More FBS Geekery

You are listening on the TX wave.

And it is shown to transmit many harmonics that changes by selecting different NC channel.

But..

How many does the detector process on the RX at any given time and on any given NC channel.

That is the question and may hold the clue to the better gold respons on channel 11.

Albert, there may not be a noticeable difference. I am merely speculating on what I see.
My understanding is that the 25KHz produces harmonics that reach all the way down to ~1.5KHZ. Look again at the second picture in my first post. The spikes (harmonics) are signals at the frequency listed under the waveform.
Pete

frnifo. MineLab says that they monitor 28 frequencies within the spectrum. I have no way of knowing if they look at them one at a time or all at once.
Pete

I think the processing power is a key to understand how many they process.

What kind of processor would it take to listen to 28 at the same time all the time.

I am a strong believer in that they process two at any given time and changes the two much more dramatically compared to other detectors.

So much that is is visible in sensitivity when testing both ends of the conductive scale. Both 1 Dollar silver coin and 1 Dollar gold coin ( on edge ) or a small 14k pendant.

Yeasty Wrote:
-------------------------------------------------------
> My understanding is that the 25KHz produces harmonics that reach all the way down to ~1.5KHZ.

Harmonics go up in multiples of the fundamental, not down. A 25kHz signal can produce harmonics at 50k, 75k, 100k, 125k, etc, but not at 1.5k.

> MineLab says that they monitor 28 frequencies within the spectrum.

Minelab says that? Are you sure?

Regarding the way in which the frequencies vary with channel no. : The waveform you can see on the oscilloscope traces is digitally generated, from software timing methods (though it could be digital logic chips). The timings are precise multiples of the 'basic' block - the time for a half-cycle of the 25KHz wave. All the other timings are multiples of this. The burst of 3KHz is made from 4 'basic blocks' high, 8 blocks low, 4 blocks high. This gives the precise 8-to-1 relationship between the 25KHz and the 3.125KHz signal. When you change channel, ALL the frequencies go up/down in step, as only the 'basic block' timings are changed. If the 3KHz goes up by 10%, so does the 25KHz.
Even with modern DSP processors, I find it hard to believe that '28 frequencies' are examined. There is no need for it, for one thing, and it would be too complex to really make sense of, as well.

Perhaps a better word for "harmonics" would be "sidebands".


Here's what MineLab says about FBS:
Minelab’s FBS technology simultaneously transmits 28 multiple frequencies from 1.5kHz to 100kHz.

FBS (Full Band Spectrum) simultaneously transmits, receives and analyses a full band of multiple frequencies. This provides the detector’s electronics with even more information about a target and the surrounding environment than is possible with single frequency or BBS technology.

If I had a better spectrum analyzer I suspect that the transmitted spectrum would look something like this:





Edited 1 time(s). Last edit at 02/13/2012 10:41PM by Yeasty.

Pimento Wrote:
-------------------------------------------------------
> Regarding the way in which the frequencies vary with channel no. : The waveform you can see on the
> oscilloscope traces is digitally generated, from software timing methods (though it could be
> digital logic chips). The timings are precise multiples of the 'basic' block - the time for a
> half-cycle of the 25KHz wave. All the other timings are multiples of this. The burst of 3KHz
> is made from 4 'basic blocks' high, 8 blocks low, 4 blocks high. This gives the precise 8-to-1
> relationship between the 25KHz and the 3.125KHz signal. When you change channel, ALL the
> frequencies go up/down in step, as only the 'basic block' timings are changed. If the 3KHz goes up by
> 10%, so does the 25KHz.

This is correct. Like the bellows on an acordian.

Yeasty Wrote:
-------------------------------------------------------
> Perhaps a better word for "harmonics" would be "sidebands".

Well, no, 'sideband' means something else entirely.

> Here's what MineLab says about FBS:
> Minelab’s FBS technology simultaneously transmits 28 multiple frequencies from 1.5kHz to
> 100kHz.
>
> FBS (Full Band Spectrum) simultaneously transmits, receives and analyses a full band of multiple
> frequencies.

