Functional fishfinder sidescan II

[Rickard]

While waiting...

for new material for a new sidescan array I resimulated and rebuilt my old array (the old one is described under the thread "Functional fishfinder sidescan"). The old array did not fulfill all my wishes on performance and I was close to conclude that good arrays based on the principle of unequal spacings are almost impossible.

However, I found a new way to arrange my transducers so the minimum distance between piezo centers became 51 mm. The new array has the following specifications: piezo diameter - about 38 mm, 200 KHz, four transducers, interpiezo centerdistances: 51, 103 and 189 mm.

The new array is much shorter than the old one but the expected sidelobe suppression is -8 dB. The old suppression was -4 dB. (Remember, -13 dB is the theoretical limit for unshaded arrarays). Various tests showed that the instrument (Humminbird Matrix 37) displays echos down to about -8 to -10 dB so there was hope for only minor influence from sidelobes. Indeed, this proved to be the case.

Attached are a photo of the array and some photos of a small new wreck I found when testing. The wreck lies at about 5 meters depth. I scanned from several sides.

One photo shows a boulder and some smaller rocks. Much more welldefined shadows than before. In the shadow from the boulder one can see how sidelobes contaminate the shadow at the edges. But but the shadow limits are still clear.

More scanning will follow!
/Rickard

[Rickard]

Beam forms of alternative array configurations.

Hi,

Just got the idea to show a graph to illustrate what the beam forms will be for my latest array and for two related alternatives of equally spaced arrays. Angle from center of main beam is at the x-axis and suppression in dB is on the y-axis. My latest array gives the black line. Sidelobes around -8 dB. An array with smallest possible equal spacings (51 mm center-to-center distance) would give the form shown by the red line. An equally spaced array with same total length (189 mm) as my present array is shown by the blue line. Both the equally spaced arrays have higher sidelobes in comparison to the optimized array. I have left out angles larger than 15 degrees since there are no influental sidelobes in that region. Computations are based on wave theory and the plane wave assumption. The algorithm is implemented in Graphing Calculator for Windows.

/Rickard

[Rickard]

Performance indication.

I got a result that gives some indication of the level of performance of the array. The array is hull-mounted just below the surface. Weather was calm. Target at a depth of 19 m and 90 m away. The object (a sunken barrier?) is about 30 m long. The nice triangular shadow tells that it has partly sunken down in the clay bottom. Influence from sidelobes can be seen as a fuzzy extention without shadow at the left end of the object. So, in this case sidelobe echos are visible but their shadows are filled by the mainlobe. The sidelobes did not contaminate the mainlobe shadow.

/Rickard

[Rickard]

Cork, side lobes and luck.

Hi,

The process of finding better suppression of side lobes in common fishfinder transducer arrays has reached the point where I’ve started to grind down the walls of my dear transducers. The gap between piezos will be less than ¼ of a wavelength (about 2 mm at 200 KHz). Grinding and sawing is rather simple on the transom mount high-speed Humminbird transducers I use. The plastic housing is about 3 mm thick and the space between house and piezo is filled with a hard rubber-like substance and a very soft, thin foam close to the piezo. I will glue the transducers together using urethane (Liquisole). Tests show this gives very strong and tight results. I expect the side lobe suppression to be better than -10 dB. Of course, smaller gaps will result in a shorter array (160 mm using four transducers) and a wider beam (2.4 degrees at -3 dB) but I’m prepared to pay that price for optimal suppression.

I know from tests that one can mask the piezo with a layer of cork. A thick (about 3 mm) layer of cork absorb transmitted and received sound effectively. If a rectangular section is left free from cork the transducer will work as if it had a rectangular piezo, as in ‘professional’ arrays. I have found in the simulations that rectangular shapes of the elements in an array results in dramatically reduced sensitivity to violation of the ‘not more than ¼ wavelength rule’. Actually, in the case of my transducers, it is possible to get a suppression of -9.5 dB without damaging the transducers at all! This is better than my best irregularly spaced array without cork mask. But since I have already slaughtered two of the transducers I will continue with the rest anyway.

Of course, cork will waste a lot of energy. Energy is also more dispersed as the vertical beam width is wider. But wider vertical beam also gives a wider swath.

Luck is needed to succeed with this type of project for a very specific reason. The piezos in the individual transducers must have the same orientation with respect to pooling direction. Unfortunately, manufacturers of fishfinders have, I think, no technical reason to be consistent about this. Let’s say an array is made up of two transducers with opposite pooling. The result is complete elimination of the main lobe since they will be exactly ½ wavelength out of phase! The problem can be solved by mounting one transducer ½ wavelength above or below the other transducer. And then you also need a method for assessing pooling of piezos….

