Possible limits, tricks and cheats of the flow sensors?

[jumpingfish]

any one of those situations will result in the car being disqualified. if proof of a car being disqualfied for any of the above situations is needed i will gladly give that proof moderators permitting.

You need to stop reading the ”rules” like they are some kind of absolutes, because they are not. Every thing has a tolerence.

A momentary Spike of say less then two tenths of a second, will not get a car disqualified it'll get them a warning.

That is exactly what happened in 2014 to red bull in Australia that you so often like to reference as if it's some kind of validation of your opinion. Red bull was disqualified in Australia, because after multiple warnings they did nothing to rectify the issue, insisting there monitoring methodology was better than the ffs.

Did RB get warnings during Qualify in Azerbaijan 2019 when Gasly was disqualified?
Horner said: Basically the fuel flow sensor sometimes get a little bit of oscillation and he got a good tow on his lap from I think it was a [Racing Point] so that puts it into the limiter in top gear and the resonance was just such that it dipped in and out of the limit by 0.02 of a gram or something. So in terms of performance, nothing, but obviously zero tolerance from the FIA.

https://www.racefans.net/2019/05/03/rac ... p-03-05-2/

[henry]

any one of those situations will result in the car being disqualified. if proof of a car being disqualfied for any of the above situations is needed i will gladly give that proof moderators permitting.

You need to stop reading the ”rules” like they are some kind of absolutes, because they are not. Every thing has a tolerence.

A momentary Spike of say less then two tenths of a second, will not get a car disqualified it'll get them a warning.

That is exactly what happened in 2014 to red bull in Australia that you so often like to reference as if it's some kind of validation of your opinion. Red bull was disqualified in Australia, because after multiple warnings they did nothing to rectify the issue, insisting there monitoring methodology was better than the ffs.

I think @saviour stivala is right. If the spike occurs at the fuel flow sensor and it takes fuel flow rate above 100kg/hr it is a “disqualification”. There is no tolerance above the limit.

Where he is wrong is that he doesn’t believe that there can be fluctuations downstream of the FFS which are not seen by the measuring equipment and which can go above the limit. As @gruntguru has pointed out the transient excursions are inevitable because of the discrete nature of injection events.

It does raise the issue of what happens if a FFS fails. It has been reported that they fall back on injector based measurements but since these have the fluctuations described they must use some other time base to avoid miscalculation of the average flow rate. In the end all of these measurements are averages even if the timebase is very small.

[dans79]

any one of those situations will result in the car being disqualified. if proof of a car being disqualfied for any of the above situations is needed i will gladly give that proof moderators permitting.

You need to stop reading the ”rules” like they are some kind of absolutes, because they are not. Every thing has a tolerence.

A momentary Spike of say less then two tenths of a second, will not get a car disqualified it'll get them a warning.

That is exactly what happened in 2014 to red bull in Australia that you so often like to reference as if it's some kind of validation of your opinion. Red bull was disqualified in Australia, because after multiple warnings they did nothing to rectify the issue, insisting there monitoring methodology was better than the ffs.

Did RB get warnings during Qualify in Azerbaijan 2019 when Gasly was disqualified?
Horner said: Basically the fuel flow sensor sometimes get a little bit of oscillation and he got a good tow on his lap from I think it was a [Racing Point] so that puts it into the limiter in top gear and the resonance was just such that it dipped in and out of the limit by 0.02 of a gram or something. So in terms of performance, nothing, but obviously zero tolerance from the FIA.

https://www.racefans.net/2019/05/03/rac ... p-03-05-2/

I went to look it up, and as expected the FIA has their website screwed up. the technical delegate report, the summons, and the ruling all point to the same PDF and that is the ruling PDF that gives almost no detail.

[saviour stivala]

"Otmar Szafnauer TD (who's car was running mercedes hardware and software) explained that his team's data had shown that the car gad BRIEFLY encountered unprecedented SURGE of fuel, but had corrected itself before the lap was done". "What happened was that there was a SPIKE of flow and than a TROUGH - so if you looked at it over a lap it was neutral". As is normal he had a chance of appelling his car disqualifiucation, but he didn't.

[j.yank]

There are not any fluctualtions or average flow rates in FIA rules: 100 kg/h is absolute maximum. The question is how this fow rate is measured.

[gruntguru]

It is all about the duration of the "spike". There is no doubt that during a single revolution of the engine there will be multiple periods where the flow into the engine is dramatically in excess of the 100 kg/hr limit, however there will also be periods where the flow is dramatically less than 100. What is important is that during the same period, the flow through the flow sensor will be oscillating much less and will be measured as the average of any oscillations that still exist at that point.

[subcritical71]

https://www.flowmeters.co.uk/formula-1 ... ing-teams/

A few seasons ago another F1© team did not trust the fuel flow figures being returned from their engine suppliers fuel management systems. Titan designed a very lightweight oval gear flow meter to be installed safely in the fuel tank of the racing car. Designed to be immune to immersion in fuel and the very noisy electrical environment of an F1© racing car the flow meter has provided accurate flow measurement over an extended period of time.

