PI signal levels?

[Anonymous]

I am starting to seriously think about doing a PI project. But I need some numbers that I have just been guessing at so far.
I would like to know about the dynamic range of signals and signal to noise ratio. For example, I would like to know what the signal strength is for coin sized targets next to the coil, and the strength of the weakest recognizable signal from a distant coin, and the strength of the background noise.
I realize that these numbers depend on a lot of things, but I am interested in ratios, so a lot of the variables cancel out. I would like to know what these levels are at the input to the amplifier. That might be difficult to measure, so the levels at the output of the amplifier are ok as long at the gain is known and there is no AGC. The levels at the output of the integrator are not as useful because much of the noise has been filtered out by then. The more that I know about the conditions behind these numbers the more useful they will be.
If anyone has been working on a PI recently and has these numbers handy I would like to hear them.
Robert

 

[Anonymous]

Hi Robert,
Here are some quick measurements to give you an idea what we are dealing with. All tests were done using the Goldquest electronics and 11in diameter coil. The object used was a US nickel. The scope was connected to the output of the front end amplifier which is a 5534.
The main contribution of noise is rf pickup on the search coil, which in our area is predominantly am modulated 200kHz from a long wave transmitter about 60 miles away. The coil was oriented to give minimum pickup which was about 50mV pk-pk. Divide this by the gain of the front end, which is x 454 and you get 110uV.
Bringing up the nickel on the coil axis, 10mV additional deflection gives full audio response. This corresponds to 22uV of signal.
Moving the coin out so that it is just detectable gives no discernible movement on the scope, but I would guess it is about ten times less at 2 - 3uV.
The nickel in the centre of the coil gives 150mV, or 330uV at the input.
Amplifier noise (without coil) was 10mV pk-pk, equating to 22uV. As you say, all this is before any noise reduction due to sampling, integrating and filtering.
Eric.

 

[Anonymous]

Eric
Thank you, those are exactly the kind of numbers I wanted.
There were two thing in there that caught my attention. One is that the noise is so high. The other is that from your numbers I would estimate that the nickel was only about 11 to 12 inches away when it was just detectable, or about one coil diameter. Of course those two facts are related.
The good news is that these numbers imply that I may not need as many bits as I thought. The bad news is the noise.
Can you say what the bandwidth of the amplifier is?
Robert

 

[Anonymous]

Hi Robert,
The amplifier bandwidth is from dc to -3db at 20kHz.
Noise can be considerably worse in the field. In addition to any rf, you have low frequencies from movement in the earth's field and possibly 50/60Hz power line noise.
In the workshop, with the coil in its worst orientation, I get 300mV of rf noise at the amp. output. This causes no noticeable worsening of threshold stability, provided the TX/sampling pulse rates are not near synchronism with the interfering signal.
Regarding object signals, these can vary widely, depending on the size and conductivity. In the first post, I mentioned that a nickel (5.43 IACS) in the centre of the coil gives 150mV. A UK 50p coin (5.56 IACS and 28mm mean diameter)gives 300mV. A bronze disc (3.6 IACS and 40mm diameter) gives 700mV. A square stainless steel plate 50mm x 50mm (2.35 IACS) saturates the amplifier at 3500mV.
Amplitude measurements were taken at the start of the decay, when the amplifier has just come out of saturation due to the TX pulse. This corresponds to about 8uS after TXoff.
The IACS (International Annealed Copper Standard) figure is the percentage conductivity relative to annealed copper. The lower percentages show a much higher initial amplitude than better conductors would. The decay however is much faster.
The maximum air range on a nickel is 14 - 15in. You can't see the signal though, until after the integrator, when much of the noise is eliminated.
Eric.

 

[Anonymous]

Eric
If you can hear the nickel at 14 inches then I think you must be hearing signals down to 1 uV or a bit lower, even if you cannot see that low on the scope.
I do not care if something saturates on a steel plate, but I would like to be able to handle a range of signals from the smallest detectable up to large coins and rings at the coil without the A-D saturating. It sounds like, for that 11 inch coil, the range is about 1 uV to 1 mV at the input. But smaller coils should give a stronger signal with a coin at the coil.
Robert

 

[Anonymous]

Hi Robert,
Yes, you are right. I calculated the signal from the coil/object geometry (G) and got a figure of 0.79uV at 14in.
G = the cube of (1/1 + 2.545 squared), where 2.545 is h/r
Taking the figure of 330uV at h=0 you get 0.79uV when multiplied by G.
Interestingly, you can just hear the signal at an inch or so more, so the detector responds well to sub microvolt signals.
Eric.