Home Made GPR?

[srpnt]

Has anyone built their own GPR?

Here's what I'm up to: I have a search area that even with a decent GPR will take 9 years of constant searching to cover. It's dry sand (which for GPR translates to "Good depth properties") and my target (I'm not telling, so don't bother asking) is going to be buried, possibly deep. My hope is to strap a GP w/ GPSS to a railbuggy and bop around the desert like a mad man on weekends, analysing the data at home. I figure if I have a decent 'hit' I can futher investigate it. Because of this, "real time" data is not important to me.

I've been studying GPR technology, and it doesn't seem too difficult. Most of the commercially available units are expensive (I am hoping to sort it out for under $1000 bucks). These units also are designed for purposes that are not my own (like finding re-bar or landmines).

So, I've invested in "Ground Penetrating Radar" by Daniels. It's written as tech text book with lots of equations and big words to say small things. Quite boring, even for an engineer (I'm an engineer). I also had a look at the NEETs (Navy Electrical Engineering Tutorial) collection on line (quite well written and relatively easy to follow... for radar, but not GPR).
For NEETS: http://www.tpub.com/content/neets/index.htm
For waveguides, see: http://www.tpub.com/content/neets/14183/
After reading this stuff, I'm thinking: "This is do-able."

What I'm thinking is to cannibalize a microwave for the High Voltage generator and the magnetron oscillator. Slap the oscillator into a waveguide (a metal box designed in legnth and width and depth by the frequency of wave desired) and I've got a transmit antenna. I figure I can power it off a car alternator run to a decent size inverter. This has the benefits of generating a large signal (like 1Kw as opposed to a wimpy little thing strapped to my back run to a coax antenna). The drawback is I could be cooking myself as I use it, but I'll jump off that bridge when I get there. I've noticed most of the designs for GPR use smaller wavelengths (100 MHZ to GHZ levels) for better resolution. They then complain about depth, so I'm going for depth first, and then resolution when I find something. With microwave (like the oven) transmission frequencies, we're talking KHZ and thus great depth. There will be almost NO resolution, but if I can search down 50 meters, who cares!
So, I haven't gotten far with the receiver, nor the analyses software. By the looks of things, I'm going to have an antenna (waveguide box and some coax cable) run to an ADC (analogue to digital converter). This signal would be compared to a the transmit signal for time. The first receive signal will be from the transmitter. The largest receive signal will be from the ground surface (these antennas will be around 1 meter off the ground). Then a target. Then the ground below the sand. In most cases, no target. Seems simple enough to bury something in the sand for calibration (I hope).

I am on-line infrequently at best. So don't expect immediate gratification from me here.
What I'm looking for is anyone who has built their own GPR has any tips on receiver antennas, wave capture and frequency response analysis. Of course, general comments are also welcomed. No questions of Electromagnetic philosophy, and rude comments about my inept spelling, please.

 

Amazon Forum Fav 👍

Diving for Pleasure and Treasure - Grab it through Amazon!

[99thpercentile]

Can I make a few comments, having used GPR in a desert environment where the USA currently has a lot of people. Just because it is a dry sand doesn't mean that you will have good depth penetration. If the sand has any salt in it, you may have poor depth penetration. The electrical conductivity and the dielectric permittivity are the only useful numbers to know, and they are both frequency dependent. Having built a few other geophysical sensors, I can say that doing this for under $1000 will be almost impossible. I have skimmed the Daniels book, which is the only real GPR book in publication, and also read GPR for Archaeology by Conyers, which has no equations at all. Unless you have some experience with either GPR or Radar, I would probably pass on the idea. The timing systems in GPR are very accurate, measured in nano-seconds, which is not a trivial issue. From your post it seems that you want to build a sheilded system, which might be impossible at low frequencies. And speaking of frequency, in the kilohertz range you would most likely be dealing with electromagnetic induction which is governed by the diffusion equation not radar, which is governed by the wave equation. A target at 50 feet is difficult to find unless it is absolutely huge. You mentioned this project taking 9 years to complete. GPR is never used as a search technique. I would do a magnetometer survey. It is much cheaper and faster. This is just my two cents. Good luck with your project and let me know if you can do it.

 

[HiiiQ]

What I'd like to know is if somebody here could come up with a descent GPR (able to detect small tunels and metal objects @ 3 meters depth.) How much would the development cost & how long would it take. Then maybe a few of us can invest & eventually end up with our own units.

 

[gollum]

Hey SRPNT,

I know pretty much every lost treasure story involving SoCal.

Since you live in Pasadena, the only appreciable Desert Treasure Story that you could hunt using a GPR strapped to the back of an ATV, would beeeeeeee.....hhhhmmmmm.....in the vicinity of San Felipe Creek, just West of Ocotillo Wells, South of the 78? Or is it maybe the Pearl Galleon NorthWest of Borrego Springs, in Coyote Canyon? The Pearl Galleon wouldn't make a lot of sense, as the pearls would be worthless after 300 years in highly mineralized ground. If it is the 10 Oxcarts of Spanish Church Treasure, I could walk you to the area where a lot of relics of that era were found in the 60s and 70s, by an old Treasure Hunter (now dead).

