Gen90 DIY mapping/calibration thread

[Robbie Corbett]

Painful hahah. To be honest I don't have any road car stuff to compare it to.

I suppose it's okay. I've been setting the channels up in the gentuner software, saving and exporting into GEMS data analysis software. Maybe better software to use, I'll have to look into it.
The resolution of some of the channels isn't great, and I haven't worked out how to set the logging frequency.

So a nice trick to help you visualise your closed loop response is to get the AFR's dialed in pretty well under open loop.

Say you get a steady 14 (or stochiometric if your using a narrow band) at slight throttle, 1500 RPM. Disable any acceleration enrichment and enable closed loop fuel with the target set to be the same as your observed (real) AFR. You should have very little to no correction %.

Set your closed loop AFR target table (assume you will have one) to be something like this:

Then while you datalog AFR against time take the no load RPM upto 1800 (or whatever your richer column of cells starts). At this point the PID should kick in and start trying to correct the error. Your tuned fuel table is all setup for 14 still while the target AFR is at 13.5. The closed loop should start to try and add fuel.

Your data log is basically plotting the closed loop response curve:

Obvs won't be as pretty as that! but will give you a good idea regarding your P,I and D terms - as you already know (but others might not) you will adjust the P term initially to achieve the desired initial rate of correction, the D term to deal with overshoot (as it effectively reduces P as you approach the target) and I to deal with long term errors, or an offset which doesn't go away.

Because you really want the closed loop fueling response to be pretty slow you will end up with quite small P and D and as a result a larger than expected I. If your control loop is too fast you will generally oscillate around the set point (even more true with narrow band), but also because the AFR can change so rapidly you end up driving the loop mad. You want to be ignoring all the small fluctuations and deal with the slower errors which would be caused by environmental or running conditions which are currently outside of your setup OR in many cases due to a missing table which the ecu designer didn't think was worth it or didnt even know about.

For example if the ecu was able to correct for combustion temperature or fuel grade etc... most cannot which leads to a tune error which cannot have been tuned for. Closed loop AFR is good for dealing with that at lower RPM and light loads (cruise idle blah).

Hope that rambling made sense

This is also the approach I use to set closed loop idle - I want the control loop to be slow and deal with the offset error (so 300rpm error), I use the faster advance curve to smooth the fast fluctuations. The advance method gives you a nice smooth idle, the closed loop idle correction gives you an idle at the correct RPM. If you ecu used PID idle it is likely the same algorithm the programmers used for AFR so the size of the terms will be in a similar ball park (maybe).

Good luck - self tuning is fantastic fun!

 

[Robbie Corbett]

Only other advice I can think to add is to get it down to a rolling road, often you can pay by the hour to rent the dyno. I like dyno dynamics and used to pay between 40 and £75 per hour. Being able to hold loads that you rarely can reach on the road or slowly ramp RPM at a particular load and look at precisely how your timing or fueling alter torque makes life so easy.

I used to hire the dyno at Surrey Rolling Road before they moved, Austec, and Skuzzle motorsport (plus one in Guilford but I forget the name). All down south but most places are happy to earn the money during off peak hours during the week. Some will want to operate the dyno for you but ask for them to show you how (pay the extra if you have to) as its very very useful to be in control of the throttle as you you tune.

 

[scruffydubber]

Thanks for taking the time to write that @Touring_Rob. I'll work through it. Expect 500 questions coming your way 😁

 

[Nokiai]

Maybe someone could help me with wiring the lsu 4.2 wideband sensor? I have the pin connection table from ktec, even the calibration aspect, but no matter if the sensor is connected or not I get the same reading... Looked through the wiring, everything seems to be alright.

Wideband wiring

Bosch Pin No.​	Gen90 pin​	Description​
1​	21​	Wideband 2 Red IP​
2​	19​	Wideband 1 Yellow VM​
3​	80​	OX Heater ground White H -​
4​	65​	OX Heater + Grey O25​
5​	25​	Wideband 4 violet IA​
6​	22​	Wideband 3 Black UN A7​

I attached the calibration aspects for both 00X76 and 00X84 firmware versions. Maybe it will be useful for somebody.

 

[Robbie Corbett]

Guys are you sure that the ecu supports the direct hook up of wideband lambda WITHOUT a controller.

It would be quite unusual of an ECU that age to support that - usually you would need an external controller which connects to the lambda, the controller has a data output line or an output which connects to the ecu.

