# ultrasound thermin without µC

Discussion in 'General Electronics Chat' started by @lpha, May 17, 2012.

1. ### @lpha Thread Starter Member

Feb 21, 2012
34
0
Hello
It's possible to do a project like this but without µC ?

I Only need to generate sound (100Hz to 1000Hz) function of obstacle distance, more lower is the distance to the obstacle more higher is the frequency of sound.

Here is a synoptic schematic of circuit without µC.
The Tx sent ultrasound wave every 1 second if no echoes, but, in presence of obstacle, Rx signal are amplified, and transmitter send immediatly an other ultrasound wave.

At 300m/s (30cm/ms) the tx sent the second wave after 8ms, ( 120 X 2 / 30 =8ms), the spaker sent an audio sound at (1/8) X 1000 = 125Hz

At the lower distance (approx 15cm) this frequency is 1000Hz

what do you think about this idea, and what's the best electronic circuit to do this ?

File size:
32.9 KB
Views:
17
2. ### CVMichael AAC Fanatic!

Aug 3, 2007
419
18
From what I understand: You send 40KHz sound (and you keep sending) until you hear a response back, right?

I am working right now to make my own ultrasound distance sensor with a PIC microcontroller.

I am sending 20 pulses at 40KHz. Then I wait until something comes back, and measure the time taken. BUT... the 40KHz pulse that I first sent is MUCH wider when it comes back. This is because it bounces back at closest object (say 30 cm), then there is another object at 40 cm, another at 1 meter, and so on... If you watch the signal with an oscilloscope you get a better idea.
We calculate the distance to the closest object, but the echo comes back for other objects also, that is why the echo is much longer (wider) than the original pulse.

This is why you have to wait a few milliseconds before you send another pulse.

So, if you send the 40KHz sound for a long duration, then you get back an echo that is MUCH longer...

Your logic should be like this:
Send short burst sound at 40KHz
Wait until echo, calculate distance
Wait until no more sound comes back (for a few milliseconds, I think about 50 ms)
Send short burst sound again...
and so on...

3. ### @lpha Thread Starter Member

Feb 21, 2012
34
0
In my case, the ultrasound Tx and Rx is on RC airplane, and the only object who can reflect echo is the soil.
When the airplane landig, I have to hear the soil approching

for exemple :
1m : 100Hz
70cm : 300Hz
50cm : 600Hz
30cm : 800Hz
20cm : 950Hz
10cm : 1050Hz

soil is plane, and there's no other object to reflect echo (in theorically)

4. ### kubeek Expert

Sep 20, 2005
5,582
1,093
I really don´t see what ultrasound distance measurement in this way has to do with theremin. Can you explain?

DerStrom8 likes this.
5. ### @lpha Thread Starter Member

Feb 21, 2012
34
0
It's for FPV landing, I need to send a variable frequency sound via audio canal of my Tv transmitter, exactly like the youtube video, but, instead of the hand of the oprator in video, in my application, soil is the reflector, so more soil approching, more frequency of the sound is high
This system will work like an altimeter but with more precision mesurment, of course, it can't work at high altitude, but i need this system only on landing

6. ### kubeek Expert

Sep 20, 2005
5,582
1,093
That still doesn´t explain the theremin, but whatever.

Let´s say you start charging an empty capacitor with constant current at the moment you send the pulse. When the pulse is received back, stop charging the cap and sample the voltage on the cap with a sample and hold circuit. (you could use 4066 for the switches)
Then the voltage sample is proportional to distance of the object. Since you want the inverse frequency, i.e. high for closest and low for furthest, feed the sampled voltage into an opamp that will invert it and shift the level, so that you can feed it into a VCO that will generate the sound (ne566).

The biggest culprit will be detecting the arriving signal, but that can be solved with lm567.

edit: You will need to use trimpots around the opamp to get the exact frequency response you want.

7. ### CVMichael AAC Fanatic!

Aug 3, 2007
419
18
You can do all that with a simple micro (like PIC), including the generate sound... Probably you still need an opamp to amplify the echo, depends on the distace you want to detect...

8. ### @lpha Thread Starter Member

Feb 21, 2012
34
0
Have you an electric diagramm for this ?

9. ### kubeek Expert

Sep 20, 2005
5,582
1,093
No and I don´t have time now to do it, take it more as a block explanation. You could try drawing a block diagram for it and then we could discuss the details further. I am going to sleep so I´ll be back tomorrow.

