Searching \ for '[EE]: Instumentational opamp drift?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=instumentational
Search entire site for: 'Instumentational opamp drift?'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: Instumentational opamp drift?'
2002\07\26@104601 by A.J. Tufgar

flavicon
face
Hello all,
         I'm using the classic instrumentational three opamp setup with
all R = 100K except the one going between the two neg terminals which is
1K.

I was using a LM324 but as recommended by the list I tried a better CMOS
opamp and went with the LMC6482.

I'm having the same problem I was having with the LM324.  With the
LMC6482 my output voltage starts at 0.04V and goes to 0.08V in a matter
of minutes, without any changeon the transducer.

I don't think it's temprature drift as the LMC6482 has a relatively low
drift.  So anyone have any idea how I can get this annoying drift to go
away??  There must be some way to correct for this.

Thanks,
Aaron

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\07\26@112658 by Tom Messenger

flavicon
face
Aaron, what type of transducer do you have connected? Is it by any chance a
capacitive type? If so, you may have problems due to no dc path to ground
(or to a reference voltage) for your inputs of your amp.  Bias currents may
cause what you are seeing.

To separate the goat from the sheep, disconnect your transducer. Replace it
with a resistor network to get the approximate voltage your transducer
would have been generating.  Then turn on the amp and check it out.  If you
still have drift, it's the amp. If not, it's either the transducer or the
interaction of the transducer and the amp.

I'm assuming you have *carefully* checked all the wiring for your amp
circuit. And that you don't have large, leaky capacitors involved (perhaps
for signal filtering) that are causing the drift. And that you don't have
wires from the 'ducer to the amp that are unshielded and are 30 meters
long. And etc.

Good luck!
Tom M.

At 10:44 AM 7/26/02 -0400, you wrote:
>Hello all,
>          I'm using the classic instrumentational three opamp setup with
>all R = 100K except the one going between the two neg terminals which is
>1K.
>

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\07\26@114954 by Francisco Ares

flavicon
face
The temperature drift is a parameter that is considered as an input
offset temperature drift, so if you multiply its maximum value from the
datasheet by your gain and by, say, 10 to 20 degrees of temperature
difference from the moment that you turn on your circuit up to the time
it is tempeature stabilized, you will have your maximum overall output
drift. If the circuit is inside some kind of enclosure, the temperature
difference is higher.

Depending on how low is your transducer signal, you will have to look
for real low temperature drifts. This parameter is really one of the
worst pitfalls regarding op-amps.

I have already used TLC27L4 from Texas, it is the same pinout as LM324,
it has an extremely low temperature drift, but it needs a low voltage
power supply (only 16V from VSS to VDD) and it is very sensitive to ESD.
Another tradeoff is a very low gain X bandwidth.

Hope this helps.
Francisco


A.J. Tufgar wrote:

{Quote hidden}

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\07\26@121053 by Alan B. Pearce

face picon face
>I have already used TLC27L4 from Texas, it is the same pinout as LM324,
>it has an extremely low temperature drift, but it needs a low voltage
>power supply (only 16V from VSS to VDD) and it is very sensitive to ESD.
>Another tradeoff is a very low gain X bandwidth.

I have just been looking at the LMC6482, and while it has similar power
supply limitations, it claims 1.5kV ESD protection.

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\07\26@124551 by A.J. Tufgar

flavicon
face
Tom I think it's a combination of both the drift isn't as bad if I
remove and replace with a resistor network.

I'm using a simple silicon resistive gage transucer with a input
resistance of 5K and output restance of 3K.

I don't use any caps for signal filtering so it's not those.  :)

There must be some way to compensate for this drift.

