If this is your first visit, be sure to
check out the FAQ by clicking the
link above. You may have to register
before you can post: click the register link above to proceed. To start viewing messages,
select the forum that you want to visit from the selection below.
" Originally Posted by simonbaker
As to how the wires are oriented (direction of winding), I believe it doesn't matter; all that matters is the direction of the current, and you can control that by just reversing the leads if you get it wrong (before you pot the coil if you use the "grounded RX coil" design! -- because you want the grounded lead to go to the correct op amp input)."
" Reply from Andy
You and I have the same thoughts on that, if the phasing is wrong, just swap the connections to one coil only. By the way, a good reason to bring the coil and shield connections out in the end of the cable and not to connect them in the search head - ever!!!"
This is most interesting part for me.
Would be good to clear this up.
Yes, you will achieve (virtually) same effect ...but is it the same for real?
I don't have proper answer. To answer that; i think man must have more advanced measuring equipment...than i have..
I wander what Qiaozhi, Aziz... and others can say on this subject?
In the past I have tried all the different combinations for connecting the screen to a Tesoro coil. The bottom line is that the original Tesoro configuration works best for that design. In other words, making the connection inside the coil housing.
I suspect most of the problems constructors are experiencing in this area are due to some confusion on the schematic, where the RX coil connects to the pre-amp input. In some versions of the schematic there is a ground connection shown at the PCB. This connection is not there in a commercial Tesoro detector. If you connect the screens together in the coil housing AND at the RX input on the PCB, of course you will have a ground loop. Using the original Tesoro configuration, I have not had any problems with wet grass ... and we get quite a lot of that in the UK.
I figured that too.
Especially now at IGSL with Minelab coil.
Because Minelab coil is having bit different internal setup.
....
Next question would be: "As to how the wires are oriented (direction of winding)..."?
Virtually this can be solved by reversing the coil leads.... but is it the same?
If TX is oriented clockwise and RX anticlockwise? Reversing the leads will "solve" problem, yes.. but is it irrelevant, the orientation?
Next question would be: "As to how the wires are oriented (direction of winding)..."?
Virtually this can be solved by reversing the coil leads.... but is it the same?
If TX is oriented clockwise and RX anticlockwise? Reversing the leads will "solve" problem, yes.. but is it irrelevant, the orientation?
IMHO it is irrelevant. If there was a difference, then the TGSL would work differently in the Northern hemisphere than it does in the Southern hemisphere. Much like the water going down the plug hole.
Firstly, many thanks Simon, you gave me a lot to think about (exactly what I wanted by the way!) I hope I get some more replies similar to yours, thanks. I thought that I would answer each (excellent) point that you made:-
I am open on this point, but this happens already with ANY coil, in fact, if I could ever draw in Eagle CAD AND it was the correct value of impedance (unlikely), I would love to try 2 of them on an IGSL in double D format of course.....That might be an awesome thing to make and use.......untill I see one, I remain unconvinced that it will not be far better than a conventional coil......that's just me, I could easily be 100% wrong.....
What I was referring to about the inner and outer windings "opposing each other" is described here by Aziz (the "field-master") http://www.geotech1.com/forums/showthread.php?t=13858 (the link that dfbowers timely-ly dug up). This is an effect found with wide flat spiral coils but not with conventional "bundled" coils like we usually make them. The end result is that the wide spiral coil doesn't have as strong (or deep) a magnetic field.
What I believe is similar to what you say, but that the Tesla coil will work better (at the same time) for both large and small objects as it is both a large and a small coil (and a medium one too!!) and we both know that with a larger coil you get more depth and see large objects better and with a small coil you get less depth, but more sensitive to smaller objects. Now you can have both at the same time (I really do believe this!).
I like the idea but I think you can't have your cake and eat it too. The reason conventional small coils are good for small shallow objects is that they concentrate all the inductance of the coil into an intense condensed magnetic field in a small volume. The spiral coil will only have a few windings with a small radius which is not the same as if the whole coil was wound small. However, the spiral coil's magnetic field I believe is strongest inside the smallest loop, so it will not be bad for detecting small shallow objects. It's main weakness would be for deeper objects where the magnetic field won't be very strong compared to a normal coil. Again, I'm talking about an extreme, wide spiral coil -- your proposed flat coil is not extreme at all, and I think should work quite well -- probably about the same as the conventional coil I would predict.
