Coils

[Cody]

If we take a look at the illustration that Bill posted below we see a detector that is multiple frequency. The RX coil is square wave driven and we see a fundamental frequency at 5khz and third harmonic at 15khz. The third harmonic has 1/3 less amplitude than the fundamental in the illustration. So we have this problem of power in the harmonics.

What I like about the smart coils is instead of driving the detector with a square wave which would take us back to a modified type of Sovereign or Explorer they opted to do something interesting. They have two fundamental frequencies matched to the coil by using a sine wave at two different fundamental frequencies. This looks to me like both the stock coil and the new higher frequencies coils will have maximum power. That eliminates the harmonic power problem and leaves us with the advantages and disadvantages of any single frequency detector based on the operating frequency.

I think we will find the higher frequency is going to respond as any good top detector at the higher frequency and also at 7.khz. We still have the problem of ground balance to iron ferrite or wet salt but then that is life in the fast lane.

Also, pulse induction has used a single coil for TX and RX since Ben tried to kill himself flying a kite.

What do you think?

 

[Ralph Bryant]

I couldn't name them off the top of my head, but there are several PI type detectors that have used coils with separate TX and RX windings of different configurations. This supposedly helps the machines run more quietly, but there are some advantages to either single or multiple winding types. I know there are both mono and DD types available for the Minelab SD and GP gold machines with several different companies making custom accessory coils for those, some of which are literal MONSTERS. I've seen a couple used in the Australian Outback that were so large they had to be built onto skids so they could be harnessed and drug along the surface. Did I hear someone mention D E E P !!??!!

Received the Xterra-30 this morning and hope to get some initial impressions posted up by the weekend, and hopefully add to the thoughts of others about these machines. I still can't believe they got a Minelab in that little box !!

Cody, is what you are saying is that the Xterra-50 appears to be "dual frequency" but only operates or processes at one of those frequencies at a time ? Or does it actually "switch" from one to the other depending on the input from the coil ?

Ralph

 

[Cody]

"Cody, is what you are saying is that the Xterra-50 appears to be "dual frequency" but only operates or processes at one of those frequencies at a time ? Or does it actually "switch" from one to the other depending on the input from the coil."

It actually switches from on to the other depending on the input from te coil. Dual frequency from a square wave has to deal with harmonics and less power the higher the harmonics. This is really what I have been saying is the X-terra is going to pack maximum power in both frequency. I think this is really interesting. Instead of driving it was a square wave and using harmonics they are using dual sine waves, in a sense, so will have maximum power in both frequencies since they will both be fundamental frequencies.

I cannot believe they don't track other parameters in the coil such a Q but only they know for sure.

So, two frequencies, sine wave, switched. (can be expanded as follows.)

So how about in the future? Start with frequency A, divide it down to frequencies v, w, x, y, z, then multiplex the inputs to one output, use a signal from the coils, 4,2,1 in binary to instruct the mux which one to send down. All you need for this is a chip in the coil that sends the binary code to the mux to select one of the inputs, v, w, x, y, z to the output. The sine wave generator needs to run at a frequency higher than the highest frequency. Easy as falling off a log.

 

[Cody]

n/t

 

[John(Tx)]

Just felt something fly over my head! All I know is that my XT50 air tests very close to my CZ6a with just a slight edge going to the 50, which I am very pleased with. In my small test garden, which has some pieces of rusted nails, my 50 will hit the 6" buried dime with a "33" while my CZ says it is iron. I can hardly wait to see what the 18.75 coil will do on small gold. HH John

 

[Digger]

I admit that I am by no means as well versed as some of you folks on the electonic theory of detectors. And, I admire your level of expertise. But to me, in layman's terms, VFLEX offers the ability of the control box to match the parameters required of the coil being used. According to Minelab, the X-Terra will not work with any previous models of Minelab coils. That indicates to me that the software is written for 7.5kHz and 18.75kHz only. By following the startup procedures, you can get a general idea as to how it works. With the control box turned off, the chosen coil is attached. When the control box is turned on, the micro-controllers in the control box communicate with the micro-controllers in the coil over a digital data link. If no coil is attached upon startup, the LCD in the control box tells you CU, coil unplugged. If a coil not designed for the X-Terra is installed at startup, the LCD tells you CI, coil incompatible. If a coil that does not communicate properly with the control box is installed at startup, the LCD will say CE, coil error. Once you properly connect a working coil that has been designed for use with the X-Terra, the detector is turned on, and a proper data link can be established, the digital signaling between the coil and the control box informs the control box as to which coil is being used. The control box then sets the correct operating parameters for that coil. So, based solely on my simple understanding of the complexity of VFLEX, it seems to me that both the 7.5 and the 18.75 parameters are both there, waiting for instructions from the coil as to which one gets to "go to work". But, only one of them is chosen at any given time, depending on the "instructions" sent from a specific coil.
So,imho, the X-Terra is a single frequency detector that has the ability to operate at more than one frequency. But only one at a time, depending on the instructions sent over the data link, by the micro-controllers in the coil. I have to think that the control box is prepared to process either frequency at start up, simply to determine which frequency gets the "boot"(a computer pun). But, will only pass data designed for the frequency requested by the coil. Then, the control box runs the program designed to communicate with a specific coil and set the parameters designed to work with that coil. Then again, this is just a guess. I could be wrong. HH Randy

 

[bill(de)]

but think a boomerang flew over my head on rest of the posts. However i love all of it and reminds me of this but eventually iuke: ask for do a little and

Nice post Digger

HH Bill

 

[Rudy]

Cody,

A few things about the train of thought you've been posting.

