Understanding PI machines

[Wet and Dry]

I have being detecting for a while and own two White's M/Ds.(DFX-BHID) Both are VLF, now I think I would maybe like to try a PI detector. I have read all the post and I'm still not sure I understand how the machine works. Is there a book out there that will explain how they work and the proper way to set the detector up for the different ground conditions out there. I think I need to understand pulse delay ,rate etc and all the other stuff too truly got the best out of a machine. Thanks for all the help I will get.

 

[nero_design]

As far as I know, the detectors themselves generally handle the pulse delay because it's preset. The detector emits a powerful pulse in the form of an Electromagnetic burst (it does so hundreds of times per second) and the resulting response in the target's field is what the detector then a-s-s-e-s-s-e-s (word deleted by website WordFilter - Eh?!!!) . VLF detectors usually use a single frequency in the lower range (usually 20khz-3khz) and are better at discrimination compared to PI detectors. But PI detectors seek deeper into the soil and are not affected by mineralized soils as much as VLF detectors are.

For this reason, PI is good for deep targets in mineralized soils... hence they are the most popular choice for Gold Detectors. VLF detectors have better discrimination and TID so they're far better for coins and relics. VLF detectors send and receive a signal from the target via the coil whereas the PI detectors measure the speed of decay in the after-current of a target after the pulse stops.

I'm seeing some posts of a detector which apparently allows some manipulation of the pulse rate but I would suggest that this is absolutely useless as a feature since the rate of pulse ought to be much the same in each and every unit. The speed of these pulses are quite fast and any manual manipulation would likely hamper the unit rather than enhance it. As an example, you don't see Minelab offering pulse variations on their high-end PI units.

Cheers,

Marco

 

[parrott]

Marco... thanks for stepping up to the plate... I guess that there are many
different variations of PI machines as there are with the various VLFs. Also the
various VLFs range in price for a few hundred to over a thousand US $$$.

I have a question about the pulse delay... I understand the pulse delay and ground
balance settings effect each other. Also, not all pulse delays are pre-set... so changing
the delay will usually require a compensating change in the ground balance...

A minium pulse delay is said to give the strongest signal into the ground, but also
means the ground balance will need some corresponding tweeking.

A short pulse delay (uscs), while giving the strongest signal, if it is compensated
correctly with the ground signal is considered the best sensitivity to small low
conductive nuggets.

Many factors are said to be addressed in selecting the best pulse delay...
However it is said that the setting the pulse delay as close to the end of the
transmitted pulse will give the highest sensitivity to all metal targets.

This shorter pulse delay setting is considered the best for small low
conductors such as gold; while the longer pulse delays are considered better
for higher conductors.

You mention in your post... PI detectors measure the speed of the decay of the
after-current of a target after the pulse stops... the ability to regulate the pulse
rate seems to be a new direction...

Thanks for posting your thoughts regarding the understanding of a PI. I for
one am interested in all I can learn about the emerging Pulse technology. It seems
to have lots of promise... Parrott...

 

[Reg]

Hi,

I have not seen any book that details just how a PI works. However, over the last 7 years, there has been plenty of information posted on this forum that will help understand how they do work.

Now, I have uploaded a simple simulation of a single pulse and what it might look like on a PI. Keep in mind this single pulse is repeated on a PI at the pulse rate normally indicated by the pps or pulses per second. So, this simulation is repeated over and over again.

On the simulation you will see two boxes labeled target sample and ground sample. The boxes are there to represent that a small sample of the signal will be analyzed during the time the box is present. Again, this sampling is repeated each pulse and those samples are added together and stored. Now, the difference between the two boxes present is the ground sample occurs later by several usec (microseconds). So, the ground sample is delayed by a certain amount of time but samples the same decaying signal.

Also, keep in mind the spike is what happens when the current in the coil is shut off quickly. Almost immediately, that spike begins to decay towards 0V. If there is no signal from the ground or an object, then the signal will ramp down and level off at 0V before the sample is taken. If a target or the ground signal is present, then the ramp is shifted upward and this shift becomes a signal heard in the headphones.

If we wanted to change the delay of the main sample, we would simply adjust the delay control which would shift the sample box to the right. When shifted, you will see the signal from any target is less since all signals are ramping down with time. If the shift is far enough to the right, then some signals will have disappeared and will not cause any signal.

Also, when using this ground technique, both boxes will move together to the right, so as the delay is adjusted, the ground sample is adjusted also. Now, this becomes important to know because all signals are not linear, meaning as the signal decays, they are not proportional to an earlier sample. This means the ratio that worked to ground balance will not work at a later delay. This means, as the delay is adjusted, then the ground balance will most likely have to be adjusted also.

