Oops apologies--meant to send that off line!
>From: "Dennis P. O'Leary" <doleary@h...>
>Reply-To: basicstamps@egroups.com
>To: <basicstamps@egroups.com>
>Subject: RE: [noparse][[/noparse]basicstamps] somewhat OT: op-amp parameters
>Date: Sun, 29 Oct 2000 21:18:29 -0800
>
>Stever,
> As a minor addition to Tracy Allen's excellent summary, using 2 or 3 op
>amps in a multi-stage design can simplify or eliminate considerations of
>individual op amp characteristics. Configure the first amp to match the
>sensor input impedance characteristics, using a FET-input amp if necessary,
>but with only part of the overall gain. This will prevent any input offset
>of the op amp from saturating the output. Then, use follower stages for
>filtering, auto zeroing, or setting a new zero level for A/D conversion,
>along with the remainder of the required overall gain. The later stages can
>be done with low-cost op amps.
> For example, my application amplifies bipolar 5 mV signals of
>electrode-generated human eye movements, drifting over a wide range, to
>0-4.096 V for A/D conversion. Stage 1 is a FET-input instrumentation amp
>with a gain of 220. Stage 2 adds auto-rest with an additional gain of 10,
>combined with analog low-pass filtering and a pot input to offset zero to
>about 2 volts. Stage 3 is a 0-5 V rail-to-rail op amp which brackets the
>signal for input to a A/D converter. It is all controlled by a PIC, which
>also sends a pulse to a Stamp BS1 generating an auditory tone driving an
>audio amp and a speaker. One great advantage of using a PIC controller, is
>that if the eye movements drift toward either 0 or 5 volts, the program
>detects it and activates the auto-reset back toward the 2-volt center
>position. If anyone would like details of this sample-and-hold auto reset
>circuit, contact me off line.
>Dennis
>
>
Original Message
>From: Tracy Allen [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=wBzH8EQ8y0cU6pvilsufJC5IMkXGk1SbAtZi4JHK6HJU3xULJWnJI2mx_jmvckXrj2qFhqfw6Wo4WFGHJg]emesys@c...[/url
>Sent: Sunday, October 29, 2000 8:34
>To: INTERNET:basicstamps@egroups.com
>Subject: [noparse][[/noparse]basicstamps] somewhat OT: op-amp parameters
>
>
>
>Steve Roberts: sroberts@s... wrote:
> >The question concerns how to interpret some
> >op-amp characteristics when trying to select a particular device. For
>example:
> >
> >1) Input offset voltage: It seems to me that if the "system" will have an
> >offset adjustment somewhere for the final output, why should you care
>about
> >10-100 uV of offset in one or more intervening op-amps? Especially if the
> >op-amps are used at low gains? And especially if the input signals are
>large?
> >
> >2) Drift over temperature: Again, if the gain is low (say 2-10) and lots
>of
> >op-amps can do < 1uV/C, so over say 20 degrees with a gain of 10 you'd
>get
> >200uV of offset referred to the output. But this seems small if the
>output
> >is volts.
>
>Hi Steve,
>You've answered your own questions!
>
>Conversely, when the input signal is low and the gain is high, that is when
>you need the low offset. Examples would be an amplifier for a thermopile
>or an oxygen probe. The gain may have to be x1000 or more. Starting with
>a low offset chip simplifies the circuit and gets rid of a pesky offset
>adjustments. It all comes down to a comparison of the signal to the error
>sources. A 20 microvolt signal amounts to about 1 degree Fahrenheit for a
>thermocouple. That may be significant for an incubator, but not so for a
>kiln.
>
>
> >3) Input voltage noise: Typically spec'ed as peak-to-peak or RMS noise
> >voltage at 0.1 to 10 Hz. Lots of op-amps have a 0.5uV p-p noise spec. But
> >again, with modest gains and largish signals, does this matter? If so,
> >under what conditions?
>
>Again it is a comparison of signal to noise. Trying to acquire a signal
>from the stars or from a gas analyzer or a long data transmission link will
>push the limits. In your gas analyzers, you chop the signal so as to
>narrow the bandwidth and bring the signal up out of the noise at DC. What
>limits the parts per million you can detect? Is it noise limited?
