MAF Scaling - Open/Closed Loop Fueling info

[dave-]

Default settings but limit to 3v as suggested.

[Kodename47]

You have decent data so lower the dV/dt to something like 0.3/0.25. Looking over the mean/mode data I would say that the 1st rescale was pretty good. I would revert to that, smooth it out (I've added the data below) and then use the Engine Load Comps to get the last bit dialed in as you'll get a finer tune as the last bit of variance is probably due to changing manifold pressures.

It's worth using the log view tab to give you a good overview of LTFT and STFT

[Table2D]
0.898	0.938	0.977	1.016	1.055	1.094	1.133	1.172	1.211	1.25	1.289	1.328	1.367	1.406	1.445	1.484	1.523	1.563	1.602	1.641	1.68	1.719	1.758	1.797	1.836	1.875	1.914	1.953	1.992	2.031	2.07	2.109	2.148	2.188	2.227	2.266	2.305	2.344	2.383	2.422	2.461	2.5	2.578	2.773	2.969	3.203	3.438	3.711	3.906	4.063	4.297	4.492	4.727	5
0.850342095	1.29491002	1.50415138	1.726490438	1.965994823	2.230315661	2.522716665	2.83903296	3.177060843	3.533967423	3.911788416	4.321375585	4.782655239	5.308370113	5.846223021	6.423442554	7.049371672	7.681936693	8.404752493	9.174043798	9.961086225	10.83313208	11.73953028	12.67853832	13.66015129	14.69431467	15.77506762	16.89282169	18.05510235	19.27290268	20.54888573	21.87280598	23.22922325	24.62534809	26.07107887	27.5764925	29.14266014	30.76554508	32.44162388	34.17392902	35.96828251	37.82774658	41.27758583	51.36894608	63.05139932	79.5941153	96.81557007	121.6440552	139.5471259	155.2441674	180.9846387	207.9978871	244.997	296.216

[freerunner]

Some general advice:

• the default settings for max_iat is in Fahrenheit. You have to change it to a reasonable value
• if you have enough sample data, reduce dv/dt value to increase accuracy

Have a look at this example:


Some thoughts:
I assume that your "800 A02G MAF CL" is the maf scale you started with, isn't it? It definitely does not look oem-ish, since it's already off.

If you're using cruise control, make sure the engine runs at different rpms for some time to gather enough data for different maf voltages. Use the higher rpms as well. You definately want to get data for 3.20V in closed-loop(!) and that's really difficult. 3.2v is used in open-loop as well, this is where load limits come into play. Have a look at @Wayno 's thread, to understand how it's done.

I've thrown the data from your first run into the vgi tool and edited it manually to reduce the bump that vgi tool exaggerated.

This is the result, formatted for RomRaider:

[Table2D]
0.898	0.938	0.977	1.016	1.055	1.094	1.133	1.172	1.211	1.25	1.289	1.328	1.367	1.406	1.445	1.484	1.523	1.563	1.602	1.641	1.68	1.719	1.758	1.797	1.836	1.875	1.914	1.953	1.992	2.031	2.07	2.109	2.148	2.188	2.227	2.266	2.305	2.344	2.383	2.422	2.461	2.5	2.578	2.773	2.969	3.203	3.438	3.711	3.906	4.063	4.297	4.492	4.727	5.0
1.029	1.514928254206154	1.7854511567429672	2.016297366907714	2.171397164362154	2.4779897872372088	2.8723520629353185	3.213812082137341	3.6057697098129013	3.9231832487894285	4.401708205276769	4.686658995948879	5.0605817634553185	5.361162766274	5.986371252136857	6.5334286772668575	7.044416382058616	7.70689691227099	8.392221598697583	9.283143691052155	10.13198444301211	11.085427383952968	12.098986525457562	12.981107171381447	13.660391017596005	14.724515247918555	16.008930066512995	16.465358003588204	17.6245198570147	18.39369093249791	19.777754143348883	21.000733009055068	22.591672322162538	23.86109346483927	25.308604435098285	26.626383312277333	28	30	31.5	33.56470134642	35.758709892278574	37.85333820183826	41	51.24455641064684	61.940777131117194	76.20261494572648	94.527	126.111	142.083	155.297	180.36	209.266	244.997	296.216

It's not done yet, you should try to get more data. Try to drive dynamically with different loads while avoiding going open-loop.

[freerunner]

Yay! @Kodename47 beat me to it.

