Q: How to amplify 15uV, 1.5MHz signal to read on Scopemeter

[JimG]

Perhaps it is a micron sized thermocouple measuring the
temperature of a speck of radium as it radioactively decays.

Yeah, that is probably it. ;-)

Nope, nothing nearly that exotic. It's a thermcouple simulator and DC
volt/millivolt source. When calibrated, it's output is 50 ppm +/- 3
uV from 11V to ~1 uV.

You probably understand how that fits with its chopper configuration,
but I don't. I just know I was told to be prepared to pick up a
signal around 15uV, which is what I am seeing now.

Unfortunately, the circuit diagram in the service manual has a
printing error, which led me down the 1.5MHz path for awhile.

Jim

 

[John Popelish]

Looks like my crude preamp circuit is going to work. I've been able
to increase the gain to 1500, which at 1.2kHz is holding up OK on the
LF411CN op amp.

If you use two opamp stages with equal gains of about the
square root of the gain of your single stage, you will get a
much better frequency response that you are getting, now.

The source signal from the chopper amp is indeed close to 15uV p-p,
which gets amplified to almost 25mV. This is a couple of divisions on
my Scopemeter and is easily visible.

But there is still a fair amount of noise on top of the basic 1.2kHz
signal. Any further suggestions for cheap ways to reduce this noise
are welcome.

A different scope would help a lot. An old analog scope
with a phosphor screen would be better at visually averaging
many traces, so that the brightest part of the image would
be a lot smoother than you get on a scope meter. But better
digital scopes include a signal averaging choice that
mathematically averages multiple passes, and displays the
average. That would really clean it up. Perhaps you can
borrow one.

It's not clear to me if something like the Minicircuits GALI-39 that
ninja suggested is still appropriate for a 1.2kHz signal with 0.05 ohm
output impedance.

No. Dropping the frequency requirement by 1000 means that
opamps are right in their sweet spot. But you could benefit
from one that has very low voltage noise for the first gain
stage, like an LT1028 or LT1037 or OP27, which all have
lower input noise than the LF411.

LF411 with about 25 nV/Hz^.5 at 10kHz:
http://cache.national.com/ds/LF/LF411.pdf

LT1028 with about 8 nV/Hz^.5 at 10kHz and over 10 times the
gain bandwidth product:
http://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1154,C1009,C1026,P1235,D3480

LT1037 with about 2.5 nV/Hz^.5 at 10kHz and about 20 times
the gain bandwidth product:
http://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1154,C1009,C1026,P1212,D2204

OP27 with about 10 nV/Hz^.5 at 10kHz but less than 1/5th the
gain bandwidth product:
http://www.analog.com/UploadedFiles/Data_Sheets/OP07.pdf

 

[JimG]

(snip)

Though, I would
still reserve 90% of the opamp gain bandwidth for the
harmonics of the signal, so I could see the wave shape,
clearly.

(snip)

I don't yet understand what you're telling me here, but I'll do some
investigation and see if I can figure it out.

I currently have it set up for around 30 dB gain. If I am
interpreting the op amp spec's correctly, it can handle 70 dB at 1kHz.

Thanks.

Jim

 

[John Popelish]

(snip)
(snip)

I don't yet understand what you're telling me here, but I'll do some
investigation and see if I can figure it out.

I currently have it set up for around 30 dB gain. If I am
interpreting the op amp spec's correctly, it can handle 70 dB at 1kHz.

At the fundamental frequency of 1.2 kHz, lets say you want a
gain of 100. That requires at least 100*1200=120kHz gain
bandwidth product from the opamp. but to produce that gain
for the 10th harmonic requires at least 100*1200*10=1.2 MHz
gain bandwidth produce. That is the factor of 10 I was
talking about. I would not ask the opamp to produce more
than 1/10th of the fundamental gain bandwidth (gain times
the fundamental frequency of the wave) so that it had
plenty of gain to accurately also amplify enough harmonics
of the wave frequency to accurately maintain the wave shape.

