Mercedes Power Unit Hardware & Software

[turbof1]

This discussion reinforced my opinion that in F1 the ECU limp-home and shut-down modes to avoid catastrophic engine failure is not used.

Only when things really become terminal. Above all, it is still a competition and going into limp home/shut down mode means forfeiting the race. I think they'd much more prefer the risk on damage and see if they can correct the issue, which is what Mercedes did. They have confirmed there is damage (which I think will likely lead to a PU change), but the trade off is that they won the race.

But yes, there are fail safes in place. I think this happened to Renault a couple of years ago on Spa.

If lambda sensor is broken there will not be signals to ECU

Note that having an issue on a sensor does not immediately mean it is outright broken. A faulty sensor could just as well send out wrong measurements to the ECU. This is what I definitely had on my car because the car not did go into any safe mode; the check engine light just flickered on my dashboard. The lambda sensor still was active, but was faulty (the engine had a correct combustion) and both tried to change combustion (which would be a wrong correction) and send an indication to my dashboard.

It's difficult to make any reasonable guess on how this all works out in a F1 PU, which is loaded up with dozens and dozens of sensors. It takes a whole department back at the home base of Mercedes just to keep an eye out on telemetry. So definitely it is vastly complicated. And let's not forget, it doesn't have to be a sensor issue.

[saviour stivala]

This discussion reinforced my opinion that in F1 the ECU limp-home and shut-down modes to avoid catastrophic engine failure is not used.

Only when things really become terminal. Above all, it is still a competition and going into limp home/shut down mode means forfeiting the race. I think they'd much more prefer the risk on damage and see if they can correct the issue, which is what Mercedes did. They have confirmed there is damage (which I think will likely lead to a PU change), but the trade off is that they won the race.

But yes, there are fail safes in place. I think this happened to Renault a couple of years ago on Spa.

If lambda sensor is broken there will not be signals to ECU

Note that having an issue on a sensor does not immediately mean it is outright broken. A faulty sensor could just as well send out wrong measurements to the ECU. This is what I definitely had on my car because the car not did go into any safe mode; the check engine light just flickered on my dashboard. The lambda sensor still was active, but was faulty (the engine had a correct combustion) and both tried to change combustion (which would be a wrong correction) and send an indication to my dashboard.

It's difficult to make any reasonable guess on how this all works out in a F1 PU, which is loaded up with dozens and dozens of sensors. It takes a whole department back at the home base of Mercedes just to keep an eye out on telemetry. So definitely it is vastly complicated. And let's not forget, it doesn't have to be a sensor issue.

Yes, and as I just told henry, "and that is what Mercedes did".

[turbof1]

This discussion reinforced my opinion that in F1 the ECU limp-home and shut-down modes to avoid catastrophic engine failure is not used.

Only when things really become terminal. Above all, it is still a competition and going into limp home/shut down mode means forfeiting the race. I think they'd much more prefer the risk on damage and see if they can correct the issue, which is what Mercedes did. They have confirmed there is damage (which I think will likely lead to a PU change), but the trade off is that they won the race.

But yes, there are fail safes in place. I think this happened to Renault a couple of years ago on Spa.

If lambda sensor is broken there will not be signals to ECU

Note that having an issue on a sensor does not immediately mean it is outright broken. A faulty sensor could just as well send out wrong measurements to the ECU. This is what I definitely had on my car because the car not did go into any safe mode; the check engine light just flickered on my dashboard. The lambda sensor still was active, but was faulty (the engine had a correct combustion) and both tried to change combustion (which would be a wrong correction) and send an indication to my dashboard.

It's difficult to make any reasonable guess on how this all works out in a F1 PU, which is loaded up with dozens and dozens of sensors. It takes a whole department back at the home base of Mercedes just to keep an eye out on telemetry. So definitely it is vastly complicated. And let's not forget, it doesn't have to be a sensor issue.

Yes, and as I just told henry, "and that is what Mercedes did".

Obviously. Problem is that that is no indicator of the specific issue, only that they were able to somehow manage it.

[saviour stivala]

Only when things really become terminal. Above all, it is still a competition and going into limp home/shut down mode means forfeiting the race. I think they'd much more prefer the risk on damage and see if they can correct the issue, which is what Mercedes did. They have confirmed there is damage (which I think will likely lead to a PU change), but the trade off is that they won the race.

But yes, there are fail safes in place. I think this happened to Renault a couple of years ago on Spa.


Note that having an issue on a sensor does not immediately mean it is outright broken. A faulty sensor could just as well send out wrong measurements to the ECU. This is what I definitely had on my car because the car not did go into any safe mode; the check engine light just flickered on my dashboard. The lambda sensor still was active, but was faulty (the engine had a correct combustion) and both tried to change combustion (which would be a wrong correction) and send an indication to my dashboard.

