Switch to reset charge amplifier

Hi there

I have a charge amplifier, used to amplify the signal from a piezo
electric transducer, with 2nF of reference capacitance shunted with
1GOhm. I need to be able to reset the amplifier, preferably with a
solid state device (the circuit has to be very compact, and operate in
a high-vibration environment). One possibility which comes to mind is
to simultaneously short the amplifier output and input to ground. The
input is the inverting input of the amplifier, and is nominally at the
same ground potential as the non inverting input. I am considering the
use of a mosfet as the shorting switch for the input. When the mosfet
is off, its gate will be at ground potential, as will its drain
(virtual ground) and source. Is there a mechanism whereby a leakage
current could still flow from drain to source? Every nanoampere of IDS
gives 1V of offset at the output, and it should be noted that the
circuit must work at 120 deg C.

Any help appreciated.

Best regards
Geoff

Geoffrey wrote:
> 
> Hi there
> 
> I have a charge amplifier, used to amplify the signal from a piezo
> electric transducer, with 2nF of reference capacitance shunted with
> 1GOhm. I need to be able to reset the amplifier, preferably with a
> solid state device (the circuit has to be very compact, and operate in
> a high-vibration environment). One possibility which comes to mind is
> to simultaneously short the amplifier output and input to ground. The
> input is the inverting input of the amplifier, and is nominally at the
> same ground potential as the non inverting input. I am considering the
> use of a mosfet as the shorting switch for the input. When the mosfet
> is off, its gate will be at ground potential, as will its drain
> (virtual ground) and source. Is there a mechanism whereby a leakage
> current could still flow from drain to source? Every nanoampere of IDS
> gives 1V of offset at the output, and it should be noted that the
> circuit must work at 120 deg C.
> 
> Any help appreciated.
> 
> Best regards
> Geoff

One cute trick for this is to invert the power supplies to the op
amp--with suitable current limiting resistors, of course--and let the
ESD protection diodes form a diode bridge that connects the output to
both the inputs.  (I forget where I heard of that one--I certainly
didn't invent it.)

Op amps obviously aren't specified for this, so you'll have to test
yours or ask the apps folks. (They probably won't know either, but it's
worth a try.)

Also the gigohm resistor probably doesn't help you at 120C--50 pA of
bias current puts you in the shot noise limit.

Cheers

Phil Hobbs
-- 
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

Geoffrey wrote:
> Hi there
> 
> I have a charge amplifier, used to amplify the signal from a piezo
> electric transducer, with 2nF of reference capacitance shunted with
> 1GOhm. I need to be able to reset the amplifier, preferably with a
> solid state device (the circuit has to be very compact, and operate in
> a high-vibration environment). One possibility which comes to mind is
> to simultaneously short the amplifier output and input to ground. The
> input is the inverting input of the amplifier, and is nominally at the
> same ground potential as the non inverting input. I am considering the
> use of a mosfet as the shorting switch for the input. When the mosfet
> is off, its gate will be at ground potential, as will its drain
> (virtual ground) and source. Is there a mechanism whereby a leakage
> current could still flow from drain to source? Every nanoampere of IDS
> gives 1V of offset at the output, and it should be noted that the
> circuit must work at 120 deg C.
> 
> Any help appreciated.
> 
> Best regards
> Geoff

Doesn't your amp have a reset input that can be controlled by a logic
level?  Oh, come to think of it, our Kistler amps need a switch closure.
 And I think some of the newer hideous ones omit this feature...

I'd seriously consider a reed relay to avoid any leakage or charge
injection weirdness from using a MOSFET.  Then again, a MOSFET might
work.  It'd take some poring over specs, then going home to dinner and
wondering what they mean, then having the lightbulb go in the shower the
next morning.  You know what I mean?

If I go to a job interview, how am I going to explain the way I do design?


