Re: couple questions about preamp design for piezo pickup


"ChrisQuayle
Phil Allison wrote:


A bootstrap input scheme is where you have a low impedance input
amplifer and use positive feedback to bootstrap, or "follow" the input,
thereby effectively increasing the input impedance. It was used in the
early transistor record players etc, which used ordinary junction
transistors. In the original scheme, the transistor base is biased as
per normal, but the upper or lower leg of the bias network is split
into two resistors, with a capacitor connected between the transistor
emitter and the junction of the two resistors. I think that's right,
but difficult to describe when it could be drawn on paper in seconds.
The same technique can be used with op-amps by running non-inverting
and using a cap feedback between the output of the op amp and a split
input resistor to ground. If you are running the op amp at > unity
gain, you need to ensure that the overall + gain doesn't exceed unity,
otherwise the circuit can become an oscillator :-). The advantage is
that the input resistors can be normal range values, rather than
megohms.



** Megohm value resistors ARE normal range components !!

Values up to 10 M or even 22M have been generally available for decades.

Bootstrap the input only if you like lots of noise.

A FET op-amp needs no such input bootstrapping, just a simple load
resistor
and is the lowest noise option.

The OP can try 1.5 Mohms.



In fact, bootstrap circuits aren't noisy if properly designed, but ymmv.




** You have **completely** missed the technical point. The very nearly
positive feedback applied to bootstrapped input load resistor increases
the input circuit's noise level.

PLUS - the source device here is effectively a 1200pF cap which will
nearly eliminate ( by shunting to ground) the self noise of a 1.5 Mohm
load resistor. It will not similarly eliminate the noise generated by a
much lower value resistor despite it being bootstrapped.

There is simply no need for such bootstrapping with FET op-amps.

Ok, don't want to start an argumentative thread here, because that's not
what the op asked for :-).


** You are the one posting argumentative stuff.

And bad advice.



TLOxx series op amps are quite noisy at high
input impedances


** They actually have their best noise figure at high input impedances -
but the effective input impedance *includes* that of the source.



though and if you calculate the reactance of a 1200 pf cap, You get the
following figures:

10000Hz ~= 13 Kohms
1000 Hz ~= 130 Kohms
100Hz ~= 1.3 Megohms

So, at low frequencies, where flicker noise dominates, you probably won't
hear much audible effect on the noise at all.


** You have not even the faintest clue.

The noise from a resistor is " white " noise with energy proportional to
bandwidth - so the noise energy in the band below 10 kHz is the same as
that between 10 kHz and 20 kHz. The audible portion is between 1 kHz and 5
Khz where the ear is most sensitive.

That 1200 pF piezo pickup will attenuate the * audible noise range * of a
1.3 Mohm resistor by 10 to 50 times or 20 dB to 34 dB soon as it is
connected.

A higher value load resistor will result in further noise improvements ( see
below).


While you may not *need* to bootstrap using an fet op amp, there can be
benefits in terms of frequency response control and ime, you can get a
better overall noise figure.


** You simply *cannot better* the noise of a resistor with no DC bias - it
is set by a law of nature and depends only on the resistance value and
temperature.

En squared = 4.K.T.B.R

En = noise voltage.
K = Boltzman's constant.
T = temp in K
B = measurement bandwidth.
R = resistance value in ohms


In circumstances where the input source is capacitive - the higher the load
resistor value the better.

The math is simple:

Double the resistor value and its self noise increases by 3 dB.

Double the resistor value and the noise attenuation by a capacitive source
increases by dab.

Ergo - you gain 3 dB improvement for each doubling in load resistor value.



That's not to say that it's a one size fits all though. For most guitar
type circuitry, a jfet will get the job done as well, but it's nothing
like as good as an op amp in terms of linearity / distortion. That's
subjective though and the soft limiting might be just what's wanted in
some cases.


** More irrelevant stuff.

Just as silly to bootstrap a JFET input as a FET op-amp input.


My advice would be to try both approaches and see which has the better
overall sound, as this is what matters to a musician.


** More irrelevant stuff.



High impedance pickups are always difficult to do signal conditioning for,
so a bit of experimenation may be needed.


** JFET inputs are close to ideal for piezo and magnetic guitar pickups.
Used as a source follower, there is no overload issue since the input level
may be up to several volts.



BTW

Ever considered why condenser mics use hard to get resistors in the 1 to 5
Gohm range to DC bias the capsule and the input FET's gate ??




....... Phil




.

Relevant Pages
Loading