Discrimination Circuit and Ground Balance.

[muddy fingers]

I've been reading a lot on this subject but nothing has really answered my question which is "How does altering the Ground Balance on a machine affect the Discrimination Circuit?" I know that in the early days of the 1980s many machines like the Garrett ADS that I had, only had Ground Balance operating in "All Metal." When you switched into the Discriminate mode, it was often impossible to get any signal at all in heavily mineralised ground, as both the audio and meter signal were totally masked by the negative ground effect. You could see this effect on the meter, as the needle would drop to the left of the display (iron zone) and firmly remain there until the All Metal mode was re-selected.
Later in the 1980s, the Disc Circuit was affected also by the Ground Balancing adjustments made in the All Metal Mode. I have read that the Disc Circuit on some of those later machines caused the I.D reading to go up or down depending on where the Ground Balance had been set because they were using a single channel.

Now I have read, and this is where I begin to get even further confused, machines like the Blisstool and GMP, "do not use a conventional ground balance circuit." "The whole detector is cycled through a disc circuit and when you ground balance it is using a compared signal, not a single channel".


I made a couple of diagrams to try and illustrate my understanding or misunderstanding of the subject. Any comments would be very much appreciated.

[Silver Dollar]

Well basically there is 50 components in the Disc and Geb channels, if they were all exactly the
same you wouldn't loose anything between modes. Also the Disc mode is subtracting signals. Some
of those signals are good targets so the amplitude from them is lessened. A digital filter would tend
to loose less here, an analog one tends to have a broad range of influence...

[Qiaozhi]

I've been reading a lot on this subject but nothing has really answered my question which is "How does altering the Ground Balance on a machine affect the Discrimination Circuit?" I know that in the early days of the 1980s many machines like the Garrett ADS that I had, only had Ground Balance operating in "All Metal." When you switched into the Discriminate mode, it was often impossible to get any signal at all in heavily mineralised ground, as both the audio and meter signal were totally masked by the negative ground effect. You could see this effect on the meter, as the needle would drop to the left of the display (iron zone) and firmly remain there until the All Metal mode was re-selected.
Later in the 1980s, the Disc Circuit was affected also by the Ground Balancing adjustments made in the All Metal Mode. I have read that the Disc Circuit on some of those later machines caused the I.D reading to go up or down depending on where the Ground Balance had been set because they were using a single channel.

In the early days, detectors were all non-motion machines. Even today, the Viking 5 works in exactly the same way. These detectors will demonstrate excellent ferrous rejection in a air test, but fail miserably in real ground. As you stated previously (and can be seen on a meter, if fitted) lowering the coil on heavily mineralized ground causes the amplitude to be reduced. This reduction pushes the signal level way below the threshold, and consequently you cannot detect any good targets unless you constantly adjust the threshold. So, although you could add ground balance in all-metal mode, the detector will be next to useless in disc mode. Eventually the concept of motion-mode was introduced, and it was discovered that combining the signal from both the GEB and DISC channels would allow the detector to achieve both ground balance and discrimination at the same time. This, of course, assumes that the ground is relatively homogenous (the level of mineralization remains fairly constant over a reasonable area), the ground surface is flattish, and that you keep the coil parallel to the ground.

If you add VDI to a two-channel detector, the readings will only be consistent when the GEB sample is positioned over the zero crossing, and the DISC sample is over the peak of the received signal. In this case it is a simple matter of trigonometry, and the processor can calculate the arctangent of R/X, to obtain the phase-shift of the target. However, if you move the ground balance sample, the calculated phase-shift will be in error and hence the VDI will be displayed incorrectly. To correct this problem, it is necessary to add a third channel that always samples at the zero-crossing. In this way the VDI will be unaffected when the ground balance is adjusted.

For a digital detector, the DISC sample can remain at the peak of the received signal, as the target signal will depend on the calculated VDI reading.

[muddy fingers]

Thank you Qiaozhi and Silver Dollar for taking the time and trouble to reply to my question, I really appreciate it. I understand a lot better now having read your explanations but still have so very much to learn but I am trying hard and have always been very interested to learn about how detectors work. It has always been difficult to know exactly what is supposed to be happening, electronically speaking, inside some of the various detectors of the past and present, especially the analogue ones, because of the way the manufacturers advertise the capabilities of the detector which does not always match what happens in reality. "Auto Ground Balance" is an example of how the wording makes one think that the detector can do more than it actually can. I've done another diagram. Is it right this time?

[Davor]

Maybe you'd do better if you play with a polar diagram with ground and various metals mapped at their proper phases. Inability to do both GB and Disc simultaneously is mostly due to lacking a channel or two in signal path. Old TR devices had only one and you could have only one at a time.

[muddy fingers]

Thanks Davor- the polar diagram is a good idea.

[Gwil]

Hi m f

In a lot of modern detectors the answer to the question "How does altering the Ground Balance on a machine affect the Discrimination Circuit?" is "Not at all". The received signal is amplified and then split into two independent channels - the GB channel ignores the ground but indicates any target that doesn't look like ground, and the Disc. channel indicates whether the target is good or bad. (The ground, and different sorts of targets, give signals of different phases which enable them to be distinguished.) Then the outputs of the two channels are used by some more electronics to produce an audio output.

[Dean Sarelius]

Hi All,
I am curious about the mechanism by which most modern PI detectors perform ground balancing..! Typically the search head is pumped over the ground for a few seconds while the detector circuit works to balance out the ground. If you move the head up and down over the ground then you can ignore the time domain and simply either sample the phase of the return signal or simply take a sample well after the pulse has decayed but close enough to perform a differential cancellation. I am keen to look at processing this data inside the MCU so would look forward to understand a little more on this topic if anyone can add..?

[Davor]

PI struggles mostly with viscous magnetic materials in the ground that have 1/t response. It means they respond sharply at the beginning, but die out less rapidly than the targets later on. For that reason a late sample that is amplified against the main sample and subtracted from it is the means of doing ground balancing in PI.
That very effect is present in VLF as well, but you can't distinguish it as any different than the rest of the ground annoyances. Only trouble is that you may lack some GB tuning range for bad soils.

[Qiaozhi]

If you move the head up and down over the ground then you can ignore the time domain and simply either sample the phase of the return signal or simply take a sample well after the pulse has decayed but close enough to perform a differential cancellation. I am keen to look at processing this data inside the MCU so would look forward to understand a little more on this topic if anyone can add..?

There is no "phase of the return signal" in a pulse induction detector. As Davor stated, ground balance in a PI is achieved by taking a second sample a few micro-seconds (say 15us to 20us) after the main sample. Now ... if you amplify the second sample signal, and subtract it from the main sample, you can cancel the ground signal by adjusting the gain. During the pumping procedure, the user adjusts the second sample gain to cancel the ground, and then any metal target will produce in imbalance in the difference between the two samples. This is how the TDI works. Also, depending on the polarity of the difference calculation, the detector can give either a high or low tone to signify if the target is of low or high conductivity compared to the ground signal. It is also well known that this technique produces a "hole" in the target response. In other words, any good targets with a conductivity in the same range as the ground signal will also be rejected.

Don't forget that (in practice) you will need to subtract the second (GB) sample from the Earth field as well as the main sample. There are some variations on this scheme which have been discussed before, which you should be able to find with the search feature of this forum.

[Davor]

... or not.
I found the search function of this forum very uncooperative. After the forum was closed for robots, google and other search engines do not index it and finding things here became more difficult.

[greylourie]

Here is one thread : http://www.geotech1.com/forums/showt...ROL-PI-CIRCUIT
And another : http://www.geotech1.com/forums/showt...Ground-Balance