Ferrari Power Unit Hardware & Software

[roon]

The ES-H-K link could be established within the control electronics in order to avoid physical cycling of the MGU-H for energy transfer. Just link the control electronics, not the motors. K>K controller>H controller>ES. No complex MGUH cycling schemes nor turboshaft clutches needed.

The Honda data may be showing just that. Controller/capacitor activity, not MGU-H rotational speed.

[giantfan10]

PG Tech:

Loophole on hybrid part. Ferrari, who lost power during the controls between Spain and France (hence the 35 hp increase in Austria), was able to take advantage of even more aggressive mapping after the final okay of the FIA in the controls.

Why is no one else talking about this? If I understand this correctly, the FIA is not monitoring this aspect of Ferrari's ERS anymore since Austria? How is that possible? Surely they must leave the sensors on for the rest of the season. Many clues have shown that it is an engine mapping that is only a software update away from being activated or deactivated (FIA was only satisfied with Ferrari's ERS after the latter did a software update, before that they were seeing "strange things"). The fact that Ferrari lost power while monitoring was enforced makes it all the more suspicious. Surely the FIA can't be that naive.

Were you this concerned when Mercedes was winning races by 25+ seconds for the first several years of this formula?
Were you so adamant that they were doing something illegal when they had a vastly more powerfull engine with party modes that no other team had or could replicate?
Do you think its possible that Ferrari ran spec 2 in a conservative mode untill they were confident in its reliability?
Is this a case of the team that you root for has lost their power advantage so anybody that surpasses them has to be cheating?

[henry]

The ES-H-K link could be established within the control electronics in order to avoid physical cycling of the MGU-H for energy transfer. Just link the control electronics, not the motors. K>K controller>H controller>ES. No complex MGUH cycling schemes nor turboshaft clutches needed.

The Honda data may be showing just that. Controller/capacitor activity, not MGU-H rotational speed.

I think the sensor scheme may prevent this. When energy leaves the K it is recorded by the sensor at the K. When energy arrives at the ES the sensor there records it. If there is no delay between them the interpretation will be that the flow was K to ES and will be subject to the 2MJ limit.

There is a limit of 5kJ on storage within the control units. Whether that would be sufficient to change the phase of leaving/arriving events in relation to the sensor sampling frequency I don’t know.

In fact I don’t know how the ES<->K flows are policed using just 2 sensors, so maybe there are ways to divert energy through the CU interconnections and fool the measurement protocols.

[GrandAxe]

The FIA has said that the Ferrari system is very complicated and considering that the power boost is only coming to some teams after a software upgrade, the mind immediately considers AI.

A theoretical software map might include machine learning software that can turn hardware (or other software) switches on and off depending on circumstance - that way, alternate functions of components can be concealed. Such an approach would confound the devil out anyone outside the implementers of the software and hardware. Machine learning software on its own is very complicated, talk less a system that might be part software, part electronic hardware, part mechanical/hydraulic/pneumatic.

Such a system might be able to learn and react to sensor readings as well as other parameters to know when to switch modes, so that the FIA (and you or I) would never be the wiser. For instance, if all the FIA has to monitor compliance are two DC sensors, then they might be beaten if circuits are made to behave in non-obvious ways, or their functionality changes completely.

[bill shoe]

Can I just say this is delicious? Technical intrigue, unexplained advantages, some kind of innovation with clearly improved results but the best competitor minds can't figure out the means... This is what F1 is supposed to be about!!

[PhillipM]

The FIA has said that the Ferrari system is very complicated and considering that the power boost is only coming to some teams after a software upgrade, the mind immediately considers AI.

A theoretical software map might include machine learning software that can turn hardware (or other software) switches on and off depending on circumstance - that way, alternate functions of components can be concealed. Such an approach would confound the devil out anyone outside the implementers of the software and hardware. Machine learning software on its own is very complicated, talk less a system that might be part software, part electronic hardware, part mechanical/hydraulic/pneumatic.

