Ferrari Power Unit Hardware & Software

All that has to do with the power train, gearbox, clutch, fuels and lubricants, etc. Generally the mechanical side of Formula One.
User avatar
turbof1
Moderator
Joined: 19 Jul 2012, 21:36
Location: MountDoom CFD Matrix

Re: Ferrari Power Unit Hardware & Software

Post

Phil wrote:
06 Nov 2019, 12:37

Anyway - as for the sensors: It would be great to know to what tolerances these sensors operate in and how secretive the FIA are about things. Is it a sensor that has a spec, is manufactured by a specific company, perhaps even off-the-shelf where any competitor could gain knowledge on the tolerances, know-how on how it's being measured that then could be used to be exploit? Or are these sensors under top-level-secrecy and the teams can only assume?

Given that RedBull seems to suggest such a device could be built and could be in use by Ferrari makes me think the former, perhaps even to a degree that the sensors are perhaps just not that good in order to prevent manipulation on that level. That's just my take though. It's a very lucrative market and you can never be surprised to what lengths a fierce competitor might be willing to go to exploit the maximum out of the grey areas of the rules.

Either way, I wouldn't be surprised if they could be exploiting something like this or not. I just think we can not exclude Ferrari trying things to exploit advantages that may be going against the ethics of the rules.

This has been explained to death back in 2014. If and only if the official fuel sensor fails, there is a backup in place that makes use of the team's sensors.
I wasn't necessarily talking about a fuel sensor failing during the entire period. I was more thinking along the lines of failed reading at the interval point. E.g. there's data there, but the data is wrong. A wrong reading could be logged as a failed reading. Where would the threshold be at which the FIA says, ok, we've had N number of failed readings, something is up? Or would they put it down to what you suggested, e.g. normal activity due to hobling, vibrating car on a bumpy track? I quite frankly have no idea how precise these sensors are, therefore my question regarding the tolerances involved. And when we are talking sensors and electrical appliances, there are always tolerances involved (see the Vettel jump start sensor topic).
To answer your question about what kind of sensor we are talking about: It is made by SenstronicsTM Limited from 2015 onwards is called the Flowsonic Elite Fuel Flow Sensor. It was made first by Gill Sensors in 2014, but FIA switched suppliers. SenstronicTM Limited is comprised of a group of people who have worked since 2009 on the development of the fuelflow sensor. More can be found here: https://www.reventec.com/fia-homologate ... r-f1-lmp1/

The public is of course not privy to the tolerances. The teams will know the tolerances, even though neither the FIA or Senstronic will deny or confirm those tolerances. The sensor unit weights in at around 250 grams. We can deduce from that this is not a particular shielded one, although it will have been developed with keeping out some random electromagnetic interference from any devices around it.

In the first race of 2014 Red Bull did complain of wrong readings. It's not relevant here to rehussle that whole story, but just take from it there is backup system that only at the discretion of the FIA can be used. A sensor failing can mean a lot of things, from outright shutting down to giving weird readings on a regular base.

Also note that when I am talking about purposely sending electromagnetic pulses into a device to jam signals, I am not claiming devices on their own malfunction under racing conditions. I am perfectly aware all the electric devices in a F1 car work splendidly under these conditions. What I am saying (and this is completely my own opinion) is that if you have to design a way to send EM pulses into the fuelflow sensor in a none-conspicuous way, you will have a hard time to do so. Let us take the biggest rumor on that: there f1 sites out there claiming a loose outside wire would be able to send these send these precise pulses into the sensor and manipulate the readings in a precise way. I am not doubting at all you can make the wire do these pulses in exact manner, I am doubting it can reliably manipulate the readings. The pulses will have to travel a bit through material that is not intended to be traveling through (because again even though the sensor will not have lot of shielding, it is not designed to be influenced by it either)

Dans79 suggested they'd do it a much more blunt fashion and tinker directly with the sensor by putting a purposely build EM device on top of it and say the FIA would not necessarily find out. Yeah you can do that, but next to me doubting that the FIA would not notice (they do scrutineering after all), if you are going down that road why even bother. Replace the sensor with a visibly identical one that spits out incorrect readings in the first place.

