Maker Pro
Maker Pro

Fast (simple) beam break detector

I need to measure a 1cm diameter ball traveling at arround 300m/s.
Measure it to find out how fast it exactly moves.

I am considering using two laser pointers which each shine on a light
sensitive diode.
The diodes would be connected in series between the 5v bias and the
signal lead on a pc soundcard. The input lead will also be connected
to the ground lead over a high resistance resistor to pull down the
signal.

The idea is that when both diodes are lit, the signal wil be high and
a simply 44kHz wave recoding wil show a constant high level. When the
first beam is blocked, the signal goes to zero and high again when the
beam is again not obstructed. A little later it will again drop low
and go up.

I should now be able to measure the distance between the two falling
flanks and calculate the time.

Given that the ball will move through a beam in 0.03ms that is 1.4
sample (at 44kHz) where the signal will go low. That is not a lot. Can
I coult on a light sensitive diode to respond quickly enough for this
signal to be detected, or should I consider something else?
I can, if needed, place two mirrors and let the laser beam bounce
between them in such a way that the light is off the diode while the
bounding beam is blocked anywhere along its path. That could for
example give me a 10cm distance over which the light is off.

The readon I want to use a sound card and not just program my tructy
old atmel 8515 mcu, is that it will be so much more simple than having
to fiddle with displaying the measured time output from the
controller visually or by transmitting it to pc. Right now the project
is not about making a cool speed measurer, but only to be able to
measure the speed for other purposes.
 
J

Jan Panteltje

Jan 1, 1970
0
I need to measure a 1cm diameter ball traveling at arround 300m/s.
Measure it to find out how fast it exactly moves.

I am considering using two laser pointers which each shine on a light
sensitive diode.
The diodes would be connected in series between the 5v bias and the
signal lead on a pc soundcard. The input lead will also be connected
to the ground lead over a high resistance resistor to pull down the
signal.

The idea is that when both diodes are lit, the signal wil be high and
a simply 44kHz wave recoding wil show a constant high level. When the
first beam is blocked, the signal goes to zero and high again when the
beam is again not obstructed. A little later it will again drop low
and go up.

I should now be able to measure the distance between the two falling
flanks and calculate the time.

Given that the ball will move through a beam in 0.03ms that is 1.4
sample (at 44kHz) where the signal will go low. That is not a lot. Can
I coult on a light sensitive diode to respond quickly enough for this
signal to be detected, or should I consider something else?
I can, if needed, place two mirrors and let the laser beam bounce
between them in such a way that the light is off the diode while the
bounding beam is blocked anywhere along its path. That could for
example give me a 10cm distance over which the light is off.

The readon I want to use a sound card and not just program my tructy
old atmel 8515 mcu, is that it will be so much more simple than having
to fiddle with displaying the measured time output from the
controller visually or by transmitting it to pc. Right now the project
is not about making a cool speed measurer, but only to be able to
measure the speed for other purposes.
How about setting and resetting a flip-flop.
Then, when the FF is high, charge a cap via a constant current source.
Buffer and peak detect the voltage on the cap, amplifiy if needed,
and drive an old fashing analog meter 'speed'.
?
 
How about setting and resetting a flip-flop.
Then, when the FF is high, charge a cap via a constant current source.
Buffer and peak detect the voltage on the cap, amplifiy if needed,
and drive an old fashing analog meter 'speed'.

Good idea about the flip flop. I am unsure about the accuracy when
charging a capasitor that way. What kind of uncertaincy should I
expect when I am unable to calibrate the setuup with a known high
velocity.
 
J

Jan Panteltje

Jan 1, 1970
0
Good idea about the flip flop. I am unsure about the accuracy when
charging a capasitor that way. What kind of uncertaincy should I
expect when I am unable to calibrate the setuup with a known high
velocity.

Q = C.U = i.t

1% caps exist.
You can make a good current source with a couple of BJTs,
and measure it to any accuracy you like with a normal DVM.
Some opamps, and you are done.
 
J

James Arthur

Jan 1, 1970
0
Good idea about the flip flop. I am unsure about the accuracy when
charging a capasitor that way. What kind of uncertaincy should I
expect when I am unable to calibrate the setuup with a known high
velocity.

For accuracy, calibrate with a signal generator and an oscilloscope.

Cheers,
James Arthur
 
R

Rich Grise

Jan 1, 1970
0
[email protected] wrote in

How about setting and resetting a flip-flop.
Then, when the FF is high, charge a cap via a constant current source.
Buffer and peak detect the voltage on the cap, amplifiy if needed,
and drive an old fashing analog meter 'speed'.
?

Great, except the meter would be proportional to transit time, or the
reciprocal of speed. You could configure some analog multiplier in "divide"
mode, or do it with counters or, (gasp) a uC of some sort.

Or, you could draw a backwards scale on the meter, for speed; meter 0
would correspond to infinite speed, of course. ;-)

Have Fun!
Rich
 
J

Jan Panteltje

Jan 1, 1970
0
Great, except the meter would be proportional to transit time, or the
reciprocal of speed. You could configure some analog multiplier in "divide"
mode, or do it with counters or, (gasp) a uC of some sort.

