355 - Hacking the F1 TCU | Page 3 | FerrariChat

355 Hacking the F1 TCU

Discussion in '348/355' started by Wolfgang72, Jul 12, 2019.

This site may earn a commission from merchant affiliate links, including eBay, Amazon, Skimlinks, and others.

  1. taz355

    taz355 F1 Veteran
    Owner Silver Subscribed

    Feb 18, 2008
    5,965
    Indio Ca/ Alberta
    Full Name:
    Grant
    Not sure if your still reading this Wolfgang but could you control a logic for full clutch reverse such that if a person went rev,neutral,rev neutral,rev that only at this point would you get a full clutch release into reverse. This way it would be selectable
     
  2. Luigi galgano

    Luigi galgano Rookie

    Mar 19, 2020
    1
    Full Name:
    Luigi Galgano
    Very impressed with your work.
    Any updates ?
     
  3. tres55

    tres55 F1 Rookie
    Rossa Subscribed

    Sep 18, 2012
    3,495
    Canada
    The 355 F1 already has a "launch control" doesn't it? It's called race start I believe.

    Unless you mean an improved version.
     
  4. Roth

    Roth Formula Junior

    Apr 1, 2016
    433
    Pepsi Generation
    Wolfgang, a very noble idea and quite interesting. I just stumbled on to this thread. Your posts are lengthy. I only read the first three and I am lazy. My understanding is if I’m not mistaking you attempting a “hack”into the F1 TCU to make the gears shift faster. As I read the detailed process, several things come to mind. Perhaps, you can explain.


    1. My limited understanding tells me the “ECU”, “TCU” are not computers in a sense of what defines a computer in that it has a CPU, ROM, RAM, a connecting mean, an Input, Output and a storage device. What Im getting at, the TCU can not be updated via a software. The instruction or sets of logic is permanently stored on the chip in the TCU board. This means the chip with the set of instruction has to be identified, remove and replace with a chip(s) containing a tailored instruction. Assuming this is achievable, the second step is determine the information on the chip(the shift points through the different gears). Program a new chip(s) and solder it back onto the board.
    2. In my thinking, another possibility is replace the TCU altogether with your design. But this also require all the data and memory stored on the original TCU. I can only imagine the cost as this requires a facility.

    Given your occupation and yourr in depth knowledge of the matter, I have no doubt above is achievable given time an money. But if time and financial resources aren’t a factor, there another factor to address.


    My opinion. The F1 system in the 355 is proven time and time again faster than the faster human. The TCU of the F1 isn’t the problem as far as speed. It is the mechanical gears and the engine power. Every car blogger comments how the double clutch in the 458 is lighting fast and smooth. This is possible because the 458 has monster engine. The F1 only shift as fast as the engine can propel the car forward. If we were to install the 458 F1 into the 355, it wouldn’t shift any faster or smoother. The 355 engine isn’t powerful enough to utilize a “faster” F1 setup. In fact the 355 engine isn’t powerful enough to even take advantage of its first generation F1. That is the bottleneck.


    Nevertheless, I think you are on to something that will benefit the collective knowledge of the 355. I am probably wrong on my assessment. Your comment and assessment appreciated. Food for my brain. :)
     
    tgaspl17 likes this.
  5. taz355

    taz355 F1 Veteran
    Owner Silver Subscribed

    Feb 18, 2008
    5,965
    Indio Ca/ Alberta
    Full Name:
    Grant
    Not sure I agree on the above. I am unsure how engine hp has any effect on shift times.
    I would think logic would tell us that if the f1 shifted 50ms faster on perfectly equal cars then a race to sixth gear would put the faster car 0.3 seconds ahead.
    Now in 6th gear at 250 plus km/hr 0.3 seconds may only be a foot it is still a foot.
     
    PhilB likes this.
  6. taz355

    taz355 F1 Veteran
    Owner Silver Subscribed

    Feb 18, 2008
    5,965
    Indio Ca/ Alberta
    Full Name:
    Grant
    I all know is if you push down the throttle fast in race mode it spins the tires. Have not experimented to get the perfect drag start.
    There was know way I know of how to set the launch rpm and then have the computer slip the clutch at that rpm minimum.
     
  7. jmworkshop

    jmworkshop Rookie

    Jun 13, 2017
    2
    Some interesting comments there Roth.
    However I tend to agree more with Taz.
    Shift times can be improved, which would make for a better, faster car.

