McLaren Life banner
1 - 13 of 13 Posts

·
World’s fastest and quickest McLarens
Joined
·
260 Posts
Discussion Starter · #1 · (Edited)
Hello everyone! We'd like to start a short and sweet weekly Tech Tuesday article where we can highlight custom features we integrate into the OEM McLaren ECU, Calibration related questions, or just general technical discussions based on what piques your interest. Since we specialize in software and engine calibration that will generally be the main focus of these discussions. However, over the past two years of setting world records, iterating calibrations based on experiences, and listening to the wants of our customers we've learned a great deal about the mechanical side as well.

For our first Tech Tuesday we will be discussing:

Knock Safety by M-Engineering
  • Why it's useful?
    • Plain and simple it can save a motor from catastrophic failure from cylinder pressures being too high. Increasing engine torque and power output is generally done by increasing cylinder pressures. Knock can create drastic spikes in cylinder pressure in very short amounts of time. High levels of knock under high load / high cylinder pressures are typically what cause engine damage. That engine damage can come in all shapes and sizes. By monitoring knock levels we can close the throttle if dangerous levels of knock are present, potentially saving the motor from catastrophic damage and alerting the driver to a potential issue. For the sake of brevity we will not discuss what knock is in this short article. If there is interest in that topic let us know and we can discuss it in a future Tech Tuesday!
  • What does it do?
    • The Knock Engine Safety is custom logic we've written and integrated into the OEM ECU to monitor knock corrections every hundredth of a second (100Hz) and drastically lowers load / boost / throttle if dangerous levels of knock are detected.
  • How it works?
    • There are two separate ways a knock safety can be triggered.
      • large singular knock event in any one cylinder
        or
      • multiple smaller knock events across multiple cylinders
    • Since cylinder pressures will change based on boost, ignition advance, charge temp, etc... Each map slot has a calibratable knock threshold. Map slots with a higher base cylinder pressure under WOT (e.g. higher octane, more boost, etc...) will have a decreased threshold and are more likely to trigger a safety if knock is present.
  • What happens when a knock safety is triggered?
    • The load is drastically reduced and the throttle is closed
      • We chose a load level where you can still safely navigate traffic if the safety occurs.
      • After 10 seconds of low throttle pedal input the safety will reset and allow for full throttle again
  • Why do we implement it in all our McLaren calibrations?
    • Every situation is different. What we mean by that is McLarens come in all conditions. We advise that you're up to date on all routine maintenance, keep close tabs on your oil level, and decrease your oil service intervals, but we also try to mitigate as much risk as possible when you choose to tune your car. We do this with heavily vetted calibrations as a starting point, however mechanical issues can and do arise. When an issue is encountered on one of our calibrations we can generally attribute to a situational cause rather than the calibration itself. Our most common situations for a knock sensor being triggered are from bad gas or improper map slot for the situation. Improper map slot refers to an end user running a map slot that is more aggressive and designed for higher octane, with a lower octane fuel in the tank.

To better illustrate Knock Safeties here is a break down of the calibratable knock safety tables in M-Composer that were used in a Stage3 McLaren 570S on E85 that recently made 850whp.
Rectangle Font Screenshot Technology Software

The example will instantly trigger a knock safety if any single cylinder hits -10* or if 3 or more cylinders hit -2* or more of correction. 10 seconds is the cooldown time needed after an engine safety is tripped to allow full throttle operation again. And load will need to be above 100 for the knock safeties to be active. For reference, a stock McLaren 720S with a stock calibration will pull -6 to -8 degrees of timing routinely under heavy load on the super crummy 91 octane fuel found in California.

Knock Safety is just one of many custom engine safeties that M-Engineering has written and integrated into the OEM McLaren ECU all in an effort to give our end users the fastest, safest, and most reliable calibration on the market!

Thanks everyone for reading, we hope you find these informative! Engagement and questions will motivate us to keep this series going. Since this is the first Tech Tuesday any input is welcomed!

