Hacker News Reader

Comment by jsingleton | Hacker News Reader

Many laptops have these sensors, often as part of a hard drive protection system. If it detects 9.8g suddenly going to 0g then it parks the read head as it's probably just been dropped. I'm not sure if they're still present in this era of SSDs. Maybe if they weren't part of the disk itself.

They were used by the Quake-Catcher Network project from Stanford. Although from the OS.

  ...seismic network by utilizing sensors in and attached to internet-connected computers.

https://github.com/carlgt1/qcn

Edit: Yes, I meant ms⁻² not g. Been a while since I've done any physics. 9.8g is probably more like what it would sense when hitting the floor, which is a bit too late to do anything useful. At least I didn't claim it to be 6.54×10⁻¹⁰ Nm²kg⁻² (bonus points if you get this).

View on HN

> If it detects 9.8g suddenly going to 0g then it parks the read head as it's probably just been dropped.

9.8g? Is your office inside a centrifuge? :-P

To clarify what delazeur meant:

g already means "9.8 m/s²"

Thus, normal gravity is 1g and free fall is 0g. 9.8g would mean 9.8 times the gravity at Earth surface, that is "96 m/s²". You would feel almost ten times as heavy. For reference:

Jupiter "surface" gravity is 25.9 m/s²

Sun "surface" gravity is 274 m/s²

Assuming force being applied along the axis of the spine, and that it continues, you would experience tunnel vision, lightheadedness, loss of consciousness, and eventual death. People can't really handle more than 5g on that axis without a special g suit.

Lesson: Don't try to take it standing up.

9.8g is actually a perfectly reasonable deceleration value at time of impact. So if it detects 9.8g followed by 0g, it's likely that it has bounced!

It would have to be followed by 1g, not 0g. Unless the laptop bounced in space.

The accelerometer on a laptop on a ballistic trajectory should still read 0, shouldn't it?

Ah, sorry. Yes, it will, because a typical accelerometer measures not the acceleration, but the deviation in acceleration from freefall. So it will show an acceleration of 1g when standing on a table, and 0g in freefall.

Which is obviously wrong if you think of it as measuring the acceleration of the laptop. The physical acceleration of the laptop is obviously zero when it is at rest on a table. (Acceleration is the change in speed, and the laptop has zero change in speed.)

Can you build a consistent accelerometer that does what you want?

If you think about the accelerometer as a little micromechanical arm which is being deflected relative to the laptop by the force of gravity, it makes a lot more sense. When they're falling together the deflection is 0.

> Can you build a consistent accelerometer that does what you want?

I gave it some thought, and no, I don't think you can build a self-contained one. But if you allow external communications it becomes very doable (see: GPS system). Also, if you allow another input that tells the device it's orientation wrt. the gravitational field, say a magnetic field sensor, it becomes easy.

If your accelerometer measures what Wikipedia calls "coordinate acceleration" it's something observer relative so I expect (waves hands) "problems" will arise with measuring in relation to your chosen reference frame at relativistic speeds.

Maybe supposed to be 9.8 m/s^2 ?

I looked it up prior to my post, and at least the T450 has an accelerometer that's separate from the disk. agumonkey pointed out that noone has reversed the Windows driver yet, though, so no Linux support (and certainly no GRUB2 support :>)

Reply on news.ycombinator.com