1mm packet of photons

Lousy lip "sync".

Gee, I was so into the subject that I didn't notice. Why would you be interested in his lips? Mikek

Do you not find it rather disconcerting when the sync is so bad? Or does everybody's lips move before the corresponding sound?

That effect is always present to some degree in real life. It depends how far from the speaker you are.

Sylvia.

Why do you watch a youtube copy and not the original?

Regards, Mikko

I don't think it really takes a trillion frames per second. It's a stroboscopic thing, like a sampling oscilloscope. I think they use a streak tube and some sort of shutter to scan the other axis. I didn't watch the entire video, but I didn't see an explanation of how it's actually done. I suspect it takes a while to build up one of those videos.

I did a quick pan and didn't see it. I'll look again when I have time.

I actually have a use for a streak camera. They are super-expensive, so if this gets to production we may look to buy the tube somewhere and make the rest ourselves. That would be fun.

Anybody here have experience with streak tubes?

I don't know the term "streak tubes", but I built a "streak camera" somewhere around 1960-61 to record fast MHD events. It was a mirror rotating at ~20,000RPM writing onto a film strip, with a magnetic pickup to synchronize initiation of "event" to mirror position.

BTW, I'm puzzled by Edgerton's apple shot being labeled 1964... that happened while I was a student there... 1958-1962. I was in his "gallery" hallway nearly every day, and it was a featured photo, along with his famous milk drop shot. ...Jim Thompson

I put in for ~$100k for one of those gorgeous Hamamatsu streak cameras about 2006ish, to support my 200 THz antenna work, but didn't get it.

Even a mechanical streak camera can do some pretty cool stuff: 100k rpm mirror, 10 metre distance, 10 um resolution = 50 ps time resolution.

Cheers

Phil Hobbs

Some of the NMR solids probes spin a sample at 8 KHz, which is half a million RPM.

LLNL used to use spinning-mirror drum cameras for nuclear shots. They had a bunch of air-motor cameras spinning asynchronously, and they just waited until all the phase angles coincided, then they fired a shot.

Streak tubes are a huge win if you can get one. Even a Tek 7104 tube can do 10 cm in 300 ps, so even a 500-um beam width translates into 15 ps resolution as a streak camera.

Cheers

Phil Hobbs

I worked with the hamamatsu streak camera, C5680 series with these modules:

blanking unit C5680

sweep unit M5675 (also M5676 sweep unit)

streak camera C5680

digital camera C4742-95

cheers, Jamie

Sounds expensive! The streak camera will dominate the cost of our proposed system.

Hi,

The mechanical mirror wouldn't have as much gain as a streak camera does , maybe it is better to spin a linear APD array?

cheers, Jamie

Assuming the optical resolution is better than the slit width (which requires a beam of a certain size, eventually limiting the mirror RPM), the time resolution is

delta_T = (L_s)/(2*Omega_m * L_L)

where Omega_m is the mirror rotation speed in rad/s, L_s is the slit width L_L is the optical lever arm length

The factor of 2 comes from the law of reflection (the reflection rotates twice as fast as the mirror).

Now 100k rpm is 1.05*10**4 rad/s, so with a 10-m lever and a 10-um slit, you get

delta_T = 1E-5/(2.1e4*10) = 48 ps.

This is of course rather unrealistic, because getting that resolution from 10 m away requires a beam width on the mirror of the order of 50 cm. You could do it with a 100 um slit and a 5-cm beam, which would get you ~500 ps resolution--not bad for an all-mechanical contraption.

Moving the detector doesn't help you because it's the relative speed of the focal spot and the detector that matters. With 10k rad/s and 10 m, the spot crosses the detector at 200 km/s, which is pretty hard to get with a solid object!

Cheers

Phil Hobbs