Hopefully this simulation will help explain what is happening.

Reg

 

[lytle78]

Eureka - I finally get the ground balance thing! A picture is truly worth a thousand words (the approximate number of words I've read on how PI's work without getting it through my thick head.

 

[Wet and Dry]

Well I still have a long ways to go before I understand what is going on with PI machines. Thanks everyone for you input. I think I will do a lot more reading. I don't think I will be buying a PI machine in the near future. HH

 

[parrott]

Wet and Dry...

I too am a slow learner; but the PI technology is here to stay... Like
you I'm reading and asking questions... So "slog on." One step at
a time...

 

[Uncle Willy]

Reg,
Wouldn't a PI machine with a pulse delay adjustment for setting it from zero Usecs to whatever be a more versatile machine, allowing one to set it for the particular type of metal or target you were seeking? Thanks.

Bill

 

[Reg]

Hi Bill,

Yes, I would be ideal to be able to sample at 0 usec for the greatest sensitivity. Unfortunately, that isn't possible when using most coils. The simulation I posted earlier on this thread really isn't to scale but simply drawn to try to help visualize what is occurring. In reality, when the current in the coil is shut off, there is a spike generated that is close to 400V in amplitude. This signal is way too huge to be analyzed. In fact, the voltage has to get down to millivolts or less to be useful and not cause a saturation of one or more of the amplifiers. If any amplifier saturates, the sensitivity is lost. So, we have to wait until the voltage has decayed to what is commonly referred to as 0V. From that voltage, any signal from a target generally only causes a very slight increase. It is not uncommon that the weaker signals are in the microvolts. Keep in mind that a microvolt is a 1000 times smaller than a millivolt and a millivolt is 1/1000 of a volt.

Now, if extreme care is taken one can build a special concentric or DD coil that is nulled such that the spike is ignored. In such cases, this type of coil could be used to sample much sooner. However, it is also extremely difficult to build also.

Usually, an extremely sensitive PI capable of detecting smaller gold will sample down near 10 usec. It is very difficult to accomplish this setting which is why many PI's sample much later. One could sample later and then increase the sensitivity to the extreme, but then we end up with a detector that is highly susceptible to external signals and almost useless in noisy environments.

So, the 10 usec is considered a great minimum setting. By using this setting and limiting the sensitivity to a reasonable amount, one ends up with a detector like the TDI or the GS 5. Some depth is sacrificed so the detector can be used in town and other areas where there may be a higher incidence of external noise.

Reg

 

[B.T]

Minelabs GPX sends out around 5000 pulses per sec not hundreds. No changing of the timings or delays in Minelabs? Whats Smooth, extra and all the other variable settings do then?
VLF,s range from 3 to 70 odd KHZ, not just to 20 as you suggest. Minelab Eureka gold goes to 60, Gold bug 2 to just over 70, Whites GMT and Gold masters range from 48 upwards.
You suggest also that VLF,s are only good on coins and relics. A good VLF in weak to mild ground will out gun a PI on small gold as well as crystal gold and many spongy gold types up to several grams in weight. They also will find super rich fine gold specimens that PI detectors don't even know exist. Huge areas exist worldwide where a good VLF should be the machine of choice due to these facts and recommendation of a PI would be useless. There are many goldfields that only produced very small or fine gold and are therefore neglected by PI detector users but which are alive with gold waiting for a good VLF operator to come along and find. I have done this myself finding literally hundreds of surface specimens containing several grams of gold each in areas where hundreds of PI detectors have been waved in vain.
True that there are also huge areas where a PI is the big hitter and the "ultimate" machine of choice for gold nuggets in highly mineralised ground such as in the Victorian golden Triangle or the massive Western Australia's famous fields but a balance in what does and can do what and where is more involved than parroting lifted info that does not contain all the facts.

 

[Reg]

Hello Marco,

II am not sure just where you got your information but some of it, I feel needs to be addressed and corrected so it more accurately reflects what is happening and why.

As an example, you stated; "I'm seeing some posts of a detector which apparently allows some manipulation of the pulse rate but I would suggest that this is absolutely useless as a feature since the rate of pulse ought to be much the same in each and every unit. The speed of these pulses are quite fast and any manual manipulation would likely hamper the unit rather than enhance it. As an example, you don't see Minelab offering pulse variations on their high-end PI units."

I am not sure why you feel the ML is the only manufacture that knows what they are doing and if you don't see a knob, they aren't doing it. Actually, on their later models, when you auto tune the noise out, you are changing the pulse rate. The difference is, on some other detectors, this control is given to the operator so he can make a faster and more precise adjustment.

