PI SAMPLE DELAY QUESTION

[daverave]

i was doing some adjustments to my surf pi detector and thought that i would measure the sample delay with my scope....its about 25 micro seconds which i guess is not good proformance...is 25 micro seconds still good for general beach treasure hunting ??? is the white's dual field better than 25 micro seconds delay ????

[baum7154]

The Barracuda detector is generally set in the vicinity of 25us minimum sample delay, but it can go lower with a tweak of the trimmer pot. If you want to push it a bit you could try to go as low as 15us and see what happens if your coil is fast enough. It could get 'chattery' on the beach.

Regards,

Dan

[Silver Dollar]

I'm not sure if a chart was ever published to show what targets are findable at different delay settings.
For tiny specs of gold you want to go fast (and necklaces act as tiny specs) so people usually want to
sample as soon as possible. For general use and finding coins 25us might be just fine. Salt water will limit
you to 10 to 15us so getting to there might be worthwhile. A simple cap or resistor change will allow you
to sample sooner. Whether your coil is fast enough to work at that range is another story...

I have a few unfinished PI's laying around and would love a chart of what is findable at different delays
to help me set them up. Any one have such info?

[bbsailor]

The Barracuda detector is generally set in the vicinity of 25us minimum sample delay, but it can go lower with a tweak of the trimmer pot. If you want to push it a bit you could try to go as low as 15us and see what happens if your coil is fast enough. It could get 'chattery' on the beach.

Regards,

Dan

Dan,

Is the Barracuda PI coil operating at 25us shielded?

Thanks

Joseph Rogowski

[daverave]

ive tried sampling earlier but then i loose depth...so i guess the coil is the problem which is not fast enough...i increased the tx frequency which has helped some what....im in the process of making another coil with litz 24 awg 19 strand wire which i hope will give me a little better proformance hopefully.

[green]

I'm not sure if a chart was ever published to show what targets are findable at different delay settings.
For tiny specs of gold you want to go fast (and necklaces act as tiny specs) so people usually want to
sample as soon as possible. For general use and finding coins 25us might be just fine. Salt water will limit
you to 10 to 15us so getting to there might be worthwhile. A simple cap or resistor change will allow you
to sample sooner. Whether your coil is fast enough to work at that range is another story...

I have a few unfinished PI's laying around and would love a chart of what is findable at different delays
to help me set them up. Any one have such info?

I've plotted some signal strength vs distance curves in the past. The data plots similar to time constant curves, The amplitude drops 63 percent every (X) distance. A 5 inch coil I plotted dropped 63 percent every 1.2 inches. A 9.5 inch coil every 1.7 inches. Including a plot with a nickel for the target with the coil shaped round and rectangular. The amplitude should drop the coil distance constant every target time constant. A nickel with a 10 usec time constant would loose 1.8 inches for every 10 usec increase in delay for the plotted curves. The curves aren't quite straight but I think it should give an idea what to expect. Some target time constant curves aren't straight, probably have to use time constant to match the delay time. Maybe makes sense maybe not, its what I'm thinking.

[daverave]

ive just wound another coil using the litz 24 awg 19T and my inductance is around about 300 micro/h at 2 ohms...i feel like taking a couple of turns off to get the inductance lower but then i guess the magnetic field will be weaker cause of less turns,if i go down to around 250 micro/h will the magnetic field be weaker and less depth...its all very confusing at times getting all these parameters correct

[baum7154]

I believe it is shielded. The Barracuda is Satdaveuk's machine and he shields his coils with graphite. Best to ask him though.

Regards,

Dan

[green]

ive just wound another coil using the litz 24 awg 19T and my inductance is around about 300 micro/h at 2 ohms...i feel like taking a couple of turns off to get the inductance lower but then i guess the magnetic field will be weaker cause of less turns,if i go down to around 250 micro/h will the magnetic field be weaker and less depth...its all very confusing at times getting all these parameters correct

I'm thinking (maybe wrong) if you removed half the turns the resistance would be half and the current would double if the coil driver could drive it. The field strength would be the same if the coil time constant stayed the same. The time constant would be shorter so maybe the field strength would be higher with less turns at the expense of higher current.