[dave-]

Originally Posted by Kodename47

You have decent data so lower the dV/dt to something like 0.3/0.25. Looking over the mean/mode data I would say that the 1st rescale was pretty good. I would revert to that, smooth it out (I've added the data below) and then use the Engine Load Comps to get the last bit dialed in as you'll get a finer tune as the last bit of variance is probably due to changing manifold pressures.

It's worth using the log view tab to give you a good overview of LTFT and STFT

Thanks for that, will flash your scale and do some more driving to gather additional data. Hopefully the weather doesn't change too much, but night is the only time I'll be able to avoid the traffic and it's getting cold (~10degC ambient) which isn't ideal. Need a temperature controlled dyno cell!

The Engine load table hasn't been altered from stock, doing a test run with the current data in the load rescale gives MP calc changes in the final 1.01 column only. Is that what I should expect once I have data from your revised scale?

Originally Posted by freerunner

If you're using cruise control, make sure the engine runs at different rpms for some time to gather enough data for different maf voltages. Use the higher rpms as well. You definately want to get data for 3.20V in closed-loop(!) and that's really difficult. 3.2v is used in open-loop as well, this is where load limits come into play. Have a look at @Wayno 's thread, to understand how it's done.

Try to drive dynamically with different loads while avoiding going open-loop.

Avoiding the open loop scenario isn't so easy! I looked at Wayno's thread, but didn't quite see how specifically to chase 3.2V in closed loop?

Also didn't think about Celsius vs Fahrenheit. And you were correct about the scale, it isn't the OEM one. I probably should have started there but it was supplied by someone familiar with the same turbo kit I have installed. Expected it to be close enough straight up but wasn't the case.

[Kodename47]

Originally Posted by dave-

Thanks for that, will flash your scale and do some more driving to gather additional data. Hopefully the weather doesn't change too much, but night is the only time I'll be able to avoid the traffic and it's getting cold (~10degC ambient) which isn't ideal. Need a temperature controlled dyno cell!

The Engine load table hasn't been altered from stock, doing a test run with the current data in the load rescale gives MP calc changes in the final 1.01 column only. Is that what I should expect once I have data from your revised scale?

Avoiding the open loop scenario isn't so easy! I looked at Wayno's thread, but didn't quite see how specifically to chase 3.2V in closed loop?

Also didn't think about Celsius vs Fahrenheit. And you were correct about the scale, it isn't the OEM one. I probably should have started there but it was supplied by someone familiar with the same turbo kit I have installed. Expected it to be close enough straight up but wasn't the case.

Depending on your CL/OL transition settings you may not get much higher MAFv ranges in closed loop, I wouldn't worry too much. I set the max to 5v just to see the data I'm getting. If you use the comparison button you can manually adjust the scale, I did that to compare yours to OEM to allow me to make it smoother.

What logging/tuning system are you using? If not ECUtek beware of differences in pressures between logging and the table definition. Not unusual for MRP/MAP or bar/psi. Sounds like the table is in MAP bar, how are you logging pressure? As ELC is just done in closed loop I rarely see changes made to the last column, but you should have many changes elsewhere.

[dave-]

Originally Posted by Kodename47

Depending on your CL/OL transition settings you may not get much higher MAFv ranges in closed loop, I wouldn't worry too much. I set the max to 5v just to see the data I'm getting. If you use the comparison button you can manually adjust the scale, I did that to compare yours to OEM to allow me to make it smoother.

What logging/tuning system are you using? If not ECUtek beware of differences in pressures between logging and the table definition. Not unusual for MRP/MAP or bar/psi. Sounds like the table is in MAP bar, how are you logging pressure? As ELC is just done in closed loop I rarely see changes made to the last column, but you should have many changes elsewhere.

Haven't changed any of the CL/OL transition from stock.

I'm using BRZedit for changes as that is the format the base tune with injector scaling data was supplied in. Also got it for the live log ability but find the sample rate is terrible (doing something wrong obviously). I should transpose the relevant changes back to a romraider bin file and work from there as it would be easier working with the mafscale utility and info on the forums.

Differences from current file to a stock A01G below, BRZedit names the tables differently in some cases. My current tactrix config file to generate the logs is attached, thought I'd attached it in the earlier post which might have helped.