Note these are no allegations towards Ferrari in any way.
#AeroFrodo

User avatar
turbof1
Moderator
Joined: 19 Jul 2012, 21:36
Location: MountDoom CFD Matrix

Re: Ferrari Power Unit Hardware & Software

Post

dans79 wrote:
06 Nov 2019, 04:05
turbof1 wrote:
06 Nov 2019, 01:20
I already outlined my opinion on that in the post you quoted.
In my opinion what you outlined is very vague at best.
turbof1 wrote:
05 Nov 2019, 13:57
-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.
2kHz isn't exactly a high frequency or sampling rate. The average cell phone records audio at 48kHz, I have a small $400 consumer grade audio recorder in my closets that capable of 192kHz. If your using a desktop computer the vrms in it uses tiny little PWM controllers that operate at up to ~2MHz.
turbof1 wrote:
05 Nov 2019, 13:57
-This has to be done under racing circumstances. When the car hobling, vibrating and cornering around.
I don't even see why this is a point of contention. The car is a rolling computer with hundreds of sensors on it that are all susceptible to the same conditions. We never hear the teams complain that a sensor failed.
turbof1 wrote:
05 Nov 2019, 13:57
-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?
Your assuming that a team would try and hide what they are doing, and that the FIA has the technical know-how to discover what the team is doing.

Oil burning, cold and hot blowing, flexi wings, mass dampers, FRIC, etc pretty much proves the teams can regularly and fairly easily outsmart the FIA. If anything the teams do a far better job of keeping their opponents in-line with requests for rule clarifications, and the threat of protests.

Lets not forget the FIA didn't fit a secondary sensor to the Ferrari last year because it thought something was questionable. It did it, because the other teams started posing questions, stirring up the press, and applying political pressure.

Additionally, you didn't specify if you where referring to interfering with the actual sensor reading, or the transmission of the sensor reading back to the ECU.


and just for some reference to whats possible, here is an article and video about tricking a temperature sensor with an EM wave.
https://news.engin.umich.edu/2019/09/re ... -industry/

The research paper referenced. https://arxiv.org/pdf/1904.07110.pdf

Think about how this could be useful to say trick the FIA plenum temperature sensor.

If you read the paper you can see how the exploit could be used on almost any type of sensor.
Forgive me for not answering each part individually. I'm not going to dissect everything since that will lead up to a very chopped up discussion. Yes, I was vague. And on purpose; this whole drama is shrouded in partial darkness and therefore I refrain from making bold specific remarks or even forming a full opinion on certain matters. Until there is more clarity and also to hear what other people have to say (hey, I am allowed to form an opinion or even change an original opinion based on other their comments!).

As I also explained to Phill, I have no doubt you can design a device capable of sending out the pulses and get that part exactly right. What I am doubting is the interference will go down exactly as planned. Because you don't want readings above 100kg/h. You don't want suspiciously low readings either. The reading have to correspond with where you are on the track. The fuel flow sensor is not developed for outside EM control (again, the other side is that the shielding will not be exactly great either).

