Simple op-amp question - capacitance?

[Scott Stephens]

Does a simple op-amp non-inverting buffer have any effective input
capacitance? Then if low noise performance isn't an issue, is it
necessary to bother with a more expensive low-capacitance input amplifier?

Why doesn't TI spec input capacitance on the TLC272?

Any tips regarding transient & pulse simulation using the Boyle spice
macromodel for op-amps?

Thanks,

--
Scott

**********************************

DIY Piezo-Gyro, PCB Drill Bot & More Soon!

http://home.comcast.net/~scottxs/

**********************************

 

[Joerg]

Hi Scott,

Why doesn't TI spec input capacitance on the TLC272?

This is a slow op amp and usually the few pF of input capacitance don't
matter in typical applications of this amp. But I am sure that you could
contact TI and obtain this data or a SPICE model that contains it.

Regards, Joerg

 

[Scott Stephens]

Hi Scott,



This is a slow op amp and usually the few pF of input capacitance don't
matter in typical applications of this amp.

I've looked at other FET amps, LF353, TL072, et. that have around 30pf
of capacitance. Now if my source is a few pf from a hi-z
(piezo-electric) source, and I want to measure a few millivolts pulsed
across it every few hundred milliseconds, will I measure the proper
amplitude, or will the amps input C soak it up?


But I am sure that you could

contact TI and obtain this data or a SPICE model that contains it.

Been there, done that. Are the spice models accurate? The app-note TI
spec's says they are not for some parameters, such as capacitance
nonlinearities, et.

I suppose I'll have to find out empirically.

--
Scott

**********************************

DIY Piezo-Gyro, PCB Drill Bot & More Soon!

http://home.comcast.net/~scottxs/

**********************************

 

[Joerg]

Hi Scott,

I've looked at other FET amps, LF353, TL072, et. that have around 30pf
of capacitance. Now if my source is a few pf from a hi-z
(piezo-electric) source, and I want to measure a few millivolts pulsed
across it every few hundred milliseconds, will I measure the proper
amplitude, or will the amps input C soak it up?

If your source is really low in capacitance such as a small transducer
it is often better to provide a little FET follower up front. Then you
know from the data sheet what the capacitive load will be and you can
get it down to the low single digit pF range. You just have to be
careful with the layout because these low Cgs FETs are often hot rods.

I suppose I'll have to find out empirically.

Not really. I have had good responses from the major US manufacturers
such as National and TI whenever I had a question that wasn't answered
in the data sheet. Usually within a couple of days.

Regards, Joerg

 

[John Woodgate]

I read in sci.electronics.design that Scott Stephens

I've looked at other FET amps, LF353, TL072, et. that have around 30pf of
capacitance. Now if my source is a few pf from a hi-z (piezo-electric) source,
and I want to measure a few millivolts pulsed across it every few hundred
milliseconds, will I measure the proper amplitude, or will the amps input C soak
it up?

If you are using it as a non-inverting buffer, I don't think that the
effective input capacitance will be significant. Obviously there is
about 5 pF of stray capacitance. You can put a guard ring around the pad
for the non-inverting input and connect it to the inverting input to
bootstrap some of the stray capacitance out of effect.

 

[Tim Wescott]

I've looked at other FET amps, LF353, TL072, et. that have around 30pf
of capacitance. Now if my source is a few pf from a hi-z
(piezo-electric) source, and I want to measure a few millivolts pulsed
across it every few hundred milliseconds, will I measure the proper
amplitude, or will the amps input C soak it up?


But I am sure that you could



Been there, done that. Are the spice models accurate? The app-note TI
spec's says they are not for some parameters, such as capacitance
nonlinearities, et.

I suppose I'll have to find out empirically.

I know that piezo actuators are _very_ capacitive -- are piezo sensors
also? If so, you'll be saved from worry...

 

[Joerg]

Hi Tim,

I know that piezo actuators are _very_ capacitive -- are piezo sensors
also? If so, you'll be saved from worry...

It depends on size. I have dealt with many that were just a few ten pF.
But then again, a 2MHz unity gain opamp wouldn't have cut the mustard
there since the pulses to be measured were in the 100nsec range. We only
used discrete circuitry until the AD603 showed up on the door step. Some
opamps could have done it but were too expensive.

Regards, Joerg

 

[Scott Stephens]

I know that piezo actuators are _very_ capacitive -- are piezo sensors
also? If so, you'll be saved from worry...

Same stuff, just a lot less of it.

--
Scott

**********************************

DIY Piezo-Gyro, PCB Drill Bot & More Soon!

http://home.comcast.net/~scottxs/

**********************************

 

[Scott Stephens]

Hi Scott,

If your source is really low in capacitance such as a small transducer
it is often better to provide a little FET follower up front.

Considered it, want to avoid it if 'bootstrapping' can fix my problem.

