GEB on IGSL/IDX

[golfnut]

Just getting head round things..

Is it fair that the geb channel does...


1)Gives a 'quiet' detector when set right - as ground signals are approximately integrated to zero.

2)The phase shift on the GEB gating is set away from target alloys while doing GEB things - So does not contribute as much target signal level as main Disc Channel.

3)Threshold level is set to just above GEB channel level at final comparator.




As, the Disc chan gating, is aligned with target alloy phases - the Disc chan picks up more target signal than GEB?

The Disc Chan gets some ground signal on it?


Loosely speaking then, the GEB chan does not contribute valuable Target signal power - it is used for a threshold for setting the beep off?

If you made a machine without a GEB Ch and you could tolerate the ground signal on the Disc Chan you could get away without GEB Ch?


S

[Davor]

It is the other way around.

GEB channel is in fact the same as the "all metal" channel in some other detectors, only a motion compensated one, and it is giving positive response for resistive part of the ground + target response. It is why this very response is passed through to the audio stage, and not the disc response.

A role of Disc is to identify which side of the reactance the target belongs to, non-ferrous or otherwise. In IGSL both of these are used in separate channels and you get independent reading on Fe and Cu channels. They may also overlap for targets in between, e.g. targets with low conductivity. There is no need for this channel to pass a proportional signal any further because it provides only a polarity information. At low signal levels a threshold setting will not pass a GEB any further, so it is safe to crank up the gain of these channels in order to promote positive discrimination of low conductivity targets at weak responses, just as I did in my IGSL mod.

Sens. or Threshold setup simply eliminates responses lower than some sensible level, e.g. at the brink of chatter. When cranked further you lose some sensitivity, but becomes more resilient against noises.

[golfnut]

lower than some sensible level

This is the GEB ch level.?




To trip the Treshold Ive been told the comparator acts as a logic AND of Disc and GEB.


So you need GEB AND Disc to cross threshold - for a tone.

But looking at the cct it looks like either would trip tone - more of an OR function - perhaps Im wrong as the Gnd signal on Disc Chan would trip tone all the time..





Bit of an Aside, I thought earlier,

with a clean area of land for gnd bal of GEB - once set - with a suitable Threshold to match GEB setting for no tone. Threshold is above the residual GEB Ch level for the test area. No tones. (Maybe 20mV away from tone 'ON')

You now move on and and start the hunt - it is fair to assume that for half of the time the GND signal on the GEB Chan could be lower than it was on the test bed.
So the Threshold would be 'higher than it needs to be' for half of the time - not great for small stuff.

As the gnd signal varies but the Threshold pot setting on your IDX or IGSL does not - this could be improved with a Threshold level that tracks the GND signal level from the GEB Ch.

Maintaining a constant offset above the GND signal level does 2 things.

1)Removes chatter
2)Maximises Trip sensitivity - as Threshold tracks in unison with GEB Ch output.

S



S

[Davor]

If you are looking at the IDX schematic, you'll see a diode that may confuse your perception of the function that particular comparator section performs. That particular diode is a zener and it's function is preventing to rock bottom the signal to the negative rail. So it is normal that it gets reverse polarised.

So if either of the signals is below the threshold the output is pulled towards the negative rail, and if both are above - the GEB signal is passed forward as is, hence AND function.

My view of that particular section is that it is much better done in IDX than in TGSL, so I changed it entirely in my mods of the IGSL.

...

What you suggest about automatic threshold setup makes sense if you are very afraid of a bit of chatter. You should be if you are into binary response, however, with proportional tone those chatters are not that loud, and kind of make sense - they allow you to "hear" a bit of ground, and that's not bad. Among such chatters you will hear your targets as pure tones, even if they are very small. My mods make soft tones softer, and even a bit softer would be OK because your ears do the automatic adjustment thing

[golfnut]

if you are very afraid of a bit of chatter

No, chatter does not frighten me - spiders do that.