    Very interesting project, keep up the hard work.
    Any updates Wolfgang ?
     
  8. Roth

    Roth Formula Junior

    Apr 1, 2016
    433
    Pepsi Generation

    A relay race consists of a team. Each member of the team has a section to run. The first runner has a baton. At the start of the race, he runs as fast as he can to the second teammate. Once he reaches him, he hands over the baton and the second teammate runs as fast as he can to the third teammate and so on to all members until the race is complete.


    A team in a relay race function like a car transmission. The first gear has to achieve a specific threshold before it can shift to the second gear just like the first runner has to cover a measured point before he can pass the baton to the second person. How soon the first runner can pass the baton depends on how fast he can run. Likewise the shift from first gear to the second depends on how fast the engine propel the car with respect to speed and distance. What happens if at the start a driver shifts from 1st to 6th in a second or two in the 355 F1? The car will stall, the gear will likely f up or both. Reason being the car hasn’t achieved the above criteria in that time frame due to the engine lack of power. Speed and distance in time is the factor.


    Let me clarify this, the 355 engine is a masterpiece to date but we need to come to a realization it can only do so much. Trying to get a part of its setup to achieve the capability of its successor (458) three generation later is unfair. Even if the double clutch in the 458 is mated to the 355, it wont achieve the intended desire. The F1 in the 355 is not the bottle neck.
     
  9. taz355

    taz355 F1 Veteran
    Owner Silver Subscribed

    Feb 18, 2008
    5,965
    Indio Ca/ Alberta
    Full Name:
    Grant
    I agree if you are expecting performance of a 458 but can you explain how my explanation is wrong and the slower shifting car would be just as fast.
     
  10. taz355

    taz355 F1 Veteran
    Owner Silver Subscribed

    Feb 18, 2008
    5,965
    Indio Ca/ Alberta
    Full Name:
    Grant
    If you are saying that a 458 transmission ina 355 would not help its accelaration we would need the specifics of the 458 transmission in terms of drag, weight of turning gears and clutches and maybe some other criteria to calculate this. If all of those specs were exactly the same I would say the transmission would then have to improve the 355 performance, but I suffice to say in my experience it would be less fun to drive because of the lack of mechanical feel.
    This of course is a personal taste.
     
  11. Roth

    Roth Formula Junior

    Apr 1, 2016
    433
    Pepsi Generation
    I mentioned the double clutch F1 in the 458 to illustrate the insignificant shift speed discrepancy of that in the 355. Its not important how fast i can hand you the baton in a relay race. Its how fast I can run to you to hand over the baton. Thats the engine ability to bring a car to speed before a gear shift can happen. If gear shift can happen at any speed regardless of engine ability, theres mo need for gears.


    The F1 in the 355 is revolutionary. But I expect most folks have no idea its capability. If landing on the moon is one step for mankind, the 355 F1 is a giant leap for the automobile industry. Faster than the fastest human. Only out done by its double clutch successor due to mechanical limitation. And even though it can do more than 380 horses it is matted to. Short of a supercharger or turbo, nothing going to change the 355 0 to 60 time. But as you all know, speed doesn’t bring joy. Velocity does and the 355 is very good at it. :)
     
  12. taz355

    taz355 F1 Veteran
    Owner Silver Subscribed

    Feb 18, 2008
    5,965
    Indio Ca/ Alberta
    Full Name:
    Grant
    I am not sure I will ever understand exactly what you are saying because from my perspective using your example and making the handoff very poor, let’s say 5 seconds to hand off, then the handoff speed becomes Increasingly more important if you want to win the race against a very equal competitor. The handoff speed will stop the next person from being able to accelerate as soon in my mind.
    If what you are saying is that a 400hp 355 will never beat a 430f1 then yes hp is for sure more important than shift speeds.
    Otherwise I guess I will just have to agree to disagree.
     
  13. tres55

    tres55 F1 Rookie
    Rossa Subscribed

    Sep 18, 2012
    3,495
    Canada
    There is a way to do a "race start" I believe the terminology is. Something to do with a sequence of events including holding down the gear up paddle and the throttle down...then let go of the paddle and it dumps the clutch for you.

    I'm trying to find the tutorial video but can't for the life of me. It was a really old video. I'm pretty confident it was called "race start" mode.

    It's basically a launch control.
     