Aaron

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\07\26@134403 by Harold M Hallikainen

picon face
       Well, let's see. The first stage has a gain of about 201. Input offset
voltage is multiplied by this to appear at the output (the second stage
has a gain of 1). Input offset voltage is 3.8mV at temperature extremes,
giving 763mV output offset. Since this would be the output offset if only
one of the input op amps had input offset, it COULD be double that if
both have input offset in opposite directions, or 1.5276V.
       Looking at input offset voltage drift, there's no maximum listed, but
typical is 1uV/degree C. If we double this for the two input op amps,
then multiply by the gain of 201, there's 402uV / degree C of output
offset drift. A 10 degree C temperature change would give you 4mV of
output change.
       Then, there's input bias current. If the op amps are driven by a low Z
source, the output due to input bias is Ib*If. The input bias current
here is 10pA max, giving 1uV for each input op amp. Bias current appears
to be a minor contributor! Since input offset current is half this, it's
even less of a contributor.
       So, it looks like input offset voltage is the major contributor, but you
could subtract that out, since it's fixed. Input offset voltage drift
would account for about 1/10 of the drift you're seeing with a 10C
change.
       Next would be resistor temperature coefficient. Because the bias current
is so low, changes in the feedback resistors do not make significant
error contributions with bias current. However, assuming there's a common
mode voltage (and I assume there is, since you're using an
instrumentation amp), minor changes in resistance, especially in the
"third" op amp input circuit, make significant deterioration in CMRR. It
would be interesting to see if the output drifts with both inputs to the
instrumentation amp grounded. As you bring the inputs up to the
transducer output voltage, I expect you'll see output variation due to
common mode gain (which is ideally zero). So, I expect the problem is
resistor temperature coefficient which is causing variation in the common
mode gain, which is giving you an output due to the common mode input.
       My solution: buy an instrumetation amp chip! These have precisely
matched resistors on the chip.

Good luck!

Harold


On Fri, 26 Jul 2002 10:44:34 -0400 "A.J. Tufgar"
<spam_OUTtufgarajTakeThisOuTspamMUSS.CIS.MCMASTER.CA> writes:
{Quote hidden}

FCC Rules Online at http://hallikainen.com/FccRules
Lighting control for theatre and television at http://www.dovesystems.com

Reach broadcasters, engineers, manufacturers, compliance labs, and
attorneys.
Advertise at http://www.hallikainen.com/FccRules/ .


________________________________________________________________
GET INTERNET ACCESS FROM JUNO!
Juno offers FREE or PREMIUM Internet access for less!
Join Juno today!  For your FREE software, visit:
dl.http://www.juno.com/get/web/.

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\07\26@141031 by A.J. Tufgar

flavicon
face
Hmmmmmmm, I just left the circuit for about an hour and I don't think
the drift ever stops.  :P

I had it on for about 20 minutes before I left and it made it to 260 mV
when I came back it was at 350.....

I'm guessing it's probably due to impresisions in the resistors....

Aaron

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\07\26@172007 by Dave Dilatush

picon face
Aaron wrote...

>I'm using the classic instrumentational three opamp setup with
>all R = 100K except the one going between the two neg terminals which is
>1K.
>
>I was using a LM324 but as recommended by the list I tried a better CMOS
>opamp and went with the LMC6482.
>
>I'm having the same problem I was having with the LM324.  With the
>LMC6482 my output voltage starts at 0.04V and goes to 0.08V in a matter
>of minutes, without any changeon the transducer.
>
>I don't think it's temprature drift as the LMC6482 has a relatively low
>drift.  So anyone have any idea how I can get this annoying drift to go
>away??  There must be some way to correct for this.

At the gain you're trying to achieve (approximately 200), you'd
be better off buying an IC instrumentation amp from Linear Tech,
Analog Devices, or Maxim.  You'll get much better results,
because the internal resistors in these parts are much better
matched than what you can achieve by using discrete resistors.
This is important because unless you have very good matching in
your resistors, you're going to get very poor common-mode
rejection, maybe only 35-40db; this may be part of your drift
problem, unless your power supply voltage is VERY tightly
regulated.