A controller such as the AEM UEGO

 

[incy-spider]

Guys are you sure that the ecu supports the direct hook up of wideband lambda WITHOUT a controller.

It would be quite unusual of an ECU that age to support that - usually you would need an external controller which connects to the lambda, the controller has a data output line or an output which connects to the ecu.

A controller such as the AEM UEGO

I think you are correct, I was told when I was considering using wideband the same, the PLX devices wideband gauge has an output too


Sent from my iPhone using Tapatalk

 

[scruffydubber]

I use a wideband controller with a 0-5v output calibrated into a spare channel.

It is interesting there are few pins labelled up wideband on the pinout. And the option to select wideband in the software, but I couldn't find any other info.


On a separate subject does anyone know the differences between firmware versions? I think I'm on a old version and don't know where to get a newer version, and what advantages it will have.

 

[adir]

@Nokiai
Please update if you figure it out.
I do however, not sure if the gen90 will do a better job then a dedicated controller.

@scruffydubber
What firmware version you have on your ecu?
I'll check mine this week if it helps..
I did
Email ktec regards the firmware version, I'll update when I know better.

 

[scruffydubber]

@Nokiai
Please update if you figure it out.
I do however, not sure if the gen90 will do a better job then a dedicated controller.

@scruffydubber
What firmware version you have on your ecu?
I'll check mine this week if it helps..
I did
Email ktec regards the firmware version, I'll update when I know better.

Great thanks @adir

Mine is 00x71, which potentially sounds quite old.

 

[Nokiai]

I am not sure of anything. The table I posted before was given to me by ktec and they labeled the pins themselves (bosch (sensor) -> ECU). Ktec website says that the ecu has wideband input and amplifier - but if it does not work it wouldn't be the only thing that is not true on the product description.

Will update if I figure out something.

 

[scruffydubber]

I am not sure of anything. The table I posted before was given to me by ktec and they labeled the pins themselves (bosch (sensor) -> ECU). Ktec website says that the ecu has wideband input and amplifier - but if it does not work it wouldn't be the only thing that is not true on the product description.

Will update if I figure out something.

interesting the pinout I have does match your wideband pinout, only with less information. I can't see the files you attached. Have you got the software set to wideband?

 

[scruffydubber]

I have made a Version 2 ecu pinout. I'll keep updating with any information found.

Added cam sensor, boost solenoid control, and added the wideband info from above (I have put questions marks on this until proven)

Please find attached

 

[Nokiai]

interesting the pinout I have does match your wideband pinout, only with less information. I can't see the files you attached. Have you got the software set to wideband?

Yes, I sent my ecu calibration to KTEC, they sent back with modifications for wideband sensor.

 

[Nokiai]

interesting the pinout I have does match your wideband pinout, only with less information. I can't see the files you attached. Have you got the software set to wideband?

I forgot to attach the files for wideband calibrations. Now attached to this reply

 

[adir]

i got my gen90 second hand, 5 years ago.
and it is the first time I'm trying to connect, and upload new calibration.
apparently, the previous owner locked it with password.
i did " clear calibration " in order to reset the ECU , and upload my calibration. and I'm keep getting this massage
what I'm doing wrong ?

 

[incy-spider]

Cycle the power and reconnect.. if really struggling you could give Andy a call at Phoenix automotive technologies and he can team view you and sort.


Sent from my iPhone using Tapatalk

 

[adir]

third time's a charm i guess..
thx !

 

[scruffydubber]

I forgot to attach the files for wideband calibrations. Now attached to this reply

Unfortunately I can't view as I don't have any 00x84 files.

I'm sure someone else can.

 

[adir]

Mine is also 00x71
Ktec didn't reply back 😔

 

[scruffydubber]

Mine is also 00x71
Ktec didn't reply back 😔

I managed to get a reply from Ktec about upgrading the firmware from 00x71.
This was there response:

"71 is one off the later versions off the software with no real changes since then. It isn't as simple as just an update as the map is written into the version of the software."

I'm still interested what the differences are, and if they have any affect on my setup.

 

[Robbie Corbett]

Im not sure I would trust anyone talking about software who constantly spells of as off.

 

[incy-spider]

I’d imagine the reason they say there isn’t much difference is because they don’t know the features that have been added or updated.. I remember when I spoke with Andy there are a lot of little Clio specific tweaks and throttle stuff that other mappers wouldn’t know about.