  14. Skippr1999

    Skippr1999 F1 Rookie
    Silver Subscribed

    Dec 22, 2009
    4,216
    It’s called how to burn up your clutch quickly. I have it on my CS and don’t dare use it.
     
  15. Roth

    Roth Formula Junior

    Apr 1, 2016
    433
    Pepsi Generation
    We are familiar with the saying, an object in motion wants to stay in morion. This is true. Also true, an object at rest wants to stay at rest. To get an object at rest to move require a force greater.


    The weight of an average “car” is about two tons. It takes a force to get it moving from a state of rest. To get it moving faster, require a force even greater and exponentially greater with increase in speed, distance with respect to time. No car engine can ever get a car moving fast enough at the start. Hence, the car transmission.


    The purpose of a car transmission allows ease and rapid movement of a car at the start and higher speed once a it is in motion. A transmission accomplish this utilizing many different size gears. Its function characteristic is dependent on two criteria. 1) Its design and internal mechanical construction. 2) External input that tells it when to take action. Lets look at both.


    Point One, a transmission design, material and construction factor in how fast it shifts between the gears. But there is an overall ceiling how fast they shift. Between the first F1 in the 355 and its successor the dual clutch, the time discrepancy is a fraction of a second. Point Two, an external instruction dictates when a transmission shifts up or down between the gears. This instruction in embedded in the TCU. The TCU relies on distance, speed in time. It will not allow the transmission shifts from first to second unless distance, speed achieved the best situation maximizing its engine power to move quickly. Alternately, a driver can shift sooner in manual mode than the instruction in the TCU allow but shifting sooner will likely slow the car overall time because the engine did not reach maximum power output. Shifting much late has the same affect.


    So here, two factors dictating the transmission function. How fast it shifts between the gears is limited by design and construction. The time discrepancy is fractions of a second. When its allow to shift depends on speed, distance in time. Which depends on engine power. The time period is much greater from a few seconds to minutes. Now keep all this in mind and lets go back to Junior High.


    We are in a relay race. You are the first person with the baton. The horn sounds, you take off toward me. It takes you 30 seconds to reach my position. At that point you hand me the baton. The time it takes the baton to leave your hand into mine takes 1/4 a second. Now lets relay these two numbers to our beloved 355 power plant setup. It takes a 1/4 a second for the F1 to shift from 1st to 2nd gear and so forth. It takes 30 seconds for the engine to move the car to a distant and speed before the first gear can shift to the next. The two numbers are lopsided. The intention of this thread and the concern on everyones mind is make the F1 gear ****s faster. Its like enrolling your son in a school baseball program so he can be better at math. Suppose someone comes up with a contraption that reduces the F1 gear shift time by an 1/8 a second. Its insignificant. Ambient temperature, wind direct or how much you eat washes that gain away. Even if someone mate a dual clutch F1 to a 355, the number will not change maybe .01 second on the 60. The 355 F1 is more than a match for the 380 horses its attached to. I will go on to say this. If a355 F1 is mated to a 458, the 0 to 60 number won’t change much. But if a 458 engine is mated to a 355, the 0 to 60 differences would be huge. The difference is engine power. Should we do another relay race? :)


    I love the 355 for its setup, sight and sound. Wouldn’t alter anything at the cost of changing these things. :)
     
  16. tres55

    tres55 F1 Rookie
    Rossa Subscribed

    Sep 18, 2012
    3,495
    Canada
    Dumping your clutch burns your tires, not so much your clutch. It does put some stress on everything else in your driveline though for that split second the clutch engages.

    Firm shifts don't wear your clutch as much as soft shifts. It's the slow engagement of your clutch that burns it out.

    If your clutch is already slipping due to wear or whatever other reason then yes you will fry your clutch by doing race starts. Thankfully clutch jobs on the 355 aren't very expensive.
     
    Skippr1999 likes this.
  17. Wolfgang72

    Wolfgang72 Rookie

    Jul 12, 2019
    45
    Munich, Germany
    Full Name:
    Wolfgang Schmidt
    #67 Wolfgang72, Mar 30, 2020
    Last edited: Mar 30, 2020
    The TCU is flashable. This is easily possible with a debug adapter. All you have to do is open the case and retrofit a connector. Then your TCU is prepared.

    It should also be possible to flash it even without opening the case. However I have not yet dealt with this because I have found another solution with my client.