This brings me to the point where I think we will fundamentally disagree and will neither agree on in the future, so I am going to discuss this only once: You mentioned a list of "tricks" (by lack of a better word) and feel this is the same as putting a purposely build an EM device on top of the fuel flow sensor. Yes I assume that is not going to happen, and this is also where I disagree with you: the list of tricks you mentioned all sat in a grey area which most were originally allowed by the FIA, but were disallowed afterwards. In my interpretation, what you suggest is black and white against the regulations (and indeed the FIA did confirm that through the clarification by not even telling their own interpretation, but directly referring to the written regulations). Mass dampers, flex wings, FRIC, cold and hot blowing were things in a grey zone, which the FIA did allow for a time, but then changed its mind through either reinterpeting the existing regulations, or enter new regulation all together. I can't possible think how the FIA would look at an EMP device and say "well this was not our intent, but the regulations do leave an opening for that". 5.10.5. explicitly forbids that:
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.
There is random scrutineering involved. Teams cannot use a grey zone for what you think is possible, so it would be blatantly cheating. Cheating is possible, but again if we are going down that way, why bother with this elaborate ruse. Produce a sensor that looks visibly the same and let it spitout wrong readings.

I am not going to take your example of the interference on the temperature sensor into reflection. Yes I did see it, yes it is impressive. But this is not a temperature sensor, and I am not buying you can get away with hiding the shown equipment in a F1 car. Also note I am nowhere denying you can interfere the readings. I am denying you can do it exactly right all the time. I think mudflap mentioned rotor control based on EM which is very precise. That is true (and does make one to appreciate the technology behind it), but we are talking about every part in the line build for optimizing that control in the case of a helicopter. The fuel flow sensor is not build for that, and your device sending out the pulses has to be none-conspicuous, preferably having a primary purpose which is totally unrelated. Not exactly optimized/

Again, this is by no means any reflection on what Ferrari actually does.
#AeroFrodo

enri_the_red
enri_the_red
12
Joined: 03 Jul 2012, 14:12
Location: Italy

Re: Ferrari Power Unit Hardware & Software

Post

Phil wrote:
06 Nov 2019, 12:37

Anyway - as for the sensors: It would be great to know to what tolerances these sensors operate in and how secretive the FIA are about things. Is it a sensor that has a spec, is manufactured by a specific company, perhaps even off-the-shelf where any competitor could gain knowledge on the tolerances, know-how on how it's being measured that then could be used to be exploit? Or are these sensors under top-level-secrecy and the teams can only assume?
Here is the data sheet of the Sentronics FS-100-02 https://www.sentronics.com/wp-content/u ... etMSDS.pdf

User avatar
turbof1
Moderator
Joined: 19 Jul 2012, 21:36
Location: MountDoom CFD Matrix

Re: Ferrari Power Unit Hardware & Software

Post

enri_the_red wrote:
06 Nov 2019, 13:59
Phil wrote:
06 Nov 2019, 12:37

Anyway - as for the sensors: It would be great to know to what tolerances these sensors operate in and how secretive the FIA are about things. Is it a sensor that has a spec, is manufactured by a specific company, perhaps even off-the-shelf where any competitor could gain knowledge on the tolerances, know-how on how it's being measured that then could be used to be exploit? Or are these sensors under top-level-secrecy and the teams can only assume?
Here is the data sheet of the Sentronics FS-100-02 https://www.sentronics.com/wp-content/u ... etMSDS.pdf
Nice! They actually have more info made public than I assumed. Shows me :mrgreen:
#AeroFrodo

User avatar
nzjrs
60
Joined: 07 Jan 2015, 11:21
Location: Redacted

Re: Ferrari Power Unit Hardware & Software

Post

turbof1 wrote:
06 Nov 2019, 14:12
enri_the_red wrote:
06 Nov 2019, 13:59
Phil wrote:
06 Nov 2019, 12:37

Anyway - as for the sensors: It would be great to know to what tolerances these sensors operate in and how secretive the FIA are about things. Is it a sensor that has a spec, is manufactured by a specific company, perhaps even off-the-shelf where any competitor could gain knowledge on the tolerances, know-how on how it's being measured that then could be used to be exploit? Or are these sensors under top-level-secrecy and the teams can only assume?
Here is the data sheet of the Sentronics FS-100-02 https://www.sentronics.com/wp-content/u ... etMSDS.pdf
Nice! They actually have more info made public than I assumed. Shows me :mrgreen:
This is also mighty interesting. A bit more details on ultrasonic flow sensors https://www.sentronics.com/article/why- ... f-a-fluid/

It seems the sensor is inherently more noisy (that page is quite upfront about electrical and vibrational noise sources) and thus I guess the device has a lot of onboard processing to clean up the signal. I wonder if the FIA does anything more than inspecting it's log / min/max values.

