TINKERERS TIMING SCHEDULE

[nick_f]

Hi Nicolae,

The timing adjustments are made in code. If you are only running one timing 1us resolution will be ok, provided you have some analog means of fine tuning signals. More timings may be possible without further analog adjustment but I'm not 100% certain.

Cheers Mick

Hi Mick,
I am not sure why do you need such a high resolution and how are you actually adjusting the pulses. Does this mean that if you want to change a pulse width or delay, you have to reprogram the MCU?
What parameter of what pulse is so critical that you consider useful to have this kind of resolution?
I make my PIC microcontroller for the Hammerhead circuit. I haven't run it on a real circuit yet, I am just refining the code for MCU. At the moment I don't see a reason to go lower than that for any of the pulses. The widths of the sample pulses will be always equal, as it should.

Regards,
Nicolae

[Tinkerer]

G`day Tinkerer,happy new years mate.

Tinkerer i wouldnt necessarily concur with a 3.3hz freq of a target signal,Sweep speed does vary alot,its not at all uncommon for pro operaters to sweep at speeds of 1 meter per 4 seconds,also large targets can give a very broad signal down to as low as 0.5hz per meter .

Zed

A happy new year to you too, may all your dreams come true.

I totally agree about the variability of the sweep speed, but we need some kind of reference speed. 1meter/second is the standard de-mining sweep speed, so I use this as a reference. 3.3hz then refers to a coil of 30cm diameter used with a sweep speed of 1 meter/second.
If you use a coil with 50cm diameter, with the same sweep speed, it would be down to 2hz. Just reference numbers.

If I use a certain Ground Balance update time, it should be referenced to the frequency of the target signal. Thus, I make a variable Ground Balance update time.
The ideal would be to add an accelerometer to the detector, placed near the coil, that would, via the MCU, adjust the Ground Balance update time, according to the sweep speed and signal frequency.

Once we have an indication of a target, it would be better to leave the Ground Balance fixed, this is why the TINKERERS_V1 switches to pinpoint mode.

Ideally the pinpoint mode should be fully static, non-motion, do you agree with that?

I much appreciate your feedback. I can not go much out anymore myself, so I depend on the help of experienced prospectors who have the feel of real world conditions.

Tinkerer

[ZED]

A happy new year to you too, may all your dreams come true.

I totally agree about the variability of the sweep speed, but we need some kind of reference speed. 1meter/second is the standard de-mining sweep speed, so I use this as a reference. 3.3hz then refers to a coil of 30cm diameter used with a sweep speed of 1 meter/second.
If you use a coil with 50cm diameter, with the same sweep speed, it would be down to 2hz. Just reference numbers.

If I use a certain Ground Balance update time, it should be referenced to the frequency of the target signal. Thus, I make a variable Ground Balance update time.
The ideal would be to add an accelerometer to the detector, placed near the coil, that would, via the MCU, adjust the Ground Balance update time, according to the sweep speed and signal frequency.

Once we have an indication of a target, it would be better to leave the Ground Balance fixed, this is why the TINKERERS_V1 switches to pinpoint mode.

Ideally the pinpoint mode should be fully static, non-motion, do you agree with that?

I much appreciate your feedback. I can not go much out anymore myself, so I depend on the help of experienced prospectors who have the feel of real world conditions.

Tinkerer

Im not sure,i like motion,have`nt used non-motion before...Auto GB CCTS are a hard nut to crack because the ground response (nasty ground that is) also has a tendancy to have the same freq as targets and so it is very hard to differentiate between the two when setting up or designing a auto GB method.
You can compromise by setting up active and passive filters to filter out freq above say 7hz and below 3hz,but you will loose sensitivity on targets with a longer TC.

Zed

[Mechanic]

Hi Mick,
I am not sure why do you need such a high resolution and how are you actually adjusting the pulses. Does this mean that if you want to change a pulse width or delay, you have to reprogram the MCU?
What parameter of what pulse is so critical that you consider useful to have this kind of resolution?
I make my PIC microcontroller for the Hammerhead circuit. I haven't run it on a real circuit yet, I am just refining the code for MCU. At the moment I don't see a reason to go lower than that for any of the pulses. The widths of the sample pulses will be always equal, as it should.

Regards,
Nicolae

Hi Nicolae,

It is more user friendly having higher resolution pulses. For instance, lets say that the original samples were 15us each and the trim pot for earth balance was set to null . If I then make a new timing that has samples that are 45us each, the earth balance will not be exactly nulled as it was before with the 15us samples. This would be from the difference in switching speeds of the components, tolerance errors, resistors not perfectly matched, opamps not perfectly matched etc.