-- 
_____________________
Mr.CRC
crobcBOGUS@REMOVETHISsbcglobal.net
SuSE 10.3 Linux 2.6.22.17

Am 03.04.2012 16:53, schrieb Geoffrey:
> Hi there
>
> I have a charge amplifier, used to amplify the signal from a piezo
> electric transducer, with 2nF of reference capacitance shunted with
> 1GOhm. I need to be able to reset the amplifier, preferably with a
> solid state device (the circuit has to be very compact, and operate in
> a high-vibration environment).

a couple of years ago I have seen a charge amplifier for X-Ray detectors 
which used the FET in the input stage to reset the integrator in a 
tricky way. The lid of the FET's metal can (I think it was a 2N4416) was 
was opened. To reset the integrator the crystal of the FET was lit by a 
pulsed LED until Vout was zero.

Cheers,
Alexander

On Apr 3, 11:53=A0pm, "Mr.CRC" <crobcBO...@REMOVETHISsbcglobal.net>
wrote:
> Geoffrey wrote:
> > Hi there
>
> > I have a charge amplifier, used to amplify the signal from a piezo
> > electric transducer, with 2nF of reference capacitance shunted with
> > 1GOhm. I need to be able to reset the amplifier, preferably with a
> > solid state device (the circuit has to be very compact, and operate in
> > a high-vibration environment). One possibility which comes to mind is
> > to simultaneously short the amplifier output and input to ground. The
> > input is the inverting input of the amplifier, and is nominally at the
> > same ground potential as the non inverting input. I am considering the
> > use of a mosfet as the shorting switch for the input. When the mosfet
> > is off, its gate will be at ground potential, as will its drain
> > (virtual ground) and source. Is there a mechanism whereby a leakage
> > current could still flow from drain to source? Every nanoampere of IDS
> > gives 1V of offset at the output, and it should be noted that the
> > circuit must work at 120 deg C.
>
> > Any help appreciated.
>
> > Best regards
> > Geoff
>
> Doesn't your amp have a reset input that can be controlled by a logic
> level? =A0Oh, come to think of it, our Kistler amps need a switch closure=
.
> =A0And I think some of the newer hideous ones omit this feature...
>
> I'd seriously consider a reed relay to avoid any leakage or charge
> injection weirdness from using a MOSFET. =A0Then again, a MOSFET might
> work. =A0It'd take some poring over specs, then going home to dinner and
> wondering what they mean, then having the lightbulb go in the shower the
> next morning. =A0You know what I mean?
>
>
  "If I go to a job interview, how am I going to explain the way I do
design?"

No problem just be honest.  "Well I think this is the way I would
solve that problem, but it usually takes me a day or two to process
things.  Can I drop you an email if I have a better idea in the shower
tomorrow morning?"

George H.
>
> --
> _____________________
> Mr.CRC
> crobcBO...@REMOVETHISsbcglobal.net
> SuSE 10.3 Linux 2.6.22.17- Hide quoted text -
>
> - Show quoted text -

George Herold wrote:
> On Apr 3, 11:53 pm, "Mr.CRC" <crobcBO...@REMOVETHISsbcglobal.net>
> wrote:
>> I'd seriously consider a reed relay to avoid any leakage or charge
>> injection weirdness from using a MOSFET.  Then again, a MOSFET might
>> work.  It'd take some poring over specs, then going home to dinner and
>> wondering what they mean, then having the lightbulb go in the shower the
>> next morning.  You know what I mean?
>>
>>
>   "If I go to a job interview, how am I going to explain the way I do
> design?"
> 
> No problem just be honest.  "Well I think this is the way I would
> solve that problem, but it usually takes me a day or two to process
> things.  Can I drop you an email if I have a better idea in the shower
> tomorrow morning?"
> 
> George H.


Oh I see I omitted a word.

What if what I meant to say was:  "then having the lightbulb go *out* in
the shower the next morning?"


Heh heh.