    For the standard sprint 0-60 the 355F1 needs 4.5s. Shifting 50 ms faster means to gain 1%. Indeed that is not much. But suppose you are on a racetrack, let's say the Autodromo Nazionale di Monza in Italy. Your lap time is around one and a half minute and you noticed it takes you 50 gear changes per lap. Of course only half of it is relevant, because done under acceleration. But still the difference in your lap time is 25 * 50 ms = 1.25 s! That's nice, right? And remember: Achieving the same effect by adding more hp is not that easy as the relation is exponentially there, as you explained correctly before.

    Now very few will race their F355 on the track. Maybe the Challenge owners, but their cars are manual though. Most of you will drive their F355 out of driving pleasure. And exactly for that it makes a big difference if the gears are changed lightning quick, maybe even quicker than you can take your hand off the shift paddle. At least that's how my client felt.

    There is a very similar situation for the F360. Now those owners can simply install the CS TCU. There is a thread about this in the Forum (see http://www.ferrarichat.com/forum/showthread.php?t=239882). If you read it, you only hear enthusiastic voices who call the change of the TCU as one of their best upgrades.

    By the way I'll go into the F360 CS TCU in one of my next parts. I'm almost a little surprised that nobody has said anything about it yet.


    There is no feature that controls wheel spin. But there is a kind of basic race start, yes. If certain conditions are met, the transmission allows engine speeds of 4800 rpm before engaging the clutch. It is closed not before 6000 rpm and in fact it never closes completely (only up to 93% to maintain a preload).

    This is how you get there: Make sure Low Friction mode is disabled and engage gear 1. Then simply press the accelerator pedal quickly and firmly (more precisely: at least to 33% at a rate of at least 18%pt per 60ms).

    If you release the accelerator pedal again (to less than 28%) or you shift into next gear, the race start ends.
     
    Luigi_g and 601 like this.
  18. Roth

    Roth Formula Junior

    Apr 1, 2016
    433
    Pepsi Generation
    Wolf, we have to account for the small but notable differences between the 355 and 360.

    The most important, according to your link, the 360 comes with two F1 instructions from Ferrari. The 355 do not. To accomplish the intent of this thread, someone has to write that instruction and “flash” it onto a chip. The person has to know the gear and F1 operating characteristics. Making the F1 shift sooner won’t get it done for reasons I stated in earlier posts.

    Another obstacle, the 360 has about 30 more horses to play with and to some degree, its F1 mechanical has been updated. The extra horses pushes the car from point A to point B quicker which means the F1 can shift sooner. But shifting faster doesn't translate to anything if the mechanical makeup of the F1 doesn't react quick enough.

    There are three factors at play here; engine power, F1 shift speed and F1 mechanical makeup. We are focusing on the F1 shift speed neglecting the other two. It won’t do anything for the 0-60 number. And if we want to gat down and dirty, the flywheel weight also factors in. I cant be certain but to some degree all these things got refined in the 360 and 430.
     
  19. Luigi_g

    Luigi_g Rookie

    Mar 3, 2019
    1
    London
    Full Name:
    Luigi galgano
    I have read about the 360 tcu cs modification, all very interesting stuff.
    What is your take on it Wolfgang ?
     
  20. Wolfgang72

    Wolfgang72 Rookie

    Jul 12, 2019
    45
    Munich, Germany
    Full Name:
    Wolfgang Schmidt
    Part 8 of Hacking the F1 TCU

    Apart from a few highlights like the gear display mentioned before, the task soon became harder and harder until there was hardly any progress. Reverse engineering is like a logic game and it reminds me a little of Minesweeper: You have information (in the game that is your starting point) and you use combinatorical logic to proceed as long as you can. Until you reach the point where you get stuck and need a new starting point to continue. Unfortunately I had reached a point where I had no starting points anymore. There was no variable left where I could imagine its content and guess how it might be used for further processing.

    That was the right time to check out how the diagnosis works! It could tell me exactly what I was looking for: I could select one of the well-described parameters, let's say the coolant temperature, and trigger a read out. Then I could trace from which variable the data is gathered. What a rich and reliable source of information!



    Here are some details on the diagnosis. The diagnosis usually takes up a fair amount of the total ECU complexity and this is no different in our TCU. Almost 20% of the program code is related to it. Protocol stack, fault monitoring and preprocessing, DTC generation and storage and so on. It's crazy when you consider that all this is only needed in case of an error. The car would drive without just as well!

