LOOKING FOR GOLDSCAN 4 Schematic.

[ivconic]

Coil from video is 18-19cm diameter, 0.43mH/2.9 ohms...
After 40 minutes of operating the IRF840 gets only a bit warm but never hot, no need for heatsink at all.
For the whole video duration; the delay is at minimum (don't have a clue about exact value).
PPS is at maximum, I guess it is 550pps as on Alex's schematic.
I don't pay attention on such side details, my focus was only on GEB and "Disc" behavior for now.

[ivconic]

"...gemranium diodes..."

[ivconic]

[ivconic]

And finally, a couple of very important notes for those who would like to make this!
Forget gemranium diodes, total nonsense!
I guess Eric had a lot of those in stock, so he wanted to use them!

Not only are they a bad solution, but they also cause a big problem for the operation of the device and precise adjustments.
To fit all the different diodes I put a BAT86. On each diode place!
Instead of J113, which I can't find in local stores; I put J112.
All quads are TL064.
All other transistors are BC547/557.
The mosfet is IRF840.
All other passive components are of original values.
All other ic chips are according to the original schematic.
I adjusted the trimmer potentiometers according to the schematic provided by Alex.
Everything is 100% correct.

[ivconic]

[ivconic]

But make no mistake, the previous file has errors, which I didn't know at the time.
Here is the file with all the mentioned errors.

[ivconic]

Ok here is what I was talking about.
It is Eric's original pcb design.
To make less clutter on the desk; I combined both pcbs into one file, so that in JLPcb they would do it as one pcb.
But I left a "perforation" between both pcbs, in case I want to separate them.
In the end, it turned out that the detector works in an unprecedentedly stable and calm manner, which can be seen on the video.
Plus is powered by a very old bench power supply and not by a battery.
So my amazement is all the greater that the detector works so stably, calmly and cleanly!

[ivconic]

Hello ivconic,
I have the equipment lying around, it worked perfectly, if you want to have it, I will send it to you by international DHL, you pay me for the transportation only through Pay Pal, you have to give me your exact address and telephone number. So I will be happy that you check the schematic because its creator was very stingy in giving something!!!
Nice day ,Saludos from México.

Thanks for the good will!
I respect that!
But there is no need.
The total shipping costs would exceed all logic.
And there is no need, I already have 5 pcb done in JLPcb.
I don't know what to do with the other 4.

[Alexismex]

Hello ivconic,
I have the equipment lying around, it worked perfectly, if you want to have it, I will send it to you by international DHL, you pay me for the transportation only through Pay Pal, you have to give me your exact address and telephone number. So I will be happy that you check the schematic because its creator was very stingy in giving something!!!
Nice day ,Saludos from México.

[ivconic]

One thing I can confirm here; the schematic by Alex, in post #408 is definitely correct and accurate.
Because I compared the Sprint file with it and thanks to that schematic I successfully corrected all the errors on the Sprint pcb layout.
With one small note; R49 and R50 on schematic do not exist on Sprint pcb layout.
I believe it's either one of Eric's revisions... or Alex himself added it to the schematic.
In any case, it is not wrong, because by adding those resistances, the detector will get "slow motion" behavior.
Only that I would set both resistors to be 470k.
I haven't put them on my work yet, but I plan to include them as an option.
That's it for now... I'm falling from fatigue, I'm in rags!
Cheers!

[ivconic]

look attachment post #407

Pito, I will definitely measure the power of the TX one of these days and let you know.
I'm too tired to do any work now. I'm going on a short rest from this.
The thing I didn't mention here is this: the pcb layout I used to do this is based on the original GS4 from Eric.
But that Sprint Layout file was drawn by someone, I don't know who, who made a dozen mistakes, there are some missing links, there are some mistakes, and even extra components that are not on the original.
The first time it was turned on, it didn't work at all, one horror.
So I worked really hard for a couple of days to locate all the mistakes, correct them and note down everything I found that was bad. So that process made me very tired.
Give me time to rest and these days I will measure everything you are interested in and post it here.

[ivconic]

What you describe is normal for subtractive GB PI designs. GB mode always loses depth because of the signal subtraction. The GB subtraction always results in a target hole exactly at the GB point. Below the GB point are low conductors, above the GB point are high conductors. Nails fall into the high conductor range because of their permeability, so if you mute the high conductor range to get rid of nails then you will also get rid of high conductor coins and such.

There are ways to improve this, by using multi-TX periods or multi-RX sampling to create 2 or more GB channels. But this only gets rid of the target hole, it does not enable decent discrimination. You can pseudo-discriminate by using an IB coil and sampling the ON-time response to get a ferrous/non-ferrous indication. This is what ML does and it still isn't all that great, but better than nothing.

You now see the fundamental problems with PI detectors. It is what everybody is trying to solve.

Yes, that's right, nothing better than a practical example to understand the whole story.
Although, over the years I have had several PI detectors from reputable manufacturers, where I could see all of that present.
In this particular case; I'm not interested in the full "Disc scale", but only the rank group of findings that are present here in my field.
And that would be various bronze coins, silver and gold coins... and that's it.
But unfortunately, in that group of findings there are all kinds of examples of both low and high conductors... on both sides of the "hole" or at the very place of the "hole".
What can be seen on the video. The Sestertius is a much better conductor than the Antoninian as is the "thick" applique/buckle.
By the way... such a Sestertius Deus sees with VDI number 98-99! Which in most cases is the number for "thick" iron!
I read the previous threads too, followed everything. I've already heard all that from a few of you.
Of course, for a full understanding of the whole story; it was necessary to make this and to familiarize myself with everything in practice.
I don't know if there is an easy solution, I believe there isn't.
This afternoon I gave a lot of thought to the introduction of the third channel. To have two "holes", one fixed and the other "movable".
Or should I give it up and try something else, something similar to what I've already done at IGSL...
I'll see... you described the situation very well. It is difficult to find even one counterargument.

[Carl-NC]

What you describe is normal for subtractive GB PI designs. GB mode always loses depth because of the signal subtraction. The GB subtraction always results in a target hole exactly at the GB point. Below the GB point are low conductors, above the GB point are high conductors. Nails fall into the high conductor range because of their permeability, so if you mute the high conductor range to get rid of nails then you will also get rid of high conductor coins and such.

There are ways to improve this, by using multi-TX periods or multi-RX sampling to create 2 or more GB channels. But this only gets rid of the target hole, it does not enable decent discrimination. You can pseudo-discriminate by using an IB coil and sampling the ON-time response to get a ferrous/non-ferrous indication. This is what ML does and it still isn't all that great, but better than nothing.

You now see the fundamental problems with PI detectors. It is what everybody is trying to solve.

[pito]

look attachment post #407

[ivconic]

Dunno...