Such a system might be able to learn and react to sensor readings as well as other parameters to know when to switch modes, so that the FIA (and you or I) would never be the wiser. For instance, if all the FIA has to monitor compliance are two DC sensors, then they might be beaten if circuits are made to behave in non-obvious ways, or their functionality changes completely.

You don't have that sort of compute power available on the car, not even close.

[Zynerji]

The FIA has said that the Ferrari system is very complicated and considering that the power boost is only coming to some teams after a software upgrade, the mind immediately considers AI.

A theoretical software map might include machine learning software that can turn hardware (or other software) switches on and off depending on circumstance - that way, alternate functions of components can be concealed. Such an approach would confound the devil out anyone outside the implementers of the software and hardware. Machine learning software on its own is very complicated, talk less a system that might be part software, part electronic hardware, part mechanical/hydraulic/pneumatic.

Such a system might be able to learn and react to sensor readings as well as other parameters to know when to switch modes, so that the FIA (and you or I) would never be the wiser. For instance, if all the FIA has to monitor compliance are two DC sensors, then they might be beaten if circuits are made to behave in non-obvious ways, or their functionality changes completely.

You don't have that sort of compute power available on the car, not even close.

Actually, if trained in the factory, it should be more about hard drive space when installed in the car.

[saviour stivala]

As everybody knows the ES allowed energy deployment is max 4MJ while the ERS-K allowed harvesting is max 2MJ which means that ERS-K can only charge ES 50%. In turn all that means that in order to have a fully charged ES after what is allowed out of it has been deployed the ERS-H will have to compensate at least the missing 50% charge to have a fully charged ES ready to deploy on the next lap. There is differences in the capabilities to harvest between the power units turbo/ERS-H combinations. The difficulties of the ERS-K harvesting even that of what it is allowed to harvest on most tracks makes the capability of harvesting by turbo/ERS-H combination even more important. A while back on here I was arguing with another gentleman with me saying that ERS-K harvesting is only done under braking (only triggered once the brake pedal is pressed), with the other gentleman saying that ERS-K harvesting is even done when the engine is under power, today in FP1 we got the answer to our arguments. FP1 11:30 Vandoorne to pit wall “I have no deployment”. Mclaren pit wall promptly gave him some ‘recharge’ setting changes. 12:14 mclaren pit wall tell Vandoorne “don’t brake too hard against the throttle” this after Vandoorne had triggered a ‘throttle fail safe”. This throttle fail safe is triggered when applying both throttle and brake together using both feet (the drivers does that in some situations to help ERS-K harvest) the fail safe algorithm is designed to override the throttle and cut the engine by the torque coordinator, which controls the rear brake-by-wire system, the same rear brake system that triggers the ERS-K harvesting.

[PhillipM]

Makes no difference, under anything less than 100% torque demand from the pedal the K can harvest. You don't need overlap to do that.

[OO7]

The obvious question, which I assumed has an obvious answer is, is there a way to store energy that is not considered outside the rule of a single 'energy store'.

I can't see where the regs limit competitors to a single ES within the car?

[roon]

When energy leaves the K it is recorded by the sensor at the K. When energy arrives at the ES the sensor there records it. If there is no delay between them the interpretation will be that the flow was K to ES and will be subject to the 2MJ limit.

Everything else could go from the K CU to the H CU, then onto the ES.

There is a limit of 5kJ on storage within the control units. Whether that would be sufficient to change the phase of leaving/arriving events in relation to the sensor sampling frequency I don’t know.

What would the sampling frequency be? Why would it not be constant?

[OO7]

I've been considering the Honda extra harvest mode and the theory proposed by Craigy? about using the compressor and or turbine as a flywheel.

