Temporary starting problems
04-08-2017 | 03:10 PM
#51
Thread Starter
Registered
Joined: Jun 2015
Posts: 192
Likes: 0
Averaging around 5,07g/s at 800rpm'ish according to OBD2 readings.
I have bought the diagnostics cable and downloaded the software for the LPG installation. I messed a bit with the mapping on idle and looked at what it did to the fuel trims. Turns out that the original map was a bit shifted and actually calibrated for colder engine/air temperatures.
See, you can calibrate the injection time in a specific RPM range. So if say at 800rpm the injection time is originally calculated to be 2ms, you can tune it down by a percentage. Say with 10% less, you would make it 1,8ms. Now it seemed that when the engine heated up, the air gets thinner, and so the injection times decrease, as less fuel is required to mix the thinner air stoic. So the mapping had to be shiften to the left, and what do you know, last week my Long Term Fuel Trims got as high as 16%, typical being 10%, they now dropped to +0,8% to +3,2%!
Previously, I had to add a bit of throttle to get the engine to start, and it would take some seconds. Forget about the engine being hot. It was a nightmare and didn't start at all. New starter made starts easier, but hot starts still struggled without adding throttle, and as soon as a second after the engine stalled it was a pain to start it up.
Now I've been idling the engine for roughly an hour in my driveway, temps at 90C, and the engine stalled several times as I was mapping the idle by turning the LPG system on and off (switching fuels). So right now, when I start the engine hot, it gives me no trouble at all. It starts within 2 seconds as it always did, without any problems whatsoever and without adding throttle at all!
Long story short; High LTFT = bad starts. 0% LTFT is great starts!
The difference is that the LPG ECU uses its own MAP sensor and uses the OBD2 data of the o2 sensor of the car itself. The Mazda ECU uses the MAF and o2 sensor.
I have bought the diagnostics cable and downloaded the software for the LPG installation. I messed a bit with the mapping on idle and looked at what it did to the fuel trims. Turns out that the original map was a bit shifted and actually calibrated for colder engine/air temperatures.
See, you can calibrate the injection time in a specific RPM range. So if say at 800rpm the injection time is originally calculated to be 2ms, you can tune it down by a percentage. Say with 10% less, you would make it 1,8ms. Now it seemed that when the engine heated up, the air gets thinner, and so the injection times decrease, as less fuel is required to mix the thinner air stoic. So the mapping had to be shiften to the left, and what do you know, last week my Long Term Fuel Trims got as high as 16%, typical being 10%, they now dropped to +0,8% to +3,2%!
Previously, I had to add a bit of throttle to get the engine to start, and it would take some seconds. Forget about the engine being hot. It was a nightmare and didn't start at all. New starter made starts easier, but hot starts still struggled without adding throttle, and as soon as a second after the engine stalled it was a pain to start it up.
Now I've been idling the engine for roughly an hour in my driveway, temps at 90C, and the engine stalled several times as I was mapping the idle by turning the LPG system on and off (switching fuels). So right now, when I start the engine hot, it gives me no trouble at all. It starts within 2 seconds as it always did, without any problems whatsoever and without adding throttle at all!
Long story short; High LTFT = bad starts. 0% LTFT is great starts!
The difference is that the LPG ECU uses its own MAP sensor and uses the OBD2 data of the o2 sensor of the car itself. The Mazda ECU uses the MAF and o2 sensor.
04-10-2017 | 12:47 PM
#52
Thread Starter
Registered
Joined: Jun 2015
Posts: 192
Likes: 0
Current starts. Battery is still a dead one. The LPG system recorded 288 starts with a too low voltage error (sub 10V) for a split moment, until it starts building up towards 11,5 during starts.
The new mapping definitely helped, but it was just quick and I need to put some more work into it so it's really well calibrated. I will also look into the option of letting the engine run exactly stoich on WOT.