Ground Penetrating radar

[uglychicken2005]

I'm new to electronics but believe in learning by doing.
I did some searching and found a gpr project https://hackaday.io/project/175115-gprino#j-discussions-title
I posted on that site but I'm not getting any input. I'm just trying to figure out the schematics and figure out a reasonable way to approach the project.
I found most of the components for the RF schematic from mouser and believe I found an alternative to the vco because the one he was using is discontinued
Any input thoughts or ideas would be greatly appreciated

[uglychicken2005]

He used an arduino mega and a 3.5 display to plot the information.
I can post the arduino IDE code if someone wants to take a gander.
I had some thoughts on all of that but I first need the circuitry

[Bud]

Your focus should be on the antenna structure and how to minimize coupling between tx and rx antennas, which the author rightly pointed out to. You can build circuits until cows come home but without a proper antenna you cant use them.

[uglychicken2005]

He gave a layout of a vivaldi antenna, it seemed quite large... a little over 21"x24". But I guess with a low frequency... i figured I could get a single sided pcb board and cut it from that for about $35 but Would copper tape work for testing?

[KE5FX]

That's pretty cool.  The most obvious improvement would be a homodyne architecture, I'd think.  I don't see the point in using two antennas.

[Alex Eisenhut]

Don't these things need kilowatts of pulse power?

[Stray Electron]

Don't these things need kilowatts of pulse power?

  Pulse power not equal average power!  The duty cycle of the 3+ million watt Search RADAR that I worked on was much less than .1%.  The pulse was generated by a fair size bank of caps charged to about 1600 volts and dumped into a Pulse Forming network by a hydrogen gas thyratron. That was 1940s technology.

[uglychicken2005]

This runs off of 5 volts

[Alex Eisenhut]

That's what I mean, this thing barely looks like it would penetrate air, never mind the ground.
I must be missing something.

[penfold]

That's what I mean, this thing barely looks like it would penetrate air, never mind the ground.
I must be missing something.

For low depths in dry ground or sold bits of bridges (as per the test case), the low frequency will work in its favour,

I have a suspicion that having the antenna that close to ground does something with nearfield/farfield transition but I know very little about vivaldi antennas (would be interesting if the antenna was designed around the impedance of mud assuming its not the same as air's 377ohm)

Would copper tape work for testing?

Copper foil/tape should be fine so long as its adhesive backing doesn't prevent it from forming a continuously conductive plane (depending how you apply it of course)

[T3sl4co1l]

Yeah, foil or tape is fine, can even be aluminum foil on cardboard.  As long as it all connects, and the shape and dimension are right.

Which, uh, impedance will depend some on the exact width, thickness, and dielectric constant, of the traces and substrate at the feedpoint.  So that end should probably still be PCB.  But the bulk of the element areas -- where not overlapping -- are basically in air, substrate doesn't matter out there.

I have a suspicion that having the antenna that close to ground does something with nearfield/farfield transition but I know very little about vivaldi antennas (would be interesting if the antenna was designed around the impedance of mud assuming its not the same as air's 377ohm)

Heh, the impedance will surely be lower down there.  Quite a bit lower if actual mud...  You'll get a fairly strong reflection off the air-soil interface, at least from a distance.  Good question, how much near-field counts for here, it's less than a meter away in the picture.  I don't remember what works for evanescent waves / coupling.  Perhaps a shallower angle, perhaps something loaded (fill the bucket with sand?), perhaps something lower profile like a bowtie flat against the soil?

Tim

[Bud]

Don't these things need kilowatts of pulse power?

This gizmo is using a well established step frequency approach. The device sweeps through a frequency range at each lateral distance step in frequency domain, then performs an FFT transfer of the collected response to the time domain. Hence the requirement for an ultra wide band antenna. This device does not emit radar like short pulses.

[Bud]

That's what I mean, this thing barely looks like it would penetrate air, never mind the ground.
I must be missing something.

[KE5FX]

This runs off of 5 volts

So did the Cray-1 ... 

[Alex Eisenhut]

That's what I mean, this thing barely looks like it would penetrate air, never mind the ground.
I must be missing something.

Basically this is a metal detector with a rotary encoder so you can get an "a-scan" out of the thing? How deep can it really go?

[Bud]

The author said the thing was designed to use 323-910 MHz sweep range (~600MHz center frequency). Based on the above table and assuming a decent LNA is used, it would be less than 1 meter depth of penetration.

[coppercone2]

This runs off of 5 volts

So did the Cray-1 ... 

line impedance : explosive

[Alex Eisenhut]

Interesting, going to read more.

[Bud]

Don't! It is like going down the rabbit hole !    So much stuff to learn, mostly IEEE papers....

Edit: but if you do, search for "landmine detection" GPR, this is where i found most interesting publications. And probably the most useful ones.

[uglychicken2005]

Could the CVCO55CW-0400-0800 be used in place of the JTOS-850VW+ because the JTOS is discontinued? They look similar as far as I can tell but then again I don't know what I'm looking at. Also if I use smd to dip conversion breakout type boards and make my circuits on perfboard would that be okay or would that mess with the signals. Though I'm still working on making an antenna big enough to work with the low frequency so I have a ways to go and a lot to learn still...

[penfold]

Could the CVCO55CW-0400-0800 be used in place of the JTOS-850VW+ because the JTOS is discontinued? They look similar as far as I can tell but then again I don't know what I'm looking at. Also if I use smd to dip conversion breakout type boards and make my circuits on perfboard would that be okay or would that mess with the signals. Though I'm still working on making an antenna big enough to work with the low frequency so I have a ways to go and a lot to learn still...

They don't look completely dissimilar, voltage tuning range is a little different so maybe a tweak will be needed to adapt to it,

Perfboard could potentially work... you are quite lucky in that the RF signal path is mostly point to point (i.e. no discrete amplifiers or filters) so it should be possible to use very short coax connections for the RF path at least just to give the illusion of a ground plane. With the low frequency parts, star-point grounding and lots and lots of decoupling on power *should* be okay. Remember 600MHz is still an appreciably high frequency so be prepared for some unexpected results err... 'learning opportunities' no matter how you approach it.

I wonder whether it would be a good suggestion to build each "block" up as a circuit each within an individual shielding can each with its RF inputs and outputs made through SMA connectors (though BNC or any other would work) and power and control signals via feedthrough capacitors... I dunno if performance would be better or worse, but for exprimentability and configurability later on it would be a boon.

[Alex Eisenhut]

Don't! It is like going down the rabbit hole !

Do those show up on GPR?

[RoGeorge]

If the goal is to build a more sensitive metal detector, you may want to look at this Practical Electronics magazine project from Nov 1970.  The link to it was posted by Gyro in another topic here https://www.eevblog.com/forum/projects/portable-proton-magnetometer-teardown-and-tech-info-wanted/msg4079602/#msg4079602