I know 100% that the earlier software didn’t support boost control either


Sent from my iPhone using Tapatalk

 

[scruffydubber]

I still haven't got the closed loop lambda control working correctly, it's very odd. It's like the ECU is putting the sensor in error, it works initially then the feedback channels freeze at a value and it no longer works.

I've had the car setup by Chris at EFIParts, he was very very helpful and did a great job. He spent around 30minutes trying to get the closed loop working and couldn't. To prevent wasting time he moved onto the rest of the map. He suggested it might be because I had changed around the channel configuration, even though its configurable in the software he's seen it before when the hardware and firmware don't quite match up if you make some changes like that on this type of ECU.

I initially had the Lambda1 connector powering my wideband and the narrowband plugged into Lambda2. (this just suited wire lengths). I've reconfigured it so lambda 1 is the original channel and plugged into the narrowband, I'm using a brand new Bosch sensor which I can see working, but for some unknown reason the ecu doesn't want to work correctly in closed loop. The ecu measures the heater element resistance, but I can't tell if the sensor is in error or not. This is why I was keen to try a later firmware.

 

[Robbie Corbett]

I guess ecu could potentially have a fault too, usually any from this era use a CJ138 or similar IC from Bosch, I guess it could have been damaged at some point.... Do they have a dedicated 0-5V input, might be better in the long run use that with a wideband + controller...

 

[Nokiai]

I still haven't got the closed loop lambda control working correctly, it's very odd. It's like the ECU is putting the sensor in error, it works initially then the feedback channels freeze at a value and it no longer works.

I've had the car setup by Chris at EFIParts, he was very very helpful and did a great job. He spent around 30minutes trying to get the closed loop working and couldn't. To prevent wasting time he moved onto the rest of the map. He suggested it might be because I had changed around the channel configuration, even though its configurable in the software he's seen it before when the hardware and firmware don't quite match up if you make some changes like that on this type of ECU.

I initially had the Lambda1 connector powering my wideband and the narrowband plugged into Lambda2. (this just suited wire lengths). I've reconfigured it so lambda 1 is the original channel and plugged into the narrowband, I'm using a brand new Bosch sensor which I can see working, but for some unknown reason the ecu doesn't want to work correctly in closed loop. The ecu measures the heater element resistance, but I can't tell if the sensor is in error or not. This is why I was keen to try a later firmware.

I think I have the same problems. ECU reading the same value, both sensor and resistance. Even after I disconnect the sensor.

 

[AlexW]

I still haven't got the closed loop lambda control working correctly, it's very odd. It's like the ECU is putting the sensor in error, it works initially then the feedback channels freeze at a value and it no longer works.

I've had the car setup by Chris at EFIParts, he was very very helpful and did a great job. He spent around 30minutes trying to get the closed loop working and couldn't. To prevent wasting time he moved onto the rest of the map. He suggested it might be because I had changed around the channel configuration, even though its configurable in the software he's seen it before when the hardware and firmware don't quite match up if you make some changes like that on this type of ECU.

I initially had the Lambda1 connector powering my wideband and the narrowband plugged into Lambda2. (this just suited wire lengths). I've reconfigured it so lambda 1 is the original channel and plugged into the narrowband, I'm using a brand new Bosch sensor which I can see working, but for some unknown reason the ecu doesn't want to work correctly in closed loop. The ecu measures the heater element resistance, but I can't tell if the sensor is in error or not. This is why I was keen to try a later firmware.

Had basically this today on a customers ITB’d car. Narrowband lambda - works briefly then freezes. Occasionally playing with settings it would come back to life, then error again.
I wonder if the heater is failing.

I will probably map this car open loop fully - it’s pretty close as is, just wanted to make sure the road use was better.

 

[scruffydubber]

Had basically this today on a customers ITB’d car. Narrowband lambda - works briefly then freezes. Occasionally playing with settings it would come back to life, then error again.
I wonder if the heater is failing.

I will probably map this car open loop fully - it’s pretty close as is, just wanted to make sure the road use was better.

That sounds exactly the same as mine. Did you have any luck in the end?

I had initially used lambda 1 to power the wideband controller. I wondered if it failed the input as effectively the heater element resistance would be wrong.

So I set/wired the lambda 1 as it should be with a new Bosch sensor. And it still does the same. The only thing left to try on mine is adding a second narrowband lambda to see if it makes it happy.

Mines running open loop and its great. But I would like to get it working.

