detection depth

[Davor]

And how about ferrous targets?

[green]

I got side tracked again. The method Eric was using to determine optimum Tx width looked like an easier way to get target TC than what I was using. Probably knowing TC +-10% is good enough. Back to finding optimum Tx width that Eric started. Tried with spice. Target TC=500usec, 1oz copper or silver coin should be close. The coins have skin effect the spice target doesn't. Including chart, Rx amplitude is the peak reading. I'll try to repeat the test with a copper coin on the bench.

[green]

I got side tracked again. The method Eric was using to determine optimum Tx width looked like an easier way to get target TC than what I was using. Probably knowing TC +-10% is good enough. Back to finding optimum Tx width that Eric started. Tried with spice. Target TC=500usec, 1oz copper or silver coin should be close. The coins have skin effect the spice target doesn't. Including chart, Rx amplitude is the peak reading. I'll try to repeat the test with a copper coin on the bench.

Tried the 1oz copper coin at 125, 500 and 2000usec Tx width. Coin located in same location for each test. Added a 500usec TC line through each trace. The difference in amplitude between the different Tx width at 800usec looks a little higher with my test compared to spice difference. Test amplitude close to the same at 10usec(skin effect?). The decay slope isn't straight on a linear log chart until after 500usec(target TC)with Tx width four times the TC. More than 500usec with shorter Tx width. The data looks a little better after running a scope calibration, still the same 8inch figure8 Rx coil, .5A constant current Tx. Might try the 1.5inch figure8 Rx coil.

[green]

Tried the 1oz copper coin at 125, 500 and 2000usec Tx width. Coin located in same location for each test. Added a 500usec TC line through each trace. The difference in amplitude between the different Tx width at 800usec looks a little higher with my test compared to spice difference. Test amplitude close to the same at 10usec(skin effect?). The decay slope isn't straight on a linear log chart until after 500usec(target TC)with Tx width four times the TC. More than 500usec with shorter Tx width. The data looks a little better after running a scope calibration, still the same 8inch figure8 Rx coil, .5A constant current Tx. Might try the 1.5inch figure8 Rx coil.

Tried the 1.5inch figure8 Rx. About the same as 8inch figure8 Rx.

[Ferric Toes]

Hi Green - All good stuff.
I have improved my bench setup somewhat by changing the TX mosfet to a 500V device (IRFP22N50A) to shorten the avalanche period. The Rds(on) is higher at 0.23 ohms but then I have also introduced 10 ohms in series with coil. This reduces the max coil current to 0.6A and I still get adequate RX signal so that noise is not a problem. The recovery period at the preamp output is such that I can now see a clean target signal from 10uS onwards. I have a 10turn pot to control the TX width and also a 10turn wirewound pot in the coil damping network so as to optimise the RX response for earliest recovery.

Triggering the scope on TX off and setting a US quarter to give the start of the decay, just going into saturation and just after 10uS, I then adjust the TX width until the 50% amplitude point stops moving to the left on the screen graticule. Increase timebase speed and Y sensitivity to 100mV to give greater resolution. Any increase in TX width beyond that has no effect. The quarter has then reached its optimum TX width to give full magnetic saturation.

Change the trigger polarity for start of the TX and adjust vertical position and sensitivity so that the width of the TX pulse can be measured with the cursors. Doing the above, I get 600uS for the best width of the TX pulse and 120uS for the fundamental TC. I think we have to ignore the early time responses as you do on your log/lin plots, taking the fundamental TC from where the straight line starts.

I will do the silver quarter, 1oz silver, and silver 1/2 quarter and post screen shots when done.

What I would like to get is a digital scope module to use with a Windows laptop and that has both linear and logarithmic Y and X axes. If only Y then that would be good, but an X Y would be great for soil magnetics too where I need a LogY LogX display. If anyone knows of one that does not cost the earth, please let me know.

Eric.

The 1st picture shows the final measurement of the TX width. The second picture shows the approx 50% part of the waveform expanded and crossing the centre horizontal line on the screen. If I reduce the TXon time to 260uS the waveform moves one whole cm to the left. Increasing it above 600uS and it stays put, showing that the quarter is fully charged magnetically.

[Ferric Toes]

And how about ferrous targets?

I decided in 1967 that ferrous and non-ferrous identification by means of the off time response was not going to be easy. Wish I still had that Tektronix scope with waveform scanning attachment and log converter. All tube(valve) too and kept the room warm in winter. XY plotter did the drawings.

