How to make PI better

[Anonymous]

Eric,
I would appreciate it if you can look at Gary' PI schematic and give any recommendations on how to make it more sensitive to hunting gold rings at salt water beaches. It's a good detector for hunting larger jewelry and coins. I havent had much luck with small gold items and suspect its the detector. I have experimented with making coils and have pretty good results. WIthout an oscope I dont know if its sample time is low enough. Any help would be great. I'd love to buy your Goldquest but can't afford it right now. I like building myself for the learning experience. This post in no way is to say anything bad about Gary's PI. It has found many items and has great depth. Just want to improve it a bit. HH Joe

 

[Anonymous]

Gary has an update at his site on increasing the Usec and making it more sensitive.

 

[Anonymous]

Hi Joe,
Reg has been tinkering with Carl's design, could be worth a look.
Cheers
Kev

 

[Anonymous]

Hi Joe,
It is very difficult for me to look at a schematic to determine what can be done to "make a PI better". I personally would have to build it and then just try a few things to know for sure. Now, given that info, I will offer my suggestions.
First, it is my understanding that Gary has recently modified his design to increase the sensitivity of his detector to smaller gold.
If your "Gary" PI doesn't have the latest mods, I would recommend them. It appears Gary has added controls to reduce the delay and be able to adjust a couple of other things, like maybe pulse on time and maybe the actual sampling time.
The basic way to make any PI detector more sensitive to smaller gold is to; 1. reduce the sampling delay time, 2. maybe reducing the actual sampling time depending upon the initial value, 3.increasing the overall amplification of the signal, 4. and reduce noise, both internal and external.
Each of these issues can have an effect depending upon the initial conditions. Looking at Gary's design, it appears that he is already pushing the amplification of each stage quite a bit. One may be able to increase the feedback resistor from 470k to something closer to 1 meg of the preamp to help some. I might even try reducing the value of C 26 to .047 uf and see if that helps without introducing too much noise. However, it might just require another stage of amplification for best results. This would require an addon board.
Reducing the delay is one logical choice as a means to increase the sensitiivty to small gold. This technique would require the addition of minimal parts also and seems to be one of the things Gary has added in his mods. However, without a scope to see what is really happening, this could prove difficult to do and see dramatic results. One would just have to try and see what happens.
I am not sure if Gary is requiring new coils with his latest mods, but that is possible. As the delay is reduced, it is more difficult to build a coil that works properly.
I hope this helps.
Reg

 

[Anonymous]

Hi Again Joe,
After reading some of the specs on Gary's website, I will offer the following. Gary says his delay time is 36 usec. If this can be reduced significantly, say down to maybe 15 usec or less, then one should see a dramatic difference in small gold signal. However, this will also cause a greater ground signal when using a mono coil.
Gary also indicated his sample time to be 50 usec. I would recommend this be reduced also. Reducing this time should reduce the noise level without reducing the overall sensensitivity to small gold. In fact, I suspect that small gold signals will be enhanced with a shorter overall sample time.
Gary's pulse length of 165 usec is a little long for the detection of gold. I have found that shortening this pulse on time to something less than 80 usec has little negative effect on gold signals but does reducing some of the other negative signals you might get on Gary's machine.
Gary's PI is somewhat similar to the design of the Stuart PI. ON the Stuart PI, I found that reducing the pulse on time also reduced signals I would get from some rocks.
Just thought I would add this additional info.
Reg

 

[Anonymous]

REG,
Thanks for the advice and I did change c26 to .047uf and R32 to 1meg. This made it alot more sensitive to gold but a little unstable. It can be set with a steady clocking pulse then it starts speeding up by itself. This repeats over and over. Which component value should I change. Any more suggestions. Thanks,Joe

 

[Anonymous]

Hi Joe,
I was afraid something like that was going to happen. You might reduce R 32 back to the original value and see if the detector stabilizes. If so, then you might increase this resistor value some to maybe 680K or 820K and see if still remains stable. The problem is so many things interact that it is impossible to just say do this or this.
It is extremely difficult to "guess" as to what is happening or causing the audio change. If you are right on the ragged edge, then even the heating of the FET's will cause what you are experiencing. Power supply instabilities can do it too, as can slight component changes. Right now, my guess is the FET's are heating.
Now, have you installed the changes Gary built in to his modified PI, so you can adjust the various timing values? If not, I strongly recommend you do this. Some of these adjustments will greatly help the stability problem and the sensitivity to small gold.
If you have installed the pots, you might try adjusting those before making the above resistor changes.
BTW, do you have a scope? If so, that will make things much easier. If you do, then I would recommend you reduce the sample time (R9A) to something between 10 usec and 20 usec, the pulse width (R5A) to 50 usec, and the delay (R7A) down as far as possible and still have the machine still work ok. Without seeing a scope trace, the only way to know for sure on the delay is to simply reduce it a little at a time and see if the machine still works ok. If the settings of any of the above items is less than what I recommended, leave them for now. Do not increase them at this time, but let me know what they are. Drop me an email.
If the delay is too short, there will be a severe loss of sensitivity. If it is on the ragged edge, you will get something like you are experiencing now, the machine will be quite unstable, and probably drift, much like you have happening.
BTW, did you build the single, the dual coil design, or a DD design?
Reg