Or, you could draw a backwards scale on the meter, for speed; meter 0
would correspond to infinite speed, of course. ;-)

Have Fun!
Rich

Yes, good observation, so when he knows the distance, he can calculate
and write the speed on the scale.
If you reverse the meter and bias it, you can have it move to the right for
more speed if you must ;-) but it won't be linear.
Then perhaps indeed a log opamp circuit.
As flip flops come often with 2 in a package, the second one can be used
as 'signal detector', enabling the meter and lighting a LED perhaps.
Then a reset button to prepare for the next shot.

Let's have a go:
+ +
| |
e PNP e PNP
b--- b- Q3
c | c |
| Q2 | |----
-------| ---|
!trigger 1 ------------- set | |-- |
74HC74 c | | |
!trigger 2--R7 -------- reset !Q --- R1------b NPN === | R2
| e Q1 | | |
a FF1 | | | [ ]<--- current
diode pulse former /// /// | | P1 |
k D1 | |
| C1 | ///
from reset other F |
-----------------------------------------------------------------
| charge switch current mirror
|
| peak detection and hold
------------ +
opamp1 out---
----- - |
| D2 |
|--------k diode a --- meter +
|---------------------------------------- + / |
=== | opamp2 out ------R4 --- O ---->[ ] meter bias
| C2 0 / -- - | | P3
/// / ---------------- | R3 ///
0 | | |
| | -----------------|
/// | | |
4016 | gain --->[ ] P2
CMOS switch | | gain
| ///
|
|
!trigger 1 ----------- set Q -
74HC74 L1
+ -- R5------ reset !Q --- R6-------- k LED a ---- +
| FF2 shot detected light
/ reset button
| S1
/// shot detector

R1 4k7
R2 1k
R3 10k
R4 depends on meter sensitivity, 5 V for full scale... if 100 uA then 50k
R5 10k
R6 4k7
R7 4k4

P1 10k
P2 10k ?
P3 1k

C1 depends on distance (pulse-length and current)
C2 1nF poly

D1,2 1N4148

Supply +5V (use CMOS opamps).


Something like this (photo detectors not drawn).

Drawing errors?

trigger1 low will set both FF.
Pushing S1 the reset button will reset both FF.
When FF1 is set, Q1 stops conducting, and C1 is charged by Q2 with a constant current.
Opamp1 charges C2 to the peak voltage, and holds, opamp2 buffers and has some gain if needed.
The 4016 switch discharges C2 upon pressing reset.
LED L1 indicates a shot was detected.

All values are 'guess values'.
 
Thanks bt the sbject title was something about simple ;-)
The project is not about building a speed detector, it isnot even
about measuring the velocity. That is al just small stepping stones
along the way. That is why I consider using a sound cards microphone
input to make accurate meassurements over time.

The design I consider now is a R/C-flipflop two pullup resistors and
two photo sensitive diodes and the sound card. That is the level of
simplicity I look for.
That way the flipflop goes high when first beam is broken (set) and it
goes ow when the second beam (reset) is broken.

I connect set and reset each to a diode which leads to gnd. Set and
reset are also each connected to the 5v from the sound card through a
pulup resistor.
That should do it.

Reset the FF by breaking the rear beam. The diode closes and the
pullup raises reset and FF's Q goes low.
Break the front beam, the diode closes and the pullup raises set and Q
goes high until rear beam is broken again.

The F output goes to the signal in pin on the card and I should get a
high signal while the projectile is between beams.
With a distance of 1m and a velocity of 300m/s it will be a duration
of 3.3ms or 147 samples at 44kHz
 
J

Jasen Betts

Jan 1, 1970
0
Newsgroups: sci.electronics.design
From: Jasen Betts <[email protected]>
Subject: Re: Fast (simple) beam break detector
References: <d1675c0e-32f5-4961-b01e-89cf31cdce9b@v53g2000hsa.googlegroups.com>
Organization: Dis (not Dat) Organisation
Followup-To:
X-Face: ?)Aw4rXwN5u0~$nqKj`xPz>xHCwgi^q+^?Ri*+R(&uv2=E1Q0Zk(>h!~o2ID@6{uf8s;a+M[5[U[QT7xFN%^gR"=tuJw%TXXR'Fp~W;(T"1(739R%m0Yyyv*gkGoPA.$b,D.w:z+<'"=-lVT?6{T?=R^:W5g|E2#EhjKCa+nt":4b}dU7GYB*HBxn&Td$@f%.kl^:7X8rQWd[NTc"P"u6nkisze/Q;8"9Z{peQF,w)7UjV$c|RO/mQW/NMgWfr5*$-Z%u46"/00mx-,\R'fLPe.)^

I need to measure a 1cm diameter ball traveling at arround 300m/s.
Measure it to find out how fast it exactly moves.

hmm, smaller than a paintball, faster too.
I am considering using two laser pointers which each shine on a light
sensitive diode.