Eric.

[green]

Hi Green - All good stuff.
I have improved my bench setup somewhat by changing the TX mosfet to a 500V device (IRFP22N50A) to shorten the avalanche period. The Rds(on) is higher at 0.23 ohms but then I have also introduced 10 ohms in series with coil. This reduces the max coil current to 0.6A and I still get adequate RX signal so that noise is not a problem. The recovery period at the preamp output is such that I can now see a clean target signal from 10uS onwards. I have a 10turn pot to control the TX width and also a 10turn wirewound pot in the coil damping network so as to optimise the RX response for earliest recovery.

Triggering the scope on TX off and setting a US quarter to give the start of the decay, just going into saturation and just after 10uS, I then adjust the TX width until the 50% amplitude point stops moving to the left on the screen graticule. Increase timebase speed and Y sensitivity to 100mV to give greater resolution. Any increase in TX width beyond that has no effect. The quarter has then reached its optimum TX width to give full magnetic saturation.

Change the trigger polarity for start of the TX and adjust vertical position and sensitivity so that the width of the TX pulse can be measured with the cursors. Doing the above, I get 600uS for the best width of the TX pulse and 120uS for the fundamental TC. I think we have to ignore the early time responses as you do on your log/lin plots, taking the fundamental TC from where the straight line starts.

I will do the silver quarter, 1oz silver, and silver 1/2 quarter and post screen shots when done.

What I would like to get is a digital scope module to use with a Windows laptop and that has both linear and logarithmic Y and X axes. If only Y then that would be good, but an X Y would be great for soil magnetics too where I need a LogY LogX display. If anyone knows of one that does not cost the earth, please let me know.

Eric.

The 1st picture shows the final measurement of the TX width. The second picture shows the approx 50% part of the waveform expanded and crossing the centre horizontal line on the screen. If I reduce the TXon time to 260uS the waveform moves one whole cm to the left. Increasing it above 600uS and it stays put, showing that the quarter is fully charged magnetically.

[ATTACH]41802[/ATTACH][ATTACH]41803[/ATTACH]

Hi Eric,
I get about 130usec for clad quarter TC so the 600usec Tx width makes sense.

Digital scope module. I use a DS1052E scope. Wouldn't want a lower sample rate for what I'm charting. 10 or 12 bits would be better, added a amplifier x9 after preamp for some more bits. Problem is, needs second channel cutting sample rate in half. I store two recordings, no target and target. Copy into Excel. Subtract no target recording from target recording with Excel and chart. The scope has normal, average and peak. The scope screen is a lot cleaner with average, would be nice if storage stored an average. Probably other things to consider.

[Skippy]

Interesting vintage pen-plots, Eric, dating from when I was learning to walk and talk. For the benefit of Mr.Green and anyone else, the trace 1 Half-Crown is a large half-dollar-sized cupro-nickel coin, hence the short time-constant (15 - 20 usec roughly).

[green]

More charts. Charted a US clad quarter and a 1oz copper coin at 250usec and 1200usec full scale. Wanted to try the quarter with Tx width 1/4, 1, and 4 times the quarter's time constant. Used the 125usec, 500usec and 2000usec Tx width used for charting the 1oz copper coin with a 500usec TC. Wanted to see if the copper coin repeated. The quarter had a 135usec TC, a little higher than the 130usec I was thinking so the 1 and 4 times the TC were a little off. Didn't record the 1/4 times the TC for the quarter. The Tx current is in a .5A constant current control loop. Fairly straight ramp up to .5A but it still takes about 20usec. The ramp would be longer than the flat top for the 31.5usec Tx width, 1/4 of 125usec.

The copper coin did repeat. The 500 and 2000usec Tx width for the quarter charted the same as expected. Added the 250usec full scale chart to show better what is happening earlier in the decay.

The output of the two amplifiers don't meet on some of the earlier charts. Didn't know what was causing it. Did an internal scope calibration yesterday, solved the problem. The temperature in the garage was a lot warmer today and I had to do another internal calibration to get them to meet. Looks like CH1 and CH2 gains don't track with temperature but it's good the internal calibration solves it.

My charts don't always look the same. Any suggestions what should be changed or added to todays charts to make them more useful?

[green]

I decided in 1967 that ferrous and non-ferrous identification by means of the off time response was not going to be easy. Wish I still had that Tektronix scope with waveform scanning attachment and log converter. All tube(valve) too and kept the room warm in winter. XY plotter did the drawings.