It's difficult to make any reasonable guess on how this all works out in a F1 PU, which is loaded up with dozens and dozens of sensors. It takes a whole department back at the home base of Mercedes just to keep an eye out on telemetry. So definitely it is vastly complicated. And let's not forget, it doesn't have to be a sensor issue.

Yes, and as I just told henry, "and that is what Mercedes did".

Obviously. Problem is that that is no indicator of the specific issue, only that they were able to somehow manage it.

And that besides the fact that we can only talk about what they chose to push-out. One such push-outs that didn’t really made sense to me was having to “replace Bottas engine after a “hydraulic system failure”. remember that?

[turbof1]

Yes, and as I just told henry, "and that is what Mercedes did".

Obviously. Problem is that that is no indicator of the specific issue, only that they were able to somehow manage it.

And that besides the fact that we can only talk about what they chose to push-out. One such push-outs that didn’t really made sense to me was having to “replace Bottas engine after a “hydraulic system failure”. remember that?

To be fair, a "hydraulics system failure" could almost mean anything. F1 teams have been saying that for years now to the media, to say really nothing at all.

Therefore, it's best to speculate and do conjecture (in a sensible way, mind you). Nobody is making absurd conclusions, just hypetheses.

[63l8qrrfy6]

I guess my sources were correct after all.

https://www.autosport.com/f1/news/14027 ... er-problem

"The problem in Brazil was caused by a failure in the variable inlet system that caused unstable combustion," said a Mercedes spokesperson.

[henry]

Here’s a comment from Andrew Shovlin on Autosport

"You always struggle a bit more with overheating in FP1 but other than that, the car seemed to be working OK."

The question is why? Do turned down engines run hotter than engines running at full power?

[Phil]

Isn't that comment perhaps due to time of FP1 at Abu Dhabi, e.g. at 'daytime' where as FP2, QF and the race are at twilight/night?

[henry]

Isn't that comment perhaps due to time of FP1 at Abu Dhabi, e.g. at 'daytime' where as FP2, QF and the race are at twilight/night?

There is that distinct possibility.

[PlatinumZealot]

Bottas confirmed this. He said the car is set up for the colder night (cooling. Tyres. Etc.). But FP1 is in the day.

[dren]

I guess my sources were correct after all.

https://www.autosport.com/f1/news/14027 ... er-problem

"The problem in Brazil was caused by a failure in the variable inlet system that caused unstable combustion," said a Mercedes spokesperson.

I wonder how much of the tuned pulsing benefits the combustion. It must rely on it if a failure in the system caused it to be unstable.

[gruntguru]

I guess my sources were correct after all.

https://www.autosport.com/f1/news/14027 ... er-problem

"The problem in Brazil was caused by a failure in the variable inlet system that caused unstable combustion," said a Mercedes spokesperson.

I wonder how much of the tuned pulsing benefits the combustion. It must rely on it if a failure in the system caused it to be unstable.

If the runner length is not optimal for the current engine speed, the cylinder fill will be reduced resulting in a richer AFR. Undesired AFR will have an effect on combustion stability.

[dren]

I guess my sources were correct after all.

https://www.autosport.com/f1/news/14027 ... er-problem

I wonder how much of the tuned pulsing benefits the combustion. It must rely on it if a failure in the system caused it to be unstable.

If the runner length is not optimal for the current engine speed, the cylinder fill will be reduced resulting in a richer AFR. Undesired AFR will have an effect on combustion stability.

Right, but will it have such a big impact with the engine being FI?

[gruntguru]

I wonder how much of the tuned pulsing benefits the combustion. It must rely on it if a failure in the system caused it to be unstable.

If the runner length is not optimal for the current engine speed, the cylinder fill will be reduced resulting in a richer AFR. Undesired AFR will have an effect on combustion stability.

Right, but will it have such a big impact with the engine being FI?

Potentially yes. We need to remember that the combustion conditions in these engines are at the very limit on a number of parameters. I would guess things like peak cylinder pressure, rate of pressure rise, detonation, heat input to various components (exhaust valve, piston crown). A richer mixture (remember these things are running ultra-lean and a richer mixture will increase combustion temp, flame speed, detonation etc) could easily tip one or more of these parameters over the edge, necessitating compromises to operating parameters (like ignition timing) to protect the engine.

[PlatinumZealot]

I guess my sources were correct after all.

https://www.autosport.com/f1/news/14027 ... er-problem

I wonder how much of the tuned pulsing benefits the combustion. It must rely on it if a failure in the system caused it to be unstable.

If the runner length is not optimal for the current engine speed, the cylinder fill will be reduced resulting in a richer AFR. Undesired AFR will have an effect on combustion stability.

But one would think a richer burn would be more stable?

If I had to hazard a guess, I would say the VLIM influences not just the AFR but the aerodynamics inside the cylinders. I would love to go deeper but we don't have confirmation on the TJI layout inside the Mercedes cylinder.