New Scanner Feature Request

[SSpdDmon]

Could we have a WBO2 advance timer added to the scanner? Maybe it would be something we could set on a "per RPM" or "grams/sec" basis? For example, a .3 second advance when rpms or airflow is low, .25 second advance for moderate rpms or airflow, and a .2 second advance for when the motor is really huffin'. If not, a even a single field to offset the data from all of the other .pids would work. I think it would help when trying to dial in the MAF and VE tables seeing as though the PCM can switch a commanded AFR immediately, but the WBO2 won't register it immediately since it's always lagging behind.

Initially, I noticed this as the car was going into DFCO. It would take a frame or two for the WBO2 to start it's accent towards 20.9% oxygen content after the spark and fuel dropped off. Then, looking closer at WOT, it's not hard to imagine the AFR being off by 200~300RPMs - which would put you in the next cell on the VE table. If that's the case on some level across the board, then any time the commanded AFR changes or RPMs are changing in a rapid manner, the BENs are going to be wrong.

[SSpdDmon]

Anyone else notice anything like this?

[joecar]

I sometimes notice that the WBO is "lagging behind".

[dc_justin]

I don't think this is a case of the wideband being slow to respond, as most claim to be able to read fast enough to analyze individual pulses. What this is more likely is a lag due to fuel puddling/boiling. Access to transient fueling tables would give the ability to adjust this.

Just my observation, as this correlation is very strong with a 2 bar SD tune.

[SSpdDmon]

I don't think this is a case of the wideband being slow to respond, as most claim to be able to read fast enough to analyze individual pulses. What this is more likely is a lag due to fuel puddling/boiling. Access to transient fueling tables would give the ability to adjust this.

Just my observation, as this correlation is very strong with a 2 bar SD tune.
I don't doubt you on transitions and the need for an adjustment for transient fueling. But, my point is the PCM can flick a switch and in an instant, jump from commanding 14.6 to 12.5. But, the combustion process still has to happen, the exhaust still has to be flooded with the richer biproduct, and then the WBO2 has to react. In order to measure transients accurately, I think you would need to be able to account for this.

In my experience, the combustion process can take a tenth of a second or two. In my case, those couple of tenths can mean the difference of ~200rpm at WOT. Now, I can get things close just going with the data I have. But, I believe I could get there a lot faster if I could adjust for this.

The next time you go out, log your WBO2, spark advance and have DFCO on. Watch how many frames it takes for the WB to react to DFCO coming on. That's what I'm getting at...

[dc_justin]

Then you're basically saying the same thing I was. ;) The condition you're describing is essentially the same condition that causes lean spikes on rapid throttle transitions and rich spikes when letting off the throttle. It's the fuel between the injector and the combustion chamber that is causing this delay, not a wideband that is slow to react. Sure, the injectors have shut off and the spark advance has fallen, but I'll bet you can still hear the combustion process taking place.

Why doesn't filtering it out in a log work well for you?

:cheers:

[redhardsupra]

this brings up a bigger question... we got sensors on intake side, then there's the combustion in the chamber, then we get readout of WB signal. how do all these signals get synchronized? there's a definite time delay between IAT reading and a WB reading. how does EFILive approach such an issue?

[SSpdDmon]

Filtering transients doesn't always work. I've attached a picture to illustrate.

I've marked off 1/2 of a second in a log I took this morning. At point 0, I had been in full decel for 3 seconds (0% throttle) and this data is from ~1750rpm. By point 2, DFCO is fully engaged and injectors should not be fueling the engine. It takes at least 2 full tenths of a second for the WBO2 to register the change sometime inbetween points 4 and 5.

I believe this theory is rpm dependent. The engine is essentially an air pump. It takes a split second or two for the WB to respond. At 1750 rpm, the crank is turning approximately 2.9 rotations every tenth of a second. That's 3 full combustion cycles from the time fueling stopped to the point where the WBO2 reacted in my example. At higher rpms, the delay may only be one tenth or so. But, sometimes that one tenth can put you closer to another cell than you thought you were.

[dc_justin]

Filtering transients doesn't always work. I've attached a picture to illustrate.

I've marked off 1/2 of a second in a log I took this morning. At point 0, I had been in full decel for 3 seconds (0% throttle) and this data is from ~1750rpm. By point 2, DFCO is fully engaged and injectors should not be fueling the engine. It takes at least 2 full tenths of a second for the WBO2 to register the change sometime inbetween points 4 and 5.

I believe this theory is rpm dependent. The engine is essentially an air pump. It takes a split second or two for the WB to respond. At 1750 rpm, the crank is turning approximately 2.9 rotations every tenth of a second. That's 3 full combustion cycles from the time fueling stopped to the point where the WBO2 reacted in my example. At higher rpms, the delay may only be one tenth or so. But, sometimes that one tenth can put you closer to another cell than you thought you were.

That's still the same effect as what I had described. Yes it takes a bit for the WB to respond, not because the combustion process is slow, but because there is still fuel entering the chamber even after the injector has shut down and timing has dropped. This is very apparent at the onset of DFCO as it takes it a moment as typically you can feel the engine sputter a bit before combustion fully shuts down.

