Sweeter than Candy

[Anonymous]

I am experimenting using short 3-5uS high voltage pulses which puts the same energy into the coil as a 12V supply does with a long pulse of up to 200uS. This method provides short pulses that end where I = V/RL where RL is the coil resistance and any series resistance. Using this method one can see the difference in the receive signal decay for small gold nuggets etc. As the field is no longer varying at the end of either of the pulses the magnetic ground signal stays the same!!!
OK, the rest is simple, take a first sample ASAP after the short pulse and a second sample after a longer pulse. Subtract sample 2 from sample 1 and the ground gets cancelled but metal targets do not. A gain adjust of one of the samples corrects for any differences in the ground signals amplitude.
Taking things to the next level it is possible to use a series of different pulse width pairs so as to make the detector useful for targets with a longer TC. What is most interesting is that the shortest pulses while being very effective for small gold etc are way too short to detect large trash items when using this method. The addition of an iron ID circuit is all that will be needed to make it complete!!!
If there is anything new in this method it is now in the public domain!!!

 

[Anonymous]

Maybe the constant current is more important than the voltage. The current keeps up the magnetic field.
There are mosfets with 5 pins, two of then is applied for the current indicator resistor. If high voltage is used, then need to put a logical gate paralell to the indicator resistor (the ls series logical gates has very short response)
For example, yo can use 100V, the resistance of the coil is 5 ohm, the max current is 20A but you can adjust the loopback to regulate 2A through the coil. So you get short rise time, with constatn current. Do I look it correctly or it will not work?

 

[Anonymous]

The method I wrote about is the opposite to what Minelab do with their PI detectors. The Minelab detector takes advantage of the fact that if the current is still rising in the coil at the end of the transmit pulse then different length pulses will give a different decay signal from magnetic ground while metal targets will be more or less the same for either length of transmit pulse.
OK, now if the current is no longer rising then the decay signals for different length pulses caused by magnetic ground will be the same. However, each type of metal target has a time constant or TC. If the transmit pulse is equal or longer than the TC then you will receive an optimum decay signal. If the transmit pulse is shorter than the TC then the received signal decay will be different.
By transmitting pulses of different lengths It is possible to subtract the sample from a long pulse from a sample from a shorter pulse and measure the difference between the samples due to the pulse width and the targets time constant. As the decay signal from magnetic ground is the same for any pulse width then the ground signal is cancelled in the subtractive process.
As you can see, one of the most important things when using this method is to make sure that the current in the coil is no longer rising. Equally important is the ability to use a very fast pulse as some targets such as gold nuggets have a TC of around 10uS. The high voltage is used for the shortest pulses to get enough energy into the coil so a 5uS transmit pulse can be used. This allows sampling of targets with TC's at or below 10uS. The method can use and compare signals received due to four or six different transmit pulse lengths so as to be able to measure targets with longer TC's.
I don't know the effect of using a constant current transmit pulse with this method but it might be worth taking a look at, Dave. * * *

 

[Anonymous]

To read Mr. Foster's tests it looks so, that takes relatively long time while a good TC coil reach the optimal state when the current will be stable and magnetic field could be homogenous. When it occurs, after can be measure and substract the magnetic ground signal. The aim is, to reach this state faster, with overcurrent, and a current level regulation. But this is only an idea, and I have no capalibities and experiment to test it at that moment <img src="/metal/html/frown.gif" border=0 width=15 height=15 alt="">

 

[Anonymous]

I am from the old school when we used to describe things using frequency instead of time. I may also be mixing up frequency vs duty cycle ieulse Width when it comes to your discussion above. Are you saying that you are using a transmit frequency of between 2-3 mhz with a normal 50/50 duty cycle instead of the lower frequencies normally associated with PI's or are you still using the low frequency transmit frequency but shorthening it's duty cycle to 2-3us?
Thanks
HH
Beachcomber

 

[Anonymous]

Howdy Beachcomber, The short 3 - 5uS pulses are the pulse duration. The pulse repetition rate is the frequency. The method transmits a number of different length pulses in a repeating pattern. The repetition rate of the repeating pattern is not critical.

 

[Anonymous]

the higher voltage with the shorter pulse duration sounds interesting. Especially after running across that new Mosfet with the faster turnoff and lower internal capacitance and lower on resistance. I am currently building the Hammerhead PI from Carl's Forum to experiment with and may try to incorporate your idea into it. Thanks again for clarifying things.
HH
Beachcomber

 

[Anonymous]

Dave,
Is this idea really possible... I am told that problem is that no matter how sort or long the pulse is, there is still a decay time before getting a return sample, so there may be the same delay on reading both. Apparently you need to store that signal very precisely and overlap it "in-phase" to the second signal, with an extremely critical timing factor.
To be honest I don't know the first thing about electronics/physics but the idea caught my eye - it would be great is you could work out a way to get a working version to market.
Has Eric given you any feedback?

 

[Anonymous]

Mark, The receive signals are measured in about one millionth of a second (1uS). The samples are stored on capacitors in circuits called sample and hold's prior to their being measured. The decay time is got around by using a balanced searchcoil such as a dual D or a coaxial coplanar. Using a balanced searchcoil it is possible to sample the receive signal a very short time after the transmit pulse ends. Indeed it is possible to sample the receive signal during the transmit pulse on time.
As to Eric giving me feedback it was his experimentation that I am continuing. He tested the same basic method and reported his findings here on the forum. The only difference in what Eric has already done to what I am doing is that I use proportionally higher voltages for the different length transmit pulses so all the pulses provide the same energy into the coil. The longest pulse in my prototype has a voltage of 9 volts while the shortest pulse has a voltage of 80 volts.
I am still testing the method at this time. Unfortunately I have not had as much time available to me as I would like. I will post my results as soon as I finish a few more tests. So far the method looks to be a very good one. I have high hopes that it will be a winner.