Fish finder sonar: circuits, protocols and transducers

[tarantula]

Hi,

I would like to build a simple wireless fish finder. There is the famous Humminbird Smartcast, which has been taken apart here:
http://www.eetimes.com/document.asp?doc_id=1281283

Basically, the same thing is available as wireless unit:
http://www.hummingbirdfishfinderguide.com/humminbird-smartcast-rf15-fish-finder/

As a first step, I'd like to tap into the 433 Mhz signal the unit is sending. There are Chinese knockoffs that go by the name "Freecast", which implies, the protocol can be decoded. Does anybody have info about how exactly it works? I guess it must be a two way communication, as there are some settings on the main unit about the depth range etc. Or is this maybe analog modulated an the main unit does it all?

If impossible, or at a later step, I'd like to tinker around with some piezo crystals myself. However, I couldn't really find any circuit descriptions, but what's worse: I can't find the piezo transducers. Where can one buy them (Europe).

Since Dave was in the sonar industry for a while, maybe he has a few hints, or maybe he could teardown one of these units and explain it. I think a lot of people would be interested...

Btw, here are the insides of a wireless unit:
http://blog.byteb.org/2012/09/repairing-humminbird-smartcast-rf15e.html

the bottom of the PCB is +- unpopulated. Just a battery.

[LaserSteve]

Take a look at the obsolete National semiconductor LM1812 for ideas.

Steve

[helius]

Back in the dark ages, sonar fish finders were used with acoustic headphones instead of graphic displays. The reason for this was twofold: it was simpler to put the transducers in fixed positions, or manually steered by the user, rather than scan them around; and the type of DSP needed to disentangle the reflected clicks into shapes wasn't yet miniaturized. Users learned which sounds were made by different underwater targets.
An interesting aspect to this that was discovered accidentally was that with two pickups and stereo headphones, some users could "see" the underwater landscape entirely through the acoustic clicks. This is the same as blind people who are able to "see" through echolocation by flicking their tongue against the soft palate.
https://en.wikipedia.org/wiki/Human_echolocation

[tarantula]

Thanks for the tip with the chip. That might be able to serve as the amp. Still no real source for defined transducers. I was thinking about using a small array of 40 kHz capsules as parametric array to overcome their shitty directionality. I'm not shure if it will work underwater though, the speed of sound is much larger and the distance of the units might be to big to be able to fit into anything that can be cast with a fishing rod.
I really wonder what they put in these two units:
http://www.fishhunter.com/home/
https://buydeeper.com/

Both seem to provide some outstanding images. No teardowns to be found though...

[LaserSteve]

Phased arrays work just as well under water as they do in air.  So do line arrays without phasing, over reasonable frequency bands.

Steve

[LaserSteve]

Well, there is nothing magic about "Navy Type  PZT" or other materials except how you enclose it so that the waves pass through the enclosure, which usually means some form of coupling gel or soft potting.  I note  in those linked pictures steps have been taken to reflect or cancel the upward going wave from the transducer plate.

http://lmgtfy.com/?q=Navy+pzt+

These guys sell piezo materials in onesies.

http://www.steminc.com/

Remember the frequencies on their web page may be "Air" frequencies, and not "Wet" frequencies, you need to ask.

Steve

[tarantula]

Thanks a lot for the hint! I think I will also consider getting the transducer of those cheap ebay sonars and repurposing them. I also found a disc in my trash that seems to resonate at around 200 kHz, which would be pretty neat.
As a first aim, I'll check out the 433 MHz protocol of the Humminbird Smartcast transmitter. Maybe that will already reveal tinkering potential...

[tarantula]

I made some progress. The protocol consists out of 16 zeros (potentially the measured depth [in the lab] or sync pulses), one longer burst and another 16 bit value with zeros and ones. That's probably the temperature. I also found the original Humminbird patent:
https://www.google.de/patents/US6693847

Judging from that, the intensity of a target is encoded in the pulse length. Also, it seems to be pretty coarse since the return has to be super-threshold.

