Fantastic P.I front end building block

[JC1]

So

So going to have to really really understand

exactly the source and effects of noise in a

correlated double sampled system, and how

frequencies can be moved in frequency.

And all the math.

Before anyone DSPs it away.

[ivconic]

FET

"And when you are designing your front end Fet electronics save yourselfs
some trouble and a simple grounded source fet with a drain/gain resistor
based on gfs of about a gain of 20 to 30 then AC couple to the op amp more
voltage gain. The voltage gain of 20 to 30 will make all the noise spec fall
onto the fet and virtually none on the op amp. 1/A^2 of the first stage. No
feedback to the fet. or you can feedback if you wish no biggy as long as you
don't have any trouble with the oscillations."

Hey JC1; it seems that you are more familliar with FET's than me.Good! Can you take a look
on Tesoro Golden Sabre Light schematic in "TEsoro Golden Sabre" thread.
There are 2 switching fets designated as J107 (BF245....simillar). Those 2 are producing
lot of noise due switching role they are playing in design. Ocassionaly device is
"tic..tic"-ing, producing short decayed false signals, which is in a way disturbing for
user. I have found way to eliminate this by lowering Sensitivity at 50%...but dont feel
satisfied with that. Was thinking to find a way how to eliminate that fet noise....
Trully dont have a clue yet. Can you take a good look on it and suggest some solution?
Actually i am not quite sure what provocking device to act like that. Coil is not,front end
is not.....Either fets, or something in last stages after fets???
But it would be good and beneficial for me (and other TGSL makers) to eliminate fets noise
and be sure that those are not source of problem any more...
If you can...please do and post your opinion there...
Best regards!

[sharky]

Use Transistors

Hi Max,

The op27 is an old story, I have about 10 here from years
use. think they quit making em.

Why are u using costlynoisy opamps use a low noise transistors in parallel, about 8 or 3 to 4 they reduce noise levels by cancelling the random noise..

[JC1]

Fet

HI Ivconic,

Hmmm,,,, could be several things, but along the line of the

fet it is used in your oscillator as an

Amplitude Stabilisation Circuit

Now if your xmit oscillator is varying in amplitude it

shouldn't matter too much if you coil is very well null.

Which is hard to do, anyway you could try varying some of

the resistors some and see if your tic tic changes or not.

http://sound.westhost.com/project22.htm



FET - A Field Effect Transistor works quite well as a voltage controlled resistor, but has a limited peak voltage, so the level must be kept below 1V if distortion is to remain within respectable limits. This is barely acceptable for the output of this oscillator. A FET circuit was considered and discarded.

[JC1]

Square root

Why are u using costlynoisy opamps use a low noise transistors in parallel, about 8 or 3 to 4 they reduce noise levels by cancelling the random noise..

If I remember right this random noise is reduced by the

square root of the number of devices. i.e. 4 op amps

half the noise of one. So if you want factor of 4 improvement

you need 16 devices. Like to go from 4nv/rtHz to 1nv/rtHz.

Hey Sharky, there are bigger "noise" problems then

4nv/rtHz of a NE5534. Solve those, and then this stuff

becomes important.

[TheWizard]

http://www.national.com/pf/LM/LMH6624.html

[JC1]

Nice

http://www.national.com/pf/LM/LMH6624.html

Nice find Wizard. Have you used this part? does it recover

fast enough? Looks fast Good CLC502 replacement !!!

And traditional voltage feedback as they say on the

datasheet

This is definitely a good canidate. And price doesn't look too

bad either. Nice find.

AN-1604: Application Note 1604 Decompensated Operational Amplifiers
OA-12: OA-12 Noise Analysis for Comlinear Amplifiers
OA-14: OA-14 Improving Amplifier Noise for High 3rd Inte
OA-18: OA-18 Simulation SPICE Models for Comlinear's Op
OA-21: OA-21 Component Pre-distortion for Sallen Key
OA-26: OA-26 Designing Active High Speed Filters
OA-27: OA-27 Low-Sensitivity, Lowpass Filter Design
OA-28: OA-28 Low-Sensitivity, Bandpass Filter Design with Tuning Method
OA-29: OA-29 Low-Sensitivity, Highpass Filter Design with
OA-30: OA-30 Current vs. Voltage Feedback Amplifiers

{the application notes at the bottom look interesting}

{worth a read and free too}

Hey Wizard you sure this isn't the CLC502 replacement?

[JC1]

25 uAmps

Hi Wizard,

Input is kinda leaky, 25 uAmps.

That's alot of input current.

better keep resistors low.

[JC1]

AD

Hi Wizard,

The AD797 looks better.

Everybody is talking about it. Be careful comparing the noise

plots the 797 is close to 1 nV/rtHz at even 100 Hz.

The national part is there at about 10 kHz.

For PI applictions the 797 is quiter.

and has alot less input current 250 nA instead of 25000 nA.

[TheWizard]

Hi
Reality, thermal noise and OP noise is not problem.
The BIG problem is broadcast and 50..120Hz noise.

[sharky]

http://www.national.com/pf/LM/LMH6624.html

Now every body speaks of noise...I told to use LM66XX previously.....but dropped off due to cost

[sharky]

Like to go from 4nv/rtHz to 1nv/rtHz.

Hey Sharky, there are bigger "noise" problems then

4nv/rtHz of a NE5534. Solve those, and then this stuff

becomes important.

High JC1

DID U mean big noise problems as 50hz and transmitter towers??? try a balanced detector similar to TwinLoop Treasure seeker...

[ivconic]

Thanks JC1 !!!

You gave me another approach to this problem! Might be not those switching fets but the one in oscillator part! How come i didnt realised that!?
I never tested TX amplit. and its behavior!
I'll perform some tests now!

Thank you!
Best regards!

[JC1]

xmitter Fet

Hi Ivconic,

The fet in xmit is supposed to stabilized the amplitude.

even if it is wandering around it may not show up on

voltmeter. may have to use scope to see it. otherwise

take a good look around the fet at the voltages if it makes

sense. you may need too read the specs on the one you

are using and sometime even "select" a fet or adjust

the resistors around the one you have to get it to work.

or it won't really work, in which case you will need another

or thermistor if you can find that. Search internet for

oscillators wein bridge, clapp, colpitts, and some will have

fet feedback stabilization. Also try touching points of osc.

and see if it beeps at you. or replace the oscillator with

some other type you have parts to build.

[JC1]

...

High JC1

DID U mean big noise problems as 50hz and transmitter towers??? try a balanced detector similar to TwinLoop Treasure seeker...

High Sharky,

Yes to both of those, but also that the double correlated

sampling moves these external noise sources around in

frequency. particularly the movement into lower frequecies

like HZ and subHz.

Maxim.

Undersampling is a powerful tool that can be used effectively in selectedapplications. It allows an ADC to behave like a mixer in that it can takea modulated high-frequency carrier signal and create an image that islower in frequency . In this way it behaves like a downconverter.

One drawback of undersampling is that unwanted signals can appear in the desired band of interest and you cannot differentiate them from the desired signal.

http://www.maxim-ic.com/appnotes.cfm/an_pk/928