The kit has a IN4148 which cant zener from +/-4volt rails as IN4148 reverse breakdown is 75v.

Symbol
Parameter
Test Conditions
Min.
Units
VR Breakdown Voltage 75V


It is the times when there is no GND breakthru that the Threshold is too high - as the gnd drops away. (This is where you miss small signals that are there - but Threshold masks them)

If the Threshold tracked - a constant amount above the Gnd signal this gives extra depth some of the time.

All worth having.


Incidentally my Threshold pot on My IDX appears to do very little - is this normal?


S

[Davor]

There are 4148 limiter diodes in IDX, but the diode I'm referring to is 1N5223. In case you misplaced that diode and put 4148 instead, your threshold, and the rest of that comparator circuit becomes useless. If it unfortunately happened, your discrimination part becomes motion compensated all metal detector.

[simonbaker]

I'll throw in my interpretation, in case it is useful or at least creates a heated argument...

Just getting head round things..

Is it fair that the geb channel does...


1)Gives a 'quiet' detector when set right - as ground signals are approximately integrated to zero.

I would say quiet in the sense that the ground will not trigger the beep, but it would not really quiet down the EMI interference.

2)The phase shift on the GEB gating is set away from target alloys while doing GEB things - So does not contribute as much target signal level as main Disc Channel.

I agree with that for the TGSL. Targets typically produce a lower voltage pulse in the GEB channel than in the the DISC channel, depending on how far their phase is from the sync pulse of both channels.

3)Threshold level is set to just above GEB channel level at final comparator.

If you mean set just above GEB channel background noise level, that sounds right. However, when it comes to background noise (uncorrelated to the Synchronous Detector sync pulse of either channel), the noise is roughly the same in both the GEB and DISC channels, so it doesn't really matter which channel you look at.

It seems you are interested in the fact that the GEB channel has a smaller target signal, and wonder if we need to set the threshold based on that channel due to the smaller signal. However, because we really want to set the threshold as low as we can go until noise makes chatter, both channels are the same in that regard.

If there were no background noise, only "ground noise", then our goal might be to separate a target signal from the "ground signal" by setting an appropriate threshold. In that case, I guess we might look at the GEB channel to set our threshold, and if the GEB pot could be set to make the ground signal zero, then we could set our threshold to zero.

As, the Disc chan gating, is aligned with target alloy phases - the Disc chan picks up more target signal than GEB?

That is what I've observed, and it makes sense in theory to me.

The Disc Chan gets some ground signal on it?

Yes, it generally does at most settings of the DISC pot.

Loosely speaking then, the GEB chan does not contribute valuable Target signal power - it is used for a threshold for setting the beep off?

In a way -- because you need both channels to respond in order to get a beep. If the ground is causing the only signal, ideally the GEB channel will have no response or an opposite response to the DISC channel. That prevents a beep. To get a beep however, the GEB channel must get a strong enough signal to overcome the threshold. That means to me that the GEB channel can hurt the sensitivity of the detector to some good targets, because it is significantly "off-phase" with respect to most good targets. So the GEB "signal power" is the most important for the beep, because the DISC channel is always larger for good targets.

If you made a machine without a GEB Ch and you could tolerate the ground signal on the Disc Chan you could get away without GEB Ch?

I think so, yes. For example, you could make a gold hunter where the DISC channel sync pulse is optimized for detecting gold, and hope the ground signal is far enough away in phase to be suppressed -- or hope the soil material is very very even and will not make too much signal. It is something I'd like to try to see how valid the concept is.

In fact, if we could make a "phase spreader" circuit that would shift received signal phases farther apart, you could set your single channel to maximize the gold signal and suppress the ground signal... I think.


-SB

[Davor]

... and when you optimise your circuit to the max, you'll notice that your phases are ~90° apart, just as you have them now.

BTW, I noticed some inconsistency in pinpoint channel of IDX - does anyone use it at all?

[simonbaker]

... and when you optimise your circuit to the max, you'll notice that your phases are ~90° apart, just as you have them now.

