Tweet
Nice detector, cheers.
-
-
How,
If anyone is considering building the Lobo-Lite, then there is one thing that needs consideration. The sensitivity circuit needs a little work, I ran out of time and didn't fix it.
As the circuit is above the maximum sensitivity is when the pot is at about 75%. It appears that full sensitivity swamps downstream circuits. The 2k2 before the sensitivity pot needs to be increased a bit more....how much? I'm not sure and don't have the time to check....I just make sure I don't turn my pot on more than three quarters.
Here are a few air tests, target weights in grams. The 5" round would probably get a 0.02 or 0.03 at a similar distance to the 0.08 crystalline piece, I didn't have anything smaller handy to test with. (Note also an obvious mistake (sorry) on the graphs the smallest one is 0.08 grams and not 0.
This machine will see many if the flat nuggets on edge at very respectable distances too, much better than a PI, this is exactly what you want for a bedrock sniffer where many of the pieces will be sitting vertically in the cracks and crevices.
Cheers
Kev.Comment
-
It's probably difficult changing the values to the sensitivity part of the circuit. Your signal is referencing op amp ground with resistors above and below the pot. I haven't done the math/simulations though. This is just my casual observation.Originally posted by Kev View Post
Changing these resistor values could limit the available sensitivity of your front end. A better choice imho would be to implement an automatic gain control (AGC) circuit. Being somewhat engineer minded I'd put in a discrimination circuit at the same time to make this detector a bit more versatile for something like coinshooting.
I haven't waded through this entire thread - who designed this circuit i.e. where does it come from? Also, it would really help to see a video air test to see what the detector sounds like at the different sensitivity settings i.e specifically when it is being "swamped" by a signal.Comment
-
Hi technos,
You are correct it may not be as simple as changing values, but I added the 2k2 and it did help a little. What happens is that sensitivity increases as you turn the pot up, but once past the three quarter point sensitivity starts to drop off. It appears to be either loading the second input amp or upsetting the following integrator stage. I have moved on now, I am working on a couple of PIs one of Bugwhisker's and Tinkerer's so if anyone wants to solve the issue that's fine. It also needs a compressor to limit large responses, and when you first turn it on you can't have the headphones on as the battery level indicator blasts your earwax onto your eardrums.
This thread introduced the Tesoro Lobo a discriminating detector designed primarily for nugget hunting, I removed the disc, thereby making the Lobo-Lite as I needed a simple compact bedrock detector. I have been able to adjust the circuit to give maximum response without having to worry about phase relationships needed for disc, only ground balance was my concern. In fact disc is not needed if you are truly serious about nugget hunting, as many bits of ferrous junk in bedrock cracks are accompanied by nuggets and fines. Near surface junk can alert you to a hard to see crevice that has undetectable gold buried deep down beneath the junk.
With the TX set for about 18 Volt peak to peak the frequency is around 21kHz, this is a bit too high for a coin shooter, notice it only gets the nickel at 12 inches with the elliptical coil. Admittedly this is equivalent to a 7 inch round and not that bad, but the TGSL project on this forum will get a nickel at 14 inches with a home made coil, one of the reasons being its lower operating frequency.
Thank you for showing interest in the circuit techos and your comments are most welcome.
Cheers
Kev.Comment
-
Hi Kev
nice project, are there any extra C's or R's in the stock coil to worry about?Comment
-
Gid'day 6666,Originally posted by 6666 View Post
That I can't be sure of, I have some GB coils and they are locally available so cheaply that I decided to not even bother making one of my own.
I just checked the 10"x5" RLC including cable, but don't have time to evaluate any embedded caps sorry.
RX
R = 55 Ohms
L = 9.25 milliHeneries
C = 1.39 microfarads
TX
R = 2.8 Ohms
L = 890 microHeneries
C = 800 nanofarads
If you build one and the sensitivity is rubbish and GB doesn't work with the 1nF cap across the RX coil then maybe you'll need to start adding some more. The waveform is sampled on the downslope which is not the ideal according to Carl and Georges' book, but no matter how I tried I lost so much sensitivity shifting the phase 180 degrees that I just decided to leave it as it was, and just tweak for maximum signal strength. Perhaps this indicates Fisher use an embedded RX cap?
Cheers
Kev.Comment
-
PLEASE Note; the Lobo-Lite will only work with the Gold Bug, Gold Bug SE, Gold Bug DP, and Gold Bug Pro coils, or home built emulations of such, designed to operate around 19kHz.
It will NOT work with Gold Bug-2 coils (Gold Bug 11) which has an operating frequency of 71kHz, unless the filters and resonant caps are modified also.
Cheers
Kev.Comment
-
Could you please wrap up the project into a single zip archive for the rest of us to play with? the PCB layouts and stuff? I am about to make a complete LF project, eventually, but this solution gave me some ideas to shrink it a bit and release a simpler version sooner. I think LF solutions deserve more attention.Comment
-
Well done, Kev.
Thanks for sharing your working example.
Yes, please. I second Davors request. A working example like your Lobo/Diablo fusion is a nice working example of a GB'ing VLF nugget machine.Originally posted by Davor View PostComment
-
Hi Guys,
Sorry I don't have time to support this design, the time it would take me to generate placement and layout info would be way more than I have available. What I can try is to give you a simplfied version of the program I used.
