ECT scale in boost control

[Julien]

Hi guys,

I’m looking to buy a G4X Xtreme to properly map my Porsche 993 turbo.

I already bought a Motec M600 but during preliminary setup I realized the limitations of this old model. So I’m considering to ditch it for either a M150 or a more affordable G4X which I didn’t knew and looks very interesting!

(it only missed 2 EGT inputs and 2 WB lambda controllers to be perfect to me) I already have a dual amplifier for the EGTs and I’ll get 2 LinkCanLambda.. so no problem here.

Anyway, to avoid doing the same mistake twice I downloaded PCLink and I’m doing all the preliminary configuration.

My setup will be:

6 Deka Fi114961 at 380kpa, 3 coils, 2 GT28RS, flex fuel, traction control (the main goal!) And I’ll retain most of the factories sensors. Expect the MAF that will go to the bin and be replaced by a Magneti Marrelli 250kpa MAP sensor.

I set all the sensors up

The ‘problem’ I have is the ECT, which on an air cooled Porsche is taken from the surface temp of one cylinder head. The sensor is a standard NTC so no problem here but the range of use is indeed higher than what you would get with coolant. (About 120-130° once all is properly warmed) and could go up to 200°

I managed to adjust the scale for the main compensation tables but on the boost setup the ‘Boost ECT target trim and ECT Wastgate trim tables seems to have fixed x axis (from -30 to 140°) this is a little short for my application, I would need up to 200-220°

I searched through the settings ans parameters but I don’t see anywhere to increase this scale.

Edit: I also realized that the limite is 200° for the range where you can edit it..

Thanks for your help

Julien

 

[Adamw]

For the ECT trim I would suggest turning on the  "GP Boost Tables", this will give you a user configurable trim table that you can set up how you like.  I'll have to confirm with the engineers if the 200°C limit on the ECT parameter can be opened up but I suspect it can (They are on holiday for a couple of weeks).

Having said that, are you sure about the CHT's you are claiming?  I dont have wide porsche experience but have spent a fair bit of time tuning a couple of turbo porsche road cars locally and on both the CHT behave remarkably similar to a water cooled engine.  They would creep up to 130C at idle after a hard workout on a hot day, but generally ran around 100C or less when moving .   The Cyl head temp sensor is just the standard Bosch coefficient ceramic so it gets very inaccurate above about 140°C and is flat-lined by about 170°C, I cant imagine the factory would use that material if the temps were expected to run as high as 200°C.

[Julien]

Thanks for your reply Adam.

The car is actually running on it’s genuine ECU (motronic 5.2) with a custom EPROM (So I don’t have a data logging facility) I checked the CHT a few times through a Piwis and an OBD reader and I have seen 120-130°. But I never checked after a track day run (I’m only monitoring the oil temp honestly).

I found those high values on Rennlist from guys talking about stand alone ECU mapping. I guess I’ll not focus on this now, knowing it may be possible to tweak it later if necessary and wait to see on the dyno what I really get. And if the sensor gets inaccurate above 140° no need to wast time anyway, it will only be used for the warmup and then disregarded even if I would have used it as a safeguard..

if you worked on road cars I’m interested to know if you managed to work around the ABD feature in the ABS module. The TPS seems to be converted to a PWM signal with also a tension variation. I configured a GP output for it a the required 160hz, with the proper %DC but I suppose the tension can not be altered and will remains at 5v. I’ll probably disconnect this feature thanks to the TC but the idea was to make it work because I like when everything works..

[Julien]

I’m working on the wiring and I may have found a typo in your help file:

For the 3 wires solenoid ISCV, the text says to connect the open wire to the first aux output and the second to close but the diagram shows the other way round. Which one is correct?

Also for the sensors grounding and shields.

I guess A7 is for the ground and shields of both triggers, B17 for the knock sensors. And the “Gnd out” (A24 and B22) are for the T° sensors, that’s right?

[Adamw]

Yes you're right, that paragraph about the idle valve doesnt read great, it looks like there are some words missing too...  With G4X the correct order of the Aux outputs is not actually critical, as the control logic can now be reversed in the software using the "Active state" setting.  But if you ignore the text for now and connect the open pin to Aux 2 as per the drawing it will then work correctly with our default "active state" setting.  

As for the grounds - A17, B17, A24 & B22 are all internally connected to "sensor ground".  So they can all be used interchangeably.  In the Link supplied looms we have both the trigger grounds and shields all sliced together and connected to A7, and similarly on the B connector, both knock sensor grounds and their shields are all spliced together and connected to B17.   The "Gnd out's" are then used for all the other sensors - TPS, APS, MAP, pressures, temps, etc.