Eric.

[ATTACH]41804[/ATTACH] [ATTACH]41805[/ATTACH]

What I would like to get is a digital scope module to use with a Windows laptop and that has both linear and logarithmic Y and X axes. If only Y then that would be good, but an X Y would be great for soil magnetics too where I need a LogY LogX display. If anyone knows of one that does not cost the earth, please let me know.

I've been subtracting a no target recording from a target recording with Excel and then charting with linear or log XY scales. Depends on zero's and gains staying the same between the two recordings. Does cancel out offsets caused by the amplifier, detector or another target being to close. Was thinking recording a log scale might be a problem. Do you remember what the log scale was on your recordings? I added 10mv to one of the 1oz copper recordings after subtraction to simulate a 10mv offset making a log Y recording. Your traces look alright, do you remember any problems with no target zero offset. I would probably have a problem closer to zero time.

[green]

Was looking at log amplifiers. If I record the decay linear Excel will chart it linear or log scale. If I record the decay log how do I chart on a log scale or shouldn't I want to.

[Skippy]

In reply to "Eric: Do you remember what the log scale was on your recordings?"
I was wondering the same thing when looking at them yesterday. It appears that 1 inch vertical represents a doubling of signal. The smallest squares are 0.1". Hence, for example, the decay of the 1967 Penny takes about 75 usec to drop 1", indicating TC = 75 usec. I've measured 84 usec on a 13KHz VLF, so would expect 90 - 100 usec for a PI, so sort-of near, but it may be something odd, like 1.5" = doubling.

And re: your log charts: If you log the data, then record it linearly, you just print it linearly... but you'll have to pay attention to the scaling, as it will only be accurate every decade. So it will be dang hard to read off any data from the graphs. Hope that made sense..

[green]

In reply to "Eric: Do you remember what the log scale was on your recordings?"
I was wondering the same thing when looking at them yesterday. It appears that 1 inch vertical represents a doubling of signal. The smallest squares are 0.1". Hence, for example, the decay of the 1967 Penny takes about 75 usec to drop 1", indicating TC = 75 usec. I've measured 84 usec on a 13KHz VLF, so would expect 90 - 100 usec for a PI, so sort-of near, but it may be something odd, like 1.5" = doubling.

And re: your log charts: If you log the data, then record it linearly, you just print it linearly... but you'll have to pay attention to the scaling, as it will only be accurate every decade. So it will be dang hard to read off any data from the graphs. Hope that made sense..

Thanks. Was looking at log114 amplifier from TI. Looks like it has enough frequency response. VO = (0.375V) Log (I1/I2), could scale to .5volt/decade but I didn't see a good way to chart a scale I could read.

[Ferric Toes]

In reply to "Eric: Do you remember what the log scale was on your recordings?"
I was wondering the same thing when looking at them yesterday. It appears that 1 inch vertical represents a doubling of signal. The smallest squares are 0.1". Hence, for example, the decay of the 1967 Penny takes about 75 usec to drop 1", indicating TC = 75 usec. I've measured 84 usec on a 13KHz VLF, so would expect 90 - 100 usec for a PI, so sort-of near, but it may be something odd, like 1.5" = doubling.

And re: your log charts: If you log the data, then record it linearly, you just print it linearly... but you'll have to pay attention to the scaling, as it will only be accurate every decade. So it will be dang hard to read off any data from the graphs. Hope that made sense..

It's hard to remember exactly the details of something I did over 50 years ago, but this sheet might help. The initial plots were noisy as the lab was in a main road and vehicles were not so well supressed as they are now. Fluorecent lights also were a problem, so later I turned them off when running a test. The plot shown was for a magnetically mineralised soil test; hence the non-linear log plot.

Eric.

[Skippy]

Sorry, my last post was in error, I was tired.
I think my assessment of the vertical scale of Eric's curves is correct: 1" represents 2-fold gain (or drop).
The 1967 penny drops 2" in 140 usec, and hence 1" in 70 usec. So Time-constant = 70 / (0.693) = 100 usec. This compares to the 84 usec I got on my VLF, which seems to under-read TC (probably due to skin effect).
Like-wise, trace 5, the Cu-Ni half-crown takes 35 usec to fall by 3", = 11.7 usec for 1" drop. So TC = 11.7 / (0.693) = 16.8 usec. This roughly matches the 18 usec my VLF indicated.