Hacker News: zacka

New comment by zacka in "Moving fast in hardware: lessons from lab to $100M ARR"

zacka — Tue, 07 Apr 2026 16:08:23 +0000

automotive contracts are typically on 6 year cycles, so tech gets designed into a new car and it's locked in until the next vehicle generation (5-7 years depending on the automaker). year to year sales can fluctuate but are fairly predictable.

New comment by zacka in "Legendary Bose Magic Carpet Suspension Is Finally Going Global"

zacka — Sun, 04 May 2025 19:48:37 +0000

To create enough force to lift the car using linear motors requires massive copper coils and rare earth magnets, which is heavy and expensive. It’s also wasteful, because in real world drive cycles peak force is only needed for short periods of time and at relatively low linear velocities. Further, linear motors require significant packaging space which requires the chassis be designed around it. Instead, we use a much smaller motor and a high “gear ratio” using hydraulics to achieve high forces via mechanical gain. This reduces raw material cost by a factor of >10x while achieving similar performance.

New comment by zacka in "Legendary Bose Magic Carpet Suspension Is Finally Going Global"

zacka — Sun, 04 May 2025 16:02:52 +0000

Yes correct, we use electric motors to directly control an efficient compact pump to bidirectionally pump fluid, or use back electromotive force to dampen motion (and capture energy while doing so). This eliminates the need for controlled valves, but requires advanced controls to mitigate system dynamics (hydraulic losses, etc).

For ride control, machine vision helps a bit, but it’s limited due to the need to determine high fidelity road z-measurements which requires very high resolution and clarity at speed, and no visual occlusion. It helps with simple problems like bump vs hill (which is not trivial to predict just with accelerometers), but we’ve figured out how to get most of the benefit of machine vision with just accelerometers and fast actuators. GM was able to use vision to improve pothole performance, but it’s still a semi-active system that only provides fast adjustable stiffness (damping).

For pre-crash, the system uses cameras to detect a collision-path vehicle and determine an optimized impact zone for that vehicle. For example, the frame or bumper has more structural rigidity which may be desirable in certain crash scenarios.

New comment by zacka in "Legendary Bose Magic Carpet Suspension Is Finally Going Global"

zacka — Sun, 04 May 2025 09:51:05 +0000

ClearMotion founder here! Wanted to share some insights that might answer questions from other commenters and clarify what we've developed.

The ClearMotion1 system is a major leap above all tech currently on the market, with transmitted vibration reduced about 80% versus top market technologies. Here's a video comparing it on production NIO cars against luxury vehicles using semi-active (or slow active) systems others mentioned - sort of like comparing a microcontroller to a NVIDIA H100. http://bit.ly/44TDtgl

This matters especially for autonomous vehicles, where the whole point is to give people back time, and preventing motion sickness while working/reading is essential.

Our tech stack:

- Electro-hydraulic actuators that both push/pull actuate and dampen within a few milliseconds, using electric motors (not solenoid valves or special fluids). We use integrated hydraulics as a mechanical gain lever

- Predictive control software that anticipates vehicle, driver, and road dynamics

- "Infinite preview" control using crowdsourced road data with <3cm localization precision

- Software-enabled features including pre-crash posture mitigation and tire grip technology

The combination creates a "software-defined" chassis, similar to how electric power steering enabled today's driver assist features.

Our Bose acquisition was to acquire specific control software and engineering talent, but most of our IP and our production hardware/software was developed in house.

There were a few questions about durability— our system has passed 5 years of testing across millions of miles - a requirement from all of our customers like NIO and Porsche. It’s also a reason why it’s so difficult to succeed as a startup in automotive, but once you’re in, you’re locked in long-term.