For best (strongest) Tx performance, you don't really need maximum power transfer. You really just need to maximize the current into the Tx coil. That is what generates the electromagnetic induction field.

Your interpretation of the signal power in the harmonics being less than the fundamental frequency is correct and is given by the Fourier series expansion (I think you alluded to it earlier). I don't know exactly what kind of signal the Explorer generates, but I once saw an oscilloscope signal of the DFX operating in dual frequency mode and it was not a constant frequency square wave. It looked like a train of square waves, with the fundamental frequency hopping every couple of cycles, between two different periods. Although I didn't sit down and computed the Fourier spectrum, I do believe that such a Tx signal train would achieve better signal strength of the two frequencies used in the DFX. Now, in the case of the Explorer, which is claimed processes many more frequencies simultaneously, they are either producing a signal reminiscent of an FM modulated carrier, or are indeed suffering a terrific signal energy loss becasue the harmonic amplitudes decrease by the same factor as the harmonic frequency.

Lastly, in regards to your schematic for frequency division, I simply don't know of a cost effective and energy efficient way of dividing the frequency of a sine wave to do what you envision. It's easy enough to divide down a digital pulse train at low energy levels (ie. a flip flop), but I can't see how to divide down the sinusoidal transmit frequency so it is readily available for use on a cycle by cycle selection. It would be much easier to just divide down the clock that drives the ROM describing the digital representation of the sine wave being synthezised.

(For those not skilled in the electronic arts, my apologies).

 

[Varmint]

Here is my "favorite" method of sine-wave generation from a digitally clocked source. Obviously, the entire circuit is tuned for one specific frequency, otherwise the multiple-feedback filters will break into all manner of spurious oscillations.

If you happen to shop in Home Depot or other places using the IVI-Checkmate, Ingenico, or NCR branded signature capture credit card terminals, this very circuit is providing the 140KHz sine wave drive to the touch screen digitizer.

Ignore the lower group of components, this is a 3.3V generator to drive the FLASH chip used to store all manner of digital info for the OS.

The first resistor/capacitor combo convert the incoming square wave (output compare from the U-controller) to a triangle wave to remove the fast edges, then the 2 stages of multiple-feedback filters round everything out nicely. THD was below 0.10% worst case, perfectly adequate for the task at hand. The 2.5V reference is to place the sine at the mid-point of the power supply in this 5V single-rail design.

Something similar could be used with an analog mux, and since all of this is surface mount tiny stuff, the real estate consumed for the section at the top is on the order of .7sq inhces since the op-amp shown is a dual in a MSOP-10 package, and the discretes are all 0603.

I would hope that the designers would use something along this line since a ROM and a state machine to drive it would obviously be of higher cost, and still requireing a filter stage to remove the DAC "edges".

Analog electronics is NOT dead, people just tend to consider digital is the end-all be-all for electronics.

DAS

 

[Varmint]

Here is my "favorite" method of sine-wave generation from a digitally clocked source. Obviously, the entire circuit is tuned for one specific frequency, otherwise the multiple-feedback filters will break into all manner of spurious oscillations.

If you happen to shop in Home Depot or other places using the IVI-Checkmate, Ingenico, or NCR branded signature capture credit card terminals, this very circuit is providing the 140KHz sine wave drive to the touch screen digitizer.

Ignore the lower group of components, this is a 3.3V generator to drive the FLASH chip used to store all manner of digital info for the OS.

The first resistor/capacitor combo convert the incoming square wave (output compare from the U-controller) to a triangle wave to remove the fast edges, then the 2 stages of multiple-feedback filters round everything out nicely. THD was below 0.10% worst case, perfectly adequate for the task at hand. The 2.5V reference is to place the sine at the mid-point of the power supply in this 5V single-rail design.

Something similar could be used with an analog mux, and since all of this is surface mount tiny stuff, the real estate consumed for the section at the top is on the order of .7sq inches since the op-amp shown is a dual in a MSOP-10 package, and the discretes are all 0603.

I would hope that the designers would use something along this line since a ROM and a state machine to drive it would obviously be of higher cost, and still requireing a filter stage to remove the DAC "edges".

Analog electronics is NOT dead, people just tend to consider digital is the end-all be-all for electronics.

DAS

 

[Varmint]

feel free to enter the schematic as shown, use 140KHz for the SQ Wave input, (don't forget the 2.5V ref) and run it for yourself.

