Low Noise Front End for P.I.

[_Raz_]

Hi Carl,

Yup, that's correct.

Thanks for confirming it. Whew, I'm glad I wasn't leading anyone up the garden path with wrong information.

Mark

[JC1]

EE

Riddle me this.

Based on the equation for thermal resistor voltage noise.

As the resistance goes up, so does the voltage noise.

Therefore, if I pick two points on a blank printed circuit board, say three inches apart, the resistance is very high.
So then the voltage noise is very high also.

Now six inches apart is even higher. 1 meter higher yet.

So for an open circuit using air as the resistor medium, I would get almost infinite voltage noise.

What's wrong with this story?

JC1

[JC1]

EE

More noise.

Johnson noise sets a lower limit on the noise of a resistor, since it is independent of the current through the resistor

Shot noise is another well-understood form of noise, and arises from the discrete charge of the electron. It is often explained as “rain on a tin roof” since its relative magnitude decreases with increasing current.

Unexplained random noise sources not only lack a physical explanation; they also differ in two major ways from Johnson noise and Shot noise. First, unexplained sources tend to be ohmic............


http://www.phys.ufl.edu/REU/1999/reports/leon/leon.pdf

[Ferric Toes]

Hi JC1,

You may be right; but how would you measure it? As soon as you connect something, the input impedance of the device shunts it, and immediately reduces the noise to whatever the equivalent parallel resistance is. 100 megohms is only 1.3uV rtHz.

Regarding the damping resistor, which does get hot in many PI's, the search coil resistance is in parallel with it, so we are at most, only looking at a few ohms noise resistance.

Eric.

[JC1]

EE

Noise 101

http://www.edn.com/article/CA371088.html

Noise 102

http://www.edn.com/article/CA402133.html

These expressions suggest that any nonzero resistance gives rise to a noise-voltage magnitude bounded only by the measurement bandwidth. In other words, as the measurement bandwidth approaches the infinite, so does the perception of a resistor's thermal noise. This result doesn't occur in practice, though why it doesn't might not be obvious at first glance. {nother riddle, heh heh}

These selections represent just a tiny sample of the many ultra-low-noise amplifiers available from these vendors.

[JC1]

EE

Hi Eric,

Good morning!, or afternoon as the case may be.

You are right Eric, even though the coil is somewhat inductive, the coil should be shunting the damping resistor noise low frequency noise. I guess that is good?

more noise riddles.

Of course the thermal "current" noise of a resistor is the inverse of the thermal Voltage noise. Therefore, the lower the value of the resistor, the higher the current noise of the resistor.

So for a short wire, dead short, the thermal current noise would become infinite. According to the "equations".
Though the voltage noise is approaching zero.

That would make the coil quite the current generator.

Wouldn't it? heh heh

[Carl-NC]

Riddle me this.

Based on the equation for thermal resistor voltage noise.

As the resistance goes up, so does the voltage noise.

Therefore, if I pick two points on a blank printed circuit board, say three inches apart, the resistance is very high.
So then the voltage noise is very high also.

Now six inches apart is even higher. 1 meter higher yet.

So for an open circuit using air as the resistor medium, I would get almost infinite voltage noise.

What's wrong with this story?

JC1

Good question... that baffled me too when I was first exposed to noise analysis. I'll try to write something up this weekend that explains it all... it'll have to be on paper as this forum editor is too limiting.

- Carl

[JC1]

EE

Hi Eric,

If you wish to measure resistor noise, I would recommend Keithly equipment. Lots of low noise apps on this site.


Thermoelectric voltages are the most common source of
errors in low voltage measurements. These voltages arise when
different parts of a circuit are at different temperatures or when
conductors made of dissimilar materials are joined together, as in
an ordinary solder joint. For example, the thermoelectric EMF of
lead-tin solder with respect to copper is 3µV/°C.


http://keithly.com/data?asset=50287

[JC1]

EE

Resistor thermal "current" noise equation.

However, using an electrometer (or any voltmeter) on its
most sensitive range introduces noise and zero drift into the
measurement. In contrast, Johnson noise current decreases as the value of RSHUNT increases.

http://keithly.com/data?asset=6169

[JC1]

EE

Hi Carl,

Ok then, I will await your reply.

To the riddle that is.

Its all so confusing.

Back to work. Paying work that is.

[Ferric Toes]

Hi again JC1,

These expressions suggest that any nonzero resistance gives rise to a noise-voltage magnitude bounded only by the measurement bandwidth. In other words, as the measurement bandwidth approaches the infinite, so does the perception of a resistor's thermal noise. This result doesn't occur in practice, though why it doesn't might not be obvious at first glance. {nother riddle, heh heh}

The answer was in the second article - parasitic shunt capacitance . Same reason that an opamp with feedback can be short on bandwidth, if the feedback resistor is too high a value.

I'm enjoying this intensive "noise course". Learning something too.

Eric.

[Ferric Toes]

Hi Eric,

Thanks for the tip. I use much less gain in my pre-amp stage (around 50) and don't have any more gain until after the demodulators (4066). I do notice quite a bit of TX pulse noise on the receive signal. I guess I should be thinking of a second pre-amp stage to raise the signal level before the demodulators and probably more filtering on the power inputs of the devices after the demodulators.

Mark

The main reason I originally went to two opamps, instead of one, is that the scope trace showing the damping and recovery can be misleading. I found I was getting a nice clean trace, but in fact the gain of the opamp had not recovered to its full feedback bound value till rather later. I had been puzzled for a while as to why the balance of the differential integrators shifted as you went to short sample delays <15uS. This was because the opamp gain was lower in the first sample than the second. This was confirmed by injecting a 100kHz signal into the amp, along with the signal from the coil. You can clearly see the envelope narrow, as it approaches the knee of the waveform. Using the twin arrangement, I now get full gain right up to the knee.

This, of course, was running a single NE5534 at a gain of 500. At 50x that you are using, it may well be OK. To inject a signal in, I use an additional 560k resistor on to the signal input pin (in my case, pin 2). I adjust the output of a sinewave generator connected to the resistor, to give 1V pk-pk on the opamp output.

Eric.

[JC1]

EE

Hi Eric,

Heh heh, Yes he did go ahead and answer the question on the unlimited bandwidth producing unlimited noise.

Still seems like one could take a diode and rectify the voltage noise on a very large resistance (like air) that would produce a huge voltage noise and get power that could be used to power the metal detector.

What's wrong with this story.

oh the EMFs are of course a dc voltage, but still noise in any generator.

More noise

so ever hear of a transimpedance amplifier? current to voltage converter, of course you have, an inverting amp without the input resistor. Is zero ohms low enough?

so why not use that, it is used on photodiodes and the like.
and after all the eddy current in the target is a current, and a current is induced into the coil. and the input resistor is zero. sounds low noise to me.

http://www.elecdesign.com/Articles/I...ArticleID=4346

Glad your enjoying.

And that's interesting about the 5534 recovery time.

might have to build your own amp, then could also interally clamp things perhaps and have more control over recovery time then trying to find yet another op amp. and you can always get transistors. Tektronix guy told me that one time when I asked why they used transistors when there was a chip that would do that.

Oh btw, peaks in the group delay could also make the goodie signal bigger, so some peaky circuits might be ok.
though they might not look like it at first.

[Ferric Toes]

Here is a picture of the gain droop of a 5534 running at 500x. The broad band is the injected high frequency signal. This droop occurs right down to low frequencies of a few Hz. It is just easier to see with HF.

Eric.

[Ferric Toes]

For comparison, here is the response from the 5534/8055 combination. Scope on both is 2uS /div.

Eric.