Preamp

[Qiaozhi]

Discussions related to the design of the preamp.

[ivconic]

Sean:
"...Davor and Ivconic, can you take on the front end ampl please between you as a team as you BOTH have excellent ideas and I think that 1nv/Hz would be just about dandy ;-))..."

Sean, it is fair from your side, i appreciate.
But i am afraid it's not possible to do things like this, in "groups", especially online.
Different people, different habits, different conditions. Different opinions and different intentions.
I suggest you to be the one who decide and the rest of us can only throw ideas and suggestions.
That's the only way to keep the continuity.
Just an example:
if i chose front end like i posted on other topic; it will need specific power supply. So i will have to design quite different power supply, than; most of the guys here will not be happy to accept it, in place of the one already proposed.
One thing usually leads to another. At the end everybody will force something quite different on their own. And we will end up with dozen different and ALL unfinished designs; all those most probably will have nothing in common with Surfmaster at the end!
That's where "project manager" term is coming on scene!
What we need here is exactly the "PROJECT MANAGER" who will decide and manage.
Most logical choice is you Sean, of course.
You can collect all the ideas and than propose two compatible approaches; one modern (4-layer pcb, smt etc...) and one "traditional" ( single or double layer pcb, no smt, conventional analogue components)...
So, people will pick the option which they find most suitable for them.
...
Personally, i can not participate in "full time" and all the time here, i can not maintain continuity in work here. Because of my personal responsibilities and fact that i am also working on 4 other projects at the same time.
But i will be here from time to time and i can help by drawing schematics and pcb's in case you need.
Also later i can build "traditional" version very fast and test it for you here.
That you may count on, from my side.
Cheers!

[Davor]

"Can you also look as active blanking or can we get away with not needing that?"
You can't get 1nV/sqrt(Hz) with traditional series resistor approach, so blanking is absolutely required. It may be active but also passive. Moodz' approach is active and it also incorporates flyback acceleration, but fun part is that the same functionality is available with passive approach.

If n-channel mosfet is placed in series with preamp input (drain connected to coil) and gate biassed at positive rail, the input is protected from positive flyback voltage. Adding a p-channel mosfet (if needed) biassed at negative rail in between the n-channel one and a preamp input protects from charging voltage. At this moment I'm not up to seeking small mosfets for voltage, charge, and availability, so feel free to step in.

As of recently I'm mostly engaged in some other endeavours that prevent me from participating more in forum activities, so my visits are far less frequent.

[Teleno]

Common base amplifier, gain is Rc/Re, where Re is valued fo critical damping

[green]

http://www.geotech1.com/forums/showt...180#post208180

I tried what Davor said with a Bss127 and a FDV302. Couldn't get it fast enough, maybe someone that knows what they are doing could help.

[Teleno]

Common base preamp:







PI_common_base.zip

[Teleno]

[Davor]

That's spot on - you can't get input noise below 3nV/sqrt(Hz) with damping resistor connected in series - regardless of the amplifying device's capability to go much lower. 1nV/sqrt(Hz) corresponds to ~60 ohm, and that's too low for a meaningful damping resistor.
Going down with noise will require a non-inverting op amp with some sort of blanking/clamping device. Alternatively it may be a FET or a transistor, but with no capacitors in biasing circuitry to avoid long recovery. It is much easier and cheaper achieving below 1nV/sqrt(Hz) noise with discrete devices, but I have no idea what role may non-linearities have in PI. It is quite simple to demonstrate importance of front-end linearity in VLF, but things are not as straightforward with PI. Intuitively I'd say linearity must play a role in early sampling, but I may also be wrong.

[ivconic]

For a moment we could stop focusing only on mono coil design in this setup.
I don't see what's the problem to consider use of double D or coplanar coaxial coil in this design?

[Old cart]

Ivconic, I think that is great idea. If we can do a DD we may not need the protection diodes at all. The input resistor can be dropped from 1 K to maybe 100 ohms and the feedback resistor decreased correspondingly to maintain the desired gain. This would take the noise down by factor of 10. Using a TI49990 ( sadly being discontinued ) would give a front end limited mainly by resistor noise. This will also reduce ground noise. Does anybody have experience with DD's on PI detectors? What is the downside other than they must be nulled?
A concentric is also OK but I think they are harder to null. This is not a huge problem if you have a scope or a very broadband multimeter-but those are quite uncommon.

[ivconic]

Ivconic, I think that is great idea. If we can do a DD we may not need the protection diodes at all. The input resistor can be dropped from 1 K to maybe 100 ohms and the feedback resistor decreased correspondingly to maintain the desired gain. This would take the noise down by factor of 10. Using a TI49990 ( sadly being discontinued ) would give a front end limited mainly by resistor noise. This will also reduce ground noise. Does anybody have experience with DD's on PI detectors? What is the downside other than they must be nulled?
A concentric is also OK but I think they are harder to null. This is not a huge problem if you have a scope or a very broadband multimeter-but those are quite uncommon.

Exactly!
I had S/N ratio on my mind exactly when i suggested "two-loop" design instead "mono-loop".
Everything is turning again and again around S/N.
So; fundamental improvement (at any project, not only this one) is to improve S/N ratio.
Tx coil can be projected separately and RX coil also can be projected separately.
Each one to suit best specs for the particular design.
Minelab did so far most on this. Not really original idea. Already seen.
...
Concentric coil actually does not need so precise nulling as it is case with IB coil.
Coils can be additionally nulled "electronically" by manipulating with amplitude at TX or gain at RX! Ain't that sweet!

[Teleno]

Ivconic, I think that is great idea. If we can do a DD .... The input resistor can be dropped from 1 K to maybe 100 ohms and the feedback resistor decreased correspondingly to maintain the desired gain. This would take the noise down by factor of 10. ...

The Rx coil will need to be damped in any case, even more so that the Tx coil because it'll have more turns, therefore higher R, hgher L and higher C.

In the presence of a target, the Rx signal is the difference of two exponentials: the Rx coil own decay minus the decay of the target. A low damping R means the Rx coil decays slowly and sampling will need to be done later, reducing S/N.

Signals for an Rx coil balanced to 99.8% - Red: no target (Rx own decay), Blue: target (diff of exponentilas).

[Old cart]

Interesting. How do you balance the TX /RX coil? I know how to do this with DD coils but how about concentric? Ivconic mentioned electronic balance.

[Teleno]

Interesting. How do you balance the TX /RX coil? I know how to do this with DD coils but how about concentric? Ivconic mentioned electronic balance.

Figs. 9 to 13 http://www.geotech1.com/pages/metdet/info/coils.pdf

[Old cart]

Coil- help for the preamp

That is a good explanation. It would seem the DD would be easiest to null at least until Ivconic explains how to null electronically.