Far more cost effective than a lookup table!

DAS

 

[Cody]

Good post, I agree with your observations. I take some "preaching" room in pontificating in these post. I have looked at the patents and drawing for the Explorer so have a decent idea there but have not seen anything on the X-terra.

I have been under the weather this last week but am going to look at some material by Payne on coil design and Minelab's patents on coils to see what they are doing. I have never designed one so try to compare them to transformers and other coils I have worked with.

The illustrations are just an attempt to convey a concepts in visual form. I realize this is not the place to dig deep into electronics so try to keep that in mind also.

 

[Anonymous]

This frequency thing about high vs. low is a little overstated. There is no problem countering salt, and it can be handled with a slightly higher frequency. In actuality SALT is a positive signal, like metal and must be rejected. And one of my Pulse units has a Tri-planar concentric coil.
A good example of what I'm talking about is the 1280 Aquanaut, that operates below the VLF range down into the VF [voice frequency] range. By modern spin the Aquanaut operating at such a low frequency should be horrible, but results say otherwise I've got some thin band gold rings, and platinum buried, and can only barely hit them with a Sovereign. Its also happened at the beach with my partner using a Sovereign or a Compass X-100. Yet I've pulled coins up, and Minie balls with my Sovereign so deep it staggers the imagination.

 

[Anonymous]

Compass made the X-200 in the early nineties and it gave you a choice of two separate frequencies, with the same coil. The X-terra would really have been more interesting had it emitted both frequencies at one time. With a microprocessor and Minelab's knowhow I'm sure it could have been very innovative. And somehow I believe these new two detectors are the beginning of something provocative. Do you see anything on the horizon, Ralph, or Cody?

 

[Cody]

When I say maximum power transfer I am talking about the amplitude of the drive signal to the coil and the current that exists in the coil, P=IXE. Are you suggesting the amplitude of the sine wave is of no importance? It has been about 15 years since I taught this but the internal impedance of the transmitter and load impedance in a simple series circuit applies or I am missing something. I realize the coil is more complex than a simple coil with no secondary such as a detector coil so am not sure I understand. The amplitude of the sine wave applied to the coil is what drives the current in the coil. My assumption is up to a point this maximum transfer of power is an important factor.

I would appreciate an explanation so I can correct my thinking in this area. I mean this statement seriously as I think you can see I am very interested in the theory as well as the application.

 

[Ralph Bryant]

Your guess is as good as mine Vlad..... But regarding multiple (separate) frequencies from a single machine/coil combination, Minelab has had such a machine on the market for years in the form of their current Eureka Gold and previous 18000, 17000, 16000 series and Golden Hawk models that operate at a user selectable 6.4, 20, 0r 60 kHz using the "same coil". I guess that's why I wasn't really all giddy and enthused about the VFLEX/frequency idea on the face of it.

Ralph

 

[Rudy]

Cody,

What I am trying to say is that what makes the Tx side do its work is the ammount of current flowing in the Tx coil winding and since this is a battery operated machine, one wants to do it as energy efficiently as possible. In other words, maximum current at the lowest energy expenditure.

Everything else being equal, a sinusoidal current source driving the loop, may be a better approach from an energy consumption point of view, instead of a source that is impedance matched to the loop. Another benefit is that the current source approach would result in a higher overall circuit Q.

 

[Cody]

The Explorer does process 28 frequencies. I know that from going over the drawing and the patents. It is the more interesting of all the detectors I have taken a serious look at but is time domain pulse induction. I am really impressed with the operation.

All we have is frequency domain and time domain but the implementation of the process is dramatically different. When it comes to TID and display.........take your pick.

If you enjoy the technical side of the house I would look at the patents for the Explorer that include time domain, frequency domain, and coil design. I think you understand my interest is as much technical and theory as application. Like all users in the field I just want it to do a good job.

I am as interested in the Ace series as the X-terra but there is not forum for the Ace so I try to stick to the featured detector. Brands are a nuisance to me to be honest concept to me since I don't have to make money based on who's brand is on the box. I would use one from Beijing if it was a better detector.

 

[Rudy]

Varmint,

There needs to be a microcontroller in the control box to perform many of the other functions and they are available with a wide variety of embedded peripherals, like on-chip D/As and A/Ds, flash memory , ....
With these capabilities, I would think that a lower cost and component count can be achieved by generating the Tx signal through mostly digital means. Change the coil, change some of the software parameters and whoila. You now have it operating at a different frequency.

If you try to do it with mostly an analog design it gets more complicated. You now have to switch to different value resistors or capacitors, or both.

 

[Ralph Bryant]

....."I would use one from Beijing if it was a better detector.

Cody"........

HEY! I think you may have just stumbled upon the answer to my target response speed dilema. If these things are really being made in Australia, and you take the speed of light and/or electronic pulses, crunch the numbers, and if the calculations work out right, the Australian made machines should in theory be just a little faster than the Chinese machines due to the proximity of the place of origin in comparison to the U.S. made machines.



Ralph