PI MD

[green]

Thanks for the replies on salt. Looking forward to trying the figure eight Rx on wet sand to see if it cancels salt signal better than it cancels ground. Any thoughts of some tests I might try while there? Been playing with the integrator with spice. We did some sampling of white noise in another thread awhile back. Added a target signal to the noise signal. Wondering if the spice simulation makes sense. Noise(out), +-15mv. Target(out), +12, -18mv. Simulated sweeping coil over target four times. First simulation first target at .6 seconds. Second simulation first target at .4 seconds. If the threshold were set at just over 15mv to not detect noise some of the target sweeps would be missed do to the noise subtracting from the target signal. Some questions. Is the noise signal similar to what I would see if I looked at the signal just before the threshold on MPP, Surf PI and some commercial PI's? The integrator schematic is similar to what I'm using, is there a better way to do it(different filter response times, more or less stages or a different way to do it, etc.) The EF sample slows the analysis so I didn't include it. Including the target+noise zip file. Simple enough to modify it to the other examples. Something I learned awhile back when working with slow simulations. If the simulation shows blank after right clicking run, right click halt, right click run again, click out and the simulation should show as it's working. Click on the simulation trace and save and it will show simulation as it runs the next time you open it. Probably not new to a lot of users.

Saving the trace file with the zip file doesn't save trace simulation while running. Need to repeat procedure when reopening the zip file.

[green]

Thanks for the replies on salt. Looking forward to trying the figure eight Rx on wet sand to see if it cancels salt signal better than it cancels ground. Any thoughts of some tests I might try while there? Been playing with the integrator with spice. We did some sampling of white noise in another thread awhile back. Added a target signal to the noise signal. Wondering if the spice simulation makes sense. Noise(out), +-15mv. Target(out), +12, -18mv. Simulated sweeping coil over target four times. First simulation first target at .6 seconds. Second simulation first target at .4 seconds. If the threshold were set at just over 15mv to not detect noise some of the target sweeps would be missed do to the noise subtracting from the target signal. Some questions. Is the noise signal similar to what I would see if I looked at the signal just before the threshold on MPP, Surf PI and some commercial PI's? The integrator schematic is similar to what I'm using, is there a better way to do it(different filter response times, more or less stages or a different way to do it, etc.) The EF sample slows the analysis so I didn't include it. Including the target+noise zip file. Simple enough to modify it to the other examples. Something I learned awhile back when working with slow simulations. If the simulation shows blank after right clicking run, right click halt, right click run again, click out and the simulation should show as it's working. Click on the simulation trace and save and it will show simulation as it runs the next time you open it. Probably not new to a lot of users.

Saving the trace file with the zip file doesn't save trace simulation while running. Need to repeat procedure when reopening the zip file.

Looking at noise again. Been turning a led on when signal exceeds a threshold setting. Wondering if adding audio would improve detection distance. Recorded a 4grain nugget swinging on a pendulum with nugget at 3, 4, and 5inches with GEB off(10.2usec target sample) and GEB on(adjusted to cancel ground, 10.2usec target sample, 100usec GEB sample). Noise signal is similar to the spice white noise analysis. Recording is negative absolute valued where spice signal is not. Been using 10mv threshold with GEB off and 15mv threshold with GEB on. Wouldn't have missed any targets swinging at 3 or 4inches. Swinging at 5 inches, maybe half. Setting the threshold at 8mv with GEB off would probably not miss any but GEB is probably needed where gold is found. Probably have to try an audio output to see if I can detect the targets the led is missing. mickstv had a circuit in another thread that I could try. The upper scope trace, a 8in disk glued to the bottom of the swinging bottle the nugget is fastened to that blocks the light to a photo cell mounted on the coil as it swings by. Still wondering if the noise trace with other detectors is similar(shape not amplitude)that I get with my detector and spice simulation. The short glitches on the traces happen when the coil current is turned off, still there if scope probe and ground lead are connected to common.