 

[Julien]

Thanks again, all is crystal clear

 I’ll get the parts, fabricate the loom and probably come back with more questions..!

[Julien]

My XtremX is on its way..

I’m preparing the setup.

I’ll post in the G4X forum once I’ve got a serial number to show

 

I found several values for my injectors Deka 107961 (FI114961)…

Reading the B4+ forum it should have been values given in a previous revision of the help file but this seems to have gone.

Do you still have them somewhere so I can cross check?

 

I’ll use a rotary switch to control the traction control and boost level.

The switch has 4 positions, I wired a few resistors and with a pull-up of 2k I’ll get the following voltages: 4,8 - 4,0 - 3,4 - 2,6

I assigned the mini and maxi in the setup but I didn’t found how/where to assign the intermediate voltages. I could create a cal table but it doesn’t seem to be possible to link it with the switch.

This would be used for the TC slip target.

I also created a virtual aux and used a GP output ‘low boost’ to reduce the boost target (pos 0 will be no slip, low boost for road use, Pos 1 no slip full boost, pos 2 some slip, pos 3 TC Off)

 

My map sensor is linear 0,5V-4,5V ; 0-250Kpa 

I configured a linear table but I wonder if the Link has the same output ? So I could save a table..

 

Concerning the CAN bus, I’ll use 2 Lambda modules and 1 gauge. All are to be set at 1Mbit/s but since the Xtreme has 2 buses, would it be smarter to use one for the sensors and the other for the gauge?

[Adamw]

Copy of the FI114961 data below.  Note is says "microseconds", but is actually milliseconds.  

 

For the TC switch, the GP rotary switch inputs are designed for a switch with even voltage steps between the min and max positions, which yours has so this should work fine.  Example below will output positions 0-3 for your quoted voltages:

 

 

For the MAP sensor we dont currently have a built in calibration that matches that one - although I may ask the firmware team to add that as it is a pretty common one.  But for now you will have to use a custom cal.  There are 13 cal tables that can be used for linear sensors and 10 that can be used for non-linear so you shouldnt run out.  

[Julien]

Thanks Adam!

That’s pretty much how I set the rotary switch, I’m lucky for the voltage, I use leftovers! I simply put 0.05 and 4.95 for the errors to monitor the wiring..

Noted for the MAP. I saw the tables but I always try to optimize..

Any clues for the CAN? I’ll probably wire all on CAN1 but it’s always interesting to know.

[Adamw]

1 hour ago, Julien said:

Any clues for the CAN? I’ll probably wire all on CAN1 but it’s always interesting to know.

Sorry I missed that one.  Since all devices are running at the same speed I would have them all on the 1 bus.  You can have at least 8 CAN lambdas and 9 other devices on one bus.  Save the 2nd bus in case you get some other device in the future that uses a different bit rate - or you might want to use DI9/10 which share the CAN 2 pins for something one day...  

[Julien]

No apology needed! you are a great help and I understand you can miss one of so many questions!

Thanks for your time

 

[Julien]

My ECU has arrived!

I don’t know what I’ve done but trying to upload my setup I lost everything 

So I restarted all the configuration from scratch…

I also requested my distributor to provide the unlock code but I’ve no reply yet..

Anyway, I was thinking to use a pair of K-type thermocouples and I bought a dual amplifier for them but I’m considering to use EGTs from DCI engines. They look stronger and they seems to be PTC type meaning they could be plugged directly to AV inputs.

Has it been done, any recommendations on a model on which the resistance curve would be available?

[Adamw]

What do you mean by lost your configuration?  Did you not have a copy of the .pclrx saved?

I dont have much experience with OEM EGT's, but most I know of are RTD's rather than termistors.  Thermistors are typically limited to about 300°C.  The problem with RTD's is they are typically very slow.  You are talking 5-10 seconds response time Vs say 0.15s for an exposed junction TC or 0.5s for an enclosed tip TC.  

The common RTD I have seen with published calibration data is AEM 30-2052 https://www.aemelectronics.com/articles/new-rtd-type-temp-sensor

https://www.aemelectronics.com/sites/default/files/30-2052_RTD_Temp_Sensor_Instructions.pdf 

[Julien]

I open my file, try to upload it on the ECU and when I quit I’ve been asked to save.

There I made the mistake of saving under the same name since I though it was fine.. I’ll make more tests once I get the unlocking code.

Noted for the EGTs.. I’ll order a couple of TC as planned. These OEM sensors are probably installed to manage the PF regeneration so I guess the response time is not an issue… where it could be during mapping!

