Traditional Fuel Model + Flex Fuel

[Link2ThePast]

Hi guys,

I've watched the HPA TPS+Map webinar a couple times and searched to try and find a detailed enough answer but I haven't been satisfied with them enough to be able to come up with a conclusion on my own.

I have fitted 20V blacktop throttles to a Toyota 2zz engine that was previously boosted a/ running flex fuel; ECU is an ExtremeX. Because of this, the previous fuel model is set to "modelled-multifuel". The engine is still boosted as well.

I've read through the link help file and it suggests the same as HPA does in the TPS+Map webinar with the R32: TPS as y axis on the main fuel table, Lambda target with MAP/MGP as the Y axis, and then tune on open loop; As per the help file, this requires the Traditional fuel model. From what I've read, Alpha N is TPSvsRPM so Fuel main isn't an issue but it leads to my first question:

1) What options do I have for the X axis for lambda target? I know RPM but is there a reason to look at any others? If so, what are the advantages and disadvantages.

Additionally, I was doing some searching, and I also saw that I could do a secondary fuel table in Link. This pulls up the option to use interpolation between a full e-85 (2nd table) and pump gas (main fuel table). I'm assuming for the 2nd fuel table, it's TPSvsRPM as well but with values (for a tank full of E85) to get to 1 lambda. As for the ignition tables, I suspect I need 1 for pump gas and a secondary one for E85.

2) I'm confident that this is the best option but if not, please explain an a better alternative and why my assumption is wrong.

Moving on from that, it also brings up an additional table "Fuel table ratio".  In the help file it mentions a typical axis with %Ethanol, Multifuel blend ratio, or some other analog variable. 

It says it controls the influence between ignition table 1 and 2, with values ranging from 0-100; where 0=Ignition 1 with full control, 100=Ignition 2 with full control, and 50 being split evenly. The table is also a single row. (I'm interpreting these as %Offset or something similar to that)

I'm certain that I'll need both considering I need to interpolate since ethanol allows greater timing advance but this leads to my next few questions:

2) For the ratio table, what would be the best choice for the axis? %Ethanol or Multi Fuel Blend?

3) Would my values just be different ethanol percentages ranging from 0-100? How fine should they be? Would the %Offset just follow those values equally? (please see attached snip)

Thanks in advance (pun not intended)

https://imgur.com/a/49a0aMx

[Adamw]

7 hours ago, Link2ThePast said:

've read through the link help file and it suggests the same as HPA does in the TPS+Map webinar with the R32: TPS as y axis on the main fuel table, Lambda target with MAP/MGP as the Y axis, and then tune on open loop; As per the help file, this requires the Traditional fuel model.

You dont need traditional, modelled will work exactly the same.  Modelled makes more sense for blended fuels.

 

7 hours ago, Link2ThePast said:

What options do I have for the X axis for lambda target? I know RPM but is there a reason to look at any others? If so, what are the advantages and disadvantages.

Really you dont even need RPM as a general comment, load is the main factor that needs to be referenced.  However, RPM can be useful for situations such as a modified engine that doesnt like to idle at lambda 1.0 but will cruise happily at that.  Or when you have something very knock limited you can use a richer target right around the peak torque RPM to reduce knock risk, then lean it off once past peak torque so you arent giving away too much power from being excessively rich.  

 

7 hours ago, Link2ThePast said:

Additionally, I was doing some searching, and I also saw that I could do a secondary fuel table in Link. This pulls up the option to use interpolation between a full e-85 (2nd table) and pump gas (main fuel table). I'm assuming for the 2nd fuel table, it's TPSvsRPM as well but with values (for a tank full of E85) to get to 1 lambda. As for the ignition tables, I suspect I need 1 for pump gas and a secondary one for E85.

Read the flex fuel set up and tuning guide in the help file.  If you are using modelled mode then your fuel table represents VE (air flow), so in theory it shouldn't change when you change fuel.  In reality however it sometimes does change a little due to egt differences and scavenging effects etc.  I have only done a few full flex tunes as E85 is only available as an expensive race fuel here now but I found provided you have good injector data you usually dont need a 2nd fuel table enabled.  For ignition tables yes, you would typically set up 2 tables that are blended between, one is tuned on straight petrol the other on straight ethanol or at least the highest concentration you will run.  With ignition however the blend isn't linear like it is with most fuel related stuff, so you need to check it at a few concentrations in between the extremes.  For example you will have a big difference in Ign timing going from E0 to E50, but from E50 to E100 there will be little difference.

