Ferrari Power Unit Hardware & Software

[NL_Fer]

Also both straights in Texas had a hairpin at the start which wasn’t easy for the Ferrari chassis. It hurts the straight performance too.

The climb in the final part of Interlagos must suit the Ferrari better, we will see some performance again, i hope.

Horner and Marco must be laughing their asses off this week.

[RZS10]

That's a quote from AMuS:

Red Bulls Analyse beinhaltete fünf Charts und eine präzise Betriebsanleitung, wie man dem Durchflussmengen-Sensor vorgaukeln kann, es würde die erlaubte Benzinmenge die Messstelle passieren, während in Wirklichkeit mehr eingespritzt wird.

RBR provided several charts and a detailed description/instruction on how to trick the sensor.

That would imply that they managed to trick the sensor themselves, at the very least on simulation level.

On a side note, from AMuS:

[...]Ferrari drove extremely slow laps on the grid. [...]

Which is odd, since Ferrari did not have control of the speed of the formation lap [...]

Also, the advantage of the Ferraris is most pronounced in qualifying, and they never seemed particularly slow in doing their out-laps. [...]

Article is referring to free practice sessions, with most of them being in FP3..which to be honest, that adds more questions.. because FP3 has the race/quali engine fitted and then guess what else happens a few hours post FP3...
That also comfortably answers why it is only Quali they have this significant advantage.

Where did you get FP3 from?

The AMuS article mentions "extrem langsame Runden in die Startaufstellung" which literally translates to "extremely slow laps onto the starting grid" - keep in mind that no one calls the formation lap anything other than "Formationsrunde" so they're rather talking about the reconnaissance laps (of which they are allowed to make several).

This becomes clear when you further read and come across "in Austin war nur eine von insgesamt vier Runden der beiden Fahrer auffällig langsam." which means "In Austin only one of a total of four laps of both drivers was noticeably slow"

No one does four or two (if they only did 2 each for a total of 4) formation laps.

[saviour stivala]

Who from all those on the grid could show how to trick the fuel flow sensor better than RBR?.

[hansdegit]

-You'd need some sort of device that is capable of precise interference inbetween the 2000Hertz measure points. Timing has to be impecable, the interference has to be impecable.
-This has to be done under racing circumstances. When the car hobling, vibrating and cornering around.
-And it has to be done in a fashion where at the very least its main purpose is not to interfere with the fuel flow sensor. It has to be "accidental". Can you sell such a precise controlled signal and frankly impressive technology behind it as accidental?

New user here...

I'm wondering...Why not just disturb the sending of the samples? I read on forum.verstappen.nl that the device uses the CAN-bus protocol,which is easily jammed.
Being an IT-professional myself, I can imagine that there is no FIA logic in the car that actually checks that there are 2000 samples being received every second of the race. So if you jam the samples that reveil more than 100kg/hr flow rate, then the average still looks ok.

Does this sound silly?

[LM10]

If something out of the Technical Directive did put a curve on Ferrari's performance, big if, it has to be in the other 2 questions RBR asked.

Which other 2 questions were asked?

[turbof1]

If something out of the Technical Directive did put a curve on Ferrari's performance, big if, it has to be in the other 2 questions RBR asked.

Which other 2 questions were asked?

From what I gathered, RBR asked 3 questions in total. Only 1 was made public, the one about sending interference into the sensor.

[turbof1]

-You'd need some sort of device that is capable of precise interference inbetween the 2000Hertz measure points. Timing has to be impecable, the interference has to be impecable.
-This has to be done under racing circumstances. When the car hobling, vibrating and cornering around.
-And it has to be done in a fashion where at the very least its main purpose is not to interfere with the fuel flow sensor. It has to be "accidental". Can you sell such a precise controlled signal and frankly impressive technology behind it as accidental?

New user here...

I'm wondering...Why not just disturb the sending of the samples? I read on forum.verstappen.nl that the device uses the CAN-bus protocol,which is easily jammed.
Being an IT-professional myself, I can imagine that there is no FIA logic in the car that actually checks that there are 2000 samples being received every second of the race. So if you jam the samples that reveil more than 100kg/hr flow rate, then the average still looks ok.

Does this sound silly?

Hello and welcome to the boards!

I would not say silly for the theory. However, again we are not just speaking about jamming, but altering the readings. Jammed signals would have been picked up, because how many of those 2000 signals would you have to jam to get any meaningful performance out of that, that makes the risk of getting caught somewhat worthwhile? All the telemetry will have to pass through the FIA's standardized ECU and send to the pitwall and FIA's databanks. That would mean you'd have to put in fake signals throughout the standardized and probably sealed off ECU as well.

[turbof1]

Who from all those on the grid could show how to trick the fuel flow sensor better than RBR?.

That is unnecessary and not the point. Again, this is not a thread about morality or past naughty moments. This is a purely technical thread where we run hypotheses.

