Honda Power Unit Hardware & Software

[Bandit1216]

How can one use elliptical gears in a gearbox? Isn't this only possible with a chain/belt or with 1:1 ratio

[Tommy Cookers]

....Having 100% linear torque delivery to the tyres would be optimal in some sort of closed loop system where a low-latency sensor can detect and vary throttle application to ensure the optimal slip for traction and type temperature, but it'd be hard for the driver to feel exactly where the limit is if the driver was the sensor.....it may make sense to have an element in the transmission that varies the torque delivery in a nonlinear (eg. sine) pattern. The driver can then feel where the limit of adhesion is because every few feet the tyres will breach and retreat from that limit. Such a system might explain this phenomenon: https://www.motor1.com/news/39418/photo ... rol-video/
This sort of thing is used in karting, where electronics aren't permitted in things like braking systems (non-circular rotors and non-circular venting patterns on brake discs are common because they permit feel for where the limit actually is).
The only other reason I can think of why it'd do that would be to deploy energy in the ES via the H. You would pulse the H up, suck in more air, shove it through the engine like in "Q3 free" mode, just for a short period of time to let the ICE climb up into the next few tens of revs each time ...
It would also explain why Honda were having so many problems with instability, if this is their strategy for long straight

ripple rate proportional to axle cyclic rate would need to avoid frequency components exciting the transmission unfavourably
difficult over a big speed range
electric machine-sourced ripple can avoid this difficulty

if in karting a brake torque ripple is engineered - does it improve grip directly or by helping driver 'feel' ?
isn't this is the 'big bang' story/myth all over again ?
though iirc imo traditional BB (of the 1950s or 1980s) is contradicted by the 2000s crossplane Yamaha BB claims
(btw amusingly there were similarly contradictory views in 1920s steam loco design)

[henry]

Maybe they do extra harvest?
1Hz oscillation for harvesting and accumulating ... somewhere, while with 20Hz they transfer via H to K ?

I don’t think so. The harvesting relies on allowing the turbine/H/compressor assembly to accelerate increasing its kinetic energy. At 1hz that’s 500msec of acceleration which would increase the revs to well beyond 120000. I estimated a 6000rpm rise in 50msec when this was discussed earlier in the thread.

[HPD]

Ok, the source is this https://unendinginsight.wordpress.com/
We do not know what methods he uses.
We can see on the Chinese straight, the same pattern.

Bot vs Ver https://unendinginsight.files.wordpress ... vs_ver.png
Honda vs Renault https://unendinginsight.files.wordpress ... vs_ric.png
Etc.
Edit: He is a reddit user, he uses the F1 App for his data.

[Juzh]

That sawtooth is also apparent on the left side of the graph. Between T2 and T3 there are some noticeable differences. Looks like one gearshift where HAM stayed flat. The shape of the Honda gearshifts seem slow. Error in graph or was the gear change really so slow? Don't hear delayed shifts in ths video.

How was the graph made? Manual data entry from video reference? Sorry if this has been explained already.

Ok, the source is this https://unendinginsight.wordpress.com/
We do not know what methods he uses.
We can see on the Chinese straight, the same pattern.

Bot vs Ver https://unendinginsight.files.wordpress ... vs_ver.png
Honda vs Renault https://unendinginsight.files.wordpress ... vs_ric.png
Etc.
Edit: He is a reddit user, he uses the F1 App for his data.

There's pretty much only 1 method possible: OCR (optical character recognition), probably using tesseract software, rendering captured video of the telemetry into still images and then running recognition on all of them. Because the framerate of the video is known you can easily plot velocity against time and so on. With the new F1 app's very clean UI and static background on the speedometers it's now possible to do near perfect recognition with ease, therefore doubting the values in graph itself is pointless.
Also, the guy who did those graphs has proven track record, though his own analysis of the graphs he makes is sometimes up for a debate imo.

Explained in much more detail in this youtube video:
https://www.youtube.com/watch?v=VUgvhnQawbw


So far I've not yet heard of anyone reverse engineering the new f1 app to extract all the possible data. It was possible on the old days, not so much now.

[henry]

Having watched the video I have a couple of questions.

Firstly, I didn’t catch how the lap distance is arrived at.

Secondly, if the method is to extract the data from the on screen data then there is still the question of how the engine rpm get to the screen. Does anyone know how this is done? In fact how does any of the data get there? I don’t have the app.

[gruntguru]

That said, personally I think the RPM up/down seesaw on the graph is just noise in the data and nothing to do with the above hypothetical system, but everyone likes theories, so there you go.

The phenomenon seen on the graph has a frequency between 1 and 2 Hz so definitely not synchronous with any rotating shaft in the car.

[gruntguru]

Having watched the video I have a couple of questions.

Firstly, I didn’t catch how the lap distance is arrived at.