Rather that using the pressure charging system as the flywheel, why not simply have an integrated flywheel on a separate spool (like a multi-stage jet engine), that can be clutched to the pressure charging system and or the MGU-H. Under part throttle conditions in GENSET mode, the MGU-K (via crankshaft power) could be used to spin-up the flywheel and exceed the 4Mj ES limit, as it is a separate mechanical ES (not 'cell' based) and thus not subject to those energy or packaging limitations.

Unfortunately it isn’t allowed by the rules.

5.2 Other means of propulsion and energy recovery :
5.2.1 The use of any device, other than the engine described in 5.1 above, and one MGU-K, to
propel the car, is not permitted.

I forgot to add that the flywheel doesn't recover energy it stores it, so article 5.2.1 shouldn't be applicable.

[wuzak]

PG Tech:

Loophole on hybrid part. Ferrari, who lost power during the controls between Spain and France (hence the 35 hp increase in Austria), was able to take advantage of even more aggressive mapping after the final okay of the FIA in the controls.

Why is no one else talking about this? If I understand this correctly, the FIA is not monitoring this aspect of Ferrari's ERS anymore since Austria? How is that possible? Surely they must leave the sensors on for the rest of the season. Many clues have shown that it is an engine mapping that is only a software update away from being activated or deactivated (FIA was only satisfied with Ferrari's ERS after the latter did a software update, before that they were seeing "strange things"). The fact that Ferrari lost power while monitoring was enforced makes it all the more suspicious. Surely the FIA can't be that naive.

You mean they only monitor what they monitor on all cars? How dare they!

[wuzak]

...The drivers use both feet for each respective pedal. I'd expect there to be some overlap between acceleration and braking events. The software should be able to deal with this. Maybe it was a specific map/mode he was in that had issues with it.

Do you remember the race when Rosberg had a huge problems with electronics (possibly Singapore 2014 or 15? My memory isn't that great...) The engineers told him (broadcast on-air) that the brake pedal input would override and cancel the throttle input. I thought it was odd at the time because you effectively can't trail brake..? I assumed that ERS has changed the philosophy on trail braking? This is the first time I've seen it brought up since...

You may have been thinking Russia 2015?

Rosberg led early, from pole position, but then developed a problem with his throttle.

From what I recall, the spring/damper unit on the accelerator pedal broke, so that the pedal would not fully return to the top of its travel.

He tried to live with it, but it got worse, and he kept on missing corners as the engine wasn't slowing down as needed, so he retired.

Other problems:
Rosberg had an MGUH failure at he final race in 2015 at Abu Dhabi. Lag was ridiculous and he was well off teh pace, but he finished.
Both Rosberg and Hamilton had MGUK failures a the Canadian GP in 2014, when they were running 1-2. Hamilton retired with rear brake failure, Rosberg survived to finish 2nd, behind Ricciardo.

[bill shoe]

AMuS article seems to show that the Ferrari advantage is an ability to deploy the 120 kW K-boost significantly longer than other powertrains. Ferrari continues this 120 kW boosting further along longer straights than competitors, but still has to back off toward the end and therefore has similar top speed at end of straight. So not really a increase in max power, but a boost in how long you can deploy your max power.

Competitors integrate each other's K-boost (120kW x time deployed) from the GPS data and see ~ 4 MJ per lap, per the nominal wording of the regulations. Except for Ferrari which by the same type of calculations is getting ~ 4.5 MJ per lap, maybe even more. Maybe it averages out to 38 hp over a lap or over a long straight, but when it (the trick) is on they get 120kW (160 hp) for a precious few extra seconds. OK.

Intuitive assumption when looking at the data is to assume the trick extends the time after the "normal" K-boost has finished. But there are two additional alternatives-

1). The trick is used at the beginning of the straight and the extra time at the end was actually the "normal" K-boost.

2). The trick is a small 15 kW addition the entire time, and Ferrari only uses the normal K-boost at the rate of ~ 105 kW. This effectively results in a total 120 kW boost but for a longer than normal time (15 kW trick plus 105 kW normal = 120 kW apparent total). This gives close to optimal speed while totally obfuscating what is really going on.