Agreed. But don't forget that a true flat TGSL Rx coil(107 turns) "a la Tesla", assuming that the outside radius would be for a 25cm search head, made from 0.25mm copper wire (actually 0.26mm with coating) would be only a band of 2.782 cm wide or about 1.095". It would not actually look much like the picture I posted from Wiki.....The hole in the middle would be almost 20 cm across.......so its appearance would be relatively similar in many ways to a conventional coil.....7.6" inner hole and 9.84" outside diameter - but flat!
Yes I agree -- your design sounds fine, it is not extreme at all -- I don't expect it to perform noticeably better for the TGSL/IGSL, but let's find out, because that would be a nice way to make improvements.
I seriously doubt whether the effect of shielding the Tx coil, which is radiating almost 100% of the energy used in the Collpits Oscillator, has as large an effect as shielding the Rx coil which is receiving maybe 1% of that energy or less......at best. Remember the Rx coil sits in the null and what we eventually pick up has to be seriously amplified before we can use it.....but it does demonstrate to me at least, that windings shield windings according to you as well. So a flat Tesla Tx coil could even improve Tx radiation as well.....because its not "blocked" anywhere.....(Except where the Rx coil overlaps, underneath of course!!)
I don't quite follow what you are saying there, not sure. I'm looking at a signal-to-noise effect. If the TX coil is on top, it it further from the target and the target sees a slightly weaker field. But the RX coil is closer to the target and sees a proportionately stronger field from the target. So maybe the net signal is the same and the RX noise is the same (unrelated to position of the coils). If you switch the coil positions, I conjecture that the math is the same -- the signal due to one coil gets weaker but from the other gets stronger.
I still don't know about windings shielding windings. For DC currents they wouldn't. Perhaps there is a frequency dependency, where as the frequency increases, there is some shielding effect. For audio frequencies, I'm not sure it is significant.
Actually, I recall that there is the "skin effect" which can be a problem for bundled coils. It is frequency dependent. It is similar to the skin effect for a thick conductor, where the current is mostly at the surface of the wire, and the apparent resistence of the wire is higher than for DC currents. Bundling wires close together can cause a related effect. So the flat-wound coil probably helps with that also, perhaps giving a lower-resistance (higher-Q) coil. So I think the flatter we can wind without increasing "width" too much may be indeed useful.
I agree. Though I have actually no idea what the actual value of capacitance from say a TGSL Rx coil actually is.....guessing only a few pFs....but as once my IGSL is finished, I am going to build a Russian PI machine as well, I am very interested in making a good job for that too.......
I think the flat coil may have important advantages for a PI detector, which I don't know much about. I think lowering capacitance is an important goal and the flat coil seems like it should do that.
You and I have the same thoughts on that, if the phasing is wrong, just swap the connections to one coil only. By the way, a good reason to bring the coil and shield connections out in the end of the cable and not to connect them in the search head - ever!!!
That grounding of one RX lead in the search head has always perplexed me but it seems to work in practice, although dfbowers seems to have settled on an ungrounded design which he likes better. I had a concern with the ungrounded design that some static electricity could blow out the LF353 inputs -- but someone said they were protected somehow from that by design.
LOL - but you put it on the bottom anyway, "just in case?" Simper, simper!
What I was referring to about the inner and outer windings "opposing each other" is described here by Aziz (the "field-master") http://www.geotech1.com/forums/showthread.php?t=13858 (the link that dfbowers timely-ly dug up). This is an effect found with wide flat spiral coils but not with conventional "bundled" coils like we usually make them. The end result is that the wide spiral coil doesn't have as strong (or deep) a magnetic field.
The diagrams from the fieldmaster are not available, nothing there except a 404 message!! Totally useless.....sorry.
That grounding of one RX lead in the search head has always perplexed me but it seems to work in practice, although dfbowers seems to have settled on an ungrounded design which he likes better. I had a concern with the ungrounded design that some static electricity could blow out the LF353 inputs -- but someone said they were protected somehow from that by design.
Don grounds his coil shielding, but only on the supply ground, not on an input to the circuitry......perhaps Don could jump in and explain far better, Don?
regards
Andy
PS. Making the flat coils is difficult, scrapped two already......we will see what No. 3 is like tomorrow.....
Last edited by der_fisherman; 08-22-2011, 07:27 PM.