So at this point I have 3 options.
1. Pull the detector apart and reset the trimpot. However doing this prevents from changing back to your other timing without readjustment.
2. Just put up with the groaning and false noise when the coil is tilted and swept.(More on this)
3. Adjust the 2nd sample width in little nibbles until the earth field is nulled the best. Even in 125ns increments the resolution could be a bit better.

Now, another reason why earth field nulling is good. Here in Australia alot of our goldfield ground contains either a residual magnetism, or tends to concentrate the earths magnetic field around itself? Either way without even transmitting a pulse I can detect a hotrock that I have, when only using 1 sample pulse(no earth field nulling). But if I set up the second sample to subtract from the first(earth null) with no transmit there is no response. Having the null out a little bit can give false signals from magnetic ground, so ideally you don't want to put up with the groaning.

Having said that, if you only intend to use 1 timing or samples of the same width, 1us resolution will be fine, provided you have analog means to fine tune the important bits. I have not looked into the hammerhead so I don't know what provisions it has to deal with these problems. Good luck with your hammerhead, and let us know how it goes.

Cheers Mick

[nick_f]

Hi Nicolae,

It is more user friendly having higher resolution pulses. For instance, lets say that the original samples were 15us each and the trim pot for earth balance was set to null . If I then make a new timing that has samples that are 45us each, the earth balance will not be exactly nulled as it was before with the 15us samples. This would be from the difference in switching speeds of the components, tolerance errors, resistors not perfectly matched, opamps not perfectly matched etc.

So at this point I have 3 options.
1. Pull the detector apart and reset the trimpot. However doing this prevents from changing back to your other timing without readjustment.
2. Just put up with the groaning and false noise when the coil is tilted and swept.(More on this)
3. Adjust the 2nd sample width in little nibbles until the earth field is nulled the best. Even in 125ns increments the resolution could be a bit better.

Now, another reason why earth field nulling is good. Here in Australia alot of our goldfield ground contains either a residual magnetism, or tends to concentrate the earths magnetic field around itself? Either way without even transmitting a pulse I can detect a hotrock that I have, when only using 1 sample pulse(no earth field nulling). But if I set up the second sample to subtract from the first(earth null) with no transmit there is no response. Having the null out a little bit can give false signals from magnetic ground, so ideally you don't want to put up with the groaning.

Having said that, if you only intend to use 1 timing or samples of the same width, 1us resolution will be fine, provided you have analog means to fine tune the important bits. I have not looked into the hammerhead so I don't know what provisions it has to deal with these problems. Good luck with your hammerhead, and let us know how it goes.

Cheers Mick

Hi Mick,
Thank you very much for the details, I understand now the need for this fine trimming. In the case of Hammerhead circuit, there is a pot (R23) that can do the balance for the sample signals.
By default, the schematics uses equal pulse widths for the two samples and the equilibration is done via this pot. I just took mine apart and played with that pot and I am unable to get much difference from adjusting it. I need to put the 8-shaped coil, for testing inside...

Regards,
Nicolae

[Mechanic]

Hi Nicolae,

If you have some way to turn the transmitter off but still keep the receive side of things going, get a strong small magnet and wave that about 2" high across your coil at 1m/s while adjusting the pot. Find the area that has the least response.

Cheers Mick

[Woody.au]

Mick,
Looks like you have the 2000 sussed out pretty well. I like your Earth Field method of cancellation. Beats shorting 2 integration channel gain stages and wobble the coil to obtain a null. Now the Question on the magnet, do you use Fridge magnet sheeting? I have never obtained much output from it.

[Mechanic]

Hi Woody,

You still need to do each channel separately, by either shorting the jumpers, or if you are able to turn them off in the program.

I use a pretty strong earth magnet. A small speaker magnet would work too. I doubt that a fridge magnet would give much of a response. By using a strong magnet you know when you have found the right spot even with all of the other interference.

Cheers Mick

[Tinkerer]

Hi Nicolae,

It is more user friendly having higher resolution pulses. For instance, lets say that the original samples were 15us each and the trim pot for earth balance was set to null . If I then make a new timing that has samples that are 45us each, the earth balance will not be exactly nulled as it was before with the 15us samples. This would be from the difference in switching speeds of the components, tolerance errors, resistors not perfectly matched, opamps not perfectly matched etc.

So at this point I have 3 options.
1. Pull the detector apart and reset the trimpot. However doing this prevents from changing back to your other timing without readjustment.
2. Just put up with the groaning and false noise when the coil is tilted and swept.(More on this)
3. Adjust the 2nd sample width in little nibbles until the earth field is nulled the best. Even in 125ns increments the resolution could be a bit better.