I could imagine it worth a try to see if it could be encouraged to under-report (either through the sensor or the sensor+smoothing+sampling strategy inherent) based on manipulating the electrical or vibrational noise spectrum. On the basis of this fantastic post / paper I believe
dans79 wrote:
06 Nov 2019, 04:05
The research paper referenced. https://arxiv.org/pdf/1904.07110.pdf
It wouldn't be much of an investment for a team to buy a few and see how they behave in different noise conditions. Similarly, If the regulations are such that teams must use the FIA homologated sensor (i.e. they can purchase it directly) I'm sure they buy 100 and use the 2 with the lowest readings.

Dr. Acula
Dr. Acula
46
Joined: 28 Jul 2018, 13:23

Re: Ferrari Power Unit Hardware & Software

Post

Well, one thing people seem to forget, although the FIA standard fuel flow sensor is the relevent one for the fuel flow regarding the rules, it's not the only way they can determine the fuel flow.
Does anyone remember the Australian GP of 2014 where Daniel Ricciardo was disqualfied after the race because he breached the flow limit? RBR argued against it by referencing their own fuel flow sensor within the engine. Basically you can calculate the fuel flow via the injector duty cycles. That's all visible in the telemetry. So even if Ferrari would manipulate the signal from the standard fuel flow sensor, the additional fuel would show up elsewhere in the telemetry.

Snorked
Snorked
68
Joined: 16 Mar 2015, 21:00

Re: Ferrari Power Unit Hardware & Software

Post

RZS10 wrote:
06 Nov 2019, 00:18
Just from a logical standpoint ... if that solution would not be feasible at all - why would RBR have it clarified? If it was so far out of reach for an F1 team what would make RBR believe that Ferrari is doing it, thus trying to stop them from doing it via TD?

(That is ofc with the assumption that there was no misrepresentation of what actually was in the TD or what the hypothetical system RBR came up with would/could do)

AMuS implies that Red Bull / Honda (?) engineers replicated it on the test bench with the Honda.

https://www.auto-motor-und-sport.de/for ... or-gp-usa/
Red Bull now chose the preliminary stage in Austin. They played through one of the theories with a Honda engine on a test bench.

LM10
LM10
121
Joined: 07 Mar 2018, 00:07

Re: Ferrari Power Unit Hardware & Software

Post

Polite wrote:
06 Nov 2019, 12:15
One and Only wrote:
06 Nov 2019, 12:03
RZS10 wrote:
06 Nov 2019, 00:18
Just from a logical standpoint ... if that solution would not be feasible at all - why would RBR have it clarified? If it was so far out of reach for an F1 team what would make RBR believe that Ferrari is doing it, thus trying to stop them from doing it via TD?

(That is ofc with the assumption that there was no misrepresentation of what actually was in the TD or what the hypothetical system RBR came up with would/could do)
AFAIK Red Bull asked more than one question. It looks to me it was more of a brainstorming session than anything else. As someone pointed out they were probably throwing s..t against the wall to see if something sticks.
maybe.. or Mercedes gives them informations on the Ferrari PU (Sassi and Allison... for years now did that, but Mercedes doesnt want to be the one protesting to fia)

rumors from motorsport.it: Honda involved in a spy story, their ICE is a copy of the Ferrari one.. also they tried to cheat on the fuel flow this season, several times, till the RaceDirection asked em to stop (noone here remember the story of the pressure of the injection system? was against Honda.. but everyone here thinking it was for Ferrari )..
Can you link that motorsport article please?