>
> >4) Input bias current: Here I can see that low input bias current lets
>you
> >handle higher source impedances, and FET inputs are good for very high
> >source impedances. Is this the main idea here?
>
>Right.
>
> >And finally, can anyone recommend a good summary of how to select a
>device
> >based on what's so good about each of the parameters? I've got Horowitz
>and
> >Hill, and Jung's op-amp book, and that's where I get much of my
> >information, but if possible I'd like to see a consise 2-page summary of
> >how to select op-amps. Any comments and/or recommendations would be
>welcome.
>
>Don't forget the AC parameters like bandwidth and slew rate, or things like
>common mode range, or power supply specs, not to mention packaging.
>Availabilty. Cost. I often find that my choices are dictated by factors
>other than 1-4 above. The best education is to understand what each
>parameter in the data sheet means. In addition to the sources you
>mentioned (which are excellent), there are the National Semi application
>notes. I like the books by Jerald G. Graeme (Burr-Brown guru), also
>anything by Bob Pease (National) or Jim Williams (Linear Tech). It is not
>only a matter of parameters, but circuits. All the major manufacturers
>have parameter trees on their web sites, and there are industry wide sites
>like specsearch.com. But they assume you know what you are doing and know
>how to rank your requirements.
>
>You start with the input signal source, the desired output, and the other
>restraints like the available power supplies. Then you narrow down the
>selection. If you are quickly down to only a few, that means you have a
>specialized requirement. Then go down the list of parameters on the data
>sheet one by one to see if there is anything you have overlooked that could
>come back to bite you later. Then you order some samples and put them
>through the paces to see if they really work. Or maybe you have a
>jellybean requirement and almost any op-amp will do. I find that most of
>my requirements can be met by a few op amps I keep in stock. It is when
>those fall short on some parameter that it is time to go shopping again.
>
> -- Tracy
> http://www.emesystems.com
>
>
>
>
>
>
>
>
>
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.
Share information about yourself, create your own public profile at
http://profiles.msn.com.
>Reply-To: basicstamps@egroups.com
>To: <basicstamps@egroups.com>
>Subject: RE: [noparse][[/noparse]basicstamps] somewhat OT: op-amp parameters
>Date: Sun, 29 Oct 2000 21:18:29 -0800
>
>Stever,
> As a minor addition to Tracy Allen's excellent summary, using 2 or 3 op
>amps in a multi-stage design can simplify or eliminate considerations of
>individual op amp characteristics. Configure the first amp to match the
>sensor input impedance characteristics, using a FET-input amp if necessary,
>but with only part of the overall gain. This will prevent any input offset
>of the op amp from saturating the output. Then, use follower stages for
>filtering, auto zeroing, or setting a new zero level for A/D conversion,
>along with the remainder of the required overall gain. The later stages can
>be done with low-cost op amps.
> For example, my application amplifies bipolar 5 mV signals of
>electrode-generated human eye movements, drifting over a wide range, to
>0-4.096 V for A/D conversion. Stage 1 is a FET-input instrumentation amp
>with a gain of 220. Stage 2 adds auto-rest with an additional gain of 10,
>combined with analog low-pass filtering and a pot input to offset zero to
>about 2 volts. Stage 3 is a 0-5 V rail-to-rail op amp which brackets the
>signal for input to a A/D converter. It is all controlled by a PIC, which
>also sends a pulse to a Stamp BS1 generating an auditory tone driving an
>audio amp and a speaker. One great advantage of using a PIC controller, is
>that if the eye movements drift toward either 0 or 5 volts, the program
>detects it and activates the auto-reset back toward the 2-volt center
>position. If anyone would like details of this sample-and-hold auto reset
>circuit, contact me off line.
