Question to Sture

[Jan]

Sture,


What properties, parts, of the echo signal is important?


I can only think of:


1) Echo delay


Obviously needed to determine distance


2) Echo strength


Together with distance an indication of material that caused echo like sand mude etc.


Is there something else we need to make a reasonable picture ?


I can make the strength indication in 8 bit resolution


So that would give 256 strength levels.


This would be for me the easiest, as it comes out of the PIC that way.


However, if need be I could increase to 12 bit resolution (4096 steps)


I hope you will not ask for a slope indication... upsets my design a bit.. hehe


Regards Jan.

[Sture Hultqvist]

Re: Question to Sture

I'm little careful about giving direct answers about


things I'm not perfectly sure about. This is such a


case.


Echo do also have a frequency distrubution and you


need a corresponding bandwith in your reciver.


Depending on distance it should be 3-30Khz.


You also have the dynamic of your echo, not only


echo strenght by distance but also high dynamic


difference in the same echo. Higher bandwidh allows


you also to take care of higher dynamics.


Bits in the ADC: First machine I build did use only


16 levels of color and it worked well in that way.


The current machine use a 12 bit ADC wich in the end


gives 64 gray levels . The 64 levels is choosen since


a standard VGA card DAC uses 6 bits. A normal card


have higher resolution but I don't think it is


neccesery i practical use to have many bits.


Your 8 bit choose I guess should be as good as


mine "12 bit reduced to 6 bit".


The ADC card (DAS800) is delivered with drivers. I only need to set it up for collecting 1024 samples


on 2 channels with a sampling frequence, as example,


of 5 KHz. The result is placed in a array of 1024


words. This array is the sorted in the way that


left channel samples is number 0-511 and right is


512-1024. For left channel the last sampel is first


and so on.


Then each sampel value is exchanged to a grayscale


value and the last operation is to do write the


array directly into the VGA card memory.


TVG slope: First machine... the 4 bit.. used a


feedback where I integreted a number of pings in


certain distances and constructed a slope sent back


with a DAC to the reciver. The current is the


schematics I publised and works as well as the


feedback variant.


I hope this is what your question was about.


Sture

[Jan]

Re: Question to Sture

Oooppps, there is more to it than I thought ...


Could you give an educated guess on the quality of the picture I could get with just the info Distance and strength?


You see, I do not have to make the best sidescan in the world, just one that is good enough to see objects of about 20 x 20 x 20cm at a 50-100 meter distance.


Its the first sidescan I ever build and want to keep it as simple as possible.


If you say "Nay, you need that extra information" that so be it, but if it is just to enhance the picture that I would prefer to leave it out.


Regards, Jan

[bob]

Re: Question to Sture

Jan,


It occurs to me that the return will be similar to AM radio reception. there will be a 'carrier' at 500kHz (if you build it at that freq.) and as the return comes back, there will be a lower frequency signal impressed on the carrier, just like a song would be impressed on a 500kHz carrier in a radio situation.


If we think about the signal (song) independently of the carrier, the highest frequency of the signal will correspond to the smallest spacial resoluton.


For example, you said that you wanted 20cm resolution, so at 1500 m/s that converts to a maximum frequency of 7.5kHz. Sture suggested that the aplifiers be capable of 30kHz, which would be equivalent to 5cm.


Does this make sense?


Now with regard to the gain, as time goes on, the signal will get weaker, so we need to compensate for this weakness, so we get a uniform grey value regardless of distance. This is why there is a variable gain section in the receive circutry. (I think TVG stands for _Time _Variable _Gain)


Sture said that his old design would use a DAC to drive the variable gain, but that this analog circuit does the trick. I may agree with him, but then again, if we can get the PICs to do a DAC output, that might be cheaper, easier, and have a lower chip count.


-Bob

[Sture Hultqvist]

Re: Question to Sture

Hi


Sorry if I'm confused you You need only


to read signal strenght and time. The bandwidth


is giving quality. I read Bob:s comments and


he have calculated right.


Sture

[Jan]

Re: Question to Sture

Sture,


After destroying my confidence and general well being with your previous answer, you made it up again with this answer.. )


I thank you... hehe


Got a fright there..


Regards, Jan

[Sture Hultqvist]

Re: Question to Sture

Myself I try to make everything as simple as possible.


All complicated construction I made seems to never


work.


But it also seems that I have made all mistakes


that is possible in this business before I got


everything right. Or I may still do something wrong!


One more thing: I have very funny reading how you


and others work and rekognizing my old and former


problems.


Sture

[bob]

Re: Question to Sture

Sture,


Keep reading, I'm sure there are more mistakes to come!!


But that's the fun! (As long as they don't cost too much!)


-Bob

[Sture Hultqvist]

Re: Question to Sture

Funny with this compare to a AM reciver In fact


I have used a car radio for one sidescan experiment


and it worked well. The problem was to control


amplification and also that a normal AM reciver


don't have a DC output. However the resulting


images was OK.


Sture

[Patrick]

Re: Question to Sture

Finally I know something you guys are talking about ... car radios.


Patrick