Interfacing fish finder take 2

Surely not the received echo pulse. You will want to detect that at or near one of the MCU input pins.

@Jremington, thanks for your previous help.

I have found a place in the circuit where the peak to peak is at 2.5 volts. Way better.

Fundamentaly speaking, to produce the wave in the water at 200 khz, the MCU use 20 individuals squares pulse in a burst manner. Then it pause for a considerably long period (relative to the length of the 20 squares pulses) and burst again. During the pause, the big capacitor recharge and get ready to power the next burst event.
The way I understand it, is that the burst on the first coil of the variable inductor transfer the signal to the second inductor (via the magnetic field). Since there is a different number of winding in each coil, the 40 khz burst is tranformed into a 200 khz soundwave that propagate in the water.
The MCU can be set to send any desired PWM frequency (they choose 40 khz and match the variable inductor to this.). Then why the inductor is a variable one? It is only a fine tunning thing?

Also, if I change the MCU PWM then I could (in theory) modify at will the frequency of the sonar (in the transceiver range of course)? Rigth?

The coil, as part of a resonant circuit, increases the signal voltage to drive the transducer, so it should be tunable.

Are you sure that the MCU produces a 40 kHz signal? If so, it would be the fifth harmonic of that 40 kHz signal that excites the 200 kHz resonance of the tuned circuit. If you change the driving frequency away from 40 kHz, that will dramatically lower the driving voltage and reduce the range of the sonar signal.

The circuit element that prevents the high transducer driving voltage from getting into the echo pulse receiving circuitry consists of the two antiparallel diodes in the resonant circuit, followed by a 470(Ohm?) resistor, on the input to the echo pulse amplifier. The signal should be about 1400 mV peak to peak, maximum.

jremington:
The coil, as part of a resonant circuit, increases the signal voltage to drive the transducer, so it should be tunable.

Are you sure that the MCU produces a 40 kHz signal? If so, it would be the fifth harmonic of that 40 kHz signal that excites the 200 kHz resonance of the tuned circuit. If you change the driving frequency away from 40 kHz, that will dramatically lower the driving voltage and reduce the range of the sonar signal.

The circuit element that prevents the high transducer driving voltage from getting into the echo pulse receiving circuitry consists of the two antiparallel diodes in the resonant circuit, followed by a 470(Ohm?) resistor, on the input to the echo pulse amplifier. The signal should be about 1400 mV peak to peak, maximum.

A. I measure 200 khz everywhere. There is no 40 khz involved my mistake. On the oscilloscope, in the bottom of many of my pictures in the previous posts, we can see a frequency (in yellow). Some time it say <10 khz and other time it is a precise value. It change with the horizontal time axis setting. I see this like the oscilloscope sampling rate but I am not certain I really understand what it really is. If I zoom a lot on the square pulse trigger, it can change from 200 khz to 12.5 khz.
B. Then the inductor is finally only a voltage booster not a frequency converter!
C. Yes it is the correct position for lower voltage. My goal is to thrash everything following that and just feed this to the comparator like you suggest or similar like a peak detector or Schmidt trigger.
In my circuit and yours, there is a second inductor. Mine is on the top right near the MCU and there is potting compund to prevent using the tunable screw. It seem to be on the side board in your unit. Any idea what it is doing? From my scope, i see that it receive the pulse+echo and output the same on the other side to feed the MCU that produce the signal enveloppe. The voltage is different but not enough to justify the whole circuitry associated to this. It must do something else than dropping voltage?

The second inductor is the receiver tuned circuit, to select 200 kHz echo signals from noise, and probably also increases the signal voltage.

My fish finder has three inductors, but I have not tried to determine their individual functions.

Good news, the system work well. I can measure depth as low as 9 cm with good precision according to my ruler measurements and up to a 1 meter. Now I need to find a deep lake to see if I can still achieve 100m (just like the original unit was doing). With my scope, I can see the echo nicely. If it stay strong, It should work in deeper zone. If it loose strength (p-p voltage) then I will have to amplify.

I use a ESP32 to produce the PWM and directly ground the inductor. When ground, it send power to the transducer and I can tap the echo signal directly on the board and send it to a double comparator OPAMP. I fine tuned the threshold for each comparator and achieve a good amount of filtering. The output signal is then feed back to the ESP and I use the pulseIn function to determine the echo bounce back duration. I will post a schematic soon.

Will I need to remove the inductor to measure it or it is possible to know what inductor I should buy to produce the Q at 200 khz? Is this just like a potentiometer. You choose the frequency and that's it? There is also a diode and a resistor that charge the capacitor. Could we calculate this instead of unsoldering everything? I plan to build a small PCB and just connect the transducer directly. I don't want to tap the signal and carry the original unit around and my results show it is possible!

