TESORO GOLDEN SABRE

[simonbaker]

I read it too, who originally said that?
Originally Posted by kliner
I think it was said that to double the detection depth a power increase of 64 times would be required
Originally Posted by kliner
I think it was said that to double the detection depth a power increase of 64 times would be required

Andy

I think there are a lot of assumptions built into that statement you have to be aware of to interpret it correctly. But it may be true for a well designed MD that operates at the "fringe" of its magnetic field. A lousy MD may operate closer in to the coil and might be easier to double the distance with less power (maybe???).

Coil magnetic fields do not follow the inverse square law of radiation I think. But assuming they did follow inverse square law, you have to factor in that the target signal also falls off like inverse square law -- so the round-trip signal is falling off by at least the fourth power. I think someone calculated that with a loop coil, the end result is a sixth power fall off. So 2^6 gives 64, hence 64 times the power to double the distance.

I think it is not as pessimistic as it sounds though. You calculated that 5 times the power would only buy you 7.8 %. However, my calculations indicate it would buy you 30%. Not sure I computed it correctly, but I calculate:

Distance Factor = (Power Increase) ^ 1/6

So DF = 5 ^ (1/6) == 1.31 or about 31% increase.

As for shielding --- that's a crazy subject I hope to examine, but I want to study some physics first and then probably need to do some computer simulations also.

As always, your questions and insights are interesting.

Regards,

-SB

P.S. I will add that using a shorter time interval does not necessarily save power. By using a continuous signal, we can essentially integrate out noise more effectively, and gain signal to noise ratio. You can't get something for nothing. That doesn't mean we're using power as well as we can though, so design changes may be practical, including shorter time interval if in fact we're not utilizing our long interval to the fullest. Another idea is more resonant TX coils to increase current without using more power; this may have other disadvantages, but something I'm trying to explore.

[der_fisherman]

I think there are a lot of assumptions built into that statement you have to be aware of to interpret it correctly. But it may be true for a well designed MD that operates at the "fringe" of its magnetic field. A lousy MD may operate closer in to the coil and might be easier to double the distance with less power (maybe???).

Coil magnetic fields do not follow the inverse square law of radiation I think. But assuming they did follow inverse square law, you have to factor in that the target signal also falls off like inverse square law -- so the round-trip signal is falling off by at least the fourth power. I think someone calculated that with a loop coil, the end result is a sixth power fall off. So 2^6 gives 64, hence 64 times the power to double the distance.

I think it is not as pessimistic as it sounds though. You calculated that 5 times the power would only buy you 7.8 %. However, my calculations indicate it would buy you 30%. Not sure I computed it correctly, but I calculate:

Distance Factor = (Power Increase) ^ 1/6

So DF = 5 ^ (1/6) == 1.31 or about 31% increase.

As for shielding --- that's a crazy subject I hope to examine, but I want to study some physics first and then probably need to do some computer simulations also.

As always, your questions and insights are interesting.

Regards,

-SB

Not as interesting as your (and others!) answers!!!!

Thanks

Andy

[golfnut]

I dont know if we are allowed to talk of ML architecture on here. Im interested in the multi frequeny Tx.

Whether the coils are non resonant - to cover 100kHz of BW.

And I like to know if each Sub Tx frequency is listened to by a series of matching Rx 'channels'

e.g. 20 Tx ch with 20 corresponding Rx channels.

As these puppie do 14" in soil vs std singl tone IB VLF at ~6"

S

[simonbaker]

In one of the other posts entitled "Metal Detector -Upgrade for Greater Depth- 25% to 30% " I decided to dig a little deeper on what this guy was selling, and here is the only thing I could figure:http://metaldetectorsnow.co.uk/4-99-...garrett-etc-2/

Certainly not my idea .. but with the way that the TGSL reacts to EMI - robbing detection range, why not try a quick and dirty experiiment.

I clipped this off of an old PC power brick and Voila! Pretty early to tell, but it believe it DOES tend to improve detection range. I have to get out of subjective mode now and take some hard measurements. Noise measurements to follow.. Simon, you HAVE to try this..

Don

I got around to trying a crude approximation of that filter thing. There is some EMI that suddenly appeared in my workshop, so I tried wrapping the MD cable around a big piece of slotted ferrite a few times. It did not seem to affect the particular EMI I'm seeing; of course I don't know if my technique is same as that power brick thing, or if the EMI is the wrong type, or entering someplace else. I'll keep an eye on that technique though, it should block common mode EMI I would think.

Note: the cable config was ungrounded RX leads (wet grass).

Regards,

-SB

[Farside]

Hi all,
My apologies if this has been answered before...

Everyone seems to be using 0.2 - 0.25mm dia wire for their coils. Is there a reason for this?

I'm thinking of using a different guage based on some research in a totally different topic which has some uncanny similarities.

Thanks in advance

[Farside]

Sorry, mis-posted and can't delete the message. I've found the correct topic for this now.

[Dr Vel]

I dont know if we are allowed to talk of ML architecture on here. Im interested in the multi frequeny Tx.

