Experience with ECU Link

[Flint]

Hi all. I want to share my experience of working with ECU Link in comparison with other ECUs. There are a number of weaknesses in this computer and I would like Link to fix it.
1. Lambda regulation is very slow, reserves everywhere, you need to very accurately calibrate the fuel table in order to drive exactly according to the target mixture. For some ECU manufacturers, to start setting up the fuel table, it is enough to set the values approximately, and due to fast lambda regulation you can already drive close to the goal.
2. Link Fury has an outdated controller, Lambda LSU 4.9, which is a slow lambda. Why can’t you use ADV lambda and a fast controller for it?
3. flat shift tables: there are a lot of them, and if you install them all in the working window and save the layout, then the next time you start the program, the values in the tables themselves are filled with data from tables with similar names. It turns out that you can’t open them all, otherwise an error will occur in the program.
4. Slow boost control, when changing the boost pressure, you have to adjust the PID every time.

[Confused]

I'm not a Link employee, so my responses here have no bearing really, but from my experiences, I feel the need to respond on at least some of them!

 

1. Closed Loop Lambda has quite a few calibration & configuration options, so it's likely that you just haven't found the right settings yet for the specific vehicle. Distance of the sensor from the combustion chamber makes a difference at a variety of engine speeds, which is why you have the Update Rate Table, and then you can also determine how aggressive you want the correction to be based upon how far away from target it is with the Gain Control Table. That said, it is also designed to maintain small corrections to an already tuned fuel table, not be a substitute for tuning.

4. Like Closed Loop Lambda, Closed Loop Boost is designed to supplement an already tuned Open Loop boost table. Once this is tuned correctly, you should simply need to adjust your Target Boost value, and the closed loop will adjust based upon minor fluctuations.

 

Maybe you're expecting the ECU to do a large part of your job as a tuner?

That said, it does have tools to aid you here, as one example, there's Quick Tune, which can be used in conjunction with a steady state dyno to fill in your fuel table, and Quick Trim/Mixture Map which can use values from a log file to amend the fuel table.

[Adamw]

1 hour ago, Flint said:

Lambda regulation is very slow, reserves everywhere, you need to very accurately calibrate the fuel table in order to drive exactly according to the target mixture. For some ECU manufacturers, to start setting up the fuel table, it is enough to set the values approximately, and due to fast lambda regulation you can already drive close to the goal.

Most likely your control is poorly set up, if you attach a copy of your tune and a log we may be able to give you some guidance. 

 

2 hours ago, Flint said:

2. Link Fury has an outdated controller, Lambda LSU 4.9, which is a slow lambda. Why can’t you use ADV lambda and a fast controller for it?

You are misinformed.  The lambda controller in the Fury is a Bosch CJ135, their most feature rich digital lambda chipset.  There is no response speed difference between the 4.9 and the ADV sensors that I have observed.  The 4.9 has about 3 times better accuracy and 3 times less pressure sensitivity over the ADV.  The  only two advantages the ADV has is quicker heat up time and a slightly higher temperature rating on the hex.

 

2 hours ago, Flint said:

flat shift tables: there are a lot of them, and if you install them all in the working window and save the lockout, then the next time you start the program, the values in the tables themselves are filled with data from tables with similar names. It turns out that you can’t open them all, otherwise an error will occur in the program.

I dont follow what you are saying here.  Can you give us a screenshot to explain or perhaps describe the steps to reproduce the problem?  I spend a lot of time helping many different users with gear shift control and I dont remember seeing any odd behaviour in tables. 

 

2 hours ago, Flint said:

Furniture boost control, when changing the boost pressure, you have to adjust the PID every time.

Again this sounds like a setup problem with one of the fundamentals.  If you require different PID gains at different pressure set points then this means your system is non-linear.  PID control relies on the system response being linear over the range that you wish to control, this is not a requirement specific to Link ECU, it is the fundamental "golden rule" of any PID control.  What a "linear system" means in relation to boost control is for a given % change in solenoid duty cycle you should get approximately the same boost pressure increase regardless of whether you are at low boost pressure or high boost pressure.  Obviously a perfectly linear boost control system is never going to happen in real life as there are many variables working against you - solenoids have deadtime, increasing back pressure, variable mass flow and EGT, wastegate flow restrictions etc... But, what Im trying to say is if a single set of PID variables cant give you stable control over the range of pressures you wish to run, then there is an issue causing a large change in response linearity in your system.  That could be plumbing related, incorrect solenoid frequency or some mechanical or airflow limitation of the system.  As an example, if say increasing the duty cycle by 5% increases boost by say 30Kpa when you are running on the lowest boost setting, then you also want a 5% increase in DC at the highest boost setting to give an effect of about 30Kpa.     

 

[Flint]

7 hours ago, Confused said:

1. Closed Loop Lambda has quite a few calibration & configuration options, so it's likely that you just haven't found the right settings yet for the specific vehicle. Distance of the sensor from the combustion chamber makes a difference at a variety of engine speeds, which is why you have the Update Rate Table, and then you can also determine how aggressive you want the correction to be based upon how far away from target it is with the Gain Control Table. That said, it is also designed to maintain small corrections to an already tuned fuel table, not be a substitute for tuning. 

 

I'd love to show you my settings. But at the resolution the site allows, I doubt you’ll see anything. 

• Max total size: 4.67 kB

5 hours ago, Adamw said:

Most likely your control is poorly set up, if you attach a copy of your tune and a log we may be able to give you some guidance. 

How can I do this with this resolution? 

• Max total size: 4.67 kB

[castillaricardo]

9 minutes ago, Flint said:

How can I do this with this resolution? 

• Max total size: 4.67 kB

Upload the tune and log to a file sharing site like Google Drive, make the sharing permissions public and post the link here

[Flint]

Unfortunately, this is not the first year I have been setting it up. And I communicate with many respected, very experienced tuners. They all agree with me. I couldn't upload anything here.
I would like the computer to come to the target faster in the age of high technology. I understand that the cells need to be filled out correctly, but why is lambda regulation done if it doesn’t keep up with the changes? I can see if the cells are not very accurately calibrated, but in the cruise zone the CLL may be 7%, and in the boost zone it may be the opposite, and you need to add fuel, but it is still slowly adjusting and the value is still negative. Yes, you can line up some cells on a dyno, but the speed and boost on a 1000 horsepower engine change very quickly, and you won’t hold it for long to understand what’s happening at 8000 rpm, you need a fast lambda, otherwise you’ll have to hit that zone several times to understand. The quick tune definitely doesn’t hit, you still have to adjust it with your hands later.

On 10/18/2023 at 9:14 PM, castillaricardo said:

 

 

[Flint]

On 10/18/2023 at 3:29 PM, Adamw said:

You are misinformed.  The lambda controller in the Fury is a Bosch CJ135, their most feature rich digital lambda chipset.

Oke. Please look at my melody. If the chipset is so fast, maybe it is overloaded with unnecessary data flow? I know that g4x has a fast processor, unlike g4+, maybe the program is not optimized and the system is overloaded?

It was also noticed that lambda regulation via CAN is faster than the built-in controller