SimonSTI

Not super fancy, but I use a math block to filter my speed signal with the average function (av). Works really well.

If the engine appears to be running correctly when the lambda is reading 6.5 then yes I would assume the sensor has failed. 

Is there some documentation on that new feature with some guideline on how to set it up?
I've read the updated help file and look at all the table, but still not 100% clear to me how to set it up properly.
Do you have to setup the knock threshold table when using this mode?
 

There is a lockout setting to account for fuel priming delay, the next software update has code so that it remembers what the ethanol value was when it turned off and uses that value while in lockout the next time you start it.

For the blip in a road car application you are best to use a 2nd throttle target table so you can command a different amout of blip based on RPM or speed or both - so you get a less agressive blip (or non blip) when driving slow.  The gear shift function only gives you the ability to vary blip with gear pos.  Activate the 2nd E-throttle table using a GP output based on clutch and brake status or whatever else you want to lockout.  
Something like below.

Hey Guys,
I am thinking about putting a GPS module into the car to have access to actual speed info.
Most probably the ECU Master GPS to CAN module will be my choice here (unless you can recommend a better price/value alternate), which is outputting speed data in units of cm/s.
As far as I can see, the way to set it up is to configure the CAN frame to receive a CAN Frequency (CAN analogs for me are all used), then have that as an input parameter in a Math channel an convert the cm/s to km/h, then have the Math channel as the source for one of the Speed inputs (GP or LF/RF/LR/RR) with calibration=0 so the km/h value is simply passed through.
Then this speed input will be the driven wheel speed source which will be non slipping obviously so can be used for launch and traction control.
Can you guys confirm the above or suggest any alternates/improvements?
Thanks,
Béal

Always do the timing check with the light

I had a quick look over your log.  I agree it definately looks like there is some mild knock in the 5000-6000rpm region on at least a couple of cylinders.  The easiest test is generally to pull 5deg out of the cells around that area and if the noise levels flatten out then it is a strong suggestion it is knock.  
The delta threshold is getting tripped more often than it should be and in some places it looks like it is not getting tripped when it needs to be so I think the deltas still need a bit of work.    Add these channels to you log will help see what is going on. 

 
I will attach some example settings from our 2.0 JDM V11 STI.  Im not sure how relevent they will be but it might give you a feel for the trends.  There are definately some of these settings I havent experimented with much and probably need more optimising one day but the knock control in this car works well and can be relied on as is.  It is a bit erractic at around 2000RPM no matter what I do which is why I have it disabled below 2500. 
I only use ign trim as fuel doesnt seem to do much to supress knock in piston engines, it works well in rotaries though which is the main reason it is there as an option.

Awesome, it works really well. I've updated the base position table following what throttle position the car was using at each ECT temp. Logs for warm up is too big to attach, sending throttle test engine off logs, idle test once engine is at temperature and a hot restart. Also including the updated map.
PC Datalog - 2020-10-21 7;16;34 pm engine start.llgx PC Datalog - 2020-10-21 7;14;34 pm idle.llgx PC Datalog - 2020-10-21 6;57;07 pm throttle test.llgx Simon Gagnon-Subaru STI 2.35 G25-660 G4X-V1-5-2 throttle PID test Adam Value.pclx

Not super fancy, but I use a math block to filter my speed signal with the average function (av). Works really well.

I've set up a good amount of tablet displays for customers and 9 out of 10 of them always try clicking on the status window for a certain function e.g ALS etc.
I've no idea how difficult this would be to implement but think it would be fantastic for the end user.
I think this could be a great option to toggle certain functions like fuel/ignition table set,boost target,fan control over-ride even as well as the motorsport features,maybe even a launch rpm bump up/down function.
Not sure if someone may have asked previously but read through a good amount of pages tonight back to around 2018 so thought i'd put it up.
Hopefully i'm not the only one that thinks this.
Thanks