    At the same time the diagnosis of our TCU is not even particularly sophisticated. Essentially, it only detects basic electrical errors, but nothing above that level. Suppose you shift from neutral into reverse, but your request cannot be followed because, for example, one of the solenoids is stuck. This type of error is not diagnosed at all!

    I assume that the basic functions of the diagnosis are known to the forum. That is show and erase DTCs, display live data, show and re-learn adaptation and run component test. There are only a few more hidden functions worth mentioning:

    - Read out program code
    - Read out and write EEPROM data
    - Run integrity check (CPU, RAM, ROM, peripheral units)
    - Set end-of-line data

    Unfortunately, there is no way to write the program code. A shame, because it would be feasible: the hardware contains everything that is necessary for it. I guess Ferrari considered the number of units as too small and the period of time our TCU would be relevant as too short to take the risk of incorrect operation. And in fact, as far as I know, there is only one SW version for our TCU. It was build on 11th March 1997. But there has never been an update (at least until today).



    Coming back to my initial problem, I did a short research about Ferraris diagnosis system. I found out that it is called SD and that there are several generations of it. Unfortunately, the tool is expensive, poorly available and with an inappropriate license model. I therefore decided on an aftermarket system called Galileo.

    Image Unavailable, Please Login

    With this tool I was able to find many new starting points! For example, I identified the variable containing the coolant temperature and then figured out how it is used to estimate the clutch temperature and so on.
     
    tgaspl17, 601, brogenville and 2 others like this.
  21. brogenville

    brogenville Formula 3
    Rossa Subscribed

    Apr 24, 2012
    2,016
    UK
    Full Name:
    Robin
    Fascinating as ever, but how does the investigation of the diagnostics systems relate back to the original project brief to improve the shifting speed of the F1 system? Genuinely curious to learn more about your process.
     
  22. Wolfgang72

    Wolfgang72 Rookie

    Jul 12, 2019
    45
    Munich, Germany
    Full Name:
    Wolfgang Schmidt
    #72 Wolfgang72, Apr 14, 2020
    Last edited: Apr 14, 2020
    Well, I got stuck interpreting the software and simply needed more information.

    The gear shift process is the core functionality of the TCU and highly complex. Everything is involved there, throttle position, engine speed, clutch speed, vehicle speed, current gear, target gear, driving mode, clutch wear, clutch temperature... you name it. It was impossible for me to guess how all that works with the knowledge I had from just reversing the hardware. Only reversing the diagnosis gave me the necessary insight into these internal things.

    In the end the developers at Magneti Marelli were clever engineers. There was a reason why they decided that a gear shift shall work the way it does. In a way my challenge is to prove that they missed something and that gear shift still can be improved.
     
  23. brogenville

    brogenville Formula 3
    Rossa Subscribed

    Apr 24, 2012
    2,016
    UK
    Full Name:
    Robin
    This is probably too simplistic a question, but assuming the Marelli engineers did a half way decent job of the TCU firmware, but were just hamstrung by the period computing power, why not just run the same firmware in a modern ecu with a much (much) faster processor?


    Sent from my iPhone using FerrariChat.com mobile app
     
  24. Wolfgang72

    Wolfgang72 Rookie

    Jul 12, 2019
    45
    Munich, Germany
    Full Name:
    Wolfgang Schmidt
    The software is compiled for a specific processor architecture. Our TCU uses a CPU32 derivative which has no direct successor that would be more powerful.

    You could switch to another processor architecture, like TriCore, Aurix or SPC5. Or change the entire Hardware right away, like @Subarubrat suggested before. But that would mean a major redesign of a safety relevant system.

    There is one intermediate option. I planned to deal with that in the next part anyway.
     
    601 likes this.
  25. brogenville

    brogenville Formula 3
    Rossa Subscribed

    Apr 24, 2012
    2,016
    UK
    Full Name:
    Robin
    I'm clearly not coming at this from a position of authority, but if you have some faith in the work that Marelli did, then it seems like there is more to be gained from recompiling the existing firmware into a format that can be used by a modern processor architecture. Then you can throw as much computing power at it as you want? Surely it sets you up better for future upgrades too?

    I'm sure its not the work of a moment, but it sounds like you (and your customer) are this for the long haul.
     

Share This Page