Simpler PI...

[Davor]

If carefully designed ... why not. They do get you well below 1nV/sqrt(Hz) on a budget.

[ppodg]

ok, I simply hate ac amps in pi

[Aziz]

Aziz, don't use single transistors, their thermal drifts will kill you!
You need a pair in ONE case, on ONE substrate! Audio is different from md!

No, I don't think, that the temperature gradient won't be an issue. The BF459 can be mounted well for good thermal coupling using a good thermal heat sink paste.
BTW, the temp gradient is in the low frequency region and I'm going to use 1.5 kHz to 3 kHz pulse frequency. All not critical. The DC drift won't be an issue as well due to AC amplifier configuration! The gain is adjusted by a resistor divider feed back loop. It is very stable.

Another good matched transistors in the same chip die: THAT300 (four matched NPN transistor!!!!). But this is expensive and difficult to get. I'm using the poor man's solution and not my expensive LM394 chip at home (yes, I have one at hand). *LOL*

Cheers,
Aziz,
poor but not dumb! *LOL*

[ppodg]

Aziz, don't use single transistors with 318, their thermal drifts will kill you!
You need a pair in ONE case, on ONE substrate! Audio is different from pi!
And, if I remember, your bfs need large current in noiseless applications.

[Aziz]

The BF459 (0,27 € at www.reichelt.de ) seems to be very very attractive. Two (rough) matched pair of them would be enough to go below 0.5 nV/sqrt(Hz).

Aziz

Nice!, very very nice. I'm below 0.4 nV/sqrt(Hz) with two rough matched BF459 pair and load balancing circuit as shown in the PDF paper (emitter load balancing). Seems to be an ultimate amplifier circuit destined for my anti-interference "Top-Hat"(c)(r)(tm) IB coil.
I'm going to release the ultimate Laptop/Netbook/Tablet PC detector controller for the world best beating metal detector application. *LOL*
Oh sh1t!, I'm totally bankrupt! I have to wait several weeks to afford a bunch of BF459 transistors (50 - 100).
(What a great luck for Greedlab that I'm so poor! )

Cheers,
Aziz

[Aziz]

The BF459 (0,27 € at www.reichelt.de ) seems to be very very attractive. Two (rough) matched pair of them would be enough to go below 0.5 nV/sqrt(Hz).

Aziz

[Aziz]

'We need easily available & cheap BJTs in a simple & flexible KISS design'

well, bcm847bs - bjt, of course - is 0,5euro at Farnell; old lm318 well below 1euro
3-4 resistors, one cap is advisable
~1nv/hzsqrt, but you don't need it - see em noise in Europe, should be enough even in oz, we are talking about a metal detector...

Ok, BCM847bs would make up an ultra low noise application (both transistors parallel at 0.5 nV/sqrt(Hz) ).
But I can't get it at www.reichelt.de . And you can't get it easily in most countries as well.
(I could get it from German Farnell of course.)

Aziz

[ppodg]

'We need easily available & cheap BJTs in a simple & flexible KISS design'

well, bcm847bs - bjt, of course - is 0,5euro at Farnell; old lm318 well below 1euro
3-4 resistors, one cap is advisable
~1nv/hzsqrt, but you don't need it - see em noise in Europe, should be enough even in oz, we are talking about a metal detector...

[Aziz]

Aziz, try bcm847bs+lm318, see fig.3 in an222 by National, but some simple tricks are a must...

Hi ppodg,

no, I can beat that easily.
We need easily available & cheap BJTs in a simple & flexible KISS design.

I'm already below 0.5 nV/sqrt(Hz).
Hard to beat me. I'm willing to wait for the super-duper genius design before I fire up my soldering iron however.
Comeon guys, give your ultimate best.

Cheers,
Aziz

[ppodg]

Aziz, try bcm847bs+lm318, see fig.3 in an222 by National, but some simple tricks are a must...

[Aziz]

Hi guys,

ok, I have won the competition. *LOL*

I have modified the improved version of OP27 circuit in the pdf paper mentioned above ("Design of ultra low noise amplifiers" by Vojtěch Janásek).
You can make it work to the conditions listed above. It's really a nice and simple circuit and I'm going to test it soon.

Except you beat me with a much better circuit. I'm waiting...
The competition run hasn't been finished yet.
Good luck.

Cheers,
Aziz

[Aziz]

Ok guys,

here is a reasonable competition. The winner will know, that I will use his genius design in a world beating detector.

Here are the conditions for an ultra low noise amplifier:
- Single supply (+8.5 V)
- AC amplifier
- stable gain (100 - 200 x) by resistor divider via feedback loop
- < 1 nV/sqrt(Hz) noise density @1 kHz (ideally < 0.7 nV/sqrt(Hz) )
- wide band (1 kHz - 70 kHz)
- cheap and simple solution

Is there anybody who is willing to win the competition?

Aziz

[Waikiki_Sweep]

http://mujweb.cz/janascard2/PDF/Desi...amplifiers.pdf

There are really amazing solutions to build low noise low cost preamplifiers. Thanks for digging in it!

[Aziz]

http://mujweb.cz/janascard2/PDF/Design%20of%20ultra%20low%20noise%20amplifiers.pdf

[Davor]

Oh, but they are - you have a bit nasty effect, charge injection, that in fact provides LO injection into the demodulated signal path. Problem may be very small in a balanced mixer, but in most VLF-s it is not so. PIs with EF subtraction in effect have an equivalent of a balanced mixer, because the same charge gets injected at target sample, and subtracted at EF sample, in effect balanced out.
The charge injection is unavoidable with CMOS switches, there are only some that give off smaller charges, but the only way to reduce them completely is a balanced mixer.

As for 1/f noise, we are dealing with targets responses centred around 10Hz, and it is as nasty as it gets. Choppers for DC amplifiers are not exactly commonplace, but may be completely avoided if sampled signal is amplified, and autozeroed, prior to integration.