[LaserSteve]

Hum, Are you sure? They usually can show multiple target returns in one Echo.
Steve

[tarantula]

Yes. They were of course lacking in the lab setting. I'm currently waiting for a small 433 MHz Demodulator, to take a scope meter to the lake. The 500 MHz scope is a bit big to bring it to the lake ...
Under lab conditions it was:

_____0000000000000000____IIIIIIIIIII____X____0101001whatever_______ and repeat. Where I marked the X would be where I'd expect the return pulses. I was hoping to maybe see an amplitude modulation of the carrier wave, but there I was probably hoping for too much :-(
This all or none scheme makes sense for the little wrist watch gadgets, but not for the units that can display gray shades. I wonder what the trick there is:

http://busse-yachtshop.de/smartcast/smartcast-rf15.html

Oh, and there is an interesting discussion with teardown here:
http://www.baronerosso.it/forum/modelli-naviganti-per-usi-vari-pesca-altro-ecc/203594-ecoscandaglio-smart-cast-rf-15-a-3.html


Come on Dave, rip a Deeper Fishfinder apart! 

[tarantula]

So, I have received the 433 MHz Receiver and played around. My suspicion, and also the information in the patent, seem to be what's going on. There is a strange 16 bit 0 value, then a long sync pulse of 6, ms, the echoes - if any - and the water temperature in 16 bit. Here are a few screenies that I obtained by rubbing the transducer and hence fooling the receiver.
1: Normal waveform without echo
2: 16 zeros
3: sync pulse
4: water temperature (?)
5: signal with artificial echoes

I'll take it to the lake later today.

What can be deduced from the already available information is that transmission is purely digital. Hence, gray scales are probably obtained by time averaging or purely made up, like in the unit I have (see the echo shape, it has many short latency peaks).
Also, the receiver doesn't seem to do a gain correction for distance. Hence, fish close to the surface are overestimated and invisible close to the bottom.

In my unit I can also see the transducer. It has a diameter of about 20 mm, which should translate into about 40 degree beam angle at -3dB.

[tarantula]

I tested it at the lake yesterday and my suspicion was confirmed. That's exactly how it works. I couldn't verify if the first number changes, though because the range of the cheap 433 MHz receiver was just a few m and my mobile oscilloscope is garbage. But what can be confirmed is that:

1. no analog information except the quantization noise in the returns. But taken that and the relatively high repeat rate might allow some averaging to grayscale
2. the long burst is the sync pulse

After estimating the cone angle and taking a closer look at some of the competition, I suspect that the Vexilar Sonarphone:
http://store12459.americommerce.com/store/p/26-SP100-T-POD.aspx
might actually be a rip off similar design.

It's highly regarded among its users, but the shape of the whole thing is similar, it's also the same type of cheap chinese plastics, uses 125 kHz with 30 degree cone angle and suspiciously cheap. If one assumes that the Smartcast/Freecast is typically overstated with its cone angle (e.g. by using -10dB vs. -3dB), that would fit well.
I would almost bet that it is actually a very similar sonar circuit, adding an NXP SA604AD or similar for logarithmic RSSI (if even or just use the rectified return), ADC and an ESP8266 or similar and bob's your uncle.
I wonder whether that might be a route to go. In the smartcast, the amplified returns are available... Hmmmmm...
Next step for now will be a LabView implementation of the receiver, so that I can plot the echoes, an attempt at recovering return strength from the digitization noise, but first and foremost a better receiver circuit for the 433 MHz. I decided to keep the unit, so I might just tap the shore unit that came with it.

[tarantula]

So, since I had these problems with reception, instead of tweaking the 433 MHz cheapo receiver, I tapped into the original fish finder shore unit. It turns out that it uses a pretty solid 433 MHz super het receiver, soldered onto the board with a "data" labelled line coming out! Bingo! In the image it's the brown board on the main PCB.
I actually wonder why half the main board is unpopulated...
Now I just tapped the data line and routed it out of the case via BNC terminal. I might take it to the lake later to see how it's behaving!

[tarantula]

I did some testing. It seems that the first data burst is indeed some value for depth, as it varies in the low numbers if there is a true bottom and all zeros for air. However, after the sync pulse, there is aparantly a forced low period (see pic) that doesn't shrink to zero if the bottom is very close. Probably that's the reason, the device will not display less than a meter of depth even in the shallowest water.
I now wrote a little LabView program that measures the pulse width and triggers an analog acquisition upon finding the sync pulse. Sadly, it's to slow to start quick enough, therefore I use it to sync to the next pulse, losing the one that actually triggered it. The software timing jitter is then compensated for by doing a circshift.
I haven't tested it in the lake, but from what it looks like, it might work...

The first picture shows the echo in really shallow water. There is the primary return and a few additional ones, probably secondary and tertiary return from reflections. Note the forced low period in the beginning that should be much shorter.