light emitting diodes could be used instead of unless you need the extra
range that the lasers get you.

phototransistors have a higher output amplitude than the photo diodes.
The diodes would be connected in series between the 5v bias and the
signal lead on a pc soundcard. The input lead will also be connected
to the ground lead over a high resistance resistor to pull down the
signal.

that should make a sizable dip on the wave trace.
using a divider between the sensor and the soundcard so that the trace
stays inside the acceptable limits may lead to the ability to more
accurately determine the time of the event.
The idea is that when both diodes are lit, the signal wil be high and
a simply 44kHz wave recoding wil show a constant high level. When the
first beam is blocked, the signal goes to zero and high again when the
beam is again not obstructed. A little later it will again drop low
and go up.

soundcards don't read DC, with the diode illuminated you'll get a
zero signal, when it is obscured the signal will dip.
I should now be able to measure the distance between the two falling
flanks and calculate the time.
Given that the ball will move through a beam in 0.03ms that is 1.4
sample (at 44kHz) where the signal will go low. That is not a lot.

it's more than enough.
Can I coult on a light sensitive diode to respond quickly enough for this
signal to be detected, or should I consider something else?

handled properly they will respond to signals 100 times faster than that.

Bye.
Jasen
 
soundcards don't read DC, with the diode illuminated you'll get a
zero signal, when it is obscured the signal will dip.

They dont read dc? Are you saying that they can not measure a constant
voltage of for example 0.4v?

handled properly they will respond to signals 100 times faster than that.

Good to know. I think this could be a fast cheap and simple solution
 
J

Jan Panteltje

Jan 1, 1970
0
Or, you could draw a backwards scale on the meter, for speed; meter 0
would correspond to infinite speed, of course. ;-)

Oh, yes, I forgot that late last night, we will use a second 4016 switch
(it has 4 after all) to disconnect the meter when no shot is detected.
And an other one to disable the current source if no shot detected to save power :)
So, now 1 74HC4016 switch left, do not want to use it for Q1, as Ron can be
as high as 150 Ohms...
OK, save more power: Switch of P3 :)


Corrected version:

+ +
| |
e PNP e PNP
b--- b- Q3
c | c |
| Q2 | |----
-------| ---|
!trigger 1 ------------- set | |-- |
74HC74 c | | |
!trigger 2--R7 -------- reset !Q --- R1------b NPN === | R2
| e Q1 | | | charge
a FF1 | | | [ ]<--- current
diode pulse former /// /// | | P1 |
k D1 | 0 / 4016
| C1 | /------------------------
from reset other FF | 0 |
---------------------------------------------------------- | |
| charge switch current mirror /// |
| |
| peak detection and hold |
----- + + |
opamp1 out--- | 4016 |
----- - | 0 / |
| D2 | 4016 / --------------|
|--------k diode a --- meter \ 0 |
|---------------------------------------- + / \ | |
=== | opamp2 out ------R4 --- O ---0 0-->[ ] meter bias |
| C2 0 / -- - | | | P3 |
/// / ---------------- | R3 | /// |
0 | | | | |
| | -----------------| | |
/// | | | | |
4016 | gain --->[ ] P2 | |
discharge switch | | gain | |
| /// | |
| | |
| | |
!trigger 1 ----------- set Q--^-----------------------------------------------------------------
74HC74 | L1
+ -- R5------ reset !Q --- R6-------- k LED a ---- +
| FF2 shot detected light
/ reset button
| S1
/// shot detector

R1 4k7
R2 1k
R3 10k
R4 depends on meter sensitivity, 5 V for full scale... if 100 uA then 50k
R5 10k
R6 4k7
R7 4k4

P1 10k
P2 10k ?
P3 2k2

C1 depends on distance (pulse-length and current)
C2 1nF poly

D1,2 1N4148

Supply +5V (use CMOS opamps).


Something like this (photo detectors not drawn).

Drawing errors?

trigger1 low will set both FF.
Pushing S1 the reset button will reset both FF.
When FF1 is set, Q1 stops conducting, and C1 is charged by Q2 with a constant current.
Opamp1 charges C2 to the peak voltage, and holds, opamp2 buffers and has some gain if needed.
The 4016 switch discharges C2 upon pressing reset.
LED L1 indicates a shot was detected.

All values are 'guess values'.
 
J

Jasen Betts

Jan 1, 1970
0
They dont read dc? Are you saying that they can not measure a constant
voltage of for example 0.4v?

Thats right, the capacitor in the input circuit will block that.
you'll get a reading of a dip or peak if the DC voltage changes but
the DC itself will look like 0V.

So it will work for your ballistic chronograph because it will see the
changes.

Bye.
Jasen
 
Thats right, the capacitor in the input circuit will block that.
you'll get a reading of a dip or peak if the DC voltage changes but
the DC itself will look like 0V.  

So it will work for your ballistic chronograph because it will see the
changes.

Thanks for the confirmation. By now I have also read that there is a
cap on both the input and the output... makes sence for a sound card,
but is anoying when you want to "misuse" it for other things.
 
Top