Honda Power Unit Hardware & Software

[wuzak]

3 on the engine, and they'll use the last 4 on other components. Probably a new MGU-H, different ERS, maybe new controls.

Haven't they used 2 previously?

[Cold Fussion]

IIRC they used 2 for the Canadian GP and 3 for Belgium.

[alexx_88]

This token system only increases costs when you are trying to catch up. I'm pretty sure they work on different solutions for each module in parallel, as they have only one shot of getting it right.

Imo, from the limited data we have, Honda should hire engineers from Ferrari/Merc to at least get level on this year's solutions. I am very worried by what Arai said, that they want their engineers to find the solutions themselves, not bring in outside help. Seems like the perfect setup to always try to catch up. Ferrari were much more down-to-earth last year and they recovered well over the winter.

[Brian Coat]

"Honda should hire engineers from Ferrari/Merc"

Interesting.

I wonder how this would match Honda's company aims and philosophy, in general and with respect to this F1 project?

[tuj]

Does Honda have the quali mode yet?

[PlatinumZealot]

This token system only increases costs when you are trying to catch up. I'm pretty sure they work on different solutions for each module in parallel, as they have only one shot of getting it right.

Imo, from the limited data we have, Honda should hire engineers from Ferrari/Merc to at least get level on this year's solutions. I am very worried by what Arai said, that they want their engineers to find the solutions themselves, not bring in outside help. Seems like the perfect setup to always try to catch up. Ferrari were much more down-to-earth last year and they recovered well over the winter.

Easier said than done. These engines are very multidisciplinary so you cannot just hire an "engine" guy like in the old days. You are going for a specialist ina given aspect. This of course means you have to hire a guy to find out what is wrong then hire another guy to work on that problem. And that specialist might not be available. I mean is honda going going to hire a specialist for every single problem? I agree with them. Hold off on hireing people for now.. Learn the problems as much as you an first then qhen you hire the specialist you can work to him closer to his level and learn even more. Pretty standard way of doing things..

[damager21]

Honda would definitely not want to hire any from outside because the same person can then potentially join Mercedes, Ferrari or Renault and give away all the information and knowledge about Honda engine.

Honda believes that their PU design is revolutionary and hence does not want to take help from anyone including McLaren. There is nothing right or wrong about it, its all about whether Honda has the capability in-house to pull it off.

I think this is also the reason why Renault has been shying away from taking any help from Red Bull. Imagine Red Bull having all info about Renault engine and moving to Mercedes next year with that info.

[godlameroso]

A lot of what makes the power unit competitive these days is centered largely around electronic controls. I think what Honda is doing is being way too methodical for McLaren's liking. However it truly is the only way to win, what I mean is, there are so many parameters that can be exploited that unless you have in-house knowledge of all the electronics involved, you'll never so much as scratch the surface. Mercedes and Ferrari are the only teams that have the capacity to currently do this, and in terms of amount of control of technology, Mercedes is much larger than Ferrari, and so is Honda. McLaren makes chassis, that's what they're good at. Honda makes engines and develops technology, so they can basically manufacture every circuit, every blade on every turbine, the casting in every block. Always has the cutting edge in computer processing for simulations, and modeling.

I don't think McLaren's facilities really even compare to what Honda has, with the resources at hand they should definitely be able to develop something worthy. They just have to be very methodical, in a race to be the fastest they certainly take their time.

[dren]

Arai comments on the Honda PU's output: http://www.autosport.com/news/report.php/id/120561

Interesting he says they are 25hp clear of Renault but 40-50hp down on Mercedes. He also says their ICE isn't as good as Ferrari. He also stated they are staying with the same tight packaging for 2016.

He must be comparing peak numbers because it is evident the Honda PU cannot sustain that output over a full lap. Even during qualifying they are left in the dust.

[Thunder]

It's stated in the Article that Arai is only talking about ICE Numbers. And it doesn't sound too far off. Renault is believed to have a very good ERS, Honda's is (mainly the MGU-H part) clearly now even remotely where it needs to be. That helped Renault enormously in Spa.