If that is where you are looking, then you will have a couple of huge problems with GPRs. First, they like for their antennaes to be right on the ground (just like metal detectors). The probable location of what you are looking for is an area that has a TON of bushes. Second, most of that area is protected (San Sebastian Marsh). There are people who moved out there, because they know about some things that have been found in the area (mostly unpublished finds). Since there is also so much off roading activity in the area, it is heavily patrolled by BLM and Park Personnel. Locals also keep a pretty close watch. If you need any more info, just let me know.

Best,

Mike

 

[Zephyr]

I'm reading "Ground Penetrating Radar" by Daniels too (copy borrowed courtesy of ILL. I find that most of the book is basically an overview of GPR, but it does have a comprehensive listing of references (useful if you want to get more in-depth info on a particular subject.) I toyed with the idea of building my own, but it's way over my head (math does that to me )
Probably easier pick up a complete vehicle-mounted GPR system really cheap at a gov't surplus auction....

 

[srpnt]

I did warn ya'all that I am a touch slow to respond. I'm a touch tired from rebuilding a Corvair engine ('67 - 140HP, if anyone cares) on a dune buggy. I've also been reading and working on roughs of GPR (still at the cocktail napkin level). I'd be futher along, but I have to work for a living, too.


Gollum, "not telling" is not telling.
HiiiQ - 3m? Buy one. There is some links on similar GPR posts in this forum. There are GPR systems, reasonably priced if you have a specific search area. You may be able to rent them, or even borrow another TH's (for a cut of the treasure, of coarse). For any GPR unit, you will need to be trained to read the signals. They can be quite complicated.

99thpercentile,
The major problem with a magnetometer search when based on a vehicle is interference from the metal of the car. From what I understand, magnetic searches based on signal strength don't have the depth capabilities I require.
The beauty of a lower frequency GPR unit is penetration. Resolution will almost be non-existent, however, if you are looking for an aberration you only have to find the odd blip. Futher mapping using a higher frequency could then determine viability. This could be done by renting a GOOD system, but I don't think I can afford the kind of system I'd need for my search or the rental costs for the time it would take.
I understand that you would say that a shielded system would be impossible at lower frequencies, but that is only because as you decrease frequencies, you must increase waveguide size. Eventually you hit a limit on tangible and useful size (it's a design trade off). The beauty of the shielded system is cleaner signals.
The concept I'm working on, vehicle mounted, means that I'd be looking at the legnth of a car as the upper limits for shielding. The benefits of the lower frequency is that it is easier to work with in hardware (transistor rates, et al).

Using a directed antenna on a GPR can avoid major interference. With an engine you would be looking at interference rates around 20kHz to 6kHz (for a 6 cylinder engine), so using a frequency outside this range would also be beneficial. I'm thinking that puts 30kHz as the lowest possible limit. More reasonable frequencies are in the 500kHz+ range.

As for the 1000 bucks:
The $1000 is hardware. I'm not paying anyone else for their expertice, because I'm a cheap bastard.
a Test-bench with a decent digital rate and USB connection can be expensive. I've got the laptop to play with, thankfully.
Antennas can be home-made.
Receivers are just combinations of widely available circuits. (amplifiers, freqency locks, etc.) Most of the basic building blocks in a standard GPR system can be found on line or in various circuit books. The key is to lock the received signal and determine amplitude. I am thinking about using a frequency lock with phase shift analysis. I vaguely remember making one of those while as an engineering student that would lock a frequency, and as the signal shifts, it outputs a voltage level.
In theory, such translations can be done very quickly in hardware.
From there, the received signal is run through an ADC (analogue to digital converter) and compaired by time so that a calculation can be done on depth. If software is used for this sort of task, it may increase the time delay as it increases the amount of data bits that would need to be loaded to the computer. For me, I don't need real-time data, so I am comfortable with simply uploading a lot of information (including noise signals) to a computer without processing. This is the information a computer would need to make a plot. Personally, I'd perfer to look at a Matlab plot to the standard black and white hand held screen. Overlay plots would make target identification easier.
Assuming there are only 3 main signals: surface reflection, treasure, everything beneath that. There will be other stuff to cause interference and the better hardware can differentiate between the treasure and trash minimises the amount of data required.

I'm trying to run the numbers for data rates on USB (approx 33MB/s) vs. data rate requirements, but I'm getting groggy tonight. I'm looking for a time-frame of data coupled with signal strength based on the ADC (more bits on an ADC means better resolution and more probability of noticeable variation), but I think I'll give it a miss tonight.

magnetometer.... humm... I think I will look into that more (I"m not fixated on GPR, it's just seemed like the best option so far). Thank you for the advise.
-Henry
ps. here's my new TH tool:

 

[gollum]

Hey,

I wasn't asking, just figuring, to see what would be the best setup for your search area. I know most every story in SoCal, and hiked most of the areas from the Chocolates to anywhere in Anza-Borrego Desert State Park. If I'm correct, you'll need something fairly accurate to 6-8 feet minimum (the Spanish buried the cross at about 6 feet, and add for 196 years of wash and wind blown sand). If you have a good Pulse Induction Machine, you can buy or build a 48" X 48" Land Coil, and put tall lawnmower wheels on it. This will have the same probs as a Towed GPR (bushes and scrub). Unless you're over around Algodones Dunes (Glamis Area), you will have a problem with brush. North of the 78 (between Ocotillo Wells and Borrego Badlands) has a lot of open sand, but nothing worth hunting. Nothing down near the Mud Hills (Naval Bombardment Range). Everything else West of Salton Sea is either Private or State Park Land.

