Honda Power Unit Hardware & Software

[diffuser]

Wazari wrote to me back in January that they struggle with the Aramco fuel because its composition causes oil dilution issues. Because of that and some sensor & software glitches, the 626 was not reliable and Honda wasn't pushing the PU above roughly 75%. Further struggle was the management/flattening of sudden torque peaks Honda experienced with their unique harvest/deploy sequences which were the cause of the destroyed gearboxes. Also the PU by itself didn't vibrate at all up until 15k RPMs, but as the K was mounted directly to the survival cell, the whole package hit some (dis)harmonic resonance and it was difficult to cure.

As you may remember the totally straight and ultra smooth driveline of the first Lexus LS400, even a couple of degrees can cause havoc if frequencies are happen to disturb/amplify themselves in unwanted places. I would have examined the entire chassis-engine integration from a cymatic POV to see where those harmful vibrations start to form.

That's weird, because publicly he said they can't go full power until the PU mounting is redesigned.

Though January was a long time ago, that might have been a transient issue.
Seeing Stroll's Monaco crash the torque problem seems to be persistent.

I wonder if they can just simply not mount the MGU to the monocoque and only mount it to the engine. That should certainly stop transferring the vibrations. That would certainly qualify as a re-design.

Based on what we see of the AMR26B, they didn't change the mount point of the MGU-K, otherwise we would have had to make the part under the sidepods wider. It looks narrower if anything. Again this has been fixed along time ago. They were just waiting for the new PU to implement it. It was my understanding that the mounts couldn't be used on the OLD ICE. I presume the new mounts need head modifications and since they were making changes for the combustion chambers, they waited to do it all at once.

[gansaraku]

Just to be clear, these are only my personal impressions based on Honda's interviews and the publicly available technical information. I'm not claiming these are confirmed facts or that I have any insider knowledge—I'm simply trying to connect the dots.
One thing I find particularly interesting is the FIA's restriction on in-cylinder combustion sensing. Without direct cylinder pressure or combustion measurements, the ECU has to rely much more on predictive models rather than real-time combustion feedback.
If those combustion models are even slightly inaccurate, I wonder whether that could affect torque estimation and, consequently, the synchronization between the ICE, the MGU-K, engine braking and the gearbox during transient phases such as corner entry and downshifts.
I'm not suggesting that this is what caused Stroll's Monaco incident. It is simply a possible link that came to mind after reading Honda's recent comments about combustion development, reduced internal friction, revised calibration and improved drivability.
I'd be interested to hear what others think. Does this sound technically plausible, or am I overlooking something?

[diffuser]

Just to be clear, these are only my personal impressions based on Honda's interviews and the publicly available technical information. I'm not claiming these are confirmed facts or that I have any insider knowledge—I'm simply trying to connect the dots.
One thing I find particularly interesting is the FIA's restriction on in-cylinder combustion sensing. Without direct cylinder pressure or combustion measurements, the ECU has to rely much more on predictive models rather than real-time combustion feedback.
If those combustion models are even slightly inaccurate, I wonder whether that could affect torque estimation and, consequently, the synchronization between the ICE, the MGU-K, engine braking and the gearbox during transient phases such as corner entry and downshifts.
I'm not suggesting that this is what caused Stroll's Monaco incident. It is simply a possible link that came to mind after reading Honda's recent comments about combustion development, reduced internal friction, revised calibration and improved drivability.
I'd be interested to hear what others think. Does this sound technically plausible, or am I overlooking something?

"Relying much more on predictive models rather than real-time combustion feedback" means you can still push things like the ignition timing to the edge, but you have to be more conservative because you're making assumptions about what the cylinder pressure and combustion conditions actually are. That will likely mean Honda will be more conservative in the first few races with the new ICE, then gradually start pushing closer and closer to the limit as they gain confidence.

You're still allowed to run the sensors in the lab and on the dyno, so you should be able to build a very good calibration map. However, things in the real world sometimes don't behave the same way they do in the lab. If the engine starts knocking heavily, you run the risk of the ICE going boom.

I think the Stroll accident is something they had been seeing quite a lot already; it's just that there aren't many tracks where the cars end up that close to the walls. I think that when a driver pulls either paddle shifter, a signal is sent to the ICE to microsecond pause torque delivery so the transmission can synchronize the next gear and complete the shift. With all the braking, acceleration, and the much greater electrical power from the MGU-K, that pause has become much more complicated to manage. Ending a torque pause while the MGU-K is regenerating under braking means the system has to maintain a consistent braking feel for the driver. If the torque pause causes the MGU-K to reduce braking at the wrong moment, even briefly, the car can end up in the wall.

[ispano6]

The sound of the new engine is magnificent. The higher pitch signature sound of the old guard is back!

https://www.motorsportweek.com/2026/07/ ... 1-testing/