--
Scott

**********************************

DIY Piezo-Gyro, PCB Drill Bot & More Soon!

http://home.comcast.net/~scottxs/

**********************************

 

[Scott Stephens]

If you are using it as a non-inverting buffer, I don't think that the
effective input capacitance will be significant. Obviously there is
about 5 pF of stray capacitance. You can put a guard ring around the pad
for the non-inverting input and connect it to the inverting input to
bootstrap some of the stray capacitance out of effect.

Ah, yes, a technique used in capacitive sensors is to feed-back
(bootsrap?) output to the center conductor of a tri-axial cable to
neutralize its capacitance, which could easily swamp the capacitance of
the sensor.

My dumb question is, with a cheap jelly-bean bifet or cmos op amp have
that 20-30pf capacitance, or is it effectively neutralized when the
inverting long-tail transistor is driven by the output? Miller
capacitance of the non-inverting long-trail transistor eliminated by the
current source?

--
Scott

**********************************

DIY Piezo-Gyro, PCB Drill Bot & More Soon!

http://home.comcast.net/~scottxs/

**********************************

 

[Ken Smith]

Does a simple op-amp non-inverting buffer have any effective input
capacitance? Then if low noise performance isn't an issue, is it
necessary to bother with a more expensive low-capacitance input amplifier?

A very low noise JFET input op-amp can have significant input capacitance
when used as a voltage follower. Most others don't have much because it
is only the common mode capacitance which remains when a voltage follower
is well with in its band width.

The TL072 working as a voltage follower at about 16KHz displays less than
2pF input capacitance in a circuit I have in production. I haven't been
able to measure how much is the TL072 and how much is the other stuff.

Why doesn't TI spec input capacitance on the TLC272?

Would you take "cause they are stupid" as an answer? If so, lets go with
that. It is a simple theory that fits all the facts.

 

[John Larkin]

Considered it, want to avoid it if 'bootstrapping' can fix my problem.

I've seen a fet opamp spec'd to have 1 pF Cin... don't recall which
one right now, Burr-Brown maybe.

John

 

[Scott Stephens]

A very low noise JFET input op-amp can have significant input capacitance
when used as a voltage follower. Most others don't have much because it
is only the common mode capacitance which remains when a voltage follower
is well with in its band width.

Ok, that's what I thought I remembered from my fading past. I'll find
out in a day or two...

Would you take "cause they are stupid" as an answer? If so, lets go with
that. It is a simple theory that fits all the facts.

Perhaps they are scared they can't guarantee it consistently, or someone
will errantly decide it is independent of gain in common-mode applications?

--
Scott

**********************************

DIY Piezo-Gyro, PCB Drill Bot & More Soon!

http://home.comcast.net/~scottxs/

**********************************

 

[ddwyer]

Same stuff, just a lot less of it.

Two solutions;
bootstrap or transimpedance , both used for photodiode amplifiers.

A similar application.
If u use transimpedance the gain appears to be independent of input cap
until a noise analysis is made.

 

[Bill Sloman]

A very low noise JFET input op-amp can have significant input capacitance
when used as a voltage follower. Most others don't have much because it
is only the common mode capacitance which remains when a voltage follower
is well with in its band width.

The TL072 working as a voltage follower at about 16KHz displays less than
2pF input capacitance in a circuit I have in production. I haven't been
able to measure how much is the TL072 and how much is the other stuff.



Would you take "cause they are stupid" as an answer? If so, lets go with
that. It is a simple theory that fits all the facts.

Having had my fingers burnt on a similar omission from the data sheet
for the TI TLC2201, a "low noise" CMOS op-amp, with (consequently)
large input transistors and a correspondingly large (about 15pF) input
capacitance, my theory is "cause they are an unethical bunch of
marketeers, who suppress any information that make their circuits look
worse than the competitions'".

Admittedly, this *is* stupid, but it is also dishonest.

 

[Fred Bloggs]

Having had my fingers burnt on a similar omission from the data sheet
for the TI TLC2201, a "low noise" CMOS op-amp, with (consequently)
large input transistors and a correspondingly large (about 15pF) input
capacitance, my theory is "cause they are an unethical bunch of
marketeers, who suppress any information that make their circuits look
worse than the competitions'".

Admittedly, this *is* stupid, but it is also dishonest.

This sounds like a "charge amplifier" application anyway...

 

[Roy McCammon]

Does a simple op-amp non-inverting buffer have any effective input
capacitance? Then if low noise performance isn't an issue, is it
necessary to bother with a more expensive low-capacitance input amplifier?

Why doesn't TI spec input capacitance on the TLC272?

because its hard to measure and guarantee and most users
can't get the circuit capacitance low enough to be
concerned about it anyway.

 

[Roy McCammon]

Would you take "cause they are stupid" as an answer?

I'd say because the consumers generally don't ask.