and maximum ethanol content resolution. Output is also very stable is much faster to react to sensor signal change than analog designs. It is currently in use in some cars in Finland with good feedback. Installation is made as simple as possible: just pull one wire through firewall to engine bay, push pins to sensor connector frame and solder three wires to ECU.
Price would be around 100 USD shipped to US from Finland. Not sure if you have to pay taxes on top of that.
I have attached a spec sheet to this message.
I would appreciate if the discussion of the other FlexFuel solution, apparently called FiTronic, would be kept outside this thread. As far as I know it has nothing to do with this patch and it can be extremely confusing to mix the discussion about it here. Thank you.
Last edited by ast; Apr 1, 2015 at 02:25 PM.
Apr 1, 2015, 03:38 PM
#287
Dang, I just purchased all the components to make my own frequency to voltage adapter. If I knew you were going to sell one I would have waited. Is the one you are selling any different schematically than this one? (Photo attached)
Last edited by bpk1337; Apr 2, 2015 at 08:12 AM. Reason: Changed "digital" to read "frequency to voltage" for clarification.
Apr 2, 2015, 04:38 AM
#291
Newbie
iTrader: (9)
Join Date: Nov 2013
Location: earth
Posts: 73
Likes: 0
Received 0 Likes
on
0 Posts
Nice to see that some people are using FreeFuel in the US and are happy with it.
I wonder if anybody in the US would be interested in adapter / converter designed for FreeFuel to convert sensor signal to ECU analog input. It's a fully digital adapter designed by me. Unlike analog solutions output voltage range is optimized to allow using safety features built into the FreeFuel patch and maximum ethanol content resolution. Output is also very stable is much faster to react to sensor signal change than analog designs. It is currently in use in some cars in Finland with good feedback. Installation is made as simple as possible: just pull one wire through firewall to engine bay, push pins to sensor connector frame and solder three wires to ECU.
Price would be around 100 USD shipped to US from Finland. Not sure if you have to pay taxes on top of that.
I wonder if anybody in the US would be interested in adapter / converter designed for FreeFuel to convert sensor signal to ECU analog input. It's a fully digital adapter designed by me. Unlike analog solutions output voltage range is optimized to allow using safety features built into the FreeFuel patch and maximum ethanol content resolution. Output is also very stable is much faster to react to sensor signal change than analog designs. It is currently in use in some cars in Finland with good feedback. Installation is made as simple as possible: just pull one wire through firewall to engine bay, push pins to sensor connector frame and solder three wires to ECU.
Price would be around 100 USD shipped to US from Finland. Not sure if you have to pay taxes on top of that.
Apr 2, 2015, 01:31 PM
#292
Newbie
Thread Starter
This design you posted is based on analog conversion component LM2917N. I have actually designed a very similar circuit with the same chip earlier that we have used on some cars. This design you posted has 100k resistor in series with the output which is way higher output impedance that I would like to use. Let's say you connect the output to input that has 1 megaohm input impedance. This results in huge 10% error. My design used 1k output impedance to minimize the possibility for error. This is what I call analog converter design.
Advantage of this type of converter is that it's simple DIY build.
Disadvantages are:
- output voltage range is limited and it's not using the full range of the ECU input
- PWM output frequency if very low and it needs to have big RC at the output to get reasonable low output ripple voltage. This also means that reaction time will be slow, my design was about 1 second from 0 to 90% of the correct value
- not really protected against possible electrical hazards that can occur in cars, although LM2917N seems pretty robust even in automotive environment.
My new design that I call a digital converter / adapter is using a microcontroller and software to count the pulses from sensor and convert this to analog output voltage. It basically does the same job but with better resolution, less ripple voltage, faster reaction, certain software filtering to avoid problems with wire that is starting to fail and easier installation.
My new design is done so that gasoline maps are for 5% ethanol content or less and ethanol maps are for 85% ethanol content or more. The converter will output solid 0.3V when ethanol content is 0-5%, linear 0.3V - 4.7V when ethanol content is between 5% and 85% and solid 4.7V when ethanol content is 85-100%. When converter detects disconnected (or dead sensor) it will output 0V. When converter detects sensor giving fault out signal it will output 5V.
FreeFuel patch has safety features when voltage is below set value (say 0.15V) or above set value (say 4.85V). In this case FreeFuel patch will use Ethanol injector scaling and Ethanol fuel map but Gasoline ignition map. With E85 this means that ignition timing will be retarded. With gasoline this means that engine will run very rich, but it will still be safe and could save you an engine as there's no possibility of running lean or advance ignition in fault condition.