Thanks again Adam!

[Adamw]

That shouldnt have messed up your map.  If you load a map from your PC into the ecu then the map in the ecu will match what is on your PC (except if the map was from a different model ecu then it would be updated to match the ecu you are loading it into - ie some ecus have more analog inputs etc).  So if you then hit save, it would save the file that you had just loaded into the ecu back to your PC.  

Note, you should also always do a "store" after loading or making changes to the ecu.  This writes the changes from the temporary ECU memory into the permanent memory.  If you dont store the changes will be lost when you turn off ignition.  

[Julien]

I suppose I messed when I tried to store my file in the ECU.

When I quite I guess the software asked me to save the ECU memory on the PC. And there I made the mistake of saving on my original file, replacing my work by the empty memory of the blank ECU.

I should have given another name to check.. mistake I won’t do twice!

The fuel and ignition maps are empty at the moment since I’ve no base maps for my engine. I only worked on the configuration which I’ve redone by now, so it not catastrophic!

[Julien]

I unlocked the ECU successfully

I found out why I lost my configuration: when you connect the ECU, the loaded configuration replace what you have on screen without letting you the opportunity to save! 
Now all is right and I’ll not do the same mistake twice.

I only powered the ECU from a stabilized power supply. So indeed, I’ve a lot of faults. The CE light status is blinking (not regularly) The parameter is set as ‘On When Fault Codes’ shouldn’t it be steady?

[Adamw]

39 minutes ago, Julien said:

The CE light status is blinking (not regularly) The parameter is set as ‘On When Fault Codes’ shouldn’t it be steady?

Engine running = On when fault codes means it will be constant on when the engine is running, it will flash codes only when the engine isnt running.  

[Julien]

Noted, all good then!

The genuine loom has been removed from the car and the new one is being built..

i have to order a few more parts and connectors.. my next post will be on the G4X forum once we powered up..

thanks again Adam!

[Julien]

On 12/25/2021 at 10:30 PM, Adamw said:

For the ECT trim I would suggest turning on the  "GP Boost Tables", this will give you a user configurable trim table that you can set up how you like.  I'll have to confirm with the engineers if the 200°C limit on the ECT parameter can be opened up but I suspect it can (They are on holiday for a couple of weeks).

Having said that, are you sure about the CHT's you are claiming?  I dont have wide porsche experience but have spent a fair bit of time tuning a couple of turbo porsche road cars locally and on both the CHT behave remarkably similar to a water cooled engine.  They would creep up to 130C at idle after a hard workout on a hot day, but generally ran around 100C or less when moving .   The Cyl head temp sensor is just the standard Bosch coefficient ceramic so it gets very inaccurate above about 140°C and is flat-lined by about 170°C, I cant imagine the factory would use that material if the temps were expected to run as high as 200°C.

Hi Adam,

we did start mapping yesterday, what a pleasure to drive the car again!

We only did a road test to rough-hew the fuel map. I had 150°C on the head (never got higher but I wonder if it’s not a sensor limitation) then we added a little more advance and it dropped to 135-145° depending of the load. The sensor is 3 years old. And outside temp was about 26°.

With the engine air deflectors it’s not easy to confirm the temp with laser thermometer. Ill see if I can take some measurements when on dyno (hopefully Friday). I wonder if I shouldn’t build a custom sensor from a thermocouple to get a better accuracy if I’ll really run at these temps….

The Porsche you did were on standard cams? 

While we are talking temps, what’s your recommendations for EGT sensors placement? I fitted mines few inches before the turbo. My tuner said his more used to have them after…

[Adamw]

15 hours ago, Julien said:

I wonder if I shouldn’t build a custom sensor from a thermocouple to get a better accuracy if I’ll really run at these temps….

You can also get thermistors that work to higher temps but they are just much less common.  We used to make a little one which was good to 200°C at my old work:  https://msel.co.nz/msel-m6-stainless-temperature-sensor/  You could possibly fit something like this with a brass adapter.  I think I have seen some that work to 300C but cant remember where.  A thermocouple is an option but then you have the hassle of an amplifier as well.  

 

15 hours ago, Julien said:

The Porsche you did were on standard cams? 

I suspect so, hard to know since they were late '80s or early '90s era cars but they werent particularly powerful and generally looked to be unmolested.

 

16 hours ago, Julien said:

While we are talking temps, what’s your recommendations for EGT sensors placement? I fitted mines few inches before the turbo. My tuner said his more used to have them after…

I dont use them very often, the few cases I have had a car with them fitted it as been for antilag safety.  My logic would say pre-turbine gives you the best picture of how the engine is working, and how far away from melting things you are.  Post turbine only really tells you how efficient the turbine is working - but you would need to know pre-turbine temp as well.  