Practical tests will have to wait until next summer.

Regards
Rickard

[Rickard]

Comparison with HB 981SI.

Hi,

I have used a Humminbird 981SI and scanned two of the targets I presented above in reports on results with the fishfinder array. Of, course the "triangle" (space ship...?) and the boulder have much better definition in the 981SI images but it is definitely possible to see some details in the fishfinder images. The frequency used in the 981SI was 455 kHz. The triangle rests deep, 29 m. Note that it was scanned at 90 m with the fishfinder array and only about 50 m with 981SI. The most important difference between the systems is the noise from sidelobes in the fishfinder images. I hope this is solved in the new array. It will take a few weeks before I have any results on that.

Regards
Rickard

[Rickard]

The SI transducer works with the Matrix 37!

Hi again,

I simply have to report that the SI transducer that came with the Humminbird 981/987SI models works very good with the Matrix 37. The sideviews in the M37 use the 455 kHz elements in the QuadraBeam transducer. The same frequency is used in all the new Humminbird sidescanning models. Therefore I made a test with the modern transducer (XHS 9 SI 160 T) connected to the M37 (after consulting HB on the risk for damages). The connectors and pin assignments are the same and the system worked fine with standard settings. But you can't freeze or capture the screen in sideview mode so photos of the screen have to be taken while the image scrolls down. Thus the attached photos are rather blurred. In reality the image is surprisingly good and not at all far behind the image on the 981SI. Naturally, the size of the sideviews sets a definite limit to resolution. The target is the same very small (5 m) wreck shown in posts above. The left view shows the shoreline.

After these testings my old M37 got new life. The results leads to some wishes however, an option to select either of the views to cover the whole screen width and a mode that uses the whole screen from top to bottom for the sideviews. And also an option to freeze the image. But the model is probably too old to get any more updates at that level. I think there are some other HB models with the same potential to operate with the SI transducers (737, and some of the 500 series models) so perhaps it could pay investing in developed options.

Rickard

[Rickard]

A final solution.

Dear sonar friends,

The latest version of a fishfinder array, as described above, is now finnished and tested.

Four Quadrabeam transducers have been grinded and cutted down to allow for a inter-piezo gaps less than 1/4 of a wavelength at 200 kHz (2 mm). The photos show the array and also the testing arrangement with a sideimaging transducer connected to a Humminbird 981SI and the experimental transducer mounted together to allow for parallel scanning on the same targets.

I will show images and photos showing the scanning results in later posts. But in general, the array worked as expected. No leaks or other problems after 8 hours of running at highest output settings.

The last image shows the expected beam form of the array. The sidelobe suppression seems to be about - 13 dB. This is the dream level for unshaded arrays.

Regards,
Rickard

[Rickard]

Testing the optimal transducer with the M37 and 981c SI.

Hi again,

Some data on the experimental transducer: array length 158 mm. Vertical aperture (width between cork strips) 9 mm. This gives a longitudinal beamwidth of 2.4 degrees and vertical 45 degrees at -3 dB at 200 kHz. This array is not a version of an unequally spaced array as the ones I built before. It has the formal features of an optimal array. Such an array is impossible to arrange with fishfinder transducers because of too thick walls. Using Quadrabeam transducers is a waste of money but they were the ones available in my case. If anyone have plans for making his own array the cheapest possible puck types will do.

First pair of images shows a small wreck. First image shows the M37 with the experimental array and the second is a snapshot from the HB 981c SI run at 455 kHz. The results from the experimental array show problems with signal strength but shadows are clean because of well suppressed sidelobes. The images are taken simultaneously.

Second pair shows a car resting on it's side. Images are taken simultaneously.

Third pair shows a wreck scanned with the experimental array connected to the 981c SI and the same wreck scanned with the 981's original SI transducer at 455 kHz. The 2D view in the 981 normally used for downlooking fishfinding is used here for sideimaging at 200 kHz.

The following three pictures show results when the experimental array is connected to the HB 981c SI when scanning on the same wreck as in the first pair. These are snapshots downloaded to the instruments memory card. Very convenient and gives better image quality. The impression is that the 981 has less problems with signal strength than the M37 has.

Last picture is taken with the 981 and the experimental array and shows a sunken concrete breakwater outside the marina.


Rickard