8 hours ago, Link2ThePast said:

For the ratio table, what would be the best choice for the axis? %Ethanol or Multi Fuel Blend?

Multi fuel blend is generally the better option as then 0-100% matches your min and max ethanol content.  So for example if you tuned the first table with E0 and the second table with E80, then you want to be using 100% table 2 when you are running on E80.  If you had ethanol content on the axis then when you had E80 in the tank you would be pulling numbers only 80% of the way between table 2 and table 1.  

[koracing]

I personally have found it's more often required to have two different fuel tables than to have just one.  Getting the first pump gas table as good as possible as a copy and paste (export/import) starting point for the second table is always advisable as it reduces the total time and tuning effort down the road for sure.  After tuning is completed it's pretty rare that I've seen second ethanol based fuel table look just like the first one.  I think the flow through injectors and the VE involved with extra mass of fuel generating more enthalpy (in turbocharged applications of course) seems to change up the on-boost fueling requirements a bit. 

[Link2ThePast]

Thanks for getting back to me you guys. I do have a few more questions stemming from the answers now though.

On 2/16/2023 at 7:50 PM, Adamw said:

Modelled makes more sense for blended fuels.

1) I thought so as well, but why does the Help file suggest having open loop correction on? To do so, I'd have to change to traditional. 

2)Would setting the CLL mode to Off in Modelled-Multi Fuel have the same effect as having open loop correction on in the Traditional model?

On 2/16/2023 at 7:50 PM, Adamw said:

Really you dont even need RPM as a general comment, load is the main factor that needs to be referenced.  However, RPM can be useful for situations such as a modified engine that doesnt like to idle at lambda 1.0 but will cruise happily at that.  Or when you have something very knock limited you can use a richer target right around the peak torque RPM to reduce knock risk, then lean it off once past peak torque so you arent giving away too much power from being excessively rich.  

That makes sense. Thanks for the insight.

 

On 2/16/2023 at 7:50 PM, Adamw said:

If you are using modelled mode then your fuel table represents VE (air flow), so in theory it shouldn't change when you change fuel.

I should've been more clear, I was referring to if I changed to a traditional model. But based on your previous answers I see it is now unnecessary. 

 

On 2/16/2023 at 7:50 PM, Adamw said:

With ignition however the blend isn't linear like it is with most fuel related stuff, so you need to check it at a few concentrations in between the extremes.  For example you will have a big difference in Ign timing going from E0 to E50, but from E50 to E100 there will be little difference.

Thanks for that tidbit! That didn't occur to me that might be the case

On 2/16/2023 at 7:50 PM, Adamw said:

Multi fuel blend is generally the better option as then 0-100% matches your min and max ethanol content.  So for example if you tuned the first table with E0 and the second table with E80, then you want to be using 100% table 2 when you are running on E80.  If you had ethanol content on the axis then when you had E80 in the tank you would be pulling numbers only 80% of the way between table 2 and table 1.

3) So %Ethanol is a literal % but MultiFuel Blend is relative %; similar to MGP and MAP yes? 

 

On 2/17/2023 at 3:25 AM, koracing said:

Getting the first pump gas table as good as possible as a copy and paste (export/import) starting point for the second table is always advisable as it reduces the total time and tuning effort down the road for sure.

Good to hear my laziness is actually effective this time; this is exactly what I was planning to do. 

[Adamw]

7 hours ago, Link2ThePast said:

I thought so as well, but why does the Help file suggest having open loop correction on? To do so, I'd have to change to traditional. 

2)Would setting the CLL mode to Off in Modelled-Multi Fuel have the same effect as having open loop correction on in the Traditional model?

In traditional equation the open loop lambda multiplier is optional, in modelled mode open loop is always in effect.  CLL is not relevant, it uses the same table as a target, but it has no effect on the open loop lambda multiplier in the background.  

A rough example of what the open loop multiplier does; 

Lets say your lambda target was 1.00 and you had tuned the fuel table so that the measured lambda was sitting bang on target at 1.00, and lets say at a particular operation condition the injector PW to achieve this correct lambda was 10.0ms.  Now say you change the target to 1.05 (5% leaner), with open loop on, the ecu knows that the new target requires 5% less fuel to achieve so it reduces the injector PW to 9.5ms.  The measured lambda should follow the target with no other user adjustment.  In contrast if open loop was off and you changed the target by the same amount, the inj PW wouldnt change at all as the target is no longer part of the fuel calculation, you would have to manually go to the correct cell in the fuel table and reduce that by 5% to get the measured lambda to match the new target.