Also, you indicated in another post that you're processing the
signal from a silicon resistive bridge pressure transducer; and
that when you disconnect the transducer and replace it with a
resistor network, your drift improves.  Be aware that these
silicon sensors- at least the ones without any on-chip signal
conditioning- can have REALLY terrible offset and span
temperature shifts.

Just as an experiment, try placing your circuit in a temperature
chamber (or a box with a small light bulb or other heater) and
see what temperature dependencies you have, with and without the
sensor.  Compare the results against the sensor's specs; it may
well be that you're getting as good as what can be expected.

DD

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\07\27@095159 by Peter L. Peres

picon face
On Fri, 26 Jul 2002, A.J. Tufgar wrote:

>Hello all,
>          I'm using the classic instrumentational three opamp setup with
>all R = 100K except the one going between the two neg terminals which is
>1K.
>
>I was using a LM324 but as recommended by the list I tried a better CMOS
>opamp and went with the LMC6482.
>
>I'm having the same problem I was having with the LM324.  With the
>LMC6482 my output voltage starts at 0.04V and goes to 0.08V in a matter
>of minutes, without any changeon the transducer.
>
>I don't think it's temprature drift as the LMC6482 has a relatively low
>drift.  So anyone have any idea how I can get this annoying drift to go
>away??  There must be some way to correct for this.
>
>Thanks,
>Aaron

It's normal for analog circuits to drift during the first few minutes.
Books mention 30 minutes wait time before the circuit is considered
stable.

Peter

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2002\07\27@161354 by John

flavicon
face
Hello Aaron,  Alan, Francisco & PIC.ers,

I designed, built & twiddled many of these traditional `3-op-amp'
inst. amp. ccts going back years to the early '80s.
Started with (? I think) LH0044 parts and finally used what was
available & pretty cheap, the PMI  OP07.
Best results obtained when I troubled to purchase a significant
qty. of the feedback resistors (27k 1% mf types) and then did my
own screening of these for consistency using the Fluke 8062
4 1/2 digit meter I had at the time. Chucked out the outlying values
& segregated the rest into `sets' of 6 which agreed to the least sig.
place.

Even so, I never got real satisfaction regarding zero drift.
The best I could do was give the machine operator the means to bias
the inputs by `fine' and `coarse' panel pot. adjustments... shuudder...
Some competing instruments at the time even had reams of counters
and gate logic to perform `auto-zero' on the amps. by biasing the inputs
on push-button demand from the operator.
The sensors are generally 20mV FSD strain gauge bridges.

Since you are using a PIC after the I/Amp ( you *are*,,, mm..?)
this requires you to use the onboard a/d, or alternatively an external
ADC package to achieve all this.
*If* your application sensor is a strain gauge bridge, or for that matter
any
device with real low output range, then I would highly
recommend you have a good look at Analog Devices AD7715 or 7730.
These are 16bit / 24 bit parts with integrated I/Amp, programmable gain
and SPI serial output. They are aimed at this type of job.
The AD7715 goes in my latest product line & it's cleared up a lot of the
old-world problems for me.
Not cheap (~$10 apiece here) but worth it.


       bestos,   John

>Date:    Fri, 26 Jul 2002 12:48:59 -0300
>From:    Francisco Ares <.....francisco.aKILLspamspam@spam@ALTEC.COM.BR>
>Subject: Re: [EE]:  Instumentational opamp drift?
>
>The temperature drift is a parameter that is considered as an input
>offset temperature drift, so if you multiply its maximum value from the
>datasheet by your gain and by, say, 10 to 20 degrees of temperature

<snippery>

{Quote hidden}

e-mail from the desk of John Sanderson, JS Controls.
Snailmail:          PO Box 1887, Boksburg 1460, Rep. of South Africa.
Tel/fax:            Johannesburg  893 4154
Cellphone no:   082 741 6275
email:                jsandspamKILLspampixie.co.za
Manufacturer & purveyor of laboratory force testing apparatus, and related
products and services.

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


More... (looser matching)
- Last day of these posts
- In 2002 , 2003 only
- Today
- New search...