[saviour stivala]

Fuel flow spikes/flactuations/averages possibilities. There is no spikes/flactuations or averages when it comes to policing the maximum fuel flow rate permitted by the rules. If there is any spikes/flactuations/averages in fuel flow anywhere in the system, and that includes downstream of the FFS, the FFS sees them all because what passes through the FFS must end-up in combustion chasmbers, and what finally pushes what has flowed through the FFS into the combustion chambers are the injectors, so there is no way that the injectors will flow more than what is permitted without the FFS not seeing it. An example. RB car on a long straight in top gear gets a good tow and it goes into the rev limitter and momentarly goes over the flow limit by just 0.02 gram. That fuel flow went through the FFS as well as through the fuel injectors. In short what went through the FFS endeed up into the combustion chambers. That car got disqualified for breaching the fuel flow rules, afterwards the team changed some godes so as to prevent that from repeating itself. The result: they were served with another penalty for changing the godes while under park ferme.

[subcritical71]

Fuel flow spikes/flactuations/averages possibilities. There is no spikes/flactuations or averages when it comes to policing the maximum fuel flow rate permitted by the rules. If there is any spikes/flactuations/averages in fuel flow anywhere in the system, and that includes downstream of the FFS, the FFS sees them all because what passes through the FFS must end-up in combustion chasmbers, and what finally pushes what has flowed through the FFS into the combustion chambers are the injectors, so there is no way that the injectors will flow more than what is permitted without the FFS not seeing it. An example. RB car on a long straight in top gear gets a good tow and it goes into the rev limitter and momentarly goes over the flow limit by just 0.02 gram. That fuel flow went through the FFS as well as through the fuel injectors. In short what went through the FFS endeed up into the combustion chambers. That car got disqualified for breaching the fuel flow rules, afterwards the team changed some godes so as to prevent that from repeating itself. The result: they were served with another penalty for changing the godes while under park ferme.

0.02 gram in what amount of time? That is the whole point I don't believe you are understanding. The limit is 100kg/hr, nobody disputes this not even you. Does that get divided into the nanosecond (1 billionth of a second) level or the second level for rules enforcement. I first find it strange that they specify the limit in hours vs. the unit that they are actually enforcing. If its the nanosecond level then there would be different strategies vs if it were at the second level. The other thing to keep in mind is that just because the sensor can refresh at multi kHz levels does not mean it is communicating to the FIA logger at that rate either due to the data rate implemented by the sensor or other technical considerations such as existing and higher priority CANBUS traffic.

[subcritical71]

0.02 gram in what amount of time? That is the whole point I don't believe you are understanding. The limit is 100kg/hr, nobody disputes this not even you. Does that get divided into the nanosecond (1 billionth of a second) level or the second level for rules enforcement. I first find it strange that they specify the limit in hours vs. the unit that they are actually enforcing. If its the nanosecond level then there would be different strategies vs if it were at the second level. The other thing to keep in mind is that just because the sensor can refresh at multi kHz levels does not mean it is communicating to the FIA logger at that rate either due to the data rate implemented by the sensor or other technical considerations such as existing and higher priority CANBUS traffic.

To clarify this comment. If the limit is published as 100kg/hr I would not expect another limit of 27.77778g/sec to be the enforcement trigger. I would expect it to be 100kg/hr. In other words as long as I average 27.777778g/sec over that hour I am ok. We know however the FIA are not doing it this way but they are vague on their actually enforcement resolution (is it per lap, per sector, per second, per millisecond, per nanosecond, etc)? It would be nice to know definitively what the limit is and the timescale.

[nzjrs]

Fuel flow spikes/flactuations/averages possibilities. There is no spikes/flactuations or averages when it comes to policing the maximum fuel flow rate permitted by the rules. If there is any spikes/flactuations/averages in fuel flow anywhere in the system, and that includes downstream of the FFS, the FFS sees them all because what passes through the FFS must end-up in combustion chasmbers, and what finally pushes what has flowed through the FFS into the combustion chambers are the injectors, so there is no way that the injectors will flow more than what is permitted without the FFS not seeing it. An example. RB car on a long straight in top gear gets a good tow and it goes into the rev limitter and momentarly goes over the flow limit by just 0.02 gram. That fuel flow went through the FFS as well as through the fuel injectors. In short what went through the FFS endeed up into the combustion chambers. That car got disqualified for breaching the fuel flow rules, afterwards the team changed some godes so as to prevent that from repeating itself. The result: they were served with another penalty for changing the godes while under park ferme.

