Honda Power Unit Hardware & Software

[restless]

Something I just found on autosport forum

Honda was unable to use qualifying mode at the Japan GP. Suzuka Special also seems to have prepared preparatory qualifying mode, but it has been forced to abandon its use by strengthening regulations from FIA, and in Suzuka where the power effect reaches as large as 0.2 sec / 10 kW, this disadvantage never hurts It was not small.


http://members.f1-life.net/race/59662/

So it seems Honda qualy mode was "oil-burning" ?!

[PhillipM]

I don't compare to any other cars - just changing the power output of one car with the same configuration.

Well, I don't know what you were doing, but I just setup a couple of cars to dial in a bit of correlation from min/max/cornering speeds at Austin, once I was fairly repeatable I then selected one car, dialled down 300bhp, and it cost it 4 SECONDS PER LAP.
Error between keyboard and chair?

[ziggy]

Something I just found on autosport forum

Honda was unable to use qualifying mode at the Japan GP. Suzuka Special also seems to have prepared preparatory qualifying mode, but it has been forced to abandon its use by strengthening regulations from FIA, and in Suzuka where the power effect reaches as large as 0.2 sec / 10 kW, this disadvantage never hurts It was not small.


http://members.f1-life.net/race/59662/

So it seems Honda qualy mode was "oil-burning" ?!

Yup. It was talked about in this thread after Suzuka. It was also confirmed by AMuS, MST, Honda, Taffin, ... It all comes together now, but sadly some people here don't wan't to hear or read about it. Honda was a little to late to the party as it was and will be further restricted in 2018. Oh, it wasn't oil burning per se, but rather a mix of 3 different additives.

It's still unclear how exactly it is done. I think Honda introduced the concept in Monza...

Edit: link: https://translate.google.com/translate? ... 47718.html

[j.yank]

I don't compare to any other cars - just changing the power output of one car with the same configuration.

Well, I don't know what you were doing, but I just setup a couple of cars to dial in a bit of correlation from min/max/cornering speeds at Austin, once I was fairly repeatable I then selected one car, dialled down 300bhp, and it cost it 4 SECONDS PER LAP.
Error between keyboard and chair?

What cars you had setup? Open wheeler with typical chassis specifications of present day Formula 1 cars?

[DFX]

I don't compare to any other cars - just changing the power output of one car with the same configuration.

Well, I don't know what you were doing, but I just setup a couple of cars to dial in a bit of correlation from min/max/cornering speeds at Austin, once I was fairly repeatable I then selected one car, dialled down 300bhp, and it cost it 4 SECONDS PER LAP.
Error between keyboard and chair?

What cars you had setup? Open wheeler with typical chassis specifications of present day Formula 1 cars?

Make a simulation with GP2 cars, which have +600hp, taking account the aero differences and run a comparison with the current F1's. If your calculations were right they wont be too far a part, yet lap times arent even close.

If you take for example monza where you have less aero influence

f1 qualy lap ON RAIN was: 1.34.xx
f2 qualy lap on dry was: 1.30.xx

Race
F1: 1.23.xx
F2: 1.34.xx

There are a lot of other performance differences, but those wouldnt account for more than 5s a lap.

[j.yank]

Well, I don't know what you were doing, but I just setup a couple of cars to dial in a bit of correlation from min/max/cornering speeds at Austin, once I was fairly repeatable I then selected one car, dialled down 300bhp, and it cost it 4 SECONDS PER LAP.
Error between keyboard and chair?

What cars you had setup? Open wheeler with typical chassis specifications of present day Formula 1 cars?

Make a simulation with GP2 cars, which have +600hp, taking account the aero differences and run a comparison with the current F1's. If your calculations were right they wont be to far a part, yet lap times arent even close.

If you take for example monza where you have less aero influence

f1 qualy lap ON RAIN was: 1.34.xx
f2 qualy lap on dry was: 1.30.xx

Race
F1: 1.23.xx
F2: 1.34.xx

There are a lot of other performance differences, but those wouldnt account for more than 5s a lap.

From what I see in these simulations (I don't make any calculations by myself) the aero differences, tire width, tire friction, sensitivity, etc have huge impact on the lap time. In this way you cannot compare real cars - you don't know so many very important parameters. That's why this is useless to compare GP2 and F1 results on any circuit, or even f1 with another f1 car. BTW, the simulations show huge drop in max speeds when you drop the power, but not such huge drop in the lap times. Interestingly, in the last qualification the speed trap of Alonso is almost equal to the speed trap of Verstappen, however his lap time is almost one second slower. It will be good to hear how can be explained this only relaying that Honda engine has huge deficit that explains their current lack of performance? If they had such huge deficit it will be shown on speed traps, too.