>Dennis
>
>
Original Message
>From: Tracy Allen [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=wBzH8EQ8y0cU6pvilsufJC5IMkXGk1SbAtZi4JHK6HJU3xULJWnJI2mx_jmvckXrj2qFhqfw6Wo4WFGHJg]emesys@c...[/url
>Sent: Sunday, October 29, 2000 8:34
>To: INTERNET:basicstamps@egroups.com
>Subject: [noparse][[/noparse]basicstamps] somewhat OT: op-amp parameters
>
>
>
>Steve Roberts: sroberts@s... wrote:
> >The question concerns how to interpret some
> >op-amp characteristics when trying to select a particular device. For
>example:
> >
> >1) Input offset voltage: It seems to me that if the "system" will have an
> >offset adjustment somewhere for the final output, why should you care
>about
> >10-100 uV of offset in one or more intervening op-amps? Especially if the
> >op-amps are used at low gains? And especially if the input signals are
>large?
> >
> >2) Drift over temperature: Again, if the gain is low (say 2-10) and lots
>of
> >op-amps can do < 1uV/C, so over say 20 degrees with a gain of 10 you'd
>get
> >200uV of offset referred to the output. But this seems small if the
>output
> >is volts.
>
>Hi Steve,
>You've answered your own questions!
>
>Conversely, when the input signal is low and the gain is high, that is when
>you need the low offset. Examples would be an amplifier for a thermopile
>or an oxygen probe. The gain may have to be x1000 or more. Starting with
>a low offset chip simplifies the circuit and gets rid of a pesky offset
>adjustments. It all comes down to a comparison of the signal to the error
>sources. A 20 microvolt signal amounts to about 1 degree Fahrenheit for a
>thermocouple. That may be significant for an incubator, but not so for a
>kiln.
>
>
> >3) Input voltage noise: Typically spec'ed as peak-to-peak or RMS noise
> >voltage at 0.1 to 10 Hz. Lots of op-amps have a 0.5uV p-p noise spec. But
> >again, with modest gains and largish signals, does this matter? If so,
> >under what conditions?
>
>Again it is a comparison of signal to noise. Trying to acquire a signal
>from the stars or from a gas analyzer or a long data transmission link will
>push the limits. In your gas analyzers, you chop the signal so as to
>narrow the bandwidth and bring the signal up out of the noise at DC. What
>limits the parts per million you can detect? Is it noise limited?
>
> >4) Input bias current: Here I can see that low input bias current lets
>you
> >handle higher source impedances, and FET inputs are good for very high
> >source impedances. Is this the main idea here?
>
>Right.
>
> >And finally, can anyone recommend a good summary of how to select a
>device
> >based on what's so good about each of the parameters? I've got Horowitz
>and
> >Hill, and Jung's op-amp book, and that's where I get much of my
> >information, but if possible I'd like to see a consise 2-page summary of
> >how to select op-amps. Any comments and/or recommendations would be
>welcome.
>
>Don't forget the AC parameters like bandwidth and slew rate, or things like
>common mode range, or power supply specs, not to mention packaging.
>Availabilty. Cost. I often find that my choices are dictated by factors
>other than 1-4 above. The best education is to understand what each
>parameter in the data sheet means. In addition to the sources you
>mentioned (which are excellent), there are the National Semi application
>notes. I like the books by Jerald G. Graeme (Burr-Brown guru), also
>anything by Bob Pease (National) or Jim Williams (Linear Tech). It is not
>only a matter of parameters, but circuits. All the major manufacturers
>have parameter trees on their web sites, and there are industry wide sites
>like specsearch.com. But they assume you know what you are doing and know
>how to rank your requirements.
>
>You start with the input signal source, the desired output, and the other
>restraints like the available power supplies. Then you narrow down the
>selection. If you are quickly down to only a few, that means you have a
>specialized requirement. Then go down the list of parameters on the data
>sheet one by one to see if there is anything you have overlooked that could
>come back to bite you later. Then you order some samples and put them
>through the paces to see if they really work. Or maybe you have a
>jellybean requirement and almost any op-amp will do. I find that most of
>my requirements can be met by a few op amps I keep in stock. It is when
>those fall short on some parameter that it is time to go shopping again.
>
> -- Tracy
> http://www.emesystems.com
>
>
>
>
>
>
>
>
>
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.
Share information about yourself, create your own public profile at
http://profiles.msn.com.