Do you acknowledge any relevance of Mr Nyquist to these discussions of cheating and enforcement of the fuel flow sensor? https://en.wikipedia.org/wiki/Nyquist%E ... ng_theorem

(@others, this is not to suggest that the previously hypothesized exploits of the FFS were simply aliasing tricks, but to check if SS is on the same page wrt. measurement at all)

[dans79]

The other thing to keep in mind is that just because the sensor can refresh at multi kHz levels does not mean it is communicating to the FIA logger at that rate either due to the data rate implemented by the sensor or other technical considerations such as existing and higher priority CANBUS traffic.

Early on in the turbo hybrid era Gill provided the sensors, I'm not sure if they have for the last two seasons though.
https://www.gillsc.com/newsitem/26/gill ... -formula-1

This is the last Gill sensor.
https://www.gillsc.com/products/flow-se ... w-meter-2/


An interesting comment from the press release.

Since its original 2014 release the fuel flow meter now features an increased sampling rate of 2.2 KHz and a faster start up time of only three seconds.

Some interesting comments from the spec sheet.

2.2 kHz internal measurement rate
CAN outputs at 100Hz, with filtering

(8) Cavitation and entrained gas can cause meter damage and spurious measurement
results, this must be avoided by appropriate system design and flow meter operation.

Just in the quotes above, you can see two ways of potentially cheating and you can see the rate at which the data is logged can be different than the sampling rate.

[nzjrs]

Just in the quotes above, you can see two ways of potentially cheating and you can see the rate at which the data is logged can be different than the sampling rate.

Re: 2.2khz vs 100hz logging, I think I read it has on-device min/max logging, but the rest of your point I agree.

I think, given the ability to generate different fuel flow rates fast enough (let's call this a FF profile, and let's say changing the FF rate with temporal resolution at ~10khz is sufficient), it is a 1 month master-student-level project to do an automated parameter sweep through different FF profiles and extract the on-device logged min/max fuel flow to look for disagreements with the *true* rate.

I'm not sure how challenging the generation of varying fuel flow is (mechanically, possibly also control-wise / electro-mechanically). I'm trying to emphasize how easy it would be to investigate the 'accuracy' of the FIA mandated FFS (including it's inherent sampling and filtering) once you had an adjustable fuel circuit and FIA FFS on a bench.

[subcritical71]

The other thing to keep in mind is that just because the sensor can refresh at multi kHz levels does not mean it is communicating to the FIA logger at that rate either due to the data rate implemented by the sensor or other technical considerations such as existing and higher priority CANBUS traffic.

Early on in the turbo hybrid era Gill provided the sensors, I'm not sure if they have for the last two seasons though.
https://www.gillsc.com/newsitem/26/gill ... -formula-1

This is the last Gill sensor.
https://www.gillsc.com/products/flow-se ... w-meter-2/


An interesting comment from the press release.

Since its original 2014 release the fuel flow meter now features an increased sampling rate of 2.2 KHz and a faster start up time of only three seconds.

Some interesting comments from the spec sheet.

2.2 kHz internal measurement rate
CAN outputs at 100Hz, with filtering

(8) Cavitation and entrained gas can cause meter damage and spurious measurement
results, this must be avoided by appropriate system design and flow meter operation.

Just in the quotes above, you can see two ways of potentially cheating and you can see the rate at which the data is logged can be different than the sampling rate.

Thanks for that summary. You can see immediately that there has to be some conditioning going on... 2.2kHz sampling frequency with only a 100Hz (or ~0.01sec) CAN bus frequency update... Is the sensor reporting the min/mean/average/max/?? at that 100Hz frequency? I did look at another FFS for motorsport application which also outputted a tamper flag. This was used to indicate electromagnetic interference was higher than a threshold and would flag it, even if intermittent. Seems this is a known exploit of these sensors.

[gruntguru]

Fuel flow spikes/flactuations/averages possibilities. There is no spikes/flactuations or averages when it comes to policing the maximum fuel flow rate permitted by the rules. If there is any spikes/flactuations/averages in fuel flow anywhere in the system, and that includes downstream of the FFS, the FFS sees them all because what passes through the FFS must end-up in combustion chasmbers, and what finally pushes what has flowed through the FFS into the combustion chambers are the injectors, so there is no way that the injectors will flow more than what is permitted without the FFS not seeing it.

For anyone that understands electronic fuel injection, there is absolutely no question that at full load:

• The flow into the engine fluctuates above and below an average of 100 kg/hr and the fluctuations are significant - probably +- 50 kg/hr or more.

• There is sufficient "damping"/"accumulation" in the high pressure section of the fuel system, to virtually eliminate these fluctuations from appearing in the low pressure section, ie the flow meter.

• The lowest frequency within these fluctuations is of the order of 250 Hz (odd-firing V6) and the readings from the flow meter would be low-pass filtered to remove any oscillations above some lower frequency, say 100 Hz. With a sampling rate of 2.2 kHz the flow meter data cannot reliable identify frequencies above 1.1 kHz anyway. Further, the shape of the fuel flow waveform could not be reliably interpreted above perhaps 200 Hz.