Reason: Forgot stuff
What I was referring to about the inner and outer windings "opposing each other" is described here by Aziz (the "field-master") http://www.geotech1.com/forums/showthread.php?t=13858 (the link that dfbowers timely-ly dug up). This is an effect found with wide flat spiral coils but not with conventional "bundled" coils like we usually make them. The end result is that the wide spiral coil doesn't have as strong (or deep) a magnetic field.
The diagrams from the fieldmaster are not available, nothing there except a 404 message!! Totally useless.....sorry.
That grounding of one RX lead in the search head has always perplexed me but it seems to work in practice, although dfbowers seems to have settled on an ungrounded design which he likes better. I had a concern with the ungrounded design that some static electricity could blow out the LF353 inputs -- but someone said they were protected somehow from that by design.
Don grounds his coil shielding, but only on the supply ground, not on an input to the circuitry......perhaps Don could jump in and explain far better, Don?
regards
Andy
PS. Making the flat coils is difficult, scrapped two already......we will see what No. 3 is like tomorrow.....
Sorry those diagrams are missing. He put it in words however:
"By reducing the complexity of the coil for a better view (reducing the number of windings), you can see, why flat spiral coils are not effective. Between each two windings there occur canceling regions. The generated magnetic field from the outer winding will be slightly cancelled by the next inner winding. Also, this causes the flattening of the magnetic fields above the coil..."
Again, that mainly applies to wide flat coils -- your "slim flat coil" design should not suffer from that effect much at all.
Don is using the "differential input" design which doesn't ground any RX leads at all and theoretically eliminates common mode signals best. I'm going to hazard a guess that the Tesoro engineers grounded the RX lead to prevent a huge common mode spike that would overwhelm the op amp.
(However, I'm not sure the Tesoro design would be successful in all cases -- I would think a static burst would be neutralized on the grounded lead, but the ungrounded lead would be isolated from the grounded lead by the inductance of the RX coil and still zap one of the inputs of the op amp maybe. Still a puzzle to me.)
You would need to change the number of turns on the coil in order to make this useful for analyzing the field around a metal detector coil. Unfortunately, with the current design, it would need a different measurement coil for each TX frequency. For example, using a 1" (25.4mm) circular coil to measure the 14kHz TX signal from a TGSL would require 35 turns (assuming the same top value of 1uT, and an output voltage of 500uV).
In order to use the same coil for all tests, you would need a microcontroller to scale the results correctly according to frequency and magnetic field intensity.
Sorry those diagrams are missing. He put it in words however:
"By reducing the complexity of the coil for a better view (reducing the number of windings), you can see, why flat spiral coils are not effective. Between each two windings there occur canceling regions. The generated magnetic field from the outer winding will be slightly cancelled by the next inner winding. Also, this causes the flattening of the magnetic fields above the coil..."
Again, that mainly applies to wide flat coils -- your "slim flat coil" design should not suffer from that effect much at all.
Don is using the "differential input" design which doesn't ground any RX leads at all and theoretically eliminates common mode signals best. I'm going to hazard a guess that the Tesoro engineers grounded the RX lead to prevent a huge common mode spike that would overwhelm the op amp.
(However, I'm not sure the Tesoro design would be successful in all cases -- I would think a static burst would be neutralized on the grounded lead, but the ungrounded lead would be isolated from the grounded lead by the inductance of the RX coil and still zap one of the inputs of the op amp maybe. Still a puzzle to me.)
-SB
I don't believe that I originally did anything different than what Ivconic documented in the "TGSL Complete Details" in terms of grounding. I was scratching my head though the first time I went outside and the grass was covered with dew. Sweeping through the grass was just too noisy. The "wet grass" configuration for sure fixed the problem.
Now Ivconic suggested that we use Belden 8723, just like the older Whites cables. The only difference I see is that there are two individually shielded pairs instead of all 4 in a bundle. I think the grounding scheme would effectively be the same regardless of USB-2 or 8723. Maybe I don't see the difference yet.
The only real difference I see is that the Rx pair is completely sheilded from the Tx pair with 8723. .. But, by connecting the ends of the coils to the shields in the coil shell we have only 3 "conductors". It would be the same as using a twin audio cable with the Tesoro standard wiring configuration. Not so with the USB-2 cable. Maybe my "wet grass" setup is only necessary for USB-2 cables!
Sorry those diagrams are missing. He put it in words however:
"By reducing the complexity of the coil for a better view (reducing the number of windings), you can see, why flat spiral coils are not effective. Between each two windings there occur canceling regions. The generated magnetic field from the outer winding will be slightly cancelled by the next inner winding. Also, this causes the flattening of the magnetic fields above the coil..."