This digital design is similar to ECA, except that ECA output range is 0V for 0% ethanol and 5V for 100% ethanol. I'n not sure what happens with ECA if the sensor fails, wire is disconnected or similar. Somehow I remember that it would set the output to 50% which would lead to running lean with E85 but I may be wrong.
Apr 15, 2015, 01:58 PM
#294
Hi,
This design you posted is based on analog conversion component LM2917N. I have actually designed a very similar circuit with the same chip earlier that we have used on some cars. This design you posted has 100k resistor in series with the output which is way higher output impedance that I would like to use. Let's say you connect the output to input that has 1 megaohm input impedance. This results in huge 10% error. My design used 1k output impedance to minimize the possibility for error. This is what I call analog converter design.
Advantage of this type of converter is that it's simple DIY build.
Disadvantages are:
- output voltage range is limited and it's not using the full range of the ECU input
- PWM output frequency if very low and it needs to have big RC at the output to get reasonable low output ripple voltage. This also means that reaction time will be slow, my design was about 1 second from 0 to 90% of the correct value
- not really protected against possible electrical hazards that can occur in cars, although LM2917N seems pretty robust even in automotive environment.
My new design that I call a digital converter / adapter is using a microcontroller and software to count the pulses from sensor and convert this to analog output voltage. It basically does the same job but with better resolution, less ripple voltage, faster reaction, certain software filtering to avoid problems with wire that is starting to fail and easier installation.
My new design is done so that gasoline maps are for 5% ethanol content or less and ethanol maps are for 85% ethanol content or more. The converter will output solid 0.3V when ethanol content is 0-5%, linear 0.3V - 4.7V when ethanol content is between 5% and 85% and solid 4.7V when ethanol content is 85-100%. When converter detects disconnected (or dead sensor) it will output 0V. When converter detects sensor giving fault out signal it will output 5V.
FreeFuel patch has safety features when voltage is below set value (say 0.15V) or above set value (say 4.85V). In this case FreeFuel patch will use Ethanol injector scaling and Ethanol fuel map but Gasoline ignition map. With E85 this means that ignition timing will be retarded. With gasoline this means that engine will run very rich, but it will still be safe and could save you an engine as there's no possibility of running lean or advance ignition in fault condition.
This digital design is similar to ECA, except that ECA output range is 0V for 0% ethanol and 5V for 100% ethanol. I'n not sure what happens with ECA if the sensor fails, wire is disconnected or similar. Somehow I remember that it would set the output to 50% which would lead to running lean with E85 but I may be wrong.
This design you posted is based on analog conversion component LM2917N. I have actually designed a very similar circuit with the same chip earlier that we have used on some cars. This design you posted has 100k resistor in series with the output which is way higher output impedance that I would like to use. Let's say you connect the output to input that has 1 megaohm input impedance. This results in huge 10% error. My design used 1k output impedance to minimize the possibility for error. This is what I call analog converter design.
Advantage of this type of converter is that it's simple DIY build.
Disadvantages are:
- output voltage range is limited and it's not using the full range of the ECU input
- PWM output frequency if very low and it needs to have big RC at the output to get reasonable low output ripple voltage. This also means that reaction time will be slow, my design was about 1 second from 0 to 90% of the correct value
- not really protected against possible electrical hazards that can occur in cars, although LM2917N seems pretty robust even in automotive environment.
My new design that I call a digital converter / adapter is using a microcontroller and software to count the pulses from sensor and convert this to analog output voltage. It basically does the same job but with better resolution, less ripple voltage, faster reaction, certain software filtering to avoid problems with wire that is starting to fail and easier installation.
My new design is done so that gasoline maps are for 5% ethanol content or less and ethanol maps are for 85% ethanol content or more. The converter will output solid 0.3V when ethanol content is 0-5%, linear 0.3V - 4.7V when ethanol content is between 5% and 85% and solid 4.7V when ethanol content is 85-100%. When converter detects disconnected (or dead sensor) it will output 0V. When converter detects sensor giving fault out signal it will output 5V.
FreeFuel patch has safety features when voltage is below set value (say 0.15V) or above set value (say 4.85V). In this case FreeFuel patch will use Ethanol injector scaling and Ethanol fuel map but Gasoline ignition map. With E85 this means that ignition timing will be retarded. With gasoline this means that engine will run very rich, but it will still be safe and could save you an engine as there's no possibility of running lean or advance ignition in fault condition.
This digital design is similar to ECA, except that ECA output range is 0V for 0% ethanol and 5V for 100% ethanol. I'n not sure what happens with ECA if the sensor fails, wire is disconnected or similar. Somehow I remember that it would set the output to 50% which would lead to running lean with E85 but I may be wrong.
