Square wave Pulse Induction

[mikebg]

Hello,
Hey Carl it is the "old school " very neat and effective ...that i love it like the world of vacuum tube beautiful ...
Deemon, i enjoy very much your electronic background and also your drawing , When i see your post it is like when i read "a real good and excellent book" and your kindness to give us your high electronic knowledge...in this day it is very unusual in this world of " the big brother".... Orwell's.....

Alex, how do you feel the idea instead vacuum tubes, to design square wave metal detectors with spark gap, vibrating interrupter or motor interrupter?
It is more neat and effective because we will spare a battery - no need to heat up filaments of vacuum tubes:
http://www.geotech1.com/forums/showt...793#post101793

[deemon]

Alex, how do you feel the idea instead vacuum tubes, to design square wave metal detectors with spark gap, vibrating interrupter or motor interrupter?
It is more neat and effective because we will spare a battery - no need to heat up filaments of vacuum tubes:
http://www.geotech1.com/forums/showt...793#post101793

Hey , Mike , what a problem ? When I need to use vacuum tubes - I use vacuum tubes .... or maybe the law restricts us to do it ?
And when I need to make this power chain , I decided to use transistors - is it a wrong solution , what do you think ? Maybe you advise me to use China-made IC's here ?

[Aziz]

Nice design indeed. Good old electronics school.

As I like the most simple KISS designs, I won't touch it even with pliers. *LOL*
I like the coil current "dancing up and down" and not staying at some level for some time.

Keep on the good work deemon.
Aziz

[Carl-NC]

Alex, how do you feel the idea instead vacuum tubes, to design square wave metal detectors with spark gap, vibrating interrupter or motor interrupter?

Believe it or not, I'm working on such a beast.

[mikebg]

Hey , Mike , what a problem ? When I need to use vacuum tubes - I use vacuum tubes .... or maybe the law restricts us to do it ?
And when I need to make this power chain , I decided to use transistors - is it a wrong solution , what do you think ? Maybe you advise me to use China-made IC's here ?

Dmitry, I use a transistor in input of my RX because it generates less noise than an (low noise) opamp.
I agree that the problem is not in the components that implements the principle of operation, but the principle itself. If the principle (block diagram) is made correctly, it can work even with ancient components.

For example, the principle of induction balanced TWIN LOOP used in 1915 for cleaning the land from scrap metal, can detect metal pieces at a depth of 40 cm. The operator can identify targets according to the timbre of the sound of them and easily calculate depth of target.

http://www.geotech1.com/forums/showt...852#post165852

But let not forget that we live in the second decade of the 21st century. Electronics back in the 20th century progressed so that it was able to realize the full block diagram of an apparatus in one chip. Now designers try even at home PCs to get rid of any moving parts: fan, hard disc, CD drive. It is very important the designer to analyze the best operating principle, and after block diagram is estimated, to think of what components will be best to realize it.

[deemon]

But let not forget that we live in the second decade of the 21st century. Electronics back in the 20th century progressed so that it was able to realize the full block diagram of an apparatus in one chip. Now designers try even at home PCs to get rid of any moving parts: fan, hard disc, CD drive. It is very important the designer to analyze the best operating principle, and after block diagram is estimated, to think of what components will be best to realize it.

OK , man , show me that "single chip" , please I mean the chip , on which I can implement my power conversion circuitry . With all this current reverse stuff , etc . Maybe I'd been sleeping too long and missed something ? And don't forget that we are now talking about transmitter , not about receiver . MD transmitter schematics , just because its nature , must be more similar to a kind of power supply unit - it's something like "quasi-resonant power converter" , not like "low noise preamp" . Of course , I have a preamp in my device too ... but I'll talk about it a little later .

[mikebg]

ANCIENT SQUARE WAVE TECHNOLOGY

Believe it or not, I'm working on such a beast.

ANCIENT SQUARE WAVE TECHNOLOGY
Carl, I believe that you will soon open a museum for the history and development of metal detecting technology. However I do not know how many floors will the building of the museum. Secure your collection will occupy two floors.

If you have a suitable building, you can order Bulgarian masters to make for museum excellent operating copies of three different metal detectors, designed and built at the end of the 19th century. There is a fourth but not operating model from 1902 with PFN and a polarized relais. It can not operate because local masters don't know how to connect several windings of polarized relais. Here is the principle of TX (relais bobbin connecton not shown):

[deemon]

By the way , this ancient relay-controlled device looks like not a square wave , but mostly like a short bipolar-pulse technology . This LC network is in fact a non-terminated "delay line" ..... and the delay time of this line ( multiplied by 2 ) will define the pulse duration . Interesting solution

[Carl-NC]

If you have a suitable building, you can order Bulgarian masters to make for museum excellent operating copies of three different metal detectors, designed and built at the end of the 19th century. There is a fourth but not operating model from 1902 with PFN and a polarized relais. It can not operate because local masters don't know how to connect several windings of polarized relais. Here is the principle of TX (relais bobbin connecton not shown):

I'm accumulating the proper parts to build a replica of Bell's 1881 detector. For the museum, of course.