[godlameroso]

This is that same methodicalness I'm talking about. They wasted over half the season sorting out the ICE, which still isn't as good as the top engines. But what about the ERS? They're methodically going over each component and until they finish that one they stay with it, only the PU's are interrelated systems so focusing on one component in isolation doesn't work, or might have unintended consequences not taking other components into consideration.

[ME4ME]

Arai comments on the Honda PU's output: http://www.autosport.com/news/report.php/id/120561

Interesting he says they are 25hp clear of Renault but 40-50hp down on Mercedes. He also says their ICE isn't as good as Ferrari. He also stated they are staying with the same tight packaging for 2016.

He must be comparing peak numbers because it is evident the Honda PU cannot sustain that output over a full lap. Even during qualifying they are left in the dust.

That also gives an inside on Renault VS Mercedes than. 75hp deficit for the ICE. The ERS on the other hand I believe are pretty even, going by Red Bull's performance at SPA especially.

[henry]

turbo speed transitions ie 'time constant' can't be better than around 100 millisec ?
there's a lot of inertial load
(I once did this sort of design calculation for a living)

the time constant of the combined ICE and permanently slaved mgu-k will be maybe 25 millisec ?
as with N/A F1 the shifting phases will be managed by modulation of ignition, fuelling, and throttle plates
nominally the complete shift takes less than 50 millisec

so managed turbo throttling etc will be used to help the turbo behaviour in the shifting phases ?

Rotrex provide inertia values for their compressors. http://www.rotrex.com/Home/Technology/P ... nical_Data

The C38 series seems to have mass flow and pressure ratios in the right ballpark. Inertia is around 9 kg-m2
Max speed is 90000 rpm so they are a little slow for F1 but should be OK given the other assumptions that need to be made.

For MGU-H power we need to guess a power, and hence, torque rating. I will assume 60Kw. Mid way between continuous power generation, 40Kw and continuous supercharger drive, 80Kw.

If the speed is 110,000 rpm we get a torque of 5.2Nm

Assuming the motor is constant torque this gives an acceleration rate of 580 rad/s2. (5500 rpm/s)

This is just the supercharger ( and its drive). The turbine and MGU-H will potentially increase the inertia but this may be compensated by overdriving the MGU-H.

This seems relatively slow to me. The regulations say that the max. delay from torque demand to torque output is 50ms.

If this is the case some control strategies will be needed for drivability.
Keeping the turbo spinning seems good but then what target speed should it spin at?
Alternatively the MGU-K could fill in the delays but if this is open loop how much fill in should be applied?
Or maybe a combination.

Honda seem to have this sorted while Renault seemed to have real problems in Australia, perhaps because of insufficient time to tune whatever strategy they use.

[gruntguru]

Henry. On turbo acceleration and anti-lag.
- You can bet the F1 compressor wheel has a significantly lower inertia than most commercially available units.
- Turbine will have higher inertia than the compressor.
- Over the normal operating range it would be safe to assume the MGUH limit in "motor" mode will be constant power (120 kW) - not constant torque.
- The turbine will provide as much (or more) power to accelerate the compressor than the MGUH in most situations - even transients
- MAP has a fairly close correlation to compressor RPM so they probably run the shaft fast enough to ensure a target minimum MAP is available instantly when the throttle is advanced.
- Although MAP is probably 3.0 bar or more at max power, the mixture at that pressure is quite lean. This means the crankshaft power will not reduce significantly at lower boosts - say 2.5 bar abs (there is still sufficient airflow to combust the entire 100kg/hr fuel limit). What I am saying is the off-throttle turbo speed required for adequate anti-lag is probably significantly lower than max power turbo rpm (100k+).

[Abarth]

Henry. On turbo acceleration and anti-lag.
- You can bet the F1 compressor wheel has a significantly lower inertia than most commercially available units.
- Turbine will have higher inertia than the compressor.
- Over the normal operating range it would be safe to assume the MGUH limit in "motor" mode will be constant power (120 kW) - not constant torque.
[...]

I recall rosberg which had no MGU-H action anymore in the last race of 2014. He said he had to start accelerating almost before the corner and estimated a 2s lag.
So I believe MGU-H spooling is pretty important for driveability.