[richardn]

I think the FIA missed a trick. If Ferrari were using constant frequency - variable volume fuel pumps phase locked to sensor reading pulses, they could have downloaded a software update to vary the frequency of measurement during the practice session and got the distinctive Ferrari engine whine to play a tune...

[hansdegit]

Hello and welcome to the boards!

thank you. I have lurked for almost a year now.

I would not say silly for the theory. However, again we are not just speaking about jamming, but altering the readings. Jammed signals would have been picked up, because how many of those 2000 signals would you have to jam to get any meaningful performance out of that, that makes the risk of getting caught somewhat worthwhile? All the telemetry will have to pass through the FIA's standardized ECU and send to the pitwall and FIA's databanks. That would mean you'd have to put in fake signals throughout the standardized and probably sealed off ECU as well.

[media]https://www.gillsc.com/products/flow-se ... w-meter-2/[/media]

According to this (outdated) spec sheet, the communication is encrypted which makes it practically impossible to insert fake readings. But I can imagine that jammed signals cannot be interpreted by the ECU and hence are never transmitted to the pitwall, but immediately discarded instead.
As long as the number of dropped samples stays within reasonable limits, I see no reason for alarm.

It sounds very simple, and according to Binotto, you'd have to be very stupid for not doing this

[Gilles27Kimi7]

Meanwhile, while by no any means conclusive, we have some data:


https://f1ingenerale.com/f1-power-unit- ... -qualcosa/

[MtthsMlw]

Meanwhile, while by no any means conclusive, we have some data:
https://f1ingenerale.com/?attachment_id=97413
https://f1ingenerale.com/?attachment_id=97412
https://f1ingenerale.com/f1-power-unit- ... -qualcosa/

Very cool graphics. Thanks for sharing.
Haven't seen the delta that nicely displayed before.
Looks pretty conclusive to me, let's wait for Brasil nonetheless

[turbof1]

Hello and welcome to the boards!

thank you. I have lurked for almost a year now.

I would not say silly for the theory. However, again we are not just speaking about jamming, but altering the readings. Jammed signals would have been picked up, because how many of those 2000 signals would you have to jam to get any meaningful performance out of that, that makes the risk of getting caught somewhat worthwhile? All the telemetry will have to pass through the FIA's standardized ECU and send to the pitwall and FIA's databanks. That would mean you'd have to put in fake signals throughout the standardized and probably sealed off ECU as well.

https://www.gillsc.com/products/flow-se ... w-meter-2/

According to this (outdated) spec sheet, the communication is encrypted which makes it practically impossible to insert fake readings. But I can imagine that jammed signals cannot be interpreted by the ECU and hence are never transmitted to the pitwall and immediately discarded instead.
As long as the number of dropped samples stays within reasonable limits, I see no reason for alarm.

It sounds very simple, and according to Binotto, you'ld have to be very stupid for not doing this

Gills does not produce the sensors anymore. The in 2015 formed company SentronicsTM Limited does.

I don't think (but can be wrong) the data is immediately being discarded. Back in 2014 they were monitoring Ricciardo's fuel flow with the stewards explicitly contacting Red Bull to reduce fuel flow a couple of times throughout the race, and indeed he got later disqualified based on said data in front of the International Tribunal. I think a copy is being kept. But it is guessing for anybody on that front I think. We can neither exclude or confirm anything.

Lol @Binotto xD.

[Tzk]

So if you jam the samples that reveil more than 100kg/hr flow rate, then the average still looks ok.

Does this sound silly?

i'd rather assume that the FIA logging system logs actual readings (measured flow + timestamp) instead of an average value. So if you mess with the signal it'll be noted in the datalogs.

Also the technical regulations state that such a system is forbidden (article 5.10.5):

Any device, system or procedure the purpose and/or effect of which is to increase the flow
rate or to store and recycle fuel after the measurement point is prohibited.

So yes, it does sound silly

[Sieper]

So if you jam the samples that reveil more than 100kg/hr flow rate, then the average still looks ok.

Does this sound silly?

i'd rather assume that the FIA logging system logs actual readings (measured flow + timestamp) instead of an average value. So if you mess with the signal it'll be noted in the datalogs.

Also the technical regulations state that such a system is forbidden (article 5.10.5):

Any device, system or procedure the purpose and/or effect of which is to increase the flow
rate or to store and recycle fuel after the measurement point is prohibited.

So yes, it does sound silly

No it doesn’t jamming (or modifying) a signal about the flow measurement is totally not included in above quoted rule. That is just about increasing the flow rate. They are not doing that, they use the normal system, they just obscure measurements about that. And likely with deniability. Electrical interference, just bad luck for us. If, in fact that is true of course,

Now it is no longer allowed, strictly outruled, to have the signal accidently be blocked.