This program will not print out artwork at 100% so I print it to a pdf printer at 200% and then when I print the pdfs out onto mylar I reduce them by 50%. I have included the artwork pdfs and you will be able to use the program to find out the component values and placement data from there.
Here's how to begin.
Extract the archive in the 'C' root, Lobo-Lite dir will be generated.
Start Tar3001.exe and click on the Target shield when it appears to speed things up a bit
It should give the option of loading the Lobo-Lite_V1.T3000 file
When the scheme opens press crtl + K keys and change the schematic window background colour to a darker one, hopeless trying to see it otherwise.
I had to strip the libraries to make it small enough to upload. If you have a later version of Target or another PCB Pool program you maybe able to import this file and generate gerbers, please make them freely available to others if you do so.
Soldering components like connectors and electrolytics on the topside of the board can be a challenge since the holes are not plated through. I usually just put plenty high quality no-clean flux on the leads, place the lead ends of the component just through the hole and solder the top and bottom sides. Heating the lower solder joint for a while will usually allow enough heat to migrate to the upper joint for you to push the component all the way home, and it should still have a good joint on the topside.
Cheers
Kev.
P.S. The file is 3.9MB and the limit is 1 so I'll need to find a file host that I'm comfortable with....back soon.Comment
-
Howdy,
I have the files uploaded, but I've discovered yet another issue, the program was designed for XP and will not run on later systems unless you run it in compatibility mode. If you get an error message when trying to run the executable, right click on it and select "Properties", then the "Compatibility" tab, and then you should see something like, "Run this program in compatibility mode for XP" etc., select that and it may come to life.......
If the board file doesn't appear in the start assistant window, click "Open existing file", and it should appear.
Have fun.
Kev.
http://ubuntuone.com/6JvHIul5wcuEuYkDdfz1LkComment
-
It can be opened with PCB-Pool current version of Target 3001 software. It complains about some missing libraries, but shows a PCB as you printed it. Unfortunately it can't be transformed to any other format. At least not from the free version.
Anyway, it's a start. Thanks.Comment
-
Hi,
Pitty about that Davor. They have a script to convert to Eagle, but that would be with full versions of the programs.
http://server.ibfriedrich.com/wiki/ibfwikien/index.php?title=Convert_TARGET_3001!_to_Eagle
You could try importing the Lobo-Lite library into your program.
The board file has a TLC27M2CD listed for the integrator, but I used the LF442 with great results, but then again, the former may operate better with the sensitivity arrangement as it stands?
Also a BOM can be generated through the File - Input/Ouput Formats menu, some of the parts may have the Element14 pt#s I used.
Cheers
Kev.Comment
-
Gi'day,
I see in my haste that quite a few critical components don't have Si part numbers on the board files, and to achieve similar results as me, and even better, then it's important to use the most suitable parts in these situations, such as 2% PPS caps for the tank etc., I'll edit the board file and make it reflect more what my final design used and upload it when I get time.
Cheers
Kev.Comment
-
Originally posted by Kev View Post
You know it's really difficult to know what this circuit is doing (and wrong with the sensitivity) without having it in hand. Measurements really need to be taken with an oscilloscope. The problem could be any number of things. One thing I found really troubling is with the 7660 inverter chip. Looking at the datasheet I see a very high ripple on the V minus (500mv) at oscillator frequency. I consider that a major flaw. But I do not know if it is on here without building the circuit.
I think that when you say "compressor" circuit, I think that may be the same as me saying "AGC" circuit.
Yes this is true. But we know that the shallower depth here is mostly a function of frequency here (besides the coil). I have more than one idea concerning this. I was thinking of a detector like the Garrett infinium (http://www.kellycodetectors.com/garr...infiniumls.htm). Something that could be a very lightweight "backpack" detector without all the weight. But you are right if you want a primarily nugget hunter. But I keep on thinking of circuit improvements, and one thing that might be very interesting is a "dual channel" ground/iron reject. It would be nice to find a way to reject bullets, but this might not possible in a nugget detector (as you noted there could be gold underneath).With the TX set for about 18 Volt peak to peak the frequency is around 21kHz, this is a bit too high for a coin shooter, notice it only gets the nickel at 12 inches with the elliptical coil. Admittedly this is equivalent to a 7 inch round and not that bad, but the TGSL project on this forum will get a nickel at 14 inches with a home made coil, one of the reasons being its lower operating frequency.
I'm like you and do not have much time to spend on this circuit at the moment. But the things that could be done to improve the design;
1) Fix the sensitivity with an AGC circuit. This will prevent detector being "swamped". Make sure each stage has proper gain. To this end a different phase detector might be better.
2) Redo the V- minus circuit and power circuit. Find different way to make V minus so 7660 can be removed. Redo power circuit so detector can operate at lower voltage and use fewer batteries.
3)I like the ground reject, but maybe it could be improved when phase detector is replaced.
4) I don't like how the tone is derived by the CD4024 as this might put harmonics back into circuit. Generate tone indirectly. But maybe this is not an issue here.
So by time I have done all this, it's a different detector
.
If you decide to use a different board, or want to redo the detector, give me a good price so I might take measurements from it (as I do not have time to spend making the board right now).
Regards
technosComment
Comment