User avatar
dans79
267
Joined: 03 Mar 2013, 19:33
Location: USA

Re: Ferrari Power Unit Hardware & Software

Post

turbof1 wrote:
06 Nov 2019, 13:33
As I also explained to Phill, I have no doubt you can design a device capable of sending out the pulses and get that part exactly right. What I am doubting is the interference will go down exactly as planned. Because you don't want readings above 100kg/h. You don't want suspiciously low readings either. The reading have to correspond with where you are on the track. The fuel flow sensor is not developed for outside EM control (again, the other side is that the shielding will not be exactly great either).
The first thing you need to consider, is that car already has 3 rather substantial EMI generators, namely the MGU-K MGU-H, and the CE. All of them will give off various amounts of emi, and operate at substantial frequencies. Not to mention the emi frequency and intensity will change, based on engine modes, rpm, voltage, current etc etc. It could be possible that no one unit affects the sensor, but when two or more components are in specific modes rpm ranges etc, the combined emi patterns yields the desired effect. Even the Ferrari ES being treated like 2 units from last year could play a part, if the CE it switching at the right frequencies. All of this could be played down as an incidental side-effect.
201 105 104 9 9 7

richardn
richardn
2
Joined: 24 Aug 2018, 11:45

Re: Ferrari Power Unit Hardware & Software

Post

"If you hear the Ferrari engines they sound much much different and there's a quintessential electric whine to it much more prominent than a turbo whine. " was a quote from Sierra117 on a Mercedes thread.

Could this be the noise of something pulsing in time with fuel sensor measurements? Has anybody analysed the sound to see if there is a 2kHz component that doesn't vary with engine revs?

User avatar
dans79
267
Joined: 03 Mar 2013, 19:33
Location: USA

Re: Ferrari Power Unit Hardware & Software

Post

Snorked wrote:
06 Nov 2019, 15:51
RZS10 wrote:
06 Nov 2019, 00:18
Just from a logical standpoint ... if that solution would not be feasible at all - why would RBR have it clarified? If it was so far out of reach for an F1 team what would make RBR believe that Ferrari is doing it, thus trying to stop them from doing it via TD?

(That is ofc with the assumption that there was no misrepresentation of what actually was in the TD or what the hypothetical system RBR came up with would/could do)

AMuS implies that Red Bull / Honda (?) engineers replicated it on the test bench with the Honda.

https://www.auto-motor-und-sport.de/for ... or-gp-usa/
Red Bull now chose the preliminary stage in Austin. They played through one of the theories with a Honda engine on a test bench.
When translated to English that doesn't make sense.
Roughly speaking, the flow rate sensor signal is manipulated to be able to inject gas at intervals when it is not being measured.
If you are going to go to the trouble of manipulating the sensors signal, then only injecting the extra fuel between readings is idiotic. The entire point of manipulating the sensor reading would be so that its reads a lower flow value than their actually is, thus allowing you inject fuel whenever you want.
Last edited by dans79 on 06 Nov 2019, 16:22, edited 1 time in total.
201 105 104 9 9 7

User avatar
turbof1
Moderator
Joined: 19 Jul 2012, 21:36
Location: MountDoom CFD Matrix

Re: Ferrari Power Unit Hardware & Software

Post

dans79 wrote:
06 Nov 2019, 15:59
turbof1 wrote:
06 Nov 2019, 13:33
As I also explained to Phill, I have no doubt you can design a device capable of sending out the pulses and get that part exactly right. What I am doubting is the interference will go down exactly as planned. Because you don't want readings above 100kg/h. You don't want suspiciously low readings either. The reading have to correspond with where you are on the track. The fuel flow sensor is not developed for outside EM control (again, the other side is that the shielding will not be exactly great either).
The first thing you need to consider, is that car already has 3 rather substantial EMI generators, namely the MGU-K MGU-H, and the CE. All of them will give off various amounts of emi, and operate at substantial frequencies. Not to mention the emi frequency and intensity will change, based on engine modes, rpm, voltage, current etc etc. It could be possible that no one unit affects the sensor, but when two or more components are in specific modes rpm ranges etc, the combined emi patterns yields the desired effect. Even the Ferrari ES being treated like 2 units from last year could play a part, if the CE it switching at the right frequencies. All of this could be played down as an incidental side-effect.
You are right, these are sources. And they can cause interference. The question does remain if you can get the exact interference you want? This could also end up with interference within interference, not getting the exact pattern you want. We also have to assume the fuel flow sensor is also alteast built to shield from EMI through normal operation from all the devices you mentioned.