[DFX]

From what I see in these simulations (I don't make any calculations by myself) the aero differences, tire width, tire friction, sensitivity, etc have huge impact on the lap time. In this way you cannot compare real cars - you don't know so many very important parameters. That's why this is useless to compare GP2 and F1 results on any circuit, or even f1 with another f1 car. BTW, the simulations show huge drop in max speeds when you drop the power, but not such huge drop in the lap times. Interestingly, in the last qualification the speed trap of Alonso is almost equal to the speed trap of Verstappen, however his lap time is almost one second slower. It will be good to hear how can be explained this only relaying that Honda engine has huge deficit that explains their current lack of performance? If they had such huge deficit it will be shown on speed traps, too.

My point is that 300hp is way to much to be accounted only for the times you have shown.
Dont know how this simulation take into account electric energy deployment aswell, which is a important factor.

Alonso didnt participate in q2 or q3 so it's not a fair comparison. Compare their times in q1, he made a purple second sector. Also, in the race you can see how big was the gap in the main straight, not even being able to cope with Sauber's Ferrari 2016 PU.

[PhillipM]

Either way, they still show, over a couple of configs (to simulate lower end as well as higher end cars), that in Austin, using the same software you use, that the laptime deficit for -300bhp is around 4+ seconds per lap. Not less than a second. Even a basic rule of thumb check should be telling you that running at 2/3rds of the horsepower of everyone else is not going to result in 99% of the laptime.
In reality it'd be even more of a difference than that because you'd struggle to keep on top of your tyres and would have to compromise your setup even further to generate tyre scrub/drag and other tricks to keep them hot.

Interestingly it does indicate that Honda are still about 80bhp down in race trim compared to Vettel in full power mode - but then we didn't really get to see what it could do at Austin given Alonso's retirement, and I had to try to pick laps where there was similar traffic so it's not very accurate, maybe I'll try some from Mexico later, but it's very simple/basic software so it's probably not worth trying.

[vinuneuro]

Interestingly, in the last qualification the speed trap of Alonso is almost equal to the speed trap of Verstappen, however his lap time is almost one second slower. It will be good to hear how can be explained this only relaying that Honda engine has huge deficit that explains their current lack of performance? If they had such huge deficit it will be shown on speed traps, too.

How much downforce they run plays a significant role in speed traps. It is almost impossible to judge power differences based on top speeds alone because of this.

If McLaren designed the car with similar aero DF vs. drag philosophy as other teams and setup the car with similar wing levels during weekends, they would produce even faster laptimes in quali but would be sitting ducks (more than they already are) on the straights on Sunday. Lack of power compromises many things.

[PlatinumZealot]

So Honda scraped through by not increasing the turbo speed like everyone else. Regardless of what they say that must have been at the expense of quite a bit of performance and I think it indicates that they absolutely dread running the turbo faster due to resonances or instability issues.

Going back to the topic of potential bearing misalignment due to housing thermal deformation - in large industrial turbomachinery it is common to perform a catenary curve optimisation by intentionally misaligning the bearings so that they align under thermal and gravitational loads (rotors are so heavy that they distort the bearings significantly).
I haven't seen this applied to automotive turbochargers but maybe it is something worth exploring in F1 to alleviate thermal effects.

They can easily change the compressor wheel to a bigger one what was it am 8 percent increase in speed was required? They can use an 8 percent broader wheel. Yeah it will be a little heavier but less speed is safer at this point.

If they change the turbo they get a penalty - no?

Yes. I was prepared for that question. I suppose the 35 place and 20 place grid drops could be construed to be a compressor wheel change!

[j.yank]

Either way, they still show, over a couple of configs (to simulate lower end as well as higher end cars), that in Austin, using the same software you use, that the laptime deficit for -300bhp is around 4+ seconds per lap. Not less than a second. Even a basic rule of thumb check should be telling you that running at 2/3rds of the horsepower of everyone else is not going to result in 99% of the laptime.
In reality it'd be even more of a difference than that because you'd struggle to keep on top of your tyres and would have to compromise your setup even further to generate tyre scrub/drag and other tricks to keep them hot.

Interestingly it does indicate that Honda are still about 80bhp down in race trim compared to Vettel in full power mode - but then we didn't really get to see what it could do at Austin given Alonso's retirement, and I had to try to pick laps where there was similar traffic so it's not very accurate, maybe I'll try some from Mexico later, but it's very simple/basic software so it's probably not worth trying.