Now, another reason why earth field nulling is good. Here in Australia alot of our goldfield ground contains either a residual magnetism, or tends to concentrate the earths magnetic field around itself? Either way without even transmitting a pulse I can detect a hotrock that I have, when only using 1 sample pulse(no earth field nulling). But if I set up the second sample to subtract from the first(earth null) with no transmit there is no response. Having the null out a little bit can give false signals from magnetic ground, so ideally you don't want to put up with the groaning.

Having said that, if you only intend to use 1 timing or samples of the same width, 1us resolution will be fine, provided you have analog means to fine tune the important bits. I have not looked into the hammerhead so I don't know what provisions it has to deal with these problems. Good luck with your hammerhead, and let us know how it goes.

Cheers Mick

Hi Mick,

I find the idea of nulling the earth's field response, using a test target, in this case a magnet, excellent.
It would be good if we could define the test target a bit more.
1) I assume the magnet is of the ceramic kind.
2) How large or strong the magnet needs to be will depend on the aggressiveness of the magnetic ground that the detector needs to balance.
3) If we use a standard size and shape magnet, we could change the distance to the coil to simulate different magnetic ground situations.

When designing a new model of metal detector, we could include in the design a cct, that stops the TX and allows the various tuning procedures in the field.

I think we could all benefit by further exploring the idea.

Tinkerer

[ZED]

Hi Mick,

I find the idea of nulling the earth's field response, using a test target, in this case a magnet, excellent.
It would be good if we could define the test target a bit more.
1) I assume the magnet is of the ceramic kind.
2) How large or strong the magnet needs to be will depend on the aggressiveness of the magnetic ground that the detector needs to balance.
3) If we use a standard size and shape magnet, we could change the distance to the coil to simulate different magnetic ground situations.

When designing a new model of metal detector, we could include in the design a cct, that stops the TX and allows the various tuning procedures in the field.

I think we could all benefit by further exploring the idea.

Tinkerer

Now thats a good idea !

Zed

[gef12]

Just a point of note here..
WHen I was out in the field t/sing my 2000 ... I had the GB earth field problem .. I noticed that when facing Nth/Sth and swinging E/W that I was getting more earth field noise than facing E/W and swinging coil Nth/Sth...
As you are swing across the fields
So when tuning with the magnet do we use it with the Nth/Sth poles of the magnet axis in body line and swing coil or swing magnet E/W or vici verca ...
Just curious

[Mechanic]

Hi Gef,

I don't think it really matters. You are canceling a slow moving strong magnetic field.

Tinkerer,
The magnet I have is of the steel kind. It is very strong. When I put it on a thick piece of steel I can't pull it off. The idea is to set the earth balance as close to perfect as possible, then magnetic soils and earth field will all be nulled. It is also worth pointing out that just because we can cancel magnetic response, this is only slow changing magnetic response. This kind of soil/ rocks will have a fast magnetic decay when pulsed with a coil which is picked up as either a short T/c response or a long tc response and or anywhere in between. This is where the trick of ground balance comes in.

Cheers Mick

[Tinkerer]

Ground Balance

Thanks Mick for the detailed explanation.
I understand now, that it is not just the earths field, but a complete Ground Balance, for the worst possible ground that you do.

When you Ground Balance for the worst possible ground and use this setting in mild ground. How does it affect the sensitivity in mild ground?

Tinkerer

[Mechanic]

Thanks Mick for the detailed explanation.
I understand now, that it is not just the earths field, but a complete Ground Balance, for the worst possible ground that you do.

When you Ground Balance for the worst possible ground and use this setting in mild ground. How does it affect the sensitivity in mild ground?

Tinkerer

Hi Tinkerer,

Not quite ground balance, just magnetic field cancellation and some emi cancellation too . The gound balance works by taking 2 late samples when it assumed that most target signals have decayed. At this point the ground will still be decaying with the first sample taken late after the large tx and the second sample taken late after the short tx. The result of these 2 samples is subtracted from each target channel. The amplitude of the result is fine tuned using the 10 turn GB pot for each channel, and then added and fet to the lowpass filter and averaged.

I hope this helps!

Cheers Mick

[revmaster]

Hi Tinkerer,

Not quite ground balance, just magnetic field cancellation and some emi cancellation too . The gound balance works by taking 2 late samples when it assumed that most target signals have decayed. At this point the ground will still be decaying with the first sample taken late after the large tx and the second sample taken late after the short tx. The result of these 2 samples is subtracted from each target channel. The amplitude of the result is fine tuned using the 10 turn GB pot for each channel, and then added and fet to the lowpass filter and averaged.

I hope this helps!

Cheers Mick

Hi Mick,

Good to see lot of development going across the forum.
Could you please elaborate more on ground balancing method.
Do you mean to send 2 pulses first one big pulse for signal and the second one very small to read ground signal.
I tried this, it doesn't works that good but lot of ground signal is still there.

Cheers,
Michael.