Maybe -- is ferrite/black sand 90° from gold? That would be lucky.

-SB

[Davor]

Well, yeah, they actually are quite apart. I found this picture by googling around, and it is accurate enough for this purpose.
Now, take a good look on the picture and notice how all the metals have response above the horizontal line, e.g. all of them are above zero - that's your GEB criterion and a perfect explanation of what "all metal" indication does, and why GEB is used for tone representation of the target, and not Disc.
And most importantly - why GEB is adjusted to the minimum ground response, and why you get very near there by joggling a piece of ferrite near your coil.

[simonbaker]

Well, yeah, they actually are quite apart. I found this picture by googling around, and it is accurate enough for this purpose.
Now, take a good look on the picture and notice how all the metals have response above the horizontal line, e.g. all of them are above zero - that's your GEB criterion and a perfect explanation of what "all metal" indication does, and why GEB is used for tone representation of the target, and not Disc.
And most importantly - why GEB is adjusted to the minimum ground response, and why you get very near there by joggling a piece of ferrite near your coil.

Yes, that works out quite nicely for small gold objects (unless you're in the wet sand at the ocean). Makes me want to try a single channel gold detector design. I guess we can do that by just jumpering the GEB channel high on our current TGSL/IGSL models. If it works OK, might be fun to make an even simpler PCB for single channel gold hunter.

-SB

[Davor]

On IGSL you can adjust discrimination of Fe and Cu channels to overlap in that particular area, and when you boost the Disc channels as I did, you get practically a third channel with a funny tone. It works for me perfectly. I just love it. IGSL rocks

BTW, joggling with GEB channel by much is not a good idea, and at the end it would confuse users even further.

[simonbaker]

On IGSL you can adjust discrimination of Fe and Cu channels to overlap in that particular area, and when you boost the Disc channels as I did, you get practically a third channel with a funny tone. It works for me perfectly. I just love it. IGSL rocks

BTW, joggling with GEB channel by much is not a good idea, and at the end it would confuse users even further.

But don't the GEB channels still limit the detection sensitivity? If the GEB channels phase could also be moved all the way to DISC setting (centered on gold), then you'd have full sensitivity -- but it would rely on the 90 deg phase relationship of gold to ferrite for "ground balance", so pretty special-purpose configuration -- but super sensitive. However, GEB phase range won't go that far as currently implemented, I believe. But it makes me want to put a wide range GEB control in TGSL so it can be configured for gold machine by just turning GEB knob.

-SB

[satdaveuk]

But don't the GEB channels still limit the detection sensitivity? If the GEB channels phase could also be moved all the way to DISC setting (centered on gold), then you'd have full sensitivity -- but it would rely on the 90 deg phase relationship of gold to ferrite for "ground balance", so pretty special-purpose configuration -- but super sensitive. However, GEB phase range won't go that far as currently implemented, I believe. But it makes me want to put a wide range GEB control in TGSL so it can be configured for gold machine by just turning GEB knob.

-SB

Thats interesting Simon see what you come up with

Regards

[Qiaozhi]

But don't the GEB channels still limit the detection sensitivity? If the GEB channels phase could also be moved all the way to DISC setting (centered on gold), then you'd have full sensitivity -- but it would rely on the 90 deg phase relationship of gold to ferrite for "ground balance", so pretty special-purpose configuration -- but super sensitive. However, GEB phase range won't go that far as currently implemented, I believe. But it makes me want to put a wide range GEB control in TGSL so it can be configured for gold machine by just turning GEB knob.

-SB

But then the GB channel would not be doing its job of ground balancing. Also, if you simply move the GB sample pulse to the same position as the DISC sample, you will just end up with two DISC channels connected in parallel.

In another post, someone suggested that a detector could be built by only using one channel, i.e. the DISC channel. If you try this experiment you will find that it works ok in an air test, but will react horribly to ferrite, and will be unusable in the field.