Theoretical Discussion: F1 2026 Variable Compression & Transient Torque Spikes (MBT vs KLSA)

Hi everyone, and thanks to Andre for the recent webinar. It sparked a specific question about the upcoming 2026 F1 Power Unit Regulations, and I'd love the community's take.

The Scenario: The 2026 rules mandate a geometric compression ratio of 16:1 (measured at ambient). However, due to the extreme thermal expansion of the materials at race temperature (~105°C+), the effective running compression ratio reportedly shifts closer to 18:1.

The Question: In a high-boost hybrid application, does this dynamic shift from 16:1 to 18:1 create a "blind spot" between the mapped MBT (Maximum Brake Torque) and the actual Knock Threshold?

Specifically, as the engine creates that unmapped efficiency gain (expanding to 18:1), would we expect to see a Transient Torque Spike at the rear wheels before the knock control strategy has time to pull timing/KLSA?

I am investigating this as a potential cause for the rear-axle instability and ride-height variance we are seeing in early 2026 simulations.

I'd appreciate any thoughts on the thermodynamics of this "Transient Torque" phenomenon.

Thanks,

Timothy Harmon, CISSP

I think F1 engines will not be knock limited. They use fuels with sufficient octane properties to avoid knock (even at 18:1 and high boost). I would be very surprised to find knock control in the noisy environment of an F1 engine. The engines are warmed up before starting, so they are likely at high compression from the moment they fire up.

If they are knock limited, the ignition mapping has taken that into account.

I would bet that the integration of additional electric motor deployment (torque) is what they are tuning now.

David, excellent insight regarding the fuel chemistry. You’re likely right that the specialized fuel blend raises the knock threshold enough to mask the combustion instability of the 18:1 shift.

Your point about the Electric Motor Deployment is the 'Eureka' moment. If the ICE output is thermally consistent (as you suggest), but the 350kW MGU-K map is fighting the chassis dynamics during that 50/50 blend phase, we end up with the same symptom: Unmapped Torque Variance.

My validator calculates the Vertical Energy (>100J) relative to the total Longitudinal G. So, whether the torque spike is Thermal (ICE) or Digital (MGU-K), the tool flags the resulting platform instability.

Ideally, we want to know which one caused the squat. Do you think the current MGU-K maps have enough resolution to smooth out those transitions, or is that where the 'Integration' battle is happening?

I would be wondering if the energy deployment hit some hard limit (like the 350kW, max battery battery draw, or maybe even a thermal limit), and instead of limiting power deployment, they cut it. So the effect is like a hard-cut revlimiter, torque on / torque off trying to maintain an average maximum value.

I bet it's not a map resolution, but more a deployment strategy / algorithm. I worked on an electric race vehicle, and one of the first things I did to control traction, was in implement a rate limit to how fast a torque request could ramp up.

I bet the MGU-K can change it's torque faster then the ICE as they try to balance the torque between the two sources.

David, that ‘Hard-Cut’ visualization (Torque On / Torque Off) perfectly aligns with the ‘Unpredictability’ drivers are reporting in the medium-speed corners.

If the MGU-K torque response (ms) is indeed faster than the ICE compensation (100ms+), hitting that 350kW hard limit would create a step-change in longitudinal force.

From a vehicle dynamics standpoint, that sudden ‘Torque Vacuum’ would instantly unload the anti-squat geometry, causing the rear ride height to overshoot (heave) just as the driver is trying to stabilize the platform.

I suspect you nailed it—it’s not just a map resolution issue; it’s a Rate of Change (Jerk) issue at the deployment limit. My validation logic watches for the resulting Vertical Energy spike (>100J), but catching the specific ‘Hard Cut’ signature in the torque request stream sounds like the upstream fix.

Brilliant insight on the ‘Rate Limit’ implementation. That sounds like the missing control law.

But I'm sure it's not that simple... Let's see how they get on this week.

I'm late to this party and it appears there3's nothing more for me to add. I will say that to David's point on the knock limitation, in a tech video I watched when the whole variable comp issue came out, that it was stated the engine weren't knock limited even at 18:1. Admittedly this was not from inside one of the PU manufacturers so you always need to take these sort of statements with a grain of salt.