The point I was trying to get at earlier is if you can get the transient fueling tables accurate in the PCM, then the amount of left over fuel in the intake runner should be minimized and reduce this delay. Personally, I'd rather have the time be put towards solving the root cause than into altering the scan tool. Then again, I have a strong interest in getting my boost fueling to come in line, as this situation below is always "scary".

http://www.marketitright.com/ss/tuning/transient-blip.gif

[SSpdDmon]

I respectfully disagree and still think they're two seperate issues.

Your log shows your commanded AFR never changed from 11.25 (aside from the first few frames). Yes, when you roll into the throttle, you see a lean spike from a fueling transition. I don't doubt that for 1 second. The issue I'm talking about is different and even appears in your log. Look back at the beginning where your commanded AFR drops from 14.7 to 11.25. Your TP% appears to be flat at ~10% (no transition). Yet, when the commanded AFR changes, it takes a few frames for the WBO2 to follow. There's no left-over fuel in the chamber there - it's adding fuel. The PCM can send an electronic signal to the injector that says, "Ok, the next time you fire, you're going to stay open longer than the last time because we need more fuel." It's that instantaneous. That richer mixture is injected, but has to burn and be expelled from the cylinder before it hits the WB.

Believe me, I'm not trying create problems or even solve them. But, in order to solve problems, you need accurate data. Like I said before, garbage in...garbage out.

Even another way to think about it. Go to McD's, grab a drinking straw and make a spitball like most of us used to in high school. :) Then, pick your target - something close and let 'er rip. Now, the signal your brain gives represents the PCM. Your diaphragm pushing air out of your lungs represents the injectors. The air passing through your mouth can be the combustion process. And finally, the spit ball can represent not just the exhaust, but the exhaust from the tail end of the previous combustion process still in the header and therefore the start of the new mixture's exhaust. Your target is then the WB. The amount of time between when your brain said, "Now!" and when the spit ball actually hit what you were aiming at is the issue I'm addressing. I believe the amount of air calculated to fire the spitball during a desired change is what you're talking about.

[dc_justin]

I respectfully disagree and still think they're two seperate issues.

Your log shows your commanded AFR never changed from 11.25 (aside from the first few frames). Yes, when you roll into the throttle, you see a lean spike from a fueling transition. I don't doubt that for 1 second. The issue I'm talking about is different and even appears in your log. Look back at the beginning where your commanded AFR drops from 14.7 to 11.25. Your TP% appears to be flat at ~10% (no transition). Yet, when the commanded AFR changes, it takes a few frames for the WBO2 to follow. There's no left-over fuel in the chamber there - it's adding fuel. The PCM can send an electronic signal to the injector that says, "Ok, the next time you fire, you're going to stay open longer than the last time because we need more fuel." It's that instantaneous. That richer mixture is injected, but has to burn and be expelled from the cylinder before it hits the WB.


I still think it's an issue of changes in fueling cause a delay in response due to extra fuel hanging around in front of the combustion chamber. My change from 14.7 to 11.25 is akin to a transient fueling condition, as the quantity of fuel is being changed. It takes time for the added fuel to evaporate/boil off the heads and enter the chamber. HPT's helpfile has a pretty good write up on this effect.

Regardless of the cause, this has sparked some thought around this and that's something we can all benefit from. I'm going to shoot Ben (EFIGUY) an email and see what his take is on the whole deal.

:cheers:

[redhardsupra]

just thinking out loud here...
where would such fuel puddle up? on the valve? wouldn't valves be hot as hell, and vaporize the fuel instantly?
considering that we're going from short to long pulse widths, there'd be a small amount of fuel puddled up, and it would have a long time to sit there (lower rpm), wouldn't it make it very likely to be all gone by the time the valve opens again?
does anyone have a timing diagram showing when valves open and when injection starts and how long it sits in the intake before it gets
swallowed in and ignited? how long does it take from exhaust valve opening to emptying out the cylinder?

[gto_in_nc]

Well, consider that a crankshaft revolution at 1000 rpm is 0.06 seconds long, at 3000 is 0.02 and at 6000 is 0.01 seconds. I'd wager (a nickel) that what you are seeing, SSpdDmon, is a function of the propagation of a wavefront of "different air" pumping through the exhaust. (Analogous to the time your spitball is in the air...) What's the distance between your exhaust port and your WB? The time for a pulse to move that far would be roughly proportional to both the length and diameter of those pipes and to the backpressure. I think.

[SSpdDmon]

Well, consider that a crankshaft revolution at 1000 rpm is 0.06 seconds long, at 3000 is 0.02 and at 6000 is 0.01 seconds. I'd wager (a nickel) that what you are seeing, SSpdDmon, is a function of the propagation of a wavefront of "different air" pumping through the exhaust. (Analogous to the time your spitball is in the air...) What's the distance between your exhaust port and your WB? The time for a pulse to move that far would be roughly proportional to both the length and diameter of those pipes and to the backpressure. I think.
My WB is in the header collector in place of the B1S1 since I'm in OL. I'd be curious to see a log from someone who has a WB in their I-pipe. Then we'd have a better idea if I'm full of it or not. :lol: