Honda Power Unit Hardware & Software

[Pany]

No way. This new pu is prooving far to be reliable and better optimised and mapped with upgraded reliable mguh and turbo. They will sort out by Suzuka not before. These complex projects cant be solved magically with good ideas only, but need long development. Meanwhile competitors are improving and updating so when this honda pu will be perfect, it will still underperform significantly with respect to the other renewed pu's. This clearly means the moment in which honda will be really competitive will be maybe in 2019 or 2020 with more confidence. This is my rational opinion and Everybody must take proper decisions according logical conclusions. No signs at avery gp of a different future. Im sorry i dreamed with you all but i just admit this dream is over

[diffuser]

In my mind this years PU should have been on the test bench in May of 2016 and in the car at 1 of the Free Practices in July. That wasn't possible cause:
1 - The token rules didn't get dropped till late spring early summer 2016 (for 2017).
2 - Even if they knew in advance that the token rules would get dropped, they couldn't test the NEW PU because of them.

So instead of having a 2017 PU ready in July it was ready in December. They're still suffering the fall out of the token system. Stuck in this catch 22 that they're having problems getting out of. Far too many of there problems are being discovered on the track.There so far behind that the lead times from development to track is just too short.

The reliability of next years PU will depend on the type of changes they need to make. Tweaks are easier to work out than reworks.

[bauc]

Will Alonso use spec 3 in Silverstone?

If he does, they will have to take penalty for new MGU-H (the last one failed in AUS)

[SameSame]

Will Alonso use spec 3 in Silverstone?

If he does, they will have to take penalty for new MGU-H (the last one failed in AUS)

They are still deciding whether or not to use an old MGU-H and Turbo with the rest of the Spec 3 upgrade.

https://www.motorsport.com/f1/news/hond ... nda+engine

[hurril]

Will Alonso use spec 3 in Silverstone?

If he does, they will have to take penalty for new MGU-H (the last one failed in AUS)

They are still deciding whether or not to use an old MGU-H and Turbo with the rest of the Spec 3 upgrade.

https://www.motorsport.com/f1/news/hond ... nda+engine

I don't understand this at all. Why don't they just switch engines at a frequency that completely matches what they can produce? They aren't in the game anyway so might as well consider every grand prix a test run of sorts instead. So get spec #3 with a brand new Turbo/MGU-h package so that they cover as much ground as possible, test wise, to be closer to on-track by the end of 2017.

The current engine seems to be able to get there.

[Pany]

totally agree

[SameSame]

If he does, they will have to take penalty for new MGU-H (the last one failed in AUS)

They are still deciding whether or not to use an old MGU-H and Turbo with the rest of the Spec 3 upgrade.

https://www.motorsport.com/f1/news/honda-dilemma-engine-silverstone-alonso-vandoorne-928267/?v=2&s=1&q=Honda+engine

I don't understand this at all. Why don't they just switch engines at a frequency that completely matches what they can produce? They aren't in the game anyway so might as well consider every grand prix a test run of sorts instead. So get spec #3 with a brand new Turbo/MGU-h package so that they cover as much ground as possible, test wise, to be closer to on-track by the end of 2017.

The current engine seems to be able to get there.

That would be the most logical option, especially considering Hungary coming up. It makes one wonder if in fact they have anything new to test. This Spec 3 is the update that has been planned since pre-season testing; if that’s all they could produce in 5 months then there’s not much hope for an upgrade per GP.

[bigblue]

if that’s all they could produce in 5 months then there’s not much hope for an upgrade per GP

FYI : repeating a bunch of stuff said previously in this thread, but briefly ... there was Spec 2, there have been a whole bunch of tweaks to the MGU-H and that looks like it's still ongoing, and this Spec 3 is supposedly only part of a much bigger change which is being worked on (search out Wazari's last couple of posts), there has been one (or two ?) fuel changes, and a bunch of mapping refinements. Up to you whether you consider it 'fast enough' or not, but it isn't just Spec 3.

[SameSame]

if that’s all they could produce in 5 months then there’s not much hope for an upgrade per GP

FYI : repeating a bunch of stuff said previously in this thread, but briefly ... there was Spec 2, there have been a whole bunch of tweaks to the MGU-H and that looks like it's still ongoing, and this Spec 3 is supposedly only part of a much bigger change which is being worked on (search out Wazari's last couple of posts), there has been one (or two ?) fuel changes, and a bunch of mapping refinements. Up to you whether you consider it 'fast enough' or not, but it isn't just Spec 3.

If you bothered to read my post in context; I was specifically referring to hardware upgrades, that fall under the controlled components list, per GP.

[hurril]

They are still deciding whether or not to use an old MGU-H and Turbo with the rest of the Spec 3 upgrade.

https://www.motorsport.com/f1/news/hond ... nda+engine

I don't understand this at all. Why don't they just switch engines at a frequency that completely matches what they can produce? They aren't in the game anyway so might as well consider every grand prix a test run of sorts instead. So get spec #3 with a brand new Turbo/MGU-h package so that they cover as much ground as possible, test wise, to be closer to on-track by the end of 2017.

The current engine seems to be able to get there.

That would be the most logical option, especially considering Hungary coming up. It makes one wonder if in fact they have anything new to test. This Spec 3 is the update that has been planned since pre-season testing; if that’s all they could produce in 5 months then there’s not much hope for an upgrade per GP.

Well I (still!) hold Honda very high and have nothing but awe for what they are able to put out. But perhaps you are right in that, perhaps they are not able to produce items that are both "spec #3" and "reliable" at the same time.

[bigblue]

If you bothered to read my post in context; I was specifically referring to hardware upgrades, that fall under the controlled components list, per GP.

Sorry, wasn't obvious to me, thought you meant not much has happened in general - was just commenting that quite a lot has been going on, though admittedly without a really big performance change yet.

[SameSame]

If you bothered to read my post in context; I was specifically referring to hardware upgrades, that fall under the controlled components list, per GP.

Sorry, wasn't obvious to me, thought you meant not much has happened in general - was just commenting that quite a lot has been going on, though admittedly without a really big performance change yet.

No worries.

Is there a reason so much is spoken of the Honda PU mapping? I never hear Ferrari or Merc speaking about how much more performance they still have to unlock after an update through mapping alone. Renault just recently spoke about it, albeit being more reliability related than a completely new map. What could be the reason they take longer to get it right?

And on a side note, what language is used for ECU mapping? C?

[Craigy]

And on a side note, what language is used for ECU mapping? C?

The current ECU is a TAG-320 (standard across the field). The tool to code it is simulink, which is used with matlab. Simulink produces the C that's then compiled and run on the TAG-320.

There's a lot more to the code systems than just "here's the tool **write stuff**...here's the code" as the whole thing is built to run simulations while the cars are on track and during development cycles.
Hardware in the loop simulation doesn't use bespoke test code, it uses the same stuff the cars are always running so it has to be written generically enough to do everything.

[SameSame]

And on a side note, what language is used for ECU mapping? C?

The current ECU is a TAG-320 (standard across the field). The tool to code it is simulink, which is used with matlab. Simulink produces the C that's then compiled and run on the TAG-320.

There's a lot more to the code systems than just "here's the tool **write stuff**...here's the code" as the whole thing is built to run simulations while the cars are on track and during development cycles.
Hardware in the loop simulation doesn't use bespoke test code, it uses the same stuff the cars are always running so it has to be written generically enough to do everything.

Wow, that is extremely surprising to find out that they use Simulink. I have used Simulink with hardware in the loop before; and in my experience it is only useful for prototyping. It is not very efficient at generating C, even though it is built on it and has an option to generate raw C without uploading the code to the hardware, as it's focus is to be used as a high level programming language.

Is the ECU extremely powerful?

[Craigy]

Is the ECU extremely powerful?

That's pretty subjective.

4000MIPS is really nothing these days for most commercial CPUs including the one in the laptop I'm typing on.
http://www.mclaren.com/appliedtechnolog ... t-tag-320/

Edited to add: McLaren don't advertise the actual CPU they put in the ECU. It's 32-bit, designed for embedded systems, and I think it was from Freescale (who were bought out by NXP, and NXP were bought out by Qualcomm).

As a result it's entirely possible that the McLaren-Honda ECU is running code on a CPU from a Mercedes sponsor.

In terms of actual engine coding, it's also quite feasible that the code produced by the developers goes through optimisations before (or after) being compiled.

A friend of mine has done commercial code tuning of this nature with CFD and FEA code for various organisations, including most of the F1 paddock. It takes time to get to a level where you need someone like him though - looking for the last few percent in an otherwise optimal system. It's another area where Honda coming to F1 from "cold" and using non-F1 people will have trouble gaining traction.