[bbsailor]

ive just wound another coil using the litz 24 awg 19T and my inductance is around about 300 micro/h at 2 ohms...i feel like taking a couple of turns off to get the inductance lower but then i guess the magnetic field will be weaker cause of less turns,if i go down to around 250 micro/h will the magnetic field be weaker and less depth...its all very confusing at times getting all these parameters correct

Dave,

Getting it "correct" is a good talking point because "correct" is dependent on three primary things:

1. The nature of the target: size, material, time constant to charge the target with enough energy to discharge and be detected.

2. The nature of the ground in which the targets are located, including ground response and variation in a single coil sweep. Highly responsive ground may limit the TX pulse length to minimize the ground response but if you looking for longer TC targets, there is a compromise that must be made.

3. Tools available including coil sizes, coil types (mono, DD, concentric, etc); TX and RX control settings including TX pulse length, TX frequency, delay range, and noise reduction techniques.

Getting the parameters correct is the optimization of all of the above to accomodate the target types sought and the locations in which the targets are located. Ocean beach machines are a typical example of this using PI technology, and waterproof coils with a delay range in the 15us range (plus or minus a few us) to locate coins and lost jewlery on the beach or in shallow salt water. However, coins and small gold jewlery require a different optimization strategy which means that if you find a lot of coins, then this may be a place where people have lost jewlery. Then you change coils and adjust your PI machine to be more sensitive to small gold and rehunt the same area again.

The ultimate value of people talking about this stuff is to arrive at a point where a microprocessor based metal detector will first allow the user to analyze the ground and enter the type of targets sought and the machine will recommend coil sizes, TX pulse widths, sweep speeds, delay and filter settings optimized for those specified primary targets in that specific ground.

Your 300uH coil at 2 ohms has a 150us Time Constant (TC). A150us long TX pulse will rise the current to 63 percent of 6 amps (12V/2 ohms) or 3.78A. The key thing for making a fast coil is to produce a coil with the least capacitance and a properly measured damping resistor to allow the earliest sampling for seeking very small low TC gold targets. However, if you are seeking coins on the beach, a larger coil is more efficient and maybe a longer pulse (now 2 TCs long) of 300us will allow the current to rise to 86 percent of 6A or 5.16A. With the higher current, the flyback voltage will be higher and require a lower value damping resistor to allow the earliest sampling.

Don't let the flyback voltage exceed the peak MOSFET voltage by too much as it will cause the MOSFET to clamp the flyback peak, create heat in the MOSFET and slightly delay the earliest sampling point.

Getting the parameters correct means that you are doing a good balancing act by understanding the interaction of the above variables.

I hope this helps?

Joseph Rogowski

[daverave]

can you explain in more detail about time constants cause im very confused by this...i did try and increase the tx pulse but i found that too high the detector lost most of its depth...i then measured the back emf which was at 250v which was the limit for the mosfet used in the surf pi circuit....maybe if a different mosfet was used with higher rating say 400v then depth maybe better but i guess then sampling will be later and what you gain with one parameter you loose with another if that is correct ????? my detector responds well to copper type coins and silver rings but not so good on go
ld rings and us nickles using a 12" coil and search frequency of 2.6khz...i guess as you said its down to coil size for different types of targets....you also mentioned about correct damping resistor which again im confused about !!! cause are we looking for a nice clean back emf curve or is some ringing acceptable so we can sample early in the ringing ??? and maybe get better sensitivity on gold items.

[bbsailor]

can you explain in more detail about time constants cause im very confused by this...i did try and increase the tx pulse but i found that too high the detector lost most of its depth...i then measured the back emf which was at 250v which was the limit for the mosfet used in the surf pi circuit....maybe if a different mosfet was used with higher rating say 400v then depth maybe better but i guess then sampling will be later and what you gain with one parameter you loose with another if that is correct ????? my detector responds well to copper type coins and silver rings but not so good on go
ld rings and us nickles using a 12" coil and search frequency of 2.6khz...i guess as you said its down to coil size for different types of targets....you also mentioned about correct damping resistor which again im confused about !!! cause are we looking for a nice clean back emf curve or is some ringing acceptable so we can sample early in the ringing ??? and maybe get better sensitivity on gold items.