In neither of these statements do Minelab say that they "monitor 28 frequencies."

Many years ago when I started looking at how BBS works, one of the things I noticed is that Minelab consistently stated in advertising that they "transmit" 17 frequencies. No where in advertising did they ever state they actually "analyze" or "process" 17 frequencies. That omission was not a mistake, and their FBS advertising continues to avoid making that mistake. Even the second quote above artfully dodges any explicit claim of analyzing 28 frequencies... it is left to the reader's imagination that they said something they really didn't say.

In your opening post you show pics of the transmit voltage waveform. If you want a clear picture of what's going on, take a look at the transmit current waveform. If you do so, you may go "aha!"

- Carl

.



Edited 1 time(s). Last edit at 02/13/2012 11:32PM by TerraDigger.

I was going to suggest you try that, but now Carl has said it too - the current is the important thing, that's what generates the magnetic field. Current and voltage aren't in phase on a single-frequency machine, they're certainly not going to be so at 'two frequencies'. Don't forget the receive coil, either. Coils are differential transducers, that is they respond to the change in field, in a frequency-sensitive manner. Phase angle and signal level will vary at the 'two' frequencies.

Commendable as to how (logical process of reasoning/analyzation) this thread is unfolding (discovering).......... with possibly one final conclusive resultant............ but maybe from different (and emotional) angles.

And if.............. by chance................ Carl's brain is 'triggered'................. even with just one 'spark' (idea).............................(((and this is to include a tangential non-related trigger..... but stemming from all of this))).................. you can bet it will end up in final production.

The epitome of collective brain-storming. Educational tuition. All for the advancement of mankind.

My scope and spectrum pictures show the signal a target is hit with. I merely looped a wire around my X-1 probe and connected it to my scope and computer mike input. I would have gotten the same result by a loop around the main coil. If anybody can give me an alternate explanation how MineLab can sort out conductive properties of a target using 25KHz magnetic burst please chime in. And please, no more talk of proving or disproving advertizing claims. There's already another thread going on about that.



Edited 1 time(s). Last edit at 02/14/2012 12:58PM by Yeasty.

I have made a minor error in my description of the make-up of the pulse-train.
On closer examination, I realised the 25KHz burst is assymetrical, it consists of 7 1/2 cycles, not 8. As good engineering design dictates the whole pulse train should be symmetrical/balanced, there is more to the timing than I first thought. In fact, the missing half-cycle is added in before and after the 25kHz burst, as two quarter-cycle time periods.
So, with reference to Yeasty's original scope traces: Expressed in 'basic blocks' (of half-cycles of 25KHz), the waveform is:
8 blocks low, 4 1/2 block high, 15 blocks of 25kHz, 4 1/2 blocks high. Total 32 blocks, 16 high, and 16 low. The 8-to-1 frequency ratio is still valid.

Yeasty.... i noticed on a ML paper a simular pattern you produced somewhere near the end. This paper also had some general info that was just plain interesting. [www.minelab.com]

Based on the timing details above, and the scope traces posted, channel 1 has a high-freq. burst of 22560 Hz, (fundamental = 2820 Hz) and channel 11 has hf burst of 27600 Hz (fundamental = 3450 Hz). The ratio of these extreme values is 1.22 :1. The centre-frequency (chan 6) would appear to be about 25 KHz, as previously suggested.

Pete,

If you were to measure the TX current this is what you would get:



Things are looking a lot clearer, and approach 20/20 once you see how the demods are fired.

- Carl

Yeasty - using a second coil the way you have done basically makes a transformer, so your pickup loop voltage mimics the drive voltage to the main transmit-coil. To actually measure the current, you would need to insert a low-value series resistor inline with the coil drive (one Ohm, maybe?), then measure the voltage accross that. Or use a non-differential magnetic sensor near the coil, such as a Hall-effect sensor, to see the field itself..