The fact is, the ability to manipulate the pulse rate changes things enough that certain interaction with external signals can be minimized, thus noise can be reduced. That is the purpose for the ability to change pulse rate.

Next, you stated; "VLF detectors send and receive a signal from the target via the coil whereas the PI detectors measure the speed of decay in the after-current of a target after the pulse stops."

What is more accurate is the fact the signal from the target alters the decay of the coil signal on the PI so the target signal does does come via the coil on the PI also.

You also stated the following; "VLF detectors have better discrimination and TID so they're far better for coins and relics." Now, I will agree that the VLF does a far better job of identifying objects via the TID feature, but this feature begins to be limited by ground conditions and maximum depth can easily suffer in really bad ground. In other words, in many areas VLF's can suffer badly, thus leaving good targets behind. Also, the TID feature can be severely hindered by mineralization causing objects to read inaccurately.

Fortunately, Eric's GS 5 and Whites TDI can be adjusted to be used for coin and relic hunting. No, the ability to TID isn't available but a decent discrimination can be done for a range of targets. As such, the two mentioned PI's have been successfully used in certain areas to find coins and relics beyond the depths that VLF's can reliably detect them as good targets.

One word of caution on this feature and that is, it is or would appear to be a step backward to someone who normally uses a VLF with theTID feature, since there is no easy way to carefully determine just what is being detected down to specific coins on the PI at this time. However, since many of the older coins people are wanting to find are older silver and copper ones and the range is somewhat wide, then the PI can do a decent job of separating these targets from surrounding trash, such as foil, nails, etc and do so at depths that exceed those of the VLF's. Again, this feature and ability generally is directly related to the ground conditions. If the ground is very mild, the VLF may do a better job. If the ground conditions are such the ground appears to be bad to the VLF, then the PI can easily excel.

So, the gap is closing between the two detectors when it comes to being able to use the PI for specific hunting of coins and relics. No, it is not a smooth transition and some people will not like the extra noise normally associated with PI's. However, others who accept this condition that take time to fully learn the PI just may become much more of a PI user because of the PI's capability.

Reg

 

[prospector 55]

I would like to also note PI units to me are really fun to use in highly mineralized ground and when you are looking for gold that is the kind of ground you will be dealing with, without exception all of my VLF units misidentify targets in highly mineralized ground and lose depth, so they effectivly become poorly designed shovels, all of my top of the line VLFs all have problems in my mineralized ground, my TDI is my best and favorite detector, I anticipate it staying that way for along time......

 

[metero]

Hi, all,
Nice to see a forum ,where is realy possible to discuss technical problems . I learned a lot here and I steel have to learn lot. Thank you.
I did't care MD since 1979-80, when participate in the assembling of " sanbanks". In that time I did't understand well how it works exactely.
Just some time ago a friend ask me to make for him MD and I found different sourses in Internet, finaly- this forum.
Let me say what (basicly) I understood and, please, tell me , if I'm wrong.
A hi current pulse is passing thrue the coil and switched off exrtime fast. As result, a hi voltage self induction pic appears. The target within the magnetic field of the coil participates in the decay , changing R/L, and we analize this process. We take a sample of the voltage in one moment ( main sample) and store it somehow ( sample and hold, uCU, etc). Later , nearly at the end of decay, we take another samle and substract it from the main one to eliminate the ground effect. The modern PI MD take several samles and from the form of the decay curve it's possible to assume the kind of metal.

And , reading the Reg post I have some considerations.

...Yes, I would be ideal to be able to sample at 0 usec for the greatest sensitivity. Unfortunately, that isn't possible when using most coils. The simulation I posted earlier on this thread really isn't to scale but simply drawn to try to help visualize what is occurring. In reality, when the current in the coil is shut off, there is a spike generated that is close to 400V in amplitude. This signal is way too huge to be analyzed. In fact, the voltage has to get down to millivolts or less to be useful and not cause a saturation of one or more of the amplifiers. If any amplifier saturates, the sensitivity is lost. So, we have to wait until the voltage has decayed to what is commonly referred to as 0V.....
Usually, an extremely sensitive PI capable of detecting smaller gold will sample down near 10 usec. It is very difficult to accomplish this setting which is why many PI's sample much later. One could sample later and then increase the sensitivity to the extreme, but then we end up with a detector that is highly susceptible to external signals and almost useless in noisy environments. Reg

If we want to take the samles short time after the switching off, can't we take 3-4 separate samples. 1-2 from a simple devider ( sure, with according limiters) ,1-2 direct and some with amlifier - with separate ADC channels. So there wil be samples from practicaly all the curve and only the MCU ( adc) speed will be the limit.A small 8bit MC68S128QE has ADC sample speed arround 1.2-1.5 uS, I don't thing for DSP yet.