I am also thinking about Singapore, where interference is a rather big issue alltogether.

I do like this line of thought though. I might disagree, but it does open up perspectives and insights. Someone quoted senstronics on that the fuel sensor is inherently vulnerable to EMI and speculated that software is in place to clean up the noise. All in all makes a much more complicated story, and maybe also regret my choice of words regarding "bollocks".
#AeroFrodo

User avatar
RZS10
359
Joined: 07 Dec 2013, 01:23

Re: Ferrari Power Unit Hardware & Software

Post

The AMuS quote only says that they proved one of the theories on the test bench, one of the (supposedly) three that were addressed in the technical directive - it's just Schmidt making the connection that they must have tested the one that leaked, which is the one about manipulating the fuel flow sensor in some way.

Polite
Polite
18
Joined: 30 Oct 2018, 10:36

Re: Ferrari Power Unit Hardware & Software

Post

LM10 wrote:
06 Nov 2019, 15:58
Polite wrote:
06 Nov 2019, 12:15
One and Only wrote:
06 Nov 2019, 12:03

AFAIK Red Bull asked more than one question. It looks to me it was more of a brainstorming session than anything else. As someone pointed out they were probably throwing s..t against the wall to see if something sticks.
maybe.. or Mercedes gives them informations on the Ferrari PU (Sassi and Allison... for years now did that, but Mercedes doesnt want to be the one protesting to fia)

rumors from motorsport.it: Honda involved in a spy story, their ICE is a copy of the Ferrari one.. also they tried to cheat on the fuel flow this season, several times, till the RaceDirection asked em to stop (noone here remember the story of the pressure of the injection system? was against Honda.. but everyone here thinking it was for Ferrari )..
Can you link that motorsport article please?
not an article but a video review of the USA gp..

"A tenere banco è anche la polemica innescata da Verstappen sul motore della Ferrari, ma a quanto pare anche Honda non sarebbe esente da "inchieste""
"The controversy triggered by Verstappen on the Ferrari engine is the main theme, but apparently even Honda would not be exempt from inquiries"

then in the video they say that the honda ICE is so similar to the Ferrari one that cannot be fortuity and the focus is the motorhead!

User avatar
nzjrs
60
Joined: 07 Jan 2015, 11:21
Location: Redacted

Re: Ferrari Power Unit Hardware & Software

Post

dans79 wrote:
06 Nov 2019, 16:09
If you are going to go to the trouble of manipulating the sensors signal, then only injecting the extra fuel between readings is idiotic. The entire point of manipulating the sensor reading would be so that its reads a lower flow value than their actually is, thus allowing you inject fuel whenever you want.
RZS10 wrote:
06 Nov 2019, 16:19
The AMuS quote only says that they proved one of the theories on the test bench, one of the (supposedly) three that were addressed in the technical directive - it's just Schmidt making the connection that they must have tested the one that leaked, which is the one about manipulating the fuel flow sensor in some way.
Exactly yeah, I assume the general principle is they came up with a couple of proof of concept 'attacks' for manipulating the fuel flow sensor accuracy (estimate) or measurement.

In some ways, it's the easiest cheat device and cheapest performance enhancer - no special ICE magic to escape the law-of-diminishing returns, just burn a very little more fuel.