1D: Front Oxygen Sensor Rich Limit: Data
1D: MAP Scaling A Mul: Data
1D: MAP Scaling A Base: Data
1D: MAP Scaling B Mul: Data
1D: MAP Scaling B Base: Data
1D: Mass Air Flow Limit A: Data
2D: Engine Load Limit A: Data
2D: Engine Load Limit B: Data
3D: Calc Engine Load B: Data
1D: Injector Flow Scaling: Data
1D: Injector Minimum Open time : Data
1D: Injector Flow Scaling v2: Data
2D: Low Pulse Width Fuel Injector Compensation: Row
2D: Low Pulse Width Fuel Injector Compensation: Data
2D: Throttle Delta Fuel Comp (Delta) A: Data
2D: Throttle Delta Fuel Comp (Delta) B: Data
2D: Cranking Injector Pulse Width: Data
2D: PI Manifold Relative Pressure Comp: Data
2D: DI Quantity Maximum Allowed: Data
3D: GDI Pressure Target B: Data
3D: Total Injection Ratio GDI port (Cold): Data
3D: Total Injection Ratio GDI port (Warm): Data
3D: Total Injection Ratio GDI port (Hot): Data
3D: Fuel Map: Col
3D: Fuel Map: Data
3D: Cranking Fuel Injector Pulse Width A (ECT): Data
3D: Cranking Fuel Injector Pulse Width B (ECT): Data
3D: Cranking Fuel Injector Pulse Width C (ECT): Data
3D: Cranking Fuel Injector Pulse Width D (ECT): Data
3D: Cranking Fuel Injector Pulse Width E (ECT): Data
3D: Cranking Fuel Injector Pulse Width F (ECT): Data
3D: Cranking Fuel Injector Pulse Width G (ECT): Data
3D: Cranking Fuel Injector Pulse Width H (ECT): Data
3D: Cranking Fuel Injector Pulse Width I (ECT): Data
3D: Cranking Fuel Injector Pulse Width J (ECT): Data
3D: Injectory Battery Offset: Col
3D: Injectory Battery Offset: Data
1D: Initial Advance Multiplier: Data
3D: Ignition Base (Safe): Col
3D: Ignition Base (Safe): Data
3D: Ignition Base (Normal): Col
3D: Ignition Base (Normal): Data
3D: Ignition Advance: Col
2D: Idle Target A: Data
2D: Idle Target B: Data
2D: Idle Target C: Data
2D: Idle Target D: Data
2D: Idle Target E: Data
2D: Idle Target F: Data
2D: Idle Target G: Data
2D: Idle Target H: Data
2D: Idle Target I: Data
1D: Manifold Pressure Sensor Limits (CEL): Data
1D: MAP Scaling C Mul: Data
1D: MAP Scaling C Base: Data
2D: MAF Sensor Scaling: Data
2D: Front O2 Sensor Scaling A: Row
2D: Front O2 Sensor Scaling A: Data
2D: Front O2 Sensor Scaling B: Row
2D: Front O2 Sensor Scaling B: Data

[freerunner]

Originally Posted by dave-

I'm using BRZedit for changes as that is the format the base tune with injector scaling data was supplied in. Also got it for the live log ability but find the sample rate is terrible (doing something wrong obviously). I should transpose the relevant changes back to a romraider bin file and work from there as it would be easier working with the mafscale utility and info on the forums.

I used to have two instances of the vgi tool running. Doing all the copy & paste work for 2D and 3D tables between BRZedit and RomRaider in one of them via the POL table window. C&P in there is supported in different formats.

What sampling rate do you usually get?

[Kodename47]

@dave- you're logging MAP in kPa. Need to divide the logging value by 100 to get Bar as per the table.

[dave-]

From memory it was ~150ms+ freerunner with only half a dozen parameters selected. Tactrix with 20-30 is much faster.

Thanks Kodename47. Haven't been able to get out there and play around, weather is crazy cold now

[solidONE]

I'm not sure whether if this has been brought up in this thread. I just want to make sure I'm on the right page with this. The CL fueling compensation table values adjustment can cause the Cl fueling to go richer or leaner than what is indicated on the OL fuel table, yes? Changing the value to 0 would cause the CL fueling to match that of the OL fuel map? Or am I oversimplifying.

OL fuel map:


CL Fuel comp:


Resulting Cl fueling: http://datazap.me/u/solidone/z6-kc1i...mark=1574-1575


CL fuel comp adjusted:


Result: http://datazap.me/u/solidone/z7-kc1i...zoom=7483-7927

The adjustments to the OL fuel table and Cl fueling compensation did not seem to have too much if any effect on CL fuel target at 0.6 loads. Though once it switches over to OL operation fuelking is right on target.