Do not try to compare any configs simulating lower or higher end cars - just make one car profile and then apply to it different powers. On some tracks even small adjustments in the downforce, drag or tires make huge difference. That's why don't touch them - only the power. In this way you will see only the power influence. BTW, some adjustments on torque/rpm affects in big way the fuel consumption but this is another story. Just change the power factor of one configuration. However, I cannot be sure that this software is accurate, but if you look at the downforce formula you see that the ratio power/downforce is not linear . This could suggest that above some power levels the time differences cannot reach big values.

[63l8qrrfy6]

They can easily change the compressor wheel to a bigger one what was it am 8 percent increase in speed was required? They can use an 8 percent broader wheel. Yeah it will be a little heavier but less speed is safer at this point.

If they change the turbo they get a penalty - no?

Yes. I was prepared for that question. I suppose the 35 place and 20 place grid drops could be construed to be a compressor wheel change!

Well I suppose they used about 1 turbo per race anyway so why not

Increasing wheel trim increases mass as well as the CG overhang (distance to first bearing). Secondly, the allowable unbalance for a given balance quality grade is proportional to rotor mass, so unbalance will increase too.

All these give you a lower frequency forward whirl mode, higher unbalance and a fair amount of extra inertia to accelerate.

And I'm not too sure how compressor efficiency varies with wheel trim, maybe GG can elucidate .

[PhillipM]

Do not try to compare any configs simulating lower or higher end cars - just make one car profile and then apply to it different powers. On some tracks even small adjustments in the downforce, drag or tires make huge difference. That's why don't touch them - only the power. In this way you will see only the power influence. BTW, some adjustments on torque/rpm affects in big way the fuel consumption but this is another story. Just change the power factor of one configuration. However, I cannot be sure that this software is accurate, but if you look at the downforce formula you see that the ratio power/downforce is not linear . This could suggest that above some power levels the time differences cannot reach big values.

You misunderstand, the configs for high/low end cars was to dial the settings into something that would roughly approximate the results properly (because it's very basic software that doesn't let you set even simple things up correctly) - all comparisons after that were made with the same baseline car. It's still 4+ seconds per lap.

[j.yank]

Do not try to compare any configs simulating lower or higher end cars - just make one car profile and then apply to it different powers. On some tracks even small adjustments in the downforce, drag or tires make huge difference. That's why don't touch them - only the power. In this way you will see only the power influence. BTW, some adjustments on torque/rpm affects in big way the fuel consumption but this is another story. Just change the power factor of one configuration. However, I cannot be sure that this software is accurate, but if you look at the downforce formula you see that the ratio power/downforce is not linear . This could suggest that above some power levels the time differences cannot reach big values.

You misunderstand, the configs for high/low end cars was to dial the settings into something that would roughly approximate the results properly (because it's very basic software that doesn't let you set even simple things up correctly) - all comparisons after that were made with the same baseline car. It's still 4+ seconds per lap.

Try this - mass: 735; Drag: 1.05; Downforce: 5.55; Front: 1.390; Air: 1.2; Tire Radius: 0.405; Longitudinal Friction: 1.8 @ 735; Lateral Friction @ 735: 1.8; Sensitivity: 0.001

[godlameroso]

If they change the turbo they get a penalty - no?

Yes. I was prepared for that question. I suppose the 35 place and 20 place grid drops could be construed to be a compressor wheel change!

Well I suppose they used about 1 turbo per race anyway so why not

Increasing wheel trim increases mass as well as the CG overhang (distance to first bearing). Secondly, the allowable unbalance for a given balance quality grade is proportional to rotor mass, so unbalance will increase too.

All these give you a lower frequency forward whirl mode, higher unbalance and a fair amount of extra inertia to accelerate.

And I'm not too sure how compressor efficiency varies with wheel trim, maybe GG can elucidate .

Both turbine and compressor can be kept relatively compact with the right wheels. The number and shape of blades along with inducer and exducer combinations along with their taper.

Single stage turbines have been done to death, I wouldn't be surprised if the answer is in some old dusty book TC has stashed somewhere I reckon.

I wonder how many blades they have on their turbine wheel, I suppose that information is on a need to know. Less blades means more surface area for the blades to affect airflow, however there has to be a compromise to get the best surface area vs best pressure ratio vs best response vs best harmonics and a host of other technical niggles you're probably more familiar with. My guess is 6-7 blades, maybe a 74-88 trim something wild like that.