Please post a schematic of how you have connected it, and what you propose to build. At least one tuned circuit is required to build up the voltage for the transducer.

Here is the schematic I came up with.
1- What I still don't know is what will happen when I remove all the components and circuitry after C3 and C4 (below the two diagonals). I don't know if they help to achieve the echo level at the signal tap point. It would be very nice if I could cut everything just there? There is no DC in that part of the circuit (I think).
2- I could probably add capacitor and stuff around the two comparators? It work like this on my breadboard but is probably far from optimal. Coupling cap? Filter?

I doubt there would be a problem to cut those traces. The signal from the two antiparallel diodes will be more or less a square wave with maximum amplitude about +/- 0.7V, so I doubt you need two comparators.

For weak return signals from long range echos, you will probably need a stage of amplification before the comparator, and it would surely be a good idea to use the existing second tuned circuit as input to that amplifier, for selectivity and noise reduction.

I did try to amplified the signal with the first half of the opamp but it did not work and I don’t understand why. I used a basic non-inverting amplifier.
Tomorrow I will send scope data showing why two comparator was better. But not by much… 1 gain stage + a comparator would be better on the long run.

Please post a link to the part labeled "MN6002". That is not a standard designation.

The only thing that comes to my mind is an MCP6002 op amp, which would be a very poor choice for anything in a circuit like this.

Yes it is a mcp6002. Single rail low voltage thus easy to connect to my 3.3v rail and output a high voltage that can be directly read by the esp32.

The non inverting amp I tried was not amplifying. Vin = Vout on the scope. Echo p-p in = echo p-p out.

Here is the output from the first comparator: we can see some ripples in the "square" output. This tend to confuse the pulseIn function and return false measurements.

On the second comparator output, the riples are not present anymore and there is a clear low signal between the trigger and the returning echo. Even for short distance. Here about 9 cm. With shorter distance, the signal combine and there is no more low phase between the pulse and first returning echo.

And finally, this last picture show the Tap signal (yellow) and how it is at pin 2 of the mcp6002 (blue). Ideally, this is the signal I would like to amplify (the returning echo is only 600 mV).

The op amp may be oscillating, probably because it is not properly bypassed. Use 100 nF directly across the power pins.

However, the MCP6002 is a terrible choice for this circuit, because it is so slow. The gain bandwidth product is 1 MHz, so the best you can do as an amplifier is gain of 5 at 200 kHz.

For the comparator, use a genuine comparator chip like the LM393, and for an op amp, choose one with a gain bandwidth product of 10-20 MHz.

Why not just use the amplifier already built in to the fish finder?

jremington:
Why not just use the amplifier already built in to the fish finder?

The fish finder use a NPN transistor and don’t really boost the signal. It just drive the second inductor. There is no real amplification going on in the fishfinder post treatment. I think they rely more on the chip capacity to output a good RSSI signal. So like you suggest, filtering the noise out and producing a good signal enveloppe. This is more sophisticated and allow for signal analysis to find fishes or plants.

I am looking for simplicity and don’t know much about OpAmp as you can see. This is why I rather not include a second inductor and only use one dual opamp. The LM393 can work on 3.3v single rail and you say it is a good comparator and it’s not expensive (all very interesting point). Would it also be suitable (better than the MCP6002) for boosting the signal? A gain of 5 (if I can make it work!) is probably more than enough at least up to 20-30meters.

When you talk about osccilation, is it related to the comparator ripples or to my non-working opamp amplifier?

You will discover that "driving the second inductor" is a very significant, frequency selective amplification step, and is a critical part of the design.

I get back to my scope.
The transistor + inductor amplified and clean up the signal. I see a 1 volt increase. It's a gain of ±2. Channel 1, input. Channel 2 inductor out.

Can you guide me to achieve a gain of 2 or 3 on the MCP6002? I use those configuration without success.

I need to select a IF tuner inductor for the circuit and the inductor on the actual unit are not identified.
I dig on internet and find the transducer test characteristics (Q factor, resistance and test frequency). See picture.
With the equation Q = (2PiL*F) / R and the available test infos, I calculate "L" to be 16 mH.

There is a lot of variable inductor on the net and most are in uH.
1- does 16 mH look weird for a Q of 48 @ 200 khz?
2- Is there other thing than henry value that I should look for?

thank you

The LM1812 may help you a lot. I thought the transducer of Lucky fish can be use on it. You can find some info on paper "Down-Under Depth Sounder". The circuit works quite similar to Lucky fisher but they use op-amp and counter to read the water depth.

hey have you ever try with JSN-SR04T or AJ-SR04M and use the fish finder transducer? can you help me? Im currently using arduino UNO. it will be very helpful if you can help me with underwater depth sensor with arduino.