Feeler: Open Source OBC Firmware

[chikinhed]

Yes. The RPM indicator is fed from the signal to the fuel injectors, which stops when you take your foot off the gas.

1) In tasks\ObcConsum\ObcConsum.cpp there are 2 numbers you can change in the formula's, they look like this:
float litresPerHour = 0.2449 * 6 * 60 * obc.fuelCons->getDutyCycle();

This formula comes up in two places in the code, one is for the average fuel consumption, the other one is for the 'immediate' fuel consumption. Change both of them! The 0.2449 is the amount of cc/second a single injector puts out (at 100% duty cycle), the 6 is the amount of cylinders in your car. You can tweak it that way. The firmware I am working on has this configurable in the configuration file and will at some point also have it in the menu, by the way.

What exactly does the 0.2449 number correlate to as it is not the flow rate of a 19 or 21.5 lb/hr injector at 100% duty cycle.

19 lb/hr = 3.33 cc/sec 328i
21.5 lb/hr = 3.76 cc/sec M3

I need to correct for 83 lb/hr injectors. Should I just multiply 0.2449*(83/19)=1.0698 to give me the new multiplier?

[vollosso]

You will also need to configure the fuel pressure.. Those 80s probably flow more like 900 cc/min

But yeah i cant calculate that number from anything.

[chikinhed]

I was able to mess around with the openOBC today. I thought I figured out where the Oil Pressure files all go to so I cut and pasted them all to the files. I threw out the .elf and the .hex files from the master directory and ran the make.exe again to make a new .elf file and then converted it into a .hex file. I loaded into the openOBC but it appears that nothing has changed at all.

Any ideas?

Sent from my GTX3582R

[daandaman]

Chikin, the dev branch firmware on github already has the oil code in it.

[Raddee]

Or, again, apply this patch (it's not much so doable by hand): https://github.com/benemorius/openOB...f09c949973594b

Explaining why just copy/pasting the code from the blog would require alot of details about how C++ and compilers work. Also, the code from the patch is much better than the one on the blog.

[chikinhed]

I downloaded the deb branch firmware, went through the whole process and now have Oil Pressure on the openOBC but I also have rhythmic flashing dots, arrows, + and the clock PM. I'll try loading it again.

The files that I cut and pasted earlier today were from the github.

Sent from my GTX3582R

[Raddee]

I downloaded the deb branch firmware, went through the whole process and now have Oil Pressure on the openOBC but I also have rhythmic flashing dots, arrows, + and the clock PM. I'll try loading it again.

The files that I cut and pasted earlier today were from the github.

Sent from my GTX3582R

The flashing symbols on the LCD are because of the experimental symbols in the LCD driver that are not working correctly yet (currently working on analysing that).
To disable the symbols, look for this line in ObcLcd.cpp:

    spi.readWrite(0x08); //must be 0x08 to enable extra lcd symbols; 0x00 will disable them

Change it to:

    spi.readWrite(0x00); //must be 0x08 to enable extra lcd symbols; 0x00 will disable them

And the display should be back to normal (except maybe a bit more dim, because the new firmware also includes backlight dimming ).

- - - Updated - - -

What exactly does the 0.2449 number correlate to as it is not the flow rate of a 19 or 21.5 lb/hr injector at 100% duty cycle.

19 lb/hr = 3.33 cc/sec 328i
21.5 lb/hr = 3.76 cc/sec M3

I need to correct for 83 lb/hr injectors. Should I just multiply 0.2449*(83/19)=1.0698 to give me the new multiplier?

What I did is use this: http://www.witchhunter.com/injectorcalc1.php
My injectors are 17 lbs/hr (as per someone from this forum, I couldn't check it anywhere), so that's 178.5 cc/min = 0.1785 l/min. This seems more/less correct for instant compared to my IC, but my average is quite off (12+ l/100km instead of 8,3 l/100km).

But I'm not entirely sure if what I'm doing is correct. So, the data the OBC samples is fuel injector duty cycle, RPM count and road speed. If someone can write up something that should get an accurate fuel consumption from that input, I'd be much obliged and I would implement it like that in the software. I have no knowledge of how to take fuel pressure etc. into account.

[FastFreddie]

[chikinhed]

Thanks Raddee and Daandaman, I now have an openOBC with Oil Pressure. Now I want to change it to say "fuel" instead of "oil" as I am going to use it for measuring fuel pressure instead. I think I remember the spot in tasks\ObcMemo\ObcMemo.cpp where it says setDisplay("oil % 2.0f psi", obc.oilPressure->getPsi()); . Can I just change the word at that location?

Edit: Yup it worked! I now can read fuel pressure on the OpenOBC. This is child's play for you guys but this is a major break through for me

[daandaman]

Well done chikin. Tbh for me it is still voodoo too. And good vid freddie

[Raddee]

Cool, glad to see it worked You can change the text as you like, just don't touch the '%2.0f' stuff or make it longer than the size of the screen.

[chikinhed]

I'm now working on trying to get the openOBC to read ethanol content. The ethanol sensors out there output a digital signal so that they can supply ethanol content and fuel temperature in the same signal. I wouldn't even begin to think that I would be able to figure out how to code the openOBC to deal with that signal but I found a thread on the SR20 forum where a couple of guys have programmed an Arduino board to convert the digital signal to a 0-5v analog signal. The Arduinos are cheap enough that I just bought one and loading the code on it is preatty much just cut and paste easy. The analog signal will be wired to the analogIn2 on the openOBC. I went through the code for the openOBC and duplicated the oil pressure code and changed all the copied 'oilpressure' to 'ethanol' and changed the calibration mV to read the corresponding ethanol %. I ran the make.exe but it's pissed off with something I did and I don't totally understand what I need to do at this stage. I need a programmer dude to assist from this point.