 

[barnesautos]

I’m not so familiar with the Gen 90 , as I have the ME221 ecu fitted.
I run a WideBand controller with the wires from it , one feeds into the narrowband which the voltage from the controller reduces , iE this would keep the ecu happy .

Now if the gen 90 does have its own controller built in to the ecu , you expect to see values displayed on the laptop being the same at the gauge.

If this is not the case , then I would suggest setting up the wires into the narrowband with the wire from the controller and not direct from the sensor IE no controller.

Another question , are you using LSU4.2 lambda sensor
OR LSU4.9 ?

 

[Nokiai]

That sounds exactly the same as mine. Did you have any luck in the end?

I had initially used lambda 1 to power the wideband controller. I wondered if it failed the input as effectively the heater element resistance would be wrong.

So I set/wired the lambda 1 as it should be with a new Bosch sensor. And it still does the same. The only thing left to try on mine is adding a second narrowband lambda to see if it makes it happy.

Mines running open loop and its great. But I would like to get it working.

How is your ecu's calibration configured for open loop? Can you share the calibration file? I am taking the same route - making it open loop, because I cannot find a way to make it work with wideband.

 

[scruffydubber]

How is your ecu's calibration configured for open loop? Can you share the calibration file? I am taking the same route - making it open loop, because I cannot find a way to make it work with wideband.

Hi. I set the feedback rate to 0ms.
I can check the channel name, and any other info you need tomorrow

 

[Nokiai]

Hi. I set the feedback rate to 0ms.
I can check the channel name, and any other info you need tomorrow

The more info the better . If you show me the options like in the screenshot it would be great.

 

[scruffydubber]

The more info the better . If you show me the options like in the screenshot it would be great.

This is my setup. ignore the channel allocation, and the OX FB Rate is set to 0 ms as it's in open loop. Thanks

 

[bikemanpt]

Hello,

This fells a bit in off topic regarding the current thread focus on lambdas, closed loops and so on.
But since its about GEN90, here it goes.
Does anyone have ESP enabled and working with this ECU?

Cuz mine is pretty dead, althru ABS works just fine...

 

[David Stuart]

Hello,

This fells a bit in off topic regarding the current thread focus on lambdas, closed loops and so on.
But since its about GEN90, here it goes.
Does anyone have ESP enabled and working with this ECU?

Cuz mine is pretty dead, althru ABS works just fine...

I don’t but it came that way.


Sent from my iPhone using Tapatalk

 

[SuperCinq]

Hello,

I have recently purchased the Gen90 ECU new from K-Tec and I am having an issue. I will contact them on Monday but ideally if I can get the problem fixed this weekend as I am racing next weekend and time is against me.

There don't seem to be any instructions, just the help files?

I have installed the software supplied on the USB stick and believe I first need to complete the ECU connection setup, when I do this no ports are recognised and so unable to connect to the ECU

What am I doing wrong? Any one come across this issue previously?

 

[bikemanpt]

Hello,

I have recently purchased the Gen90 ECU new from K-Tec and I am having an issue. I will contact them on Monday but ideally if I can get the problem fixed this weekend as I am racing next weekend and time is against me.

There don't seem to be any instructions, just the help files?

I have installed the software supplied on the USB stick and believe I first need to complete the ECU connection setup, when I do this no ports are recognised and so unable to connect to the ECU

What am I doing wrong? Any one come across this issue previously?

Assuming you have the ftdi driver installedcheck if when you connect the cable to the usb port, a new COM port appears in the device manager, usually COM3, if so, then on ktec software select that port. Ignition must be live also.

 

[bikemanpt]

I don’t but it came that way.


Sent from my iPhone using Tapatalk

Well, let me reformulate the question, is it supposed to work? Or is it a add-on like cruise control that although it was advertised on their web page, never appeared..?

 

[SuperCinq]

Assuming you have the ftdi driver installedcheck if when you connect the cable to the usb port, a new COM port appears in the device manager, usually COM3, if so, then on ktec software select that port. Ignition must be live also.

Thanks for getting back to me, I have installed the FTDI driver supplied on the USB but when I attach the USB cable the device manager doesn't recognise or acknowledge it. I have the USB cable connected to the ECU and the ignition is switched on.

 

[Robbie Corbett]

If it's not showing as a serial port in the device manager then the issue is at the computer end not ecu end. Could also be a cable fault.

What version of Windows are you using?

 

[SuperCinq]

It's Windows 10