More info can always be found at www.M-Engineering.us or by emailing us at [email protected]

- M-Engineering Crew
 

·
Super Moderator
“I’m a Tuner”
Joined
·
1,270 Posts
@M-Engineering Without saying so, has interestingly touched on the topic of a cheap "tune in a box" from 'no name' software venders which has come full circle in my shop recently.

In my shop we have a very low mile 600LT that needs an engine because of a really cheap tune sold to him as the "house brand" tune at a local dyno tuner. (Remember you always get what you pay for! In this situation the shop sold him their highest profit margin option, which can easily be found being sold on eBay or thru other reputable vendors for just a few thousand dollars).
Basically the customer paid all the money asked of him which was over 3k (including "dyno time"), and instead of protecting the customers best interest, the retailer maximized their profits)

Unfairly, this customer trusted the local dyno tuner as an expert, and had no clue the "house brand" tune could be so poorly executed... as in NOW a year later its time to replace his entire engine! Wait... What?!??

In just 1500 miles the cheap tune has beat his rods and bearings to death.
  • Cracks in the heads or head gaskets, fails a compression leak down test
  • Low compression due to bent rods, all cylinders
  • Metal shavings found in the oil filter, Crank Shaft or Rod Bearing Material


You might ask... "How can this happen so easily?"

Here is why: Amateur tuners have a very hard time completely mastering all the controls of a McLaren ECU. The time investment can be 100's of hours per platform model.

McLaren Has how many variations of vehicle platforms? anybody remember McLaren's "Track22" business plan to have 22 variations being sold by 2022?

Math time!
100+ hours x 22 different vehicle platforms = ???? hours invested to create a top quality Tune? :unsure:

I'm not a mathmagician, But something tells me the quality result retails for a bit more than 1,000.00 to get this right! ;)



In reality, Bosch is a very clever company and they made this ECU extremely difficult to fully take control of!

So amateur tuners inevitably take a short cut! The short cut is so simple that these wanna be tuners cannot resist it!

All they have to "do" is de-sensitize the knock sensors when they push the timing to get the HP numbers they want on a dyno.

Great idea! In a static environment.... But, they forget to account for the fact that your McLaren actually lives in a dynamic environment... And the McLaren ECU knows this... that is why its so critical the ECU is allowed to remove timing when the situation warrants it!

You know... Different temps outside... altitude changes, load levels in different gears.... excellent fuel today... a crappy tank of fuel tomorrow? oops!

If your knock sensors are "numbed" to the reality's of the real world... then all that extra engine timing will do exactly what they programmed it to do... even when it shouldn't.
Unfortunately your connecting rods and bearings will be the victims of the shortcuts.


M-engineering might be the only tuner that fully understands the complexities of the real world for your McLaren Engine!
The fact that they focus on adding so many safeties really impresses me to the point that I am a die hard fan of their efforts! (disclaimer - I am also an authorized reseller of their product... so I am definitely biased!)

To be fair... There may be other companies that have invested the time, research, and long term dedication into a tuning solution for McLaren's that is on par with M-Engineering.

And if they have, then I'm not talking about those guys. Unfortunately, none of the tuners I've ever worked with will "show you" their work product, so it's impossible to know if their work is sloppy or professional until it's too late...
For me personally, I'm not a gambler. Especially when the stakes are so high.


@M-Engineering, Thanks for starting this weekly tech tip format!

I'm personally excited to continue learning and being a part of your amazing work mastering the McLaren engine ECU.


Kevin P
Supercar Garage - The McLaren Specialist
770-284-0172

Kennesaw GA
* Home of the flying McMedics... The only mobile McLaren Specialist! -We come to you!
 

·
Owner
Joined
·
978 Posts
I've been away from the tuning scene for many years and found this post terribly in interesting.