Wherever you are going, be careful, and don't let the Park Rangers get you. They watch San Sebastian Marsh area closely.

Best,

Mike

 

[Sheldius]

Resolution decreases as you use larger waves (kHz) which means you can't see differences small than X cms or meters. For kHz GPR, it would have to be a VERY large target (like many meters across) or you wouldn't even get a blip against the background.

And the FCC won't allow the GPR companies to put to much power in the US units because it messes with cell phone and radio transmissions (so they say). You might run afoul of then if you knock out a channel for a while.

There are backpack Magnetometers with GPS for $18K(and they have been used to find caves and tunnels). You could also have an airplane magnetometer (aerial mag), but that is more of mineral resource targets.

Good Luck and HH

 

[Functional]

I wonder if he decided on which approach to use? Wish I knew electronics, so I could build some of the things I've seen. I know I can solder stuff together, but mostly its been only automotive electronics.

I'd like to see someone come up with a few simple electronic circuits that a person could assemble and plug into a laptop, or PDA and let the computer software do the work. I wouldn't mind trying to build an MD, resistivity device, magnetometer, or maybe a GPR, using the sound card in a laptop.

Anyway, as for the current situation, maybe one of these links will help.

DIY Magnetometer Links:
http://www.sas.org/E-Bulletin/2004-04-02/wanderings/body.html
Actual Plans for DIY Magnetometer:
http://www.portup.com/~dfount/proton.htm

Free Engineering & Science Software Links:
http://www.measurementdevices.com/index.php?name=Web_Links&req=viewlink&cid=95

Free Engineering & Science Software:

Oasis Montaj by Geosoft Inc.
http://www.geosoft.com/
Freeware:
http://www.geosoft.com/pinfo/oasismontaj/index.asp

electrical resistivity:
http://www.earthdyn.com/resist.html
Several Links:
http://geophys.geol.msu.ru/rec_lab3.htm

Jeff Daniels' 2D/3D/4-C
Interpretation and Display Software (IDL Programming Language)
http://www.geology.ohio-state.edu/~jeff/software.htm
And:
Finite Difference Time Domain Modeling Code for GPR Modeling
(Roger Roberts' Ph.D. "C" Code)

Council for Independent Archaeology:
http://www.cix.co.uk/~archaeology/cia/resistivity/resist.htm

A Windows 95/NT Program to Download Data from the Geoscan FM36 Gradiometer and RM15 Resistivity Meter.
http://www.nd.edu/~mschurr/nddown/index.html

Snuffler - Freeware Geophysics Software:
http://www.homeusers.prestel.co.uk/aspen/sussex/snuffler.html

Processing GPR Data and FREE SOFTWARE:
http://www.du.edu/~lconyer/data_processing.htm
GPR data processing example:
http://www.du.edu/~lconyers/gpr_data_processing_example.htm
Download (49.3 Mb):
http://www.du.edu/~lconyers/petra_data_and_processing_programs.zip

Archaeological Research Resources;
Remote Sensing & Geophysical Survey,

A SOFT RECEIVER FOR VLF USE:
http://www.vlf.it/trond2/softreceiver.html
http://www.vlf.it/trond2/list.html

Using a PC with soundcard as a VLF receiver:
http://www.qsl.net/dl4yhf/vlf_rcvr.html

GPS Resources ,
GIS and Digital Mapping:
http://www.har-indy.com/Links.html

Free geophysical software by Markku Pirttijärvi:
http://www.gf.oulu.fi/~mpi/softat/

Spectrum Lab A specialized audio analyzer, filter, frequency converter and data logger that can be used as a VLF receiver. Data can be viewed as a spectrum plot or waterfall display.
http://www.ukaranet.org.uk/software/index.htm

Low Cost Geophys:
DIY Resistivity:
http://www.wiganarchsoc.co.uk/content/News_Letters/news029.htm

Mother Load Locator:
http://treasurenow.com/html/MLL.html


If all else fails, maybe using seismic imaging, (like archeologists sometimes use), would work?
Hope something here helps.

F.

 

[urbantreasure]

Not to be a thrill kill or anything and these posts are 7 years old so maybe it never happened but if the area to be searched is that big and requires GPR to possibly find something then I doubt too many people would try to horn in on the action!

 

[dabenpb]

GPR/GPSAR expert here ...

Has anyone made progress here? I have a product we use which could be applicable.

Not to be a thrill kill or anything and these posts are 7 years old so maybe it never happened but if the area to be searched is that big and requires GPR to possibly find something then I doubt too many people would try to horn in on the action!

 

[lyon46]

dabenpb, Would that be the drag along unit seen on a recent GOLD RUSH episode with the Dodge Bros? Rick