Need to measure fuel injector pulse width for MPG calculator. Cant isolate sign
GICU812
Posts: 289
As part of a larger project, I want to be able to measure the width of a fuel injector pulse, allong with the VSS (Vehicle Speed Sensor) to determine MPG.
The problem is, the injector is always powered with the engine on, it is fired when the ECM grounds the other side of the injector. So I cannot poll on a low high, as its always high, or a high low, as it never actually goes low. I always see 12V on the circuit even when the injector is firing.
Aside from intercepting the ECM signal with a transistor and using that to ground the injector, then sensing off the ECM pin, is there another way anyone can think of to monitor this?
I dont really want to go that route because im afraid of adding delay to the firing and throwing off the combustion cycle. I dont know how much delay I would actually impose, or how much the engine would actually notice, but its just doesnt seem like the best idea to me. Feel free to correct me if im wrong though.
Suggestions?
·
The problem is, the injector is always powered with the engine on, it is fired when the ECM grounds the other side of the injector. So I cannot poll on a low high, as its always high, or a high low, as it never actually goes low. I always see 12V on the circuit even when the injector is firing.
Aside from intercepting the ECM signal with a transistor and using that to ground the injector, then sensing off the ECM pin, is there another way anyone can think of to monitor this?
I dont really want to go that route because im afraid of adding delay to the firing and throwing off the combustion cycle. I dont know how much delay I would actually impose, or how much the engine would actually notice, but its just doesnt seem like the best idea to me. Feel free to correct me if im wrong though.
Suggestions?
·
Comments
What and or how are you seeing this, with a vom. if so you may not see any transition does you meter have a freq. mode try that. also you can moniter the ground side when its not grounded it high.
What would happen if I drove the stamp pin high 5v and tied it to the ECM, it would pull the pin low right? Then I would definitely get a high --> low transistion for the PULSIN. Sound plasuable?
I just hit another snag though. The speed signal. I thought it would be easy to figure out, until I started to code it.
Like I said, its measured in pulses per mile, which is easy enough to count pulses, but I need to simotaniously track time, so I can get pulses per time period to figure out miles per hour to get the miles per gallon.
Though, theoretically, the faster the pulses per time period, the shorter the pulse width, I dont know if I would be able to read off pulse width accurately enough. Hypothetically at 55MPH, and 4000 PPM, Id be seeing 61 pulses per second, 122 incidents per second ( one on and one off) or a pulse width of .008196 seconds. Does this sound plausable? Can someone reccomend a better solution?
Check out the Parallax Industrial Control manual. (page 46 shows some buffer circuits)·There is a·LOT of good stuff in this book... The opto coupler circuit (E) should get you reading those pesky 12 VDC pulses safely..
http://www.parallax.com/Portals/0/Downloads/docs/books/edu/ic.pdf
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Team K.I.S.S
Build Safe, Build Mean, Build Strong!
Post Edited (slamer) : 11/21/2007 5:10:42 PM GMT
1st lets get the speed sensor working use the count command this will give you pulses in a time frame
you can also use the freq command for this but lets start with count on one car I worked on It was a little grey wire
off the pcm. if you need to get the signal off the speed sensor then it may take more work some sensors use a sine
wave output and change the freq. Just hook up a scope and spin the wheel, On another project I did the sensor was hooked
into the transfer case.
2nd the injectors you should be able to hook up to the ground side of the injector and get a reading with a scope now
the voltage will be toooo high for the stamp and there may be some spikes because of the injectors, so you will need to
work on that also, maybe just an opto chip.
I might still try the pulsin just to see what happens, and it would use less cycles, I'll try both and see which is more accurate.
From my experience, newer VSS's tend to be hall effect sensors, so in the tranny is a spoked "gear" with x number of raised levels however wide. I've got two taps, one from raw VSS and the other from the ECM buffered feed. I dont know which I'll use yet, I'll see which is easier to work with.
I still havent had a chance to check the injector pulse again, I've been ironing out the kinks in other aspects of this project. I'll let you guys know what happens.
That being said, I think this is a great idea. But one potential issue you might want to consider is the fuel "curve" of the injector. Fuel flow out of an injector is not linear in relation to time, so you might need to have some method of either calculating this curve or a pre-determined value to reference at certain pulsewidths... Have you put any thought into how you'll compensate for this yet?
You might be able to get the injector flowed at varying duty cycles by a shop to help with this... Otherwise I think it'll be very difficult to get an accurate reading.
Aside from that, I think this would be a pretty simple little device to build and calculate. Good luck!
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Team K.I.S.S
Build Safe, Build Mean, Build Strong!
Eventually I might look into hijacking the serial communications to the dash display, but I dont know that I will have the patience or skill to decode it. Its probably also in the OBD2 data stream, allong with speed, injector pulse, and darn near every other signal in the car.... That would definitely be the way to go, but again, it would be a huge pain to decode all that data.
I tried to tap the VSS signal last night. With the scope, its apparent there is definite potential, At 43 MPH im seeing a pulse width of approxamately 20ms(or 25?), 71 MPH was 15ms, and 43 MPH was 35ms
I think I had calculated at 55MPH the pulse should be about .016 Seconds, or 16 Ms, so thats in the right area
Now I realize thats not exactly linear... I dont know how to explain that other than I was driving, so only able to glance at the scope. The accual accuracy of those numbers isnt as important as the fact that pulse width is directly related to speed as I had theorized. Also, shorter pulsewidth = more pulses per time period, so the count command will work as well.
The catch is that when I hooked my buffered VSS tap (from the ECM) to the stamp, through a resistor, I lost VSS signal in the car. I dont think I had signal in the stamp either.
I guess the pulsin command pulled the signal low. I dont know if the count command will do the same, I'll try tonight. I may just have to put a optoisolator in there, unless someone has a better suggestion.
Still waiting to get a chance to test the injector ideas again..