Whether the coils are non resonant - to cover 100kHz of BW.

And I like to know if each Sub Tx frequency is listened to by a series of matching Rx 'channels'

e.g. 20 Tx ch with 20 corresponding Rx channels.

As these puppie do 14" in soil vs std singl tone IB VLF at ~6"

S

I have played with multi frequency detection for a couple decades, as a side hobby building my own toys. What I do not see is much discussion over this subject, as well as DSP, and comb filtering. On a side note I am wondering. I have found a super deal on 100 TS904IN Quad OpAmps, see pdf for specs. What I am curious about is if anyone sees that using the two standby pins for synchronous switching (HI-Z output states) would in effect do the job of choppers instead of adding a 4066, discrete fets, or similar parts? I could be way off the mark since as is typical the data sheet is poorly written and no circuit examples are given of using the standby inputs.

I should add does anyone else remember when data sheets were more in depth and contained many example circuits?

[dfbowers]

I got around to trying a crude approximation of that filter thing. There is some EMI that suddenly appeared in my workshop, so I tried wrapping the MD cable around a big piece of slotted ferrite a few times. It did not seem to affect the particular EMI I'm seeing; of course I don't know if my technique is same as that power brick thing, or if the EMI is the wrong type, or entering someplace else. I'll keep an eye on that technique though, it should block common mode EMI I would think.

Note: the cable config was ungrounded RX leads (wet grass).

Regards,

-SB

In reading a little about these ferrite beads, it appears as though they are designed to attenuate common mode EMI in the MHz ranges. They have their uses. I see them used quite a bit in power supples and used on single ended SCSI busses inside larger computer systems.

[simonbaker]

I have played with multi frequency detection for a couple decades, as a side hobby building my own toys. What I do not see is much discussion over this subject, as well as DSP, and comb filtering. On a side note I am wondering. I have found a super deal on 100 TS904IN Quad OpAmps, see pdf for specs. What I am curious about is if anyone sees that using the two standby pins for synchronous switching (HI-Z output states) would in effect do the job of choppers instead of adding a 4066, discrete fets, or similar parts? I could be way off the mark since as is typical the data sheet is poorly written and no circuit examples are given of using the standby inputs.

I should add does anyone else remember when data sheets were more in depth and contained many example circuits?

It's a cool idea if you can reduce part count that way. Sounds worth experimenting with.

-SB

[golfnut]

op amp cct reference

Dr Vel

these are good op amp reference texts


http://www.ti.com/lit/an/sboa092a/sboa092a.pdf

http://www.ti.com/lit/an/snla140a/snla140a.pdf

steve

[Dr Vel]

Thank you Steve those are very useful books.

[goldfinder]

Regarding DSP

I worked for many years in the DSP arena and Digital Signal Processing can get extremely high return for the money. A matched filter that looks for a specific return signal or at least a signal that has close frequency components to the return signal can pick up weak signals out of noise. Finding signals in a SNR of 1 to 1000 was quite common using matched filters on a return signal. The problems arise in defining the noise spectrum and the signal spectrum. With the fast processor we have these days a DSP metal detector could likely tripple or quadruple depth.
Goldfinder

[simonbaker]

Dr Vel

these are good op amp reference texts


http://www.ti.com/lit/an/sboa092a/sboa092a.pdf

http://www.ti.com/lit/an/snla140a/snla140a.pdf

steve

Yes, nice, thanks!

[xix]

I worked for many years in the DSP arena and Digital Signal Processing can get extremely high return for the money. A matched filter that looks for a specific return signal or at least a signal that has close frequency components to the return signal can pick up weak signals out of noise. Finding signals in a SNR of 1 to 1000 was quite common using matched filters on a return signal. The problems arise in defining the noise spectrum and the signal spectrum. With the fast processor we have these days a DSP metal detector could likely tripple or quadruple depth.
Goldfinder

Hi Goldfinger,
this sounds marvellous!! how can we start such type of project?
KR,
XiX

[simonbaker]

I worked for many years in the DSP arena and Digital Signal Processing can get extremely high return for the money. A matched filter that looks for a specific return signal or at least a signal that has close frequency components to the return signal can pick up weak signals out of noise. Finding signals in a SNR of 1 to 1000 was quite common using matched filters on a return signal. The problems arise in defining the noise spectrum and the signal spectrum. With the fast processor we have these days a DSP metal detector could likely tripple or quadruple depth.
Goldfinder

The Synchronous Detector in the TGSL is really a special case of a FFT at a single frequency -- the only frequency that matters in a VLF detector, the TX frequency. It doesn't need any fancy hardware -- just a switch!

I don't see how a DSP would be the right tool for a single frequency VLF MD. It is great if you are looking for unknown signals in noise, but the beauty of a Synchronous Detector is you know exactly the frequency of the signal you are looking for -- so a full FFT is really a waste.

It may have more use with a PI detector where you have a broad spectrum broadcast and you are looking for unknown responses and want to look at the frequency domain for various reasons. The time domain still may be best for actually detecting the presence of the target though.

Regards,

-SB