Fun with laser and PMT tube, or how much is the speed of light here?

[muvideo]

When I trying to build my first avalanche pulse generator,
someone explained me that I was independently measuring
the dielectric constant of the RG58 cable I used as charge line.

This triggered the idea to measure the time of flight of a light pulse.
I was also curious if normal leds and laser diodes could be used
as very high speeds. Probably it's a basic excercise in university
lab but I've never had the luck to see the experiment live.

The requirement was to use only equipment I had around,
the pulser was working ok, I had some smd leds to test
but the problem was what to use as detector.

The detector has to be fast, sensitive and be coupled to a scope input.
Characteristics not easy to have from common components,
but I had right in the “lab” (let me use this word   ) the righ
device for the task, this :
https://www.eevblog.com/forum/product-reviews-photos-and-discussion/photomultiplier-tube-teardown-and-speed-test/

A photomultiplier tube, boxed and almost plug and play.
Once connected to a suitable power supply (rigorously high
voltage, as per vacuum devices tradition) this glass tube
is so nice that spits out a pulse ready to be coupled with
an oscilloscope. Well, almost ready, but close enough.
Ant it's sensitive and fast, at least the risetime is fast,
the pulse width isn't that fast but it's enough for the task.

Let's do the math: the light travels 30cm per nanosecond,
so with the 6m path I have (two times bench to wall)
I have 20nS flight time.
The pulse generator risetime is less than 1.5nS, and narrow:



The photomultiplier has an output pulse width of 10nS,
this is the time it takes to discharge after being hit by
light pulse,so I should have enough time to generate two pulses
well spaced apart.
The biasing network on the tube had to be slightly modified.
The test shows that the leds are fast enough, at least
one of the orange leds I had around, a red one was
much worse.
This is a led I used for testing the pmt, the capacitor
and inductor are used to maintain the led near conduction:


This is the final setup, I used a laser instead of a led,
because it easy to collimate and be reflected back
from distance.
Under the rechargeable cell there is the avalanche
pulser that feeds the laser head, one of these:
https://www.eevblog.com/forum/product-reviews-photos-and-discussion/low-cost-ebay-red-lasers-teardown/

The small glass piece in front of the laser reflects
part of the beam to the big mirror on the side of the
photomultiplier. The rest of the beam travels
straight to a mirror 3m away that reflects it
back into the photomultiplier.
This beam is late by the 6meters it travels so the PMT
receiver will see two pulses 20nS apart.


And this is the result:



The nice thing is that the laser pulse is shorter
than the distance it travels, so I imagine it like
a light string probably less than 1m long
that travels the 6m from the laser head and receiver

Now I can answer to my curiosity:
seem that also in my “lab” the light
bounces around at the usual speed

Fabio.

[Time]

Neat.  Thanks for sharing.  I also love fast pulses!

[caroper]

Now you know the distance from your bench to the wall to 5 decimal places

[muvideo]

Now you know the distance from your bench to the wall to 5 decimal places

I prefer thinking I measured the speed of light to 1 decimal place

[ejeffrey]

Nice work.  I used to do this demo for an undergraduate physics class.  Somehow measuring the speed of light in air is always more impressive than doing the same in coax or optical fiber

I am actually a little bit surprised that the LEDs work so well, I generally think of LEDs as being pretty slow compared to laser diodes, but then I never tried.  With the laser diode you also don't really need a PMT -- a standard biased 1 mm photodiode can generally be relied upon to have 1 ns rise time into a 50 ohm load, and with a laser you can often collect enough light to see directly on your scope without any amplification.

[muvideo]

I am actually a little bit surprised that the LEDs work so well, I generally think of LEDs as being pretty slow compared to laser diodes, but then I never tried. 

The first I tested was a disappointing 0603 red led,decent risetime but bad extinction, the second one was the
orange led that was very good, the pulse shape was limited by the pmt, and even a little better than
the laser, probably because the laser intensity was a little high for the pmt.
Infact the best pulse for the led was obtained with pmt biased around 900V, while
the laser gave excellent results at 700-750V, the nominal biasing is 1kV.
I have also some smd white leds, it would be interesting to test these and compare to
plain blue ones, anyone has an idea of how fast are white led's phosphors?

With the laser diode you also don't really need a PMT -- a standard biased 1 mm photodiode can generally be relied upon to have 1 ns rise time into a 50 ohm load, and with a laser you can often collect enough light to see directly on your scope without any amplification.

This is interesting, can you explain, or point to some online source?
I have are some IR detectors from remotes, but probably in the junk box I can find something better.

edit: i found a pair of hamamatsu s1223-01, i think they are too slow, I should have some sfh203 somewhere,
but also these are specified for 5nS risetime...

Fabio.

[SeanB]

The white phosphors are pretty slow, you can expect millisecond to second response times for the various component phosphors.

[muvideo]

The white phosphors are pretty slow, you can expect millisecond to second response times for the various component phosphors.

Wow, I thought that Ce:YAG were much faster.

[SeanB]

Pretty much the same phosphors are used on CFL lamps, and a slow response reduces flicker. Remember white LED's are really designed for a human eye, and the response is just required to be as fast as that.

[muvideo]

Pretty much the same phosphors are used on CFL lamps, and a slow response reduces flicker. Remember white LED's are really designed for a human eye, and the response is just required to be as fast as that.

As far as I know CFL and LED use different phosphors. Now I'm curious,
as soon as I can I'll mount one of the white leds and report back the results,
I'll also try to use a photodiode to see if I can obtain a faster detector.

Fabio.

[ejeffrey]

This is interesting, can you explain, or point to some online source?
I have are some IR detectors from remotes, but probably in the junk box I can find something better.

Basically, the thing limiting the speed of most photodiodes is their capacitance.  The N and P doped regions are separated by a depletion region, and form a parallel plate capacitor.  That capacitor plus the characteristic impedance of your load (50 ohm if you use standard lab coax and a 50 ohm scope) gives you a low pass filter.  In reality it is considerably more complicated, but that is a good first approximation.  The capacitance depends on the area of the photodiode and also on the reverse bias voltage.  A reverse bias voltage increases the thickness of the depletion region and makes the 'plates' farther apart, decreasing the capacitance and improving the speed.

edit: i found a pair of hamamatsu s1223-01, i think they are too slow, I should have some sfh203 somewhere,

Yeah, those are 3.6x3.6 mm diodes and they have too high a capacitance (20 pF) to be used at high speed.  If you look at the S5972 it has a diameter of 0.8 mm and only a capacitance of 3 pF.

IR receivers for remote controls are photo transistors which are much, much slower.  They are cheap, simple, and give a lot of gain, but you wouldn't use them in a high performance application.