"Tuning" as I have seen it has always been modifying the stock calibration tables of the OEM ECU to reduce safety margins that the manufacture put in place primarily to either (1) make the warranty/longevity goals or (2) comply with emissions requirements. Not all engines are the same and the manufacturer needs the cars to pass emissions testing, live with a bunch of different (and sometimes crap) fuels all while lasting a reasonable amount of time.

Tuners go in and remove knock margin (the amount timing is reduced to prevent onset of most pre-ignition) and lean the fuel curve (fuel is used to cool turbo and supercharged - and to a lessor degree, NA - motors) the fuel map. They also look for ways to increase boost pressures and remove safeties like differential pressure across intercoolers or torque limits that protect transmissions so as to produce more power.

This is the first time I have heard of anyone saying they are writing Tri-Core, PowerPC or other embedded processor code to replace the factory knock code. I did have a bud that wrote his own code for one of the Ford or Chrysler ECUs as they were cheap (compared to Motec and the like) while being WAY more powerful and capable but that was a complete code base rewrite. Given the running checksums and other protections in place to prevent OE code (as opposed to calibration table) modification, this is a huge undertaking and would be phenomenal if accomplished. Just debugging code to get it to run properly without a simulator would be amazing. Having it work better than the ECU manufacture's code would be a triumph.......... so far beyond my pay grade :)

The OE knock code is pretty darn good. Any reason to replace that which is already working?
 

·
Owner
Joined
·
978 Posts
WRT to all the time, energy and effort it takes to get a working set of map modifications, the problem I found with the tuning industry is it is all based on theft. The manufacturer generates a set of calibration templates every time a version of ECU firmware is compiled. These templates are what the engineers/dyno technicians use as they develop the calibrations for a particular engine and, as you might guess, are proprietary. Anyone wanting to modify calibrations has one of two choices. They can either hand build the calibration tables (there are hundreds to thousands in modern ECU code) by trial and error or "get" the template. Imagine finding 20 timing maps, modifying one value or range in one map then running the motor and trying to find when that map is being used by monitoring the change from base value for timing. Alternatively, you could de-compile the base code (into assembly then further into C) and try to reverse engineer what the code is doing so you can then identify which maps are what. Most go the template route and those templates have been stolen from the manufacturer. This original sin regarding the templates is the basis for my theft comment and not that any one tuner is a thief.

On top of the origin for the map template information, it is very difficult for any tuner to hide their work. So, if you spend hundreds of hours learning and building a good safe reliable tune the next guy with WinOLS and a BDM tool reads it off the ECU, deciphers your modifications and, viola, has the ability to use those changes in "their" tune. It's a screwy business.
 

·
World’s fastest and quickest McLarens
Joined
·
260 Posts
Discussion Starter · #7 · (Edited)
I've been away from the tuning scene for many years and found this post terribly in interesting.

"Tuning" as I have seen it has always been modifying the stock calibration tables of the OEM ECU to reduce safety margins that the manufacture put in place primarily to either (1) make the warranty/longevity goals or (2) comply with emissions requirements. Not all engines are the same and the manufacturer needs the cars to pass emissions testing, live with a bunch of different (and sometimes crap) fuels all while lasting a reasonable amount of time.

Tuners go in and remove knock margin (the amount timing is reduced to prevent onset of most pre-ignition) and lean the fuel curve (fuel is used to cool turbo and supercharged - and to a lessor degree, NA - motors) the fuel map. They also look for ways to increase boost pressures and remove safeties like differential pressure across intercoolers or torque limits that protect transmissions so as to produce more power.