[Wayno]

Originally Posted by solidONE

Changing the value to 0 would cause the CL fueling to match that of the OL fuel map?\

No

Did you read the table description and the previous threads on CL?

CL fueling is an adjustment from 14.70. It has nothing to do with the values in the OL table.

After the ECU determines the CL fuel value and the OL fuel value it may chose a point between them. Until then CL values have nothing to do with OL values.

All setting it to 0 will do is disable LTFT and/or disable CL.

[solidONE]

Originally Posted by Wayno

No

Did you read the table description and the previous threads on CL?

CL fueling is an adjustment from 14.70. It has nothing to do with the values in the OL table.

After the ECU determines the CL fuel value and the OL fuel value it may chose a point between them. Until then CL values have nothing to do with OL values.

All setting it to 0 will do is disable LTFT.

Okay that would explain my results... If it's adjusted from 14.7 then I've been going the wrong direction on my adjustments. I didn't find any details on the Cl fueling compensations. I never bothered to mess with this table myself besides testing some different settings from Kodename47 to try to stabilize LTFT in OL. Back to the drawing board.

[Wayno]

So, you have E85 in the tank?

You're going to go around in circles every time you fill up.

Can you go back to 98?


This is what you SHOULD do:

I assume you're using stock MAF sensor with same diameter MAF housing as stock.

You should have 100% 98 in the tank, use the 100% STOCK MAF scale which is very close for E0, then you can scale your injectors.

Once you have your injectors scaled, you can then do the 100% DI scaling and 100% PI scaling to tweak the injector scales if required.

THEN scale the open loop MAF.

Then you should put E85 in and re-scale the injectors for E85.


If you don't want to do that, you could do the same on E85.
You need a 100% stable blend/mixture though.
And you will need to pick a MAF scale to do the port injector scaling.
I use the CL part of the openflash scales for my E85 tunes.
Do the 100% DI scaling to scale the closed loop MAF scale.
Then do 100% PI scaling against your closed loop MAF scale to scale your port injectors.
Then combine DI/PI and do your open loop MAF scaling.


If you have a stock MAF sensor and stock MAF scale, and are seeing 20% trims, it's an injector scaling problem, not MAF scaling. Adjusting the MAF scale by that much will throw your tune out by putting you in to different load columns.


If the kit came with a different MAF scale due to having a different MAF housing diameter, then you will need to use that scale.

Note: most petrol tunes from the US and especially the openflash petrol tunes have MAF scales for E10. If the tune is for E85 though you might get lucky.

Originally Posted by dave-

Ok lets see if I can get anywhere with this, apologies in advance for the TLDR...

SBD OFT Tune
http://www.datazap.me/u/dave/brzedit-turbo
Base timing was reduced and POL made a bit richer against wideband on th dyno otherwise nothing really changed, was just to verify install.
Log was done in BRZedit, think there were too many items selected as sampling rate isn't great.

I'd gotten some sort of MAF limit error which from memory was because LTFT had hit +30 or +40, not sure exactly. Between then and now, we found out that the primary O2 bung on the downpipe was too big and the sensor probably wasn't in far enough to be reading correctly.

99.95% sure we have O2 part sorted now as I have an LC1 wideband in the front pipe before cat and there isn't a lot of difference between them until going very rich which the factory sensor isn't capable of reading? I'd have that data in my tactrix log but can't find a 2.5mm stereo adapter smaller enough to fit in the LC1 socket.

Unfortunately before fixing all that, we put in the bigger injectors in preparation for more boost and E85, rookie mistake I know.

Current Tune
This was supplied a base tune that should be suitable for my turbo kit. It has been tweaked for the ID850 injectors. Car starts perfectly cold and hot so I think we have the cranking side ok?

This is immediately after flashing the file and leaving the car idle.
http://www.datazap.me/u/dave/base-tu...ed-id850s-idle

This is cruising around the car park and down the street. There are 3 logs, second and third are because the car stalled as per the hesitation mentioned on light throttle.
http://www.datazap.me/u/dave/base-tu...d850s-cruising

I then zero'd out the 3x DI/PI tables so we'd have 100% DI to see if anyone here could see something odd in the logs. 2 logs in the link, one idle and the other a light cruise. I'm not sure if it was my imagination but the idle dip issue did seem to go away.
http://www.datazap.me/u/dave/base-tune-di-only-id850s