I've helped with development of the new knock normalised mode and done lots of Beta testing. 
The Knock Normalised Mode has the big benefit that it will adapt to increased noise level, when your engine wear over time and consequently produce more background noise. In tradional mode this will lead to false knock detection. In general speaking Knock Normalised mode will  adapt to changing background noise levels and should need less adjustment from engine to engine, once you have a good basemap worked out. It also allows better adjustment how much ignition retard is applied, based on knock level detected. 
It is meant to use all the tables, which are shown in PC LINK in this mode.
I would start with a small Knock Target Table. Break points every 1000rpm and on y-axis every 50kpa (boosted engine) should be sufficient. A filled table of 2.0 should be a good to starting point.
During transient engine operation the knock levels are going higher. In the tradional mode lockouts for delta rpm and tps where set. In Knock Normalised Mode it is meant to use all the Knock Transient Tables to keep knock control active all the time. The Knock Normalised Levels will increase even more, during transient operation, than in traditional mode. The idea with all the transient threshold tables is to increase temporarly the knock treshold, instead of just deactivating knock control. 
As a starting point fill the Knock Normalised MAP and RPM Delta Level tables with factor 1.5. 
The Knock Normalised Delta RPM and RPM threshold defines when the knock multiplicator above is applied. Setting is very dependand how stable your knock engine speed ROC and knock Map Delta channels are. If there are too noisy you can apply more filter to get this channel(s) smoothed. 
I recommend the following order for setup and tuning. 
1. setup function as suggested above
2. adjust knock noise filter (same process as in tradional mode)
3. calibrate knock target table in steady state conditions
4. calibrate all knock transient tables, so knock target is little above knock levels in all knock free transient condions. 
Notes: 
- you can calibrate knock input filter also in tradional knock mode. This setting will be the same in both modes. 
- it is normal behavior that knock normalised levels go higher during transient than in steady state operation
- as always when you calibrate knock detection system use knock ears! 
- Important knock level detected must be logged with internal logger with at least 200hz (better more) to show peak values correct in PC Link
 
 

This will make checking and setting up the knock settings without needing to plug an external knock sensor and/or extra equipment. Knock audio function is already there.
Of course this requires some bandwidth (not really much see below) but USB 3.0 or disabling logging channels to the rescue.
For 32khz mono 16bit = 32000 * 16 * 1 = 512 kbps or 64 KB/s which is not much
If the electronics required don't exist on current ECU's (i don't know about G4X, i only have G4+) you may consider implementing it in the future.
Kind regards

Can anyone confirm that mapper's method is still relevant? Nothing changed in subsequent versions that would disrupt?

Here is the logic I used, seem to work. 25s on cold start, 5s on hot restart.

This went up late last week: PC Link Downloads | Link Engine Management
Release Notes: https://www.linkecu.com/pclink/Link Engine Management PC Link G4X Updates 26112020.pdf
Known issue with the drivers, see here: https://forums.linkecu.com/topic/13145-ftd2xxdll-is-missing/

I would also make sure to use a RPM level so that your wideband doesn't just light up 20 seconds after key on if you were sitting there without the engine running. 

I have a logic set up for my per cylinder widebands. My G4+ controls a relay that powers the per cylinder wideband controller on 25 sec after the engine is running from cold (Engine Run Time > 25s), and turns it on immediately when the car is fired up AND coolant is up to temp (i.e. RPM>500, ECT>60 deg C). 
I also have a Link Lambda in my downpipe also has a 15s delay built in to it, but the ECU controls the staging of that directly so as not to get any error codes from the ECU regarding unexpected function of the wideband. The delay is whatever Link builds in to them from the factory. Not so concerned about this one coming on sooner since it is located behind the turbo.
I got the rough logic a little over a couple of years ago reading through some patent documents that were filed by GM or Suzuki (can't recall anymore) that explained how exactly they were controlling their oxygen sensors for max durability and efficiency on their passenger cars.
It's more complicated than above as Adam rightly suggested (I doubt he remembers, but Adam helped me with implementing what we could of what the OEMS are doing based on the illustrations from the patent docs I was able to dig up), as they are playing with voltage over time to effect how quickly the sensors heat up, but there were some elements of what they were doing that could be duplicated by the ECU, albeit in a less granular manner.
 
 

Thats a typo you have found there sorry, I will get that fixed up.  Corrected note below.
So yes you can use A8/Aux 8 & A11/Aux 7 for what ever you wish and you dont need to worry about A9/10.

Average knock level to help adjust gain between cylinder.
 

The RS232 comm port on the G4X is now a totally separate port from the PC connection, so it cant be used to connect to PC Link at this time.  
There are definitely some other users reporting similar connection issues with G4X to what you describe.  The engineers have been working hard on it for months trying to get to the bottom of it and some of the recent firmware updates have made improvements but I still dont think it is totally solved.  In some cases where they have been able to investigate they have found wiring issues and in others it seemed to be isolated a specific laptop so there has been no silver bullet yet.  It has been hard to diagnose as we havent been able to reproduce in-house. 
In the next PC Link release (release tentatively planned for next week) there is a totally new direct USB connection method that bypasses the virtual com port system altogether, and for the few people with problem connections that have tried it so far it has been encouraging.  I will PM you a link to try.

Average knock level to help adjust gain between cylinder.
 

Not super fancy, but I use a math block to filter my speed signal with the average function (av). Works really well.

Not super fancy, but I use a math block to filter my speed signal with the average function (av). Works really well.

Not super fancy, but I use a math block to filter my speed signal with the average function (av). Works really well.