This is the first time I have heard of anyone saying they are writing Tri-Core, PowerPC or other embedded processor code to replace the factory knock code. I did have a bud that wrote his own code for one of the Ford or Chrysler ECUs as they were cheap (compared to Motec and the like) while being WAY more powerful and capable but that was a complete code base rewrite. Given the running checksums and other protections in place to prevent OE code (as opposed to calibration table) modification, this is a huge undertaking and would be phenomenal if accomplished. Just debugging code to get it to run properly without a simulator would be amazing. Having it work better than the ECU manufacture's code would be a triumph.......... so far beyond my pay grade :)

The OE knock code is pretty darn good. Any reason to replace that which is already working?
Very good questions! We do not replace factory knock control, simply improve on the ECUs ability to disable heavy load when heavy knock is present. We do not alter the factory knock recognition or timing it pulls when knock occurs. We actually rely on that for our system to operate properly. The timing corrections dictated by the factory knock control strategy is what triggers the knock safety. We have simply integrated code to work in concert with factory logic.

With our ability to integrate our custom hyperlogging (logging over 200 parameters at over 100Hz, which we can dictate based on RAM Address) allows us to precisely monitor systems far beyond what others are capable of on the stock ECU. We can log anything that comes into, goes out of, or is calculated in the ECU. Much more than the MDS can do and at a much higher rate. We are not just limited to UDS or Standard PIDs like the other options available.

We write a tremendous amount of custom code for Tricore and have experience with custom code for SH and ST10 processors as well. Engine safeties, and knock safety in particular, are just the tip of the iceberg. In future Tech Tuesdays we will continue to address other custom code.
 

·
World’s fastest and quickest McLarens
Joined
·
260 Posts
Discussion Starter · #8 ·
WRT to all the time, energy and effort it takes to get a working set of map modifications, the problem I found with the tuning industry is it is all based on theft. The manufacturer generates a set of calibration templates every time a version of ECU firmware is compiled. These templates are what the engineers/dyno technicians use as they develop the calibrations for a particular engine and, as you might guess, are proprietary. Anyone wanting to modify calibrations has one of two choices. They can either hand build the calibration tables (there are hundreds to thousands in modern ECU code) by trial and error or "get" the template. Imagine finding 20 timing maps, modifying one value or range in one map then running the motor and trying to find when that map is being used by monitoring the change from base value for timing. Alternatively, you could de-compile the base code (into assembly then further into C) and try to reverse engineer what the code is doing so you can then identify which maps are what. Most go the template route and those templates have been stolen from the manufacturer. This original sin regarding the templates is the basis for my theft comment and not that any one tuner is a thief.

On top of the origin for the map template information, it is very difficult for any tuner to hide their work. So, if you spend hundreds of hours learning and building a good safe reliable tune the next guy with WinOLS and a BDM tool reads it off the ECU, deciphers your modifications and, viola, has the ability to use those changes in "their" tune. It's a screwy business.
Theft is a huge reason we take so many precautions to obfuscate our custom code and calibrations. We have a general protocol we follow when it comes to custom code. That simply is to write code that if the person intent to steal it is talented enough to reverse engineer it they are talented enough to write it from scratch. There is a very good reason no one has stolen our map switching or other code since we released it years ago :) Some have tried to emulate our code into McLaren but come up very short in their end product.

We come from Japanese calibration originally where reverse engineering was the only way to sort out useful calibratable tables. Those type of documents simply were not available for many of the vehicles being produced in Japan in the 2000/2010s. A little background, we all came from one of the largest tuning companies in the country where we pioneered their Porsche calibration program before venturing out on our own.

We use programs that assist in reverse engineering, it is not possible to write the custom code that we integrate into our calibrations with simple map packs or templates.

Very good insight into the industry however, you seem to have a pretty good grasp on how unscrupulous many of the calibrators are in this industry. Something that seems to be far more rampant in exotics than say Subaru or GTR. This behavior is something we are actively trying to combat by giving people the tools(M-Tuner) to datalog and verify their calibrations on their own or with a 3rd party.
 
  • Like
Reactions: ima2nr (McMedics)

·
World’s fastest and quickest McLarens
Joined
·
260 Posts
Discussion Starter · #9 ·
Yes, love an M-engineering post, right up my street @M-Engineering keep them coming!