-- 
_____________________
Mr.CRC
crobcBOGUS@REMOVETHISsbcglobal.net
SuSE 10.3 Linux 2.6.22.17

On Wed, 4 Apr 2012 08:40:41 -0700 (PDT) George Herold
<gherold@teachspin.com> wrote in Message id:
<ca148295-6632-48e3-8cc6-76ec6ec0f8b4@w5g2000vbv.googlegroups.com>:

>On Apr 3, 11:53&#4294967295;pm, "Mr.CRC" <crobcBO...@REMOVETHISsbcglobal.net>
>wrote:
>> Geoffrey wrote:
>> > Hi there
>>
>> > I have a charge amplifier, used to amplify the signal from a piezo
>> > electric transducer, with 2nF of reference capacitance shunted with
>> > 1GOhm. I need to be able to reset the amplifier, preferably with a
>> > solid state device (the circuit has to be very compact, and operate in
>> > a high-vibration environment). One possibility which comes to mind is
>> > to simultaneously short the amplifier output and input to ground. The
>> > input is the inverting input of the amplifier, and is nominally at the
>> > same ground potential as the non inverting input. I am considering the
>> > use of a mosfet as the shorting switch for the input. When the mosfet
>> > is off, its gate will be at ground potential, as will its drain
>> > (virtual ground) and source. Is there a mechanism whereby a leakage
>> > current could still flow from drain to source? Every nanoampere of IDS
>> > gives 1V of offset at the output, and it should be noted that the
>> > circuit must work at 120 deg C.
>>
>> > Any help appreciated.
>>
>> > Best regards
>> > Geoff
>>
>> Doesn't your amp have a reset input that can be controlled by a logic
>> level? &#4294967295;Oh, come to think of it, our Kistler amps need a switch closure.
>> &#4294967295;And I think some of the newer hideous ones omit this feature...
>>
>> I'd seriously consider a reed relay to avoid any leakage or charge
>> injection weirdness from using a MOSFET. &#4294967295;Then again, a MOSFET might
>> work. &#4294967295;It'd take some poring over specs, then going home to dinner and
>> wondering what they mean, then having the lightbulb go in the shower the
>> next morning. &#4294967295;You know what I mean?
>>
>>
>  "If I go to a job interview, how am I going to explain the way I do
>design?"
>
>No problem just be honest.  "Well I think this is the way I would
>solve that problem, but it usually takes me a day or two to process
>things.  Can I drop you an email if I have a better idea in the shower
>tomorrow morning?"

Sure, but wait till you get out of the shower first.

On Apr 4, 1:59=A0am, Alexander <inva...@invalid.invalid> wrote:
> Am 03.04.2012 16:53, schrieb Geoffrey:
>
> > Hi there
>
> > I have a charge amplifier, used to amplify the signal from a piezo
> > electric transducer, with 2nF of reference capacitance shunted with
> > 1GOhm. I need to be able to reset the amplifier, preferably with a
> > solid state device (the circuit has to be very compact, and operate in
> > a high-vibration environment).
>
> a couple of years ago I have seen a charge amplifier for X-Ray detectors
> which used the FET in the input stage to reset the integrator in a
> tricky way. The lid of the FET's metal can (I think it was a 2N4416) was
> was opened. To reset the integrator the crystal of the FET was lit by a
> pulsed LED until Vout was zero.
>
> Cheers,
> Alexander

That's a neat idea!  If the signal is unipolar the OP could put an LED
across it and and reset it by shinning a second LED into the first.

George H.

Many CT systems use current to frequency converters, the X-ray photodetecto=
r arrays have similar large C issues.  This way the I to v converter is alw=
ays draining the charge a bit.  The IFC had a range of picoamps to 4 ma, re=
sulting in 5 Hz to 10 Mhz, from the oscillator. This gets rid of dumping th=
e integrator.
Picker/Phillips had the patent.

Steve=20

On Apr 5, 11:31=A0am, o...@case.edu wrote:
> Many CT systems use current to frequency converters, the X-ray photodetec=
tor arrays have similar large C issues. =A0This way the I to v converter is=
 always draining the charge a bit. =A0The IFC had a range of picoamps to 4 =
ma, resulting in 5 Hz to 10 Mhz, from the oscillator. This gets rid of dump=
ing the integrator.
> Picker/Phillips had the patent.
>
> Steve

My two cents for the design would be an op-amplifier with a high
fedility gain for the feedback. A second stage may be included as a
limiter, including the diodes in a back to back configuration for
reset.

eeL