[WM6]

If you have a suitable building, you can order Bulgarian masters to make for museum excellent operating copies of three different metal detectors, designed and built at the end of the 19th century. There is a fourth but not operating model from 1902 with PFN and a polarized relais. It can not operate because local masters don't know how to connect several windings of polarized relais. Here is the principle of TX (relais bobbin connecton not shown)

Using an old telegraph sounder as metal detector is really great Bulgarian innovation.

[ODM]

Interesting outside-the-box setup, indeed! I wonder if some of those switching transistors could be replaced with magnetic components as well, saturable reactors? They make some current based control designs rather simple.

I haven't personally been much into using transformers in my circuits outside of HF/VHF, since they can't often be bought as wanted (unless you buy 1k+ at a time) but for an unique solution like this, it might be interesting. Would be nice to see you present this method on video, with oscilloscope and various targets, if you currently have it on your workbench?

[deemon]

Interesting outside-the-box setup, indeed! I wonder if some of those switching transistors could be replaced with magnetic components as well, saturable reactors? They make some current based control designs rather simple.

I haven't personally been much into using transformers in my circuits outside of HF/VHF, since they can't often be bought as wanted (unless you buy 1k+ at a time) but for an unique solution like this, it might be interesting. Would be nice to see you present this method on video, with oscilloscope and various targets, if you currently have it on your workbench?

Of course , I'll do the test , but a little later - now I have a plan to test and compare several different signal processing algorithms , with the same power chain . In theory , one of them is optimal ( but also the most complicated ) , others are sub-optimal . One algorithm is almost finished ( only on workbench ) , so the first results are coming soon . But what I need to get a comparable results is to make a proper round search coil . Now I making all experiments with the ferrite search coil ( remained from my previous experiments ) ... but I think that it cannot give the maximum sensitivity .

About the power chain - I think that circuit shown in the topic is good enough . It works stable and reliable , without any problems . By the way , the most critical parts are not the transistors , but the HV diodes ( D1 , D2 ) . They must stand 600 V pulses and be fast enough , and at the same time must have minimum possible forward drop on current 0,5-1 A . Now I use here silicon-carbide Shottky diodes CSD10060 , they have good forward drop and terribly fast . But I have one idea how to improve this switch - once I'll test it too .

[mikebg]

Using an old telegraph sounder as metal detector is really great Bulgarian innovation.

WM6, the polarized relay is not a sounder. It is used as a controlled commutator to generate square wave. The unknown circuit should operate self-oscillating, but we don't know how been connected windings to PFN circuit.

http://www.youtube.com/watch?v=ps0QMRxjn7o

The use of PFN (pulse forming network) in TX of a metal detector is the great innovation.

[deemon]

So , guys , let's continue

Now we'll talk about possible signal processing algorithms in this square wave approach . After theoretical analysis I found 3 basic algorithms that can be used here . First algorithm , being the most simple , works in "all metal" mode and utilizes the quality that I had been talking before - average coil current drop in presence of the metal near the search coil . This effect doesn't depend on the target metal type , it works very simple - our target "sucks" the energy from the coil , and we can detect it . By the way , this effect does have one VERY interesting property , I'll talk about it a little later . I'd never seen it before - it's really something new

The second algorithm is way more complicated , but it can also extract more information from the target . Its idea is very similar to my recuperative PI device , described here - http://www.geotech1.com/forums/showt...y-recuperation . And I used the very similar signal extracting technique , I mean the pulse slicing , clamping and amplification with the correlation block on the end of the signal chain . Main schematic idea may be called "distributed precision" - I mean the signal amplification in different stages , when each next stage does correct the errors of the previous one . As we can see on the block diagram - an exponential target response being separated from DC component , amplified by several gain stages and finally multiplied on the reference signals of different shapes ( like I did it before ) with averaging the final signal . This idea helps us to cancel the noise ( due to averaging ) and at the same time to find the real time constant of the target ( target discrimination ) - TC is proportional to the two channels output ratio ( for example , tilt and parabolic ) . I just tested it on the bench and it works with square wave signal as well as with serial pulses before . But of course , this algorithm can be easily implemented in digital domain also . In this case we must connect our ADC to the output of the current transformer ( TR2 ) via the single op-amp I-U converter , performing all other operations in DSP .

And finally , the third algorithm , which I called "ultimate" because it can extract maximum target information . As we can see , I use here the same compensation idea ( based on Faraday's law ) , that I described here - http://www.geotech1.com/forums/showt...3-mono-coil-IB . The only difference is that I used an "integral form" of the principle . In the previous circuit I used the derivative of the coil current to compensate the coil voltage , but now I use the coil voltage integral to compensate the coil current And I perform this compensation just in the current transformer ( which has 3 windings instead of 2 ) , so we don't need to load our signal preamp with the coil idle current wave , that can improve its noise performance . And we don't need to "rectify" the signal here ( converting it to unipolar form ) , this allows us to amplify the signal without any clamp circuits , etc , with a simple and high-quality AC amplifier , and use auto gain control ( AGC ) to prevent overload in presence of big metal targets near the coil . This allows us to push amplification to its theoretical limit and gain the maximum possible performance of the entire device . And we can do even more - we can process our signal as a continuous wave , without any "holes" in it ( due to transmit pulse blanking ) , so the "eternal" PI devices drawback is removed .

[Sergey_P]