 

[Julien]

Thanks Adam, I’ll look for such a thermistor in Europe, that will be the best solution.

Could you check with your software team if it’s possible to allow rescaling some tables that are locked and stop at 140° such as Crank Hold Time(s), Post Start Hold Time(s), Post Start Decay Time(s), Accel Cold Correction (this one is limited at 90°), Overrun Deactivation…

I also wonder if in the software the ECT value is not locked at 150° max because the sensor should have been able to show over that (with a pretty bad accuracy) but never did and 150 is too round to be true.. for water cooled engines I can understand because your head gasket will probably be gone before that value ! 

[Adamw]

On 6/6/2022 at 7:40 PM, Julien said:

Could you check with your software team if it’s possible to allow rescaling some tables that are locked and stop at 140° such as Crank Hold Time(s), Post Start Hold Time(s), Post Start Decay Time(s), Accel Cold Correction (this one is limited at 90°), Overrun Deactivation…

It wouldnt be impossible, but it would be a big job so we would only consider it if there was some evidence it was required.  I havent seen any need on any engine I have tuned yet.  

 

On 6/6/2022 at 7:40 PM, Julien said:

I also wonder if in the software the ECT value is not locked at 150° max because the sensor should have been able to show over that (with a pretty bad accuracy) but never did and 150 is too round to be true..

The ECT parameter is not limited in any way, but likely the calibration table is - certainly most of the predefined calibrations in the drop down list only span to 150°C. Since thermistor calibration is not linear we cant extrapolate beyond the bounds of the calibration table - it will stop at whatever max value is in the cal table.  If you do a custom cal that spans to 200C or whatever the sensor can reliably measure then ECT will run to that value.  

It would also pay to test your factory sensor in boiling water or something just to confirm the cal is correct too.  I have seen some other odd calibrations quoted for the Porsche cylinder head sensors but both that I tested matched the standard bosch cal.  

[Julien]

7 hours ago, Adamw said:

It wouldnt be impossible, but it would be a big job so we would only consider it if there was some evidence it was required.  I havent seen any need on any engine I have tuned yet.  

Ok, it expected it to be a minor modification.. we booked the dyno on Friday. I’ll get back to you after, once the engine will be fully mapped. But the guys from Driftworks also replaced the CHT sensor to extend it’s range for mapping there 964 turbo engine.

7 hours ago, Adamw said:

The ECT parameter is not limited in any way, but likely the calibration table is - certainly most of the predefined calibrations in the drop down list only span to 150°C. Since thermistor calibration is not linear we cant extrapolate beyond the bounds of the calibration table - it will stop at whatever max value is in the cal table.  If you do a custom cal that spans to 200C or whatever the sensor can reliably measure then ECT will run to that value.  

It would also pay to test your factory sensor in boiling water or something just to confirm the cal is correct too.  I have seen some other odd calibrations quoted for the Porsche cylinder head sensors but both that I tested matched the standard bosch cal.  

I did it in my oven last year with a thermocouple and an the result where fine compared to the standard Bosch calibration. (but I stopped at 80° because at that time I only wanted to check if it was the same and my thermocouple was limited in range)

the Bosch cal goes up to 160° Indeed the curve being what she is I agree the accuracy will not be very good at the end.

 

[Julien]

So, we spent 4h on dyno yesterday. Most is done but we had AIT issues to make proper full load adjustments.. I’ll go back with fans for the intercooler. The head temp was stuck at 150°. I drove the car one more hour on the road with light loads we had no less than 135°. It was raining on the way back so impossible to push..

I also did test the sensors while cooking a chiken today^^ I put my oven at 200° and the thermistor was showing 27ohm so I really think there is a limitation at 150° in your software. I’ll make a custom calibration using the standard Bosch NTC values and see if I can display over 150° this way (even if I’m aware the accuracy won’t be good). Since on normal use I’m around 135 at least it will be an indication. 
Anyway that’s not major problem since these temps should not be encountered very often.

What is a problem for me is the limitation on the Startup offset table which stops at 100° and is not editable. I think I can live with the other tables limited a 140° (even if 150 would have been better)

and maybe the Acell cold correction (limited at 90°)

Do you think these could be extended in a future update? (It maybe easier to extend the tables rather than to make them user editable ?)

Talking  about software/firmware update, is there a mailing list to subscribe to in order to be notified of new releases ?