AST, would it be out of line for me to ask what IC you are using on your new build? I can understand if you do not want to share. Sent you a PM also.
Apr 16, 2015, 09:53 AM
#297
Newbie
Join Date: Dec 2008
Location: Chicago
Posts: 2
Likes: 0
Received 0 Likes
on
0 Posts
I did re-enter injector scalings that were modified according to the formulas ast posted regarding adjusting the size for E100 and E0. I also tried further reducing the scalings to try and make the car run richer but no luck.
Apr 20, 2015, 08:24 AM
#298
Newbie
Join Date: Sep 2011
Location: Bavaria
Posts: 29
Likes: 0
Received 0 Likes
on
0 Posts
My new design that I call a digital converter / adapter is using a microcontroller and software to count the pulses from sensor and convert this to analog output voltage. It basically does the same job but with better resolution, less ripple voltage, faster reaction, certain software filtering to avoid problems with wire that is starting to fail and easier installation.
My new design is done so that gasoline maps are for 5% ethanol content or less and ethanol maps are for 85% ethanol content or more. The converter will output solid 0.3V when ethanol content is 0-5%, linear 0.3V - 4.7V when ethanol content is between 5% and 85% and solid 4.7V when ethanol content is 85-100%. When converter detects disconnected (or dead sensor) it will output 0V. When converter detects sensor giving fault out signal it will output 5V.
My new design is done so that gasoline maps are for 5% ethanol content or less and ethanol maps are for 85% ethanol content or more. The converter will output solid 0.3V when ethanol content is 0-5%, linear 0.3V - 4.7V when ethanol content is between 5% and 85% and solid 4.7V when ethanol content is 85-100%. When converter detects disconnected (or dead sensor) it will output 0V. When converter detects sensor giving fault out signal it will output 5V.
I started adapting an ATmega328 prototype a while ago I made to translate from MAF-Signal to PWM-Signal for the LED-boost-gauge in the 1988-1992 Galant VR4 gauge cluster. As only 1 timer/counter was necessary for thie ethanol-sensor translation I then ported it to an Attiny13 but I ran into a weird bug that caused the voltage to oscillate after a few minutes. (I think it was the internal clock source getting unstable...)
Originally I was planning only to have adjustable upper and lower "dead zones" to use the full 0-5V and signaling a faulty/disconnected sensor via an LED (or tapping the line to the check engine light), but I really like the idea of using the extreme values 0V/5V for fallback/"safety mode", I really didn't recognized the freefuel patch already has such safety features...
Due to no sensor for testing (I was feeding it frequency signals from a soundcard output...) and still rebuilding the car it should go into, I kinda lost interest and set it on ice.
I'll pick up on this as soon as the car is ready and I have a sensor for testing. I could try to clean up the code (AVR asm) and push it to my git repo, but as I already decided to start from scratch (and make it properly interrupt-driven..) it would be an already-dead codebase...
If anyone thinks about using an arduino: The arduino-IDE/libraries are mostly useless for proper timing/PWM code. The Arduino hardware (especially the miniclones) is really nice and cheap for quick prototyping when coded in C, but for bigger or more time-critical applications you should at least get rid of the arduino-bootloader. I ended up coding mostly in assembly, especially when timing is involved and interrupts are useful (so thats nearly for every project...).
Apr 28, 2015, 04:24 PM
#299
I know this is an older question, but I think he means "digital" as in using a computer to count the frequency pulses from the GM sensor rather than the LM2917 chip and analog components in a more rudimentary setup. Both units will produce a voltage output like the ZT or any other type of sensor controller that has a 0v-5v analog output for ECU data-logging.
ast,
Hey ast, don't bash my 100K resistor.
j/k. The only reason for that is smoothing of the signal output to a nice steady voltage [when coupled to the LM2917's output in my configuration]. This is really coming along, though, and I like the fact that you and others are people making digital/computerized solutions--and that's where I wanted it to go, but I didn't feel the capabilities of some off-the-shelf proto boards were up to par.
Apr 28, 2015, 04:26 PM
#300
If anyone thinks about using an arduino: The arduino-IDE/libraries are mostly useless for proper timing/PWM code. The Arduino hardware (especially the miniclones) is really nice and cheap for quick prototyping when coded in C, but for bigger or more time-critical applications you should at least get rid of the arduino-bootloader. I ended up coding mostly in assembly, especially when timing is involved and interrupts are useful (so thats nearly for every project...).