Feb 21, 2012
34
0

File size:
43.2 KB
Views:
19
11. ### kubeek Expert

Sep 20, 2005
5,582
1,093
Yes this looks pretty good, assuming the box in lower left corner is the carrier detector. You will also need two capacitors, so that c1 will be used for timing and c2 for storing the previous value.

Unfortunately the control signals for the switches will be more complicated than that, and you will use more of them.

I will assume the astable xxHz will be have short pulse width, for example 100us on and 1s off to keep the ultrasound pulse short. You will need to feed the output of the astable to an RS latch so that when the pulse starts the latch keeps it's output high and holds the charging switch on.
When the carrier detector detects the pulse, it resets latch1 to stop the cap charging. It also enables path from c1 through a buffer amplifier to c2 (with switch between amp and c2), so c2 stores the measured value. The 556 will be fed from c2 again through a buffer amplifier.

After a little delay after resetting the latch, to allow C2 to charge to the new value, C1 will be discharged through a discharge switch and will be waiting for a new pulse coming from astable xxHz.

12. ### kubeek Expert

Sep 20, 2005
5,582
1,093
Now that I think of it you will need to use an AND gate in the path from the detector, so that when TX is transmiting the detector can't stop the measurement. There may need to be some other delay involved because the receiver might resonate for a while after the tx pulse stops, outputting a false signal.

13. ### @lpha Thread Starter Member

Feb 21, 2012
34
0
ow !!!
Ok now I think that is impossible for me to draw diagram for this description

14. ### @lpha Thread Starter Member

Feb 21, 2012
34
0
there's a lot of composents and the circuit wil be too heavy :/

15. ### kubeek Expert

Sep 20, 2005
5,582
1,093
Like almost allways, doind stuff with a micro is much simpler than doing it in analog.

16. ### @lpha Thread Starter Member

Feb 21, 2012
34
0
why my first idea will not work ?
This is a new schematic who I explain precisely the principle
I only need three 555 or one 555 + 556
And double AOP IC

I suppose that there is no second echo
In all case, the circuit will be finely adjusted to eliminate second echoes and other perturabations sound, only the priencipals echoes will be converted to 5V at the output of comparator

File size:
39.6 KB
Views:
15
17. ### kubeek Expert

Sep 20, 2005
5,582
1,093
To get the 0.5ms length of the signal, the 2khz 555 needs to be in monostable configuration and not an oscillator.

Ok I think I am starting to understand how is it supposed to work. I don´t think the relation between echo and non-echo will be like that, since the 1hz oscillator will basically do nothing, and the rx will dictate the sound and tx repetition frequency.

The biggest problem I see here is that the sound coming from tx will not bother to bounce off the object, but it will go straight into the rx first. Thus the sound (if the circuit was correct for what you want) would be at the highest possible frequency all the time. There has to be something that will block the rx for some short time at beginning of each tx pulse to prevent that.

18. ### @lpha Thread Starter Member

Feb 21, 2012
34
0
circuit must be autonomous and detec continously the presence (or not) of obstacles
There's no operator to push button to generate the first wave, so this is why i used a 1Hz oscillator (to generate the first wave in place of operator)

For the rest, I have a possiblitiy to place the Tx transducer and the Rx transducer at 1m50 to 1m80 distance (extremity of plane wings), in this conditions, I think that it's impossible to directly detect the genrated wave from Tx.

19. ### kubeek Expert

Sep 20, 2005
5,582
1,093
Then I think it should work.

20. ### @lpha Thread Starter Member

Feb 21, 2012
34
0
Here is a modified diagramm, I add another monostable (but not with 555 because I need two opposit output),
I have modified the 2KHz monostable to 3Khz

But I have no idea about a reset pin of AOP, what kind of AOP ? and can I use the reset Pin in this configuration ?

Now, In my diagramm, when Tx sent during 0,33ms, the output of comparator is stay at 0V during 0,5ms, regardless to Rx wave received (direct wave or echos)

This impose a minimum detecting distance of 8,5cm, I only need a minimum of 10cm, so it's ok

EDIT : the two monostable (0,33 ms and 0,5 ms) are rising edge triggred monostables

• ###### US 5 eng.JPG
File size:
46.4 KB
Views:
10
Last edited: May 19, 2012