It would be nice to have all the different common options that people might possibly want programmed into the openOBC and then be able to select which ones you want turned on by selecting them on the SD card to make it simpler for the people that want a more plug n play set up. It is already there for the start up screen and a few other options but more is better.

[daandaman]

We will consider implementing the bifuel sensor functionality in the expansion we are working on. Could you link to the datasheet of your sensor?

[Raddee]

Please do note the OpenOBC circuitry is 3,3V for the most part - be sure to check if the inputs you intend to use can take 5V.

[chikinhed]

Hmmm. I was under the impression that the inputs were 0-5v.

I'm kind of wishing that there were more inputs.

Sent from my GTX3582R

[vollosso]

Hmmm. I was under the impression that the inputs were 0-5v. I'm kind of wishing that there were more inputs. Sent from my GTX3582R

we just started work on an expansion board. I just ordered a 4 port mac valve for openobc ebc

[chikinhed]

Nice.

Sent from my GTX3582R

[bry195]

there is a frequency to analog convertor module (board) on ebay. its cheap.

the flex fuel sensors output a frequency and a pulse width. one is fuel temp and the other is flex content. if you look at the directions for the aem flex fuel sensor it explains the conversion formulas.

[chikinhed]

The flex sensor calibration is pretty simple...

ethanol % = [ Frequency (Hz) - 50 ]

Sent from my GTX3582R

[dwonda]

I'm not trying to sound negative or anything I think it all sounds pretty cool, but is there any reason why someone would need to know the ethanol content or fuel temperature? Also does reading fuel pressure require an additional sensor to be added? And can the openOBC tap into the intake air temp sensor?

[Raddee]

Hmmm. I was under the impression that the inputs were 0-5v.

I'm kind of wishing that there were more inputs.

Sent from my GTX3582R

I just checked, the analog inputs are 3,0V (and they have a pullup). Not sure why though.
Ohwell, this wednesday we had a meeting about an expansion board for boost control and other things. Good things are on the way

[chikinhed]

I'm not trying to sound negative or anything I think it all sounds pretty cool, but is there any reason why someone would need to know the ethanol content or fuel temperature? Also does reading fuel pressure require an additional sensor to be added? And can the openOBC tap into the intake air temp sensor?

Ethanol content is much more relevant to those that have a turbo or supercharger installed. By adding 10 % ethanol to the 94 octane ethanol free fuel that we have allows me to turn the boost up by about four psi which translates to about another 60 whp or more. I can do the math and calculate how much ethanol I think I should have but the ethanol percentage can vary quite a bit from pump to pump and depending on the season. Knowing the exact ethanol content would let me know if it's safe to turn the boost up. If these guys make an openOBC boost controller then maybe it could be programmed to automatically turn the boost up or down for me.

Fuel temp may be interesting for those with a modified fuel system to see if they are heating the fuel by running multiple fuel pumps but not useful on a day to day basis.

Fuel pressure would require a pressure transducer. I am simply going to us the oil pressure code and transducer that has already been documented. I've just gone into the code and changed it to read 'Fuel Press' instead of 'oil'.

I'm not sure if intake air temp can be accessed by the openOBC. It would be nice to have some code written that had access to it that could compare it to ambient when above a certain injector duty cycle. That could trigger a warning or have boost decreased for those with water injection. You can also use it for seeing if you will have more power available as you get a 1% increase in power for every 10°F decrease in intake air temp or a decrease of 1/10 of a second in the 1/4 mile for every 10°F decrease.

Sent from my GTX3582R

[Raddee]

The OBC should be able to read the intake temp. from the DME but so far components on my OBC blow up when I try to hook it up to the DME so I can't even begin to work on software for that. Still clueless about that and the person that did the hardware design doesn't seem to respond so given up on that for now

[dwonda]

The OBC should be able to read the intake temp. from the DME but so far components on my OBC blow up when I try to hook it up to the DME so I can't even begin to work on software for that. Still clueless about that and the person that did the hardware design doesn't seem to respond so given up on that for now

I thought the OBC was already hooked up to the DME?

And Chikenhed, basically you are telling me you can get more power by adding more Ethanol? I find this fascinating because I know Ethanol has less potential energy than gasoline, and we aren't talking about ignition timing neither but boost. Does ethanol have a different ideal air to fuel ratio?

[chikinhed]

I thought the OBC was already hooked up to the DME? And Chikenhed, basically you are telling me you can get more power by adding more Ethanol? I find this fascinating because I know Ethanol has less potential energy than gasoline, and we aren't talking about ignition timing neither but boost. Does ethanol have a different ideal air to fuel ratio?

I'm quoting some one else here....

"93 Octane Gasoline = 280 Degrees C
E85 = 365 Degress C

What this means is that you can heat up and compress an intake charge with E85 a LOT more than pump fuel before it auto-ignites. In laymans terms, this means you can run higher compression ratios and higher inlet temps; a side effect of spinning a turbo or SC out of the center of it's efficiency range."

From my own experience of adding just 10-12% of e85 to a tank of 94 Chevron allowed me to tun up the boost to 24 psi without knock from 19 psi. I actually stopped at 24 psi, I may have been able to get more. Sometimes it's best not to be too greedy. Sent from my GTX3582R