In all fairness Kev it takes more than just de-sensitising the knock thresholds, there is more to it than that, but, if he had a proper crappy tune, with just this changed, thats really bad and I feel sorry for him
Thanks, we look forward to interacting with you guys more in future Tech Tuesday posts!

True in some cases, but enough small knock events at high loads / high cylinder pressures that go unaccounted for due to desensitized knock sensitivity can wreak havoc overtime, typically on bearings or rods.
 
  • Like
Reactions: batcheej

·
World’s fastest and quickest McLarens
Joined
·
260 Posts
Discussion Starter · #10 ·
@M-Engineering Without saying so, has interestingly touched on the topic of a cheap "tune in a box" from 'no name' software venders which has come full circle in my shop recently.

In my shop we have a very low mile 600LT that needs an engine because of a really cheap tune sold to him as the "house brand" tune at a local dyno tuner. (Remember you always get what you pay for! In this situation the shop sold him their highest profit margin option, which can easily be found being sold on eBay or thru other reputable vendors for just a few thousand dollars).
Basically the customer paid all the money asked of him which was over 3k (including "dyno time"), and instead of protecting the customers best interest, the retailer maximized their profits)

Unfairly, this customer trusted the local dyno tuner as an expert, and had no clue the "house brand" tune could be so poorly executed... as in NOW a year later its time to replace his entire engine! Wait... What?!??

In just 1500 miles the cheap tune has beat his rods and bearings to death.
  • Cracks in the heads or head gaskets, fails a compression leak down test
  • Low compression due to bent rods, all cylinders
  • Metal shavings found in the oil filter, Crank Shaft or Rod Bearing Material


You might ask... "How can this happen so easily?"

Here is why: Amateur tuners have a very hard time completely mastering all the controls of a McLaren ECU. The time investment can be 100's of hours per platform model.

McLaren Has how many variations of vehicle platforms? anybody remember McLaren's "Track22" business plan to have 22 variations being sold by 2022?

Math time!
100+ hours x 22 different vehicle platforms = ???? hours invested to create a top quality Tune? :unsure:

I'm not a mathmagician, But something tells me the quality result retails for a bit more than 1,000.00 to get this right! ;)



In reality, Bosch is a very clever company and they made this ECU extremely difficult to fully take control of!

So amateur tuners inevitably take a short cut! The short cut is so simple that these wanna be tuners cannot resist it!

All they have to "do" is de-sensitize the knock sensors when they push the timing to get the HP numbers they want on a dyno.

Great idea! In a static environment.... But, they forget to account for the fact that your McLaren actually lives in a dynamic environment... And the McLaren ECU knows this... that is why its so critical the ECU is allowed to remove timing when the situation warrants it!

You know... Different temps outside... altitude changes, load levels in different gears.... excellent fuel today... a crappy tank of fuel tomorrow? oops!

If your knock sensors are "numbed" to the reality's of the real world... then all that extra engine timing will do exactly what they programmed it to do... even when it shouldn't.
Unfortunately your connecting rods and bearings will be the victims of the shortcuts.


M-engineering might be the only tuner that fully understands the complexities of the real world for your McLaren Engine!
The fact that they focus on adding so many safeties really impresses me to the point that I am a die hard fan of their efforts! (disclaimer - I am also an authorized reseller of their product... so I am definitely biased!)

To be fair... There may be other companies that have invested the time, research, and long term dedication into a tuning solution for McLaren's that is on par with M-Engineering.

And if they have, then I'm not talking about those guys. Unfortunately, none of the tuners I've ever worked with will "show you" their work product, so it's impossible to know if their work is sloppy or professional until it's too late...
For me personally, I'm not a gambler. Especially when the stakes are so high.


@M-Engineering, Thanks for starting this weekly tech tip format!

I'm personally excited to continue learning and being a part of your amazing work mastering the McLaren engine ECU.


