opamp voltage follower/buffer question

[tempus]

Hey all;

I'm working on a TL072 based buffer for my piezo signal using a dual (+12/-12v) power supply. Connecting the piezo signal to the + input of the opamp and the output of the opamp to the - input gives no output, unless I drop a resistor to ground from the +input. Is this the way it's supposed to work or am I doing something wrong? I ask bercause every schematic I see of an opamp voltage follower shows no such connection to ground, and I thought the whole point was not to use a resistor so as to take advantage of the high input Z of the opamp.

[antonis]

+Input needs to see a DC path to GND (midpoint of +/- 12V)
You can use a resistor of value as high as 2M2 for TL072..
In case you need higher Zin you can bootstrap the above mentioned resistor..
(see fig. 6 below..)

[tempus]

Thanks for your reply.

I've been using a 6.8M resistor to ground, which seems to be working OK. The diagrams you attached are for a 741. I'm using a TL072, which has an input resistance of 10 trillion ohms, so the 6.8M resistor should be OK without bootstrapping right?

I get that the input needs a DC path to ground, but I do find it odd that I've never seen a schematic that includes the resistor or explains that it is also setting the input Z.

[antonis]

I've been using a 6.8M resistor to ground, which seems to be working OK. The diagrams you attached are for a 741. I'm using a TL072, which has an input resistance of 10 trillion ohms, so the 6.8M resistor should be OK without bootstrapping right?

Input impedance (resistance) of TL072, wired as a non-iverting buffer, is actually set by its non-inverting input impedance multiplied with open-loop gain, but that it's out of concern here..

The value of bootstrapped resistor is a compromise between acceptable voltage offset (V = I_bias X R_bias) due to input bias current (typical 1nA for TL072 at 65^oC working temperature) and apparent bootstraped one, in the mean of the bigger the resistor the higher the boostrapped resistor value but also the higher the input voltage offset..

For your case, a value of 1M should be more than adequate (even a 100k one should be fine), both for high Zin and low input voltage offset..
Just forget "weird" resistors values on previously posted schematic (they have to do with "equalizing" 741 significant inputs bias current..), replace R2 with a short and make R1 value 100k (say)..
(any value in the range of 10k - 1M should work fine for R1..)

I get that the input needs a DC path to ground, but I do find it odd that I've never seen a schematic that includes the resistor or explains that it is also setting the input Z.

And you'll continue not finding any as long as you search for DC op-amp configuration..
For AC op-amp configuration, resistor to GND is needed both for biasing non-inverting input to a reference voltage and providing a path for input cap discharging..

[merlinb]

I get that the input needs a DC path to ground, but I do find it odd that I've never seen a schematic that includes the resistor or explains that it is also setting the input Z.

Diagrams often assume that the signal source provides a suitable voltage for biasing the opamp, e.g. the signal source is another opamp, or whatever. With a wound pickup it would work since the pickup is a continuous piece of coiled wire to ground. But a piezo is basically a capacitor, so it provides no DC path to ground.

[tempus]

Thanks again for the replies. That clears things up a bit, especially merlinb's response about the piezo not having a DC path to ground.

Antonis: the output Z of a piezo pickup is 1.3M at 100 Hz, so I need an input Z of at least 5M.

[antonis]

ΟΚ.. use a resistor of 1M value and you'll have more than 50M bootstrapped value..

[tempus]

Is there any benefit to bootstrapping? My current design calls for a single resistor of 6.8M to ground, which is simple with a low parts count.

Also, since the piezo is basically a capacitor, do I need a DC blocking cap at the input?

[antonis]

Is there any benefit to bootstrapping? My current design calls for a single resistor of 6.8M to ground, which is simple with a low parts count.

Bootstrapping raises resistor apparent value for AC..
If you're OK with 6M8, it should be more than fine with us, also..

Also, since the piezo is basically a capacitor, do I need a DC blocking cap at the input?

It depends on cap value..
Some people don't use blocking cap for values greater than 10nF (but use a pair of clamping diodes between power supply rails and non-inverting input) where the same guys use a blocking cap for values less than 10nF (with a pair of anti-parallel diodes across inverting & non-inverting inputs)

[tempus]

Thanks again for your reply.

Bootstrapping raises resistor apparent value for AC..
If you're OK with 6M8, it should be more than fine with us, also..

Yes, I'm aware of that, but if that's the only advantage it seems like an unnecessarily complicated solution. Would it be lower noise or have any other advantage?

[PRR]

Tl072 at stage temperatures can run 100Meg bias resistor. Your 6.8M seems fine.