Pimento Wrote:
-------------------------------------------------------
> Yeasty - using a second coil the way you have done basically makes a transformer, so your pickup loop
> voltage mimics the drive voltage to the main transmit-coil.

Yup Pete, I overlooked that in your last post... Pimento is right, you can't look at the signal with a pick-up coil. You gotta measure the current; a series R works (1 ohm or less), I prefer a Tek A6302 current probe, DC-50MHz.

- Carl

If you fancy a bit of hacking, Yeasty, you can get Hall sensors out of brushless DC motors, such as 12V cooling fans in PC's, floppy disk drive motors, CDROM drives, VCR motors and more. They are usually 4 leaded SOT-23 surface mounted, fiddly, but they're 'free', so a few can be destroyed in the name of science! They need a DC voltage applied to them, and then the differential outputs can be viewed on a 'scope. Some Hall sensors are digital output, but they are uncommon.

Thanks guys. That's pretty cool. I learned something!
I'd like to ask about the way FBS does a ferrous properties measurement of a target. In this document (last page) they discuss how the received wide pulse can be used for ferrous discrimination. [www.minelab.com]
From my understanding they measure the slope of the returned wide pulse and assign a FE value.

The topic of selecting noise channel 1 or 11 to favour high /low conductors came up again on another forum. So I referred back here for the pertinent details, but ended up confusing myself.
This thread is e-Trac specific, the poster on the other forum was asking about the Explorer. I assumed they would behave the same, ie. channel 1 through to channel 11 would be the same frequencies. but....other Explorer discussions say ch1 is the highest freq, ch11 the lowest, and this eTrac thread suggests ch1 is the lowest freq, ch11 the highest. A reversed sequence. Then I re-read Yeasty's words, and his scope traces, and spectrum analysis plots, and they are all contradictory. The scope trace says ch1 is highest freq (like the Explorer), the spectrun analysis, and the text say that ch1 is lowest.....
I'm inclined to conclude that the eTrac and Explorer ARE reversed in channel number sequence, the eTrac has ch1 lowest freq -> ch11 highest, and Yeasty has labelled his scope trace image incorrectly....
But if anyone else has any ideas, please say.
(I've sent Yeasty a PM)



Edited 1 time(s). Last edit at 07/27/2015 01:28AM by Pimento.

Now that Carl is the Director of Engineering at First Texas, it should be interesting to see what's coming down the road. Between Carl and Dave J. we may see a very interesting multi-frequency machine.

How big of a real world difference in target detection are we talkin here? If the difference between the noisiest channel and quietest channel is 1/2" with a 1050 coil then maybe....that could be overcome with sweep technique and waiting for the grass to be mowed. I understand it's about getting the machine optimized to its fullest potential before factors in the field are even considered. Maybe some well done depth tests with each channel would be the final word? If Minelab were to publish said tests themselves it would put some things to rest,but nothing will ever be perfect,or in some cases,even similar. Maybe that's why they don't. Conditions,machine tuning and operator skill are hardly ever equal...which is what makes this hobby all that more fascinating.

They are reversed we noted that soon after the ET came out running them on the beach in the salt water sand.

This is really interesting! Wish Mr. Bruce Candy would chime in about FBS and the differences of FBS 2. Has anyone did these tests on a CTX?

John

@Dew: thanks for the confirmation of the reversal.
@sanddog:The transmit waveform is identical for BBS, FBS, FBS2.



Edited 1 time(s). Last edit at 07/27/2015 02:04PM by Pimento.

I have just relooked at the transmitting frequencies on both my ETrac and CTX3030. Both showed that NOISE CANCEL #1 is the lower frequency and NOISE CANCEL #11 is the higher frequency. If my previous post is different it is wrong and I apologize.

Pete

Thanks for that, Yeasty.
Just for interest, there is a thread over on Geotech1 about multi-frequency, and some chat about whether home-brewing a detector based on it is viable. Based on comments from those who know, the BBS technique seems like the one to 'copy', but don't hold your breath for any progress...