pcmhacking.net

closed loop idle is totally irrelevant to this AFR issue, all it does is uses timing and IAC to keep the idle at the desired idle speed... forget about it for this AFR offset

Yes I'm clear on that closed loop idle now and I learnt something there which is great.

But which problem are you talking about? The target afr going lean for a second when idling or the hot restart going lean?

vlad01 wrote:I wonder how it actually leans out? Can it be that the engine cools off overall but the sensors heat up?

If that is the case, then that would be an uncorrectable issue? only be able to somewhat botch it. What would be useful is another offset term based on hot start temps and time out. A correction factor not an actual offset as we already have such a table.

The charge temp is the estimated temperature of air and fuel entering the cylinder, part of the fuel calculation then uses this for air density. Having a MAT sensor in an intake runner allows the measurement to be taken as close as possible to the cylinder. The air (and fuel) is heated as it passes through the intake plenum, the coolant contribution table allows for correction to the charge temperature at various engine speeds and loads. A value of 70% in the table means the charge temperature is made up of 70% coolant and 30% MAT - ie if coolant is 80 and MAT is 40 with a 70% contribution then charge temp would be 68 degrees.

The inverse temp table just makes the ECU code faster instead of using the true ideal gas law, in a perfect world both the row labels and row data would be identical - so the calculated charge temp of say 68 degrees would be 68 degrees, instead on a BLCD bin when 68 degrees is calculated the ECU uses 67.14 degrees. It allows for some fudging when needed, but shouldnt just be messed with to bandaid a tune (not saying you are doing this, you are doing the right things getting VE and everything sorted first). What you are logging for charge temp is the calculated coolant contributed value above, not this modified value - its just used for the fuel calc.

So with no air movement the MAT sensor suffers from heat soak when the engine is off. Restarting it results in a reading that is much higher than the actual air charge temperature - closer to coolant temperature. Charge temp goes high, so once running the ECU pulls fuel as it thinks the density is much lower. The bodged inverse temp table at high temperatures (almost achievable in normal running of a healthy engine) means the fuel calc assumes a lower temp adding more fuel in post start, the MAT sensor cools off to read correctly pretty quickly. Combined with the slow decay of the initial offset AFR I think this is a fix for hot restarts.

VL400 wrote:

vlad01 wrote:I wonder how it actually leans out? Can it be that the engine cools off overall but the sensors heat up?

If that is the case, then that would be an uncorrectable issue? only be able to somewhat botch it. What would be useful is another offset term based on hot start temps and time out. A correction factor not an actual offset as we already have such a table.

The charge temp is the estimated temperature of air and fuel entering the cylinder, part of the fuel calculation then uses this for air density. Having a MAT sensor in an intake runner allows the measurement to be taken as close as possible to the cylinder. The air (and fuel) is heated as it passes through the intake plenum, the coolant contribution table allows for correction to the charge temperature at various engine speeds and loads. A value of 70% in the table means the charge temperature is made up of 70% coolant and 30% MAT - ie if coolant is 80 and MAT is 40 with a 70% contribution then charge temp would be 68 degrees.

The inverse temp table just makes the ECU code faster instead of using the true ideal gas law, in a perfect world both the row labels and row data would be identical - so the calculated charge temp of say 68 degrees would be 68 degrees, instead on a BLCD bin when 68 degrees is calculated the ECU uses 67.14 degrees. It allows for some fudging when needed, but shouldnt just be messed with to bandaid a tune (not saying you are doing this, you are doing the right things getting VE and everything sorted first). What you are logging for charge temp is the calculated coolant contributed value above, not this modified value - its just used for the fuel calc.

So with no air movement the MAT sensor suffers from heat soak when the engine is off. Restarting it results in a reading that is much higher than the actual air charge temperature - closer to coolant temperature. Charge temp goes high, so once running the ECU pulls fuel as it thinks the density is much lower. The bodged inverse temp table at high temperatures (almost achievable in normal running of a healthy engine) means the fuel calc assumes a lower temp adding more fuel in post start, the MAT sensor cools off to read correctly pretty quickly. Combined with the slow decay of the initial offset AFR I think this is a fix for hot restarts.

Yes I understand all of this but it's not stable with default values on this engine, meaning my calcVE values vary a lot from warm days and slow traffic to cold days and highway driving.

So on cold days my highway fuel usage was a lot more and constantly ran rich, then I get to the city and my AFRs were way too lean with the same VE table.

Anyway, I have 2 logs to post up and Dylan did a very good job, its the closest its been to stable. And my highway cruising AFRs to cold frosty morning was so bang on my monitors virtually show one line for target and WB lol.

There is one very cold log, then refueled and got a little heat soak and it didn't do lean, but its wasn't super hot heat soak like midday temps.

Pretty happy so far.

In other news, my seat covers, roof lining and modified front carpet is ready after several months away at the trimmers.

Holden202T wrote:

vlad01 wrote:As for this time out, is it adjustable and whats it called? Would be nice to stop it as it makes the engine stumblely at idle.

ok so max coolant temp for cold engine table is the one to dictate when it switches from the cold afr tables to the main afr tables .... you could maybe even increase this if its needing to be richer for longer.

the delay tables there are the ones you want to change times to make that 20 seconds smaller .....

tables.jpg

you'd notice it if you have the motor up to temp even, you are driving along and come to a stop, 18 seconds after you stop the target AFR will jump from main to idle table figures.

like I said I normally change this to 5 seconds for most worked motors I've tuned, I have found if its at 14.7 for too long at idle it can start to hunt, but I've never had an issue with 5 seconds ....

its probably another emissions or save fuel type of setting .... but also stock motors will run a lot leaner at idle happily ....

for reference the above tables are from $12P BLCD.

Thats what I have been looking for. It should heap a bit, idle goes a bit funny when it commands 14.7 for that 18 sec, especally hot where its normally leaner than commanded.

Thanks a heap!

I think some of this conversation could use its own thread... It would be very hard to find this info later for future reference...

I've learnt a lot in the last few pages

So here they are.

Cruise AFRs are great! other areas need a little reworking but are ok for now.

my L/100 on the highway are shithouse lol, around 8-8.5L/100

Yep, glad Im building a new engine. That should fix that problem.

Biggvl, keen to see you on Saturday, might do a comp test for shits n giggles, also weigh the car will be useful info.

vlad01 wrote:my L/100 on the highway are shithouse lol, around 8-8.5L/100

yeah positively shithouse, I don't know how you live with yourself

it is when it costs you an extra tank of fuel every week.

My old average total usage are what my best highway usage is now lol. Used to get 6.5 on the highway few years ago, not even tuned properly too.