[unitedatoms]

If you are interested in correlation based sonars look at Costas Arrays https://en.wikipedia.org/wiki/Costas_array.

And video about them with "Perfect Ping" performed on piano

Those arrays are perfect for time correlation, I tried to think of general application for this array in test metrology, like dirt cheap superfast 10GHz oscilloscopes based on slow mass produced hardware, like ethernet MAC parts with 4-bit DACs.

[tarantula]

That looks very interesting! Thank you!
I was wondering, which transducer type is used to replay them?

[tarantula]

I took the whole setup to the lake and it works. I collected some interesting images, but when I came back, it turned out that LabView didn't save them. But one I could save. It shows the sensor approaching the shore.
What is certainly cool is that there is some detail on the bottom. It was also possible to see some structure on the ground. The original unit didn't show that, it seems to average quite a bit for the bottom...
There were also some "fish" on the unit, but I'm not sure that that were real fish. On the LabView vi I couldn't see any fish bows, only sometimes there were echos between bottom and surface for one scan. But the receives signal is pretty noisy, as you can see in the plot. It's not uncommon to have random pulses.
I think I'll improve the vi and see how it goes the next time...

[ripper121]

Do you have any news about this?
I want also to hack such a cheap fishfinder and extend the Range for a RC boat.

[tarantula]

Hi,

I do have indeed some news. I kind of got stuck at the LabView interface for the cheapo fishfinder because the digital returns just turned out to be too little information. Impossible to distinguish weeds, baitfish etc. What I did now is a bit of a contrived setup: I have hacked an old R/C boat to work with aregular 2.4 GHz remote with kilometers of range. Then I attached a tow point and a Deeper fish finder via a short piece of (floating) fishing line. Now I can troll the lake for fish and then cast where they are. But currently it's winter in Germany, so there is little fish out anyway. What I'm currently doing is putting a cheap used Android phone in the boat and use it in the Deeper boat mode. In that way I also get a depth map of the lake. Since the Deeper can only connect to one phone (I believe...) and for range reasons I'll use my FPV camera from my quadrocopter and a portable video monitor to get a view of the screen of the phone plus FPV from the phone! Should be pretty nice, the camera has plenty of range...
Apart from this there are two options:
- built one's own sonar. With the help of the Smartcast patent this should be possible. I'm currently playing around with a slightly improved circuit which uses a transformer instead of the single coil of the Smartcast. But due to my job taking a lot of time, it'll probably be a bit until I get this running.
- the route which I haven't pursued yet, which is tapping into the original analog return from the Smartcast module and digitizing that. In combination with an SA604AD RSSI signal that could be fairly sensitive.
For transmission one might be able to us the 3DR Radio, a module which provides a serial link to quadrocopters but should also be open to other types of data.

[Piki1989]

Hi,

I would like to propose some solution regarding visualization of frames provided by finder sonar.
You can basically use software provided from Blue Robotics for ping sonar, which is called ping viewer:
https://github.com/bluerobotics/ping-viewer (for compilation you need to install Qt libraries 5.15 with Qt Charts), but you can use also installable version of this SW.
https://docs.bluerobotics.com/ping-viewer/
For injecting data I'm using simple python script provided here:
https://github.com/bluerobotics/ping-python/blob/master/tools/ping1d-simulation.py

My question is:
Did you manage to reverse eng messages payload?
Especially for such data as:
*distance,
*confidence,
*scan start,
*scan lenght,
*fishes

[gadgeteering]

I know this is an old posting but it has been very helpful with the project that I am currently working on.

I thought I would share some of the information that I have been able to find through my investigations.

Using RTL_433 I have been able to capture some of the transmissions of data from the Lucky fish finder, using PulseView



" alt="" class="bbc_img" />

Like tarantula has state there seems to be a 6ms sync pulse and the other pulse are PWM 800us long

[CaptDon]

I haven't read every line of every reply, however, if this hasn't been covered in another reply, It is difficult to see very shallow depths due to time constraints. You can't receive while transmitting so the system is blind for at least the time of the transmit pulse. The actual 'time' of the pulse will be perhaps 5 to 10 full squarewaves of the transmit frequency which on fish finders tends to be between 100Khz up to 200Khz. Much like radar, there is probably the equivalent of 'Sensitivity Time Control' where the gain is very low at first and then ramps up for the distant objects. Best wishes for your project.