Again, that mainly applies to wide flat coils -- your "slim flat coil" design should not suffer from that effect much at all.
Don is using the "differential input" design which doesn't ground any RX leads at all and theoretically eliminates common mode signals best. I'm going to hazard a guess that the Tesoro engineers grounded the RX lead to prevent a huge common mode spike that would overwhelm the op amp.
(However, I'm not sure the Tesoro design would be successful in all cases -- I would think a static burst would be neutralized on the grounded lead, but the ungrounded lead would be isolated from the grounded lead by the inductance of the RX coil and still zap one of the inputs of the op amp maybe. Still a puzzle to me.)
-SB
To answer my own puzzle - working on the theory that grounding one RX lead is for common mode spike protection - I had doubts, thinking that the inductance of the RX coil isolated the other lead from quick spikes and the spikes would zap the op amp on the ungrounded lead anyway. However, maybe the C6 capacitor across the op amp inputs acts as a "lightening arrestor" in that case, and shunts a voltage spike over to the grounded lead. So perhaps the common mode spike protection theory for grounding one RX lead is still in play.
BTW - I originally tried Belken cable and still have some around. However, I wasn't getting good results with my prototypes; my impression was that all that shielding was affecting the null, or maybe it was acting like a target and making the configuration very sensitive to cable vibration. But it must be good because Ivconic has good success with it. I also tried audio cables and had similar lackluster results. But these were early days when many things were iffy in my circuits and coils, so maybe should try again. All I know is my best results so far have involved USB cables (although the connectors seem very noisy if you tweak them).
You would also think that individually shielded pairs would help you get the best null, but my intuition gets fooled all the time in this business.
What I was referring to about the inner and outer windings "opposing each other" is described here by Aziz (the "field-master") http://www.geotech1.com/forums/showthread.php?t=13858 (the link that dfbowers timely-ly dug up). This is an effect found with wide flat spiral coils but not with conventional "bundled" coils like we usually make them. The end result is that the wide spiral coil doesn't have as strong (or deep) a magnetic field.
The diagrams from the fieldmaster are not available, nothing there except a 404 message!! Totally useless.....sorry.
That grounding of one RX lead in the search head has always perplexed me but it seems to work in practice, although dfbowers seems to have settled on an ungrounded design which he likes better. I had a concern with the ungrounded design that some static electricity could blow out the LF353 inputs -- but someone said they were protected somehow from that by design.
Don grounds his coil shielding, but only on the supply ground, not on an input to the circuitry......perhaps Don could jump in and explain far better, Don?
regards
Andy
PS. Making the flat coils is difficult, scrapped two already......we will see what No. 3 is like tomorrow.....
Plus, I made 2 videos in an experiment where I made two coils, One with the coils connected to the shield inside the coil shells and one with the shields and ground wires connected only to ground on the PCB.
Plus see these two videos The first is the "wet grass" configuration:
Small addition, maybe will help somehow:
Musketeer TS1000 coil is shielded with some material, most probably graphite spray, and further connected to TX shield (part of TX coaxial).
RX coil is not connected to coil shield nor anything else in coil , but is leaded via separated coaxial to pcb and is connected to pcb gnd (at Musketeer).
So TX leads are "carrying" ground and shielding.
OK.
At IGSL i had doubts and choices, how to properly connect RX leads.
So i connected them just on RX pins on pcb, without additional connection to pcb gnd. Checked behavior and seems it is working fine.
Next; i connected additionally one RX input to pcb gnd - again seems to work fine.
Whole testing was done in workshop, on a bench.
Later, with second setup i went outdoor and tested behavior on real soil.
If Senses turned to maximum than few and rare pikes occurs. But really rare, not often and not so disturbing.
Not even close so disturbing as on videos from post above.
Also noticed that pikes can occur only when coil is fast risen from soil, otherwise don't.
But when Senses (both) lowered to "6" or "7" than no pikes at all.
Depths are pretty good even when both Senses stays at that position.
For example; i have buried 2cm coin at one place, at 24cm depth.
It is very easy detected even when coil is some 10cm over the soil surface.
No doubts, higher pitch indicate "colored" target.
Next check i will do; is with first setup (RX leads are no way connected to pcb gnd) on real soil at same place with coin.
Last days i had other things to do, so i will check that very soon.
Comment