Secondly, if the method is to extract the data from the on screen data then there is still the question of how the engine rpm get to the screen. Does anyone know how this is done? In fact how does any of the data get there? I don’t have the app.

The data he exports to excel does not include distance so he must integrate the velocity data in excel.

Assume that the data displayed in the app is all sourced from the team's telemetry.

[gruntguru]

With the new F1 app's very clean UI and static background on the speedometers it's now possible to do near perfect recognition with ease, therefore doubting the values in graph itself is pointless.

Agreed. The only grey area is time measurement. The technique assumes the on-screen data is synchronised to real time so if the flow of data is "surging" there will be errors. Those errors would also show in the speed trace so not a likely source for the rpm oscillation.

[henry]

The average slip at that speed would be in the vicinity of 1% or 2% so a 1% oscillation would be significant.

It would be good if someone more knowledgeable on tyres can chip in here. The tractive force would be about 6.7 kN (assuming 330 km/hr and 600kW). Normal force on the rear tyres would be about 4.5 kN plus aero - say 15 kN total?

The most likely source of a significant variation in the tractive force is the MGU-K. If peak power is 600kW (ICE + K) then the minimum (ICE-K) would be 400kW. So 1% slip caused by 200kW would put the two slip values at 2% and 3%. I simply eyeballed the RPM so it might be 0.8% variation for slip values of 1.6 and 2.4%. I don’t know how realistic they would be.

With a 50-50 bang-bang strategy that would give a net gain to the ES of the order of 60kW, charge at 180kW, discharge at 60kW.

This could be mixed with an extra harvest strategy during the charge phase, this would reduce the average charge. Depending on the mark space of the extra harvest the average ES gain might be at 45 to 50kW. Apologies to @roon if this was your point.

This might be a fair bargain with lower losses on the electrical side than the mechanical, and the opportunity to do something more useful with that energy than move air.

[roon]

By slip do you guys mean clutch or tire? Not sure what you're accrediting to me, Henry.

Could this be a way to put heat into the rear tires? Cycle tire slip moments on the straights. Maybe VES simply had too many red bulls. Would pulsing the accelerator pedal at 1 hz produce this effect? Without noticeable affect to speed.

In the T2-T3 section of that Verstappen-Bottas graph there are large symmetrical peaks and valleys. If not data error, the trailing sides are odd, assuming they're gearshifts. They should be vertical lines on the trailing side if they were normal gearshifts.

[Tommy Cookers]

The phenomenon seen on the graph has a frequency between 1 and 2 Hz so definitely not synchronous with any rotating shaft in the car.

2 Hz seems rather like a surge frequency of the car via its transmission of force to ground
1 Hz seems a bit low for this

[henry]

By slip do you guys mean clutch or tire? Not sure what you're accrediting to me, Henry.

Could this be a way to put heat into the rear tires? Cycle tire slip moments on the straights. Maybe VES simply had too many red bulls. Would pulsing the accelerator pedal at 1 hz produce this effect? Without noticeable affect to speed.

In the T2-T3 section of that Verstappen-Bottas graph there are large symmetrical peaks and valleys. If not data error, the trailing sides are odd, assuming they're gearshifts. They should be vertical lines on the trailing side if they were normal gearshifts.

I mean tyre slip since I think that amount of clutch slip would kill the clutch, assuming there really is slip.

Apologies to you and @restless who’s post I wrongly remembered as yours.

The average slip for the tyres would be less so it might reduce their temperature.

I’m not sure this could be initiated by accelerator pedal movement, but it might if the driver taught themselves to do it in a simulator.

The gearshift data is interesting. The trailing edge of the upshifts won’t be vertical because of the inertia of PU but they do seem quite slow. @tommy cookers has pointed out a number of times that the MGU-K can be used to slow the PU during upshifts so you’d expect the Rev drop to be pretty rapid. I wish these cars didn’t have so many gears, the data would be much clearer.

It’s noticable that Verstappen double shifts down during braking to keep the revs high. Or maybe his fingers twitch like his foot

[Tommy Cookers]

we know the clutch has for years been slipped in a finely controlled way (to minimise the disruptive effects of shifts)
eg during upshifts clutch torque transmission capability is varied at about 8 millisec periodicity
this work dumped as slip friction is quite small

just guessing, cyclic clutch slipping at eg 1 Hz would involve far more slip work, and produce far more heat and wear

[kasio]

Here is a good video if you want to analyze more.
The Mercedes RPM obviously looks more stable than the Honda RPM. But I still think that it is within the normal range.

Max q3 https://streamable.com/r2tbc
Ham q3 https://streamable.com/nld6t

Here You are. The speed/rpm variation is only the first time he is passing the stright(what we see in telemetry as well then). The second time he has better speed and its not variating. So it now even more looks like it is harvesting.