Dave,

The nature of large copper coins puts their TC in the 100us range so that a TX pulse of about 300us fully charges the mass of the coin's metal to release eddy currents when the TX current is suddenly turned off. The quick discharge of the TX charge current releases the eddy current energy stored during the TX pulse that is detected in the RX window right after the delay time set on the PI delay setting. The optimum value damping resistor will allow you to move the RX sample window as close to the TX turn off point to begin detecting the residual target eddy currents just after the set delay time. If the delay were set at 15us then the RX sample window would detect targets from about 16us to 30us but during the initial 15us delay, low TC targets like gold nuggets, some gold jewlery and some metals will be decaying quickly during the 15us delay time and may not have much energy current during the RX period. Only targets that are on the edge of being detected will be mostly affected by small tweeks in reducing the delay.

The energy in the flyback pulse needs to be damped so that the turn off current and resulting flyback pulse is damped to zero volts as soon as possible. The closer the turn off slope is to 90 degrees vertical the better. Since all coils have some capacitance in the form of the construction techniques, including the coax capacitance, coil wire turn to turn capacitance, shielding capacitance, and MOSFET COSS will store energy and form a resonant frequency that extends the time it takes to fully damp the flyback pulse. That is why it is better to fine tune the coil and damping resistor to the PI machine for optimum results.

The value of the damping resistor (Rd) determines the discharge slope of the turn off current. This also form the discharge TC of the coil. A 300uH coil with a 1000 ohm Rd will have a discharge TC of 300/1000 or 0.3us which is optimum for targets with TCs 5X higher or 1.5us. The optimum turn off pulse TC is 5 X faster than the target TC.

When the MOSFET clamps for a few microseconds to keep the flyback peak at the MOSFET voltage, the time during which the pulse is flat topped (clamped) will cause a delay in the time when the earliest sample can occur.

When you are using a TX pulse in the 3K PPS range, you will be integrating many samples and the RX sensitivity will be somewhat affected by the speed of your sweep. Try speeds of 1.2M sweep width per 2 seconds and then try 1.2M sweep width in 1 second to see if the sweep speed affects sensitivity.

On my high speed PI machines, I get better range on U.S. nickels than on U.S. laminated quarters.

Joseph Rogowski

[daverave]

talking about sweep speed...in air test by passing target across coil slowly i find sensitivity is better....i can see now that time constant is time taken to charge/discharge the targets...im still a little unsure about correct damping...cause surely to get a 90 angle slope there must be some ringing present ????

[Qiaozhi]

talking about sweep speed...in air test by passing target across coil slowly i find sensitivity is better....i can see now that time constant is time taken to charge/discharge the targets...im still a little unsure about correct damping...cause surely to get a 90 angle slope there must be some ringing present ????

The idea that metal targets are "charged/discharged" is incorrect. This is a misconception which implies the target is somehow charged with eddy currents during the TX on-time, and this charge is then released.

If you were to apply a voltage to the TX coil for a sufficiently long time (such that the current flowing in the coil has reached a steady-state dc condition) there will be zero eddy currents flowing in the target. It is only at the point when the TX pulse is turned off, and the magnetic field around the coil collapses, that eddy currents start to form in the target. These currents will rapidly reach a peak, and then start to exponentially decay. You need to have a rapidly changing magnetic field to create eddy currents.

Note that eddy currents are also generated in the target during the initial turn-on of the TX pulse; and again these will die away swiftly as the coil current flattens out. These currents are not responsible for any "charging" of the target. In fact, if still present in the target when the TX pulse turns off, they will have a negative impact on the target response.

[Carl-NC]

can you explain in more detail about time constants cause im very confused by this...i did try and increase the tx pulse but i found that too high the detector lost most of its depth...

As you increase the TX pulse width the flyback response slows down, so probably what happened was that the sample point (15us or whatever) became overloaded. If you increase the sample delay things will start working again.