Kevin P
Supercar Garage - The McLaren Specialist
770-284-0172

Kennesaw GA
* Home of the flying McMedics... The only mobile McLaren Specialist! -We come to you!
Thanks Kevin for the kind words! We very much enjoy working with you because of your critical thinking, dedication to quality work, and your ability to diagnose issues far and beyond a typical manufacturer's diagnostic flow chart.

You touched on so many good points in this post. We're always bummed to see a McLaren engine on the fritz but luckily we were able to tag team this engine and diagnose the issue with empirical data before it ended up with a catastrophic failure in need of a new short block. Much much cheaper to rebuild a motor without a giant hole in the side of it.

Nice work and look forward to the end result when we get this bad boy back to it's full glory.
 
  • Like
Reactions: ima2nr (McMedics)

·
World’s fastest and quickest McLarens
Joined
·
260 Posts
Discussion Starter · #12 ·
Typically
@M-Engineering

I have recently come across post lambda adaptation issues due to an exhaust leak, could you shed some more light on what that would do to the AFR?

I would assume overly rich.

Typically exhaust leaks will form a Venturi, think like a perfume bottle. As the exhaust gases passes over the leak it will pull in fresh air. This will skew the oxygen sensor readings lean, if the leak is pre o2 sensors, which results in the closed loop system adding more fuel in to compensate to hit the target lambda. However the o2 reading is not actually representative of the fuel and air masses being combusted. Fuel trims work in short term and long term variants. Short term trims are instantaneous and adjust on the fly. If short term trims are consistently adding or removing fuel that will be then be transferred into long term fuel trims. If long term fuel trims become too far out of range (Typically +/-25%) then a DTC will be thrown for system to rich or lean respectively. Short term and long term trims are multiplied against each other typically to give you a final trim value.

Long story short, yes you are correct it would likely result in a rich mixture in the cylinder but it will be reported to the o2 sensor as near lambda target. Fuel Trims will be adding fuel in this case. Check your long term and short term trims and this should give you a good indication if something is up. STFT and LTFT are standard PIDs available with a basic scan tool, albeit at a much slower refresh rate than that of M-Tuner.

Hope this helps!
 

·
Registered
Joined
·
465 Posts
Typically



Typically exhaust leaks will form a Venturi, think like a perfume bottle. As the exhaust gases passes over the leak it will pull in fresh air. This will skew the oxygen sensor readings lean, if the leak is pre o2 sensors, which results in the closed loop system adding more fuel in to compensate to hit the target lambda. However the o2 reading is not actually representative of the fuel and air masses being combusted. Fuel trims work in short term and long term variants. Short term trims are instantaneous and adjust on the fly. If short term trims are consistently adding or removing fuel that will be then be transferred into long term fuel trims. If long term fuel trims become too far out of range (Typically +/-25%) then a DTC will be thrown for system to rich or lean respectively. Short term and long term trims are multiplied against each other typically to give you a final trim value.

Long story short, yes you are correct it would likely result in a rich mixture in the cylinder but it will be reported to the o2 sensor as near lambda target. Fuel Trims will be adding fuel in this case. Check your long term and short term trims and this should give you a good indication if something is up. STFT and LTFT are standard PIDs available with a basic scan tool, albeit at a much slower refresh rate than that of M-Tuner.

Hope this helps!
Thanks for your reply @M-Engineering I have attached a picture below, I actually should have mentioned that it was the post cat o2s being effected. And here is the result...While you are here Do you remember trying to help me with my misfire issues in the following thread... Bad Hesitation under heavy throttle, help

I wasnt actually aware that the post cat o2s do actually trim fuel, I managed to reach 3-4% enrichment which for a post cat o2, is quiet a bit of an adjustment, I was under the impression that only the primary o2s trim the fuel, not the case clearly... See below and was the cause of my mix being too rich to the stage it wasnt able to ignite and would simply misfire badly and wouldnt go any further and also caused havoc on my STFTs because of the misfire unburnt oxygen passing by the primary o2.

Rectangle Slope Parallel Font Electric blue
 
1 - 13 of 13 Posts
Top