 

[JimG]

At the fundamental frequency of 1.2 kHz, lets say you want a
gain of 100. That requires at least 100*1200=120kHz gain
bandwidth product from the opamp. but to produce that gain
for the 10th harmonic requires at least 100*1200*10=1.2 MHz
gain bandwidth produce. That is the factor of 10 I was
talking about. I would not ask the opamp to produce more
than 1/10th of the fundamental gain bandwidth (gain times
the fundamental frequency of the wave) so that it had
plenty of gain to accurately also amplify enough harmonics
of the wave frequency to accurately maintain the wave shape.

Understood - thanks.

I take that to mean that I should limit the gain to around 300:

300 * 1.2kHz * 10 = 3600kHz

Gain bandwidth of the op amp looks like it's 4MHz at 25C.

Jim

 

[John Popelish]

Understood - thanks.

I take that to mean that I should limit the gain to around 300:

300 * 1.2kHz * 10 = 3600kHz

Gain bandwidth of the op amp looks like it's 4MHz at 25C.

That is what I am suggesting. If you want more gain than
that, cascade more amplifiers.

 

[JimG]

That is what I am suggesting. If you want more gain than
that, cascade more amplifiers.

Will do. Many thanks.

Jim

 

[Phil Allison]

"JimG"

But there is still a fair amount of noise on top of the basic 1.2kHz
signal. Any further suggestions for cheap ways to reduce this noise
are welcome.

** Errrrrrr - a LP filter ??

Fuckwit IT geek.



......... Phil

 

[Ninja]

On Fri, 26 Oct 2007 15:25:37 -0700, "Ninja"


The output impedance of the calibrator is very low - less than 0.05
ohm. Does the low impedance make the preamp you suggested more, or
less, feasible?

I realize that the thread has taken in a different direction now that the
bandwidth requirement has been reduced by a factor of 1000, but for the sake
of completeness . . .

The thermal noise floor (voltage wise) is proportional to the square root of
the impedance level. A signal from a low impedance source would be competing
with a lower noise floor, so maintaining a good signal to noise ratio is
easier.

Congratulations on your success. There's nothing like rolling up your
sleeves and digging into a problem to further your understanding.

 

[Daniel Mandic]

Fuckwit IT geek.

Turn Off

 

[Phil Allison]

"Daniel Mandic"

Turn Off

** Go stick your head in a gas oven.




........ Phil

 

[Daniel Mandic]

** Go stick your head in a gas oven.




....... Phil

fuch off Moron. I would step you into before my Ass sees the burning...


begone.... (anti-Nazi)

 

[Daniel Mandic]

** Go stick your head in a gas oven.




....... Phil

latinic fucktard

crawl back to your moronic earth-hole )))))


You make me laugh , you are fooni.



(when your comments show your electronic judiciousness,
)))))))))))))))



Keep on Clown.



Best Regards,

Daniel Mandic

 

[Daniel Mandic]

"Daniel Mandic"




** Go stick your head in a gas oven.




....... Phil

How can you be so Styrian? (County in Austria)


I have been fishing 20punds Carps wit' a 6mil synthetic core ))))

show me the same ;.)))))))))))))!



C'mon, you are a totally looooooooooseeeeeeer. You could not even match
in the Basics with me.

Phil........ )))))))))))


(yeah..... fill'em)

 

[Phil Allison]

"Daniel Mandic"



http://www.yates.com.au/Products/Pe...eholdPests/RatsakDoubleStrengthRTURatBait.asp






....... Phil

 

[Suzy]

Turn Off

Apologies for this Aus (?) character. Most of us are friendly people who can
discuss things. He is an exception. Most of us have blocked him and only see
his rubbish if quoted!

 

[John Fields]

Apologies for this Aus (?) character. Most of us are friendly people who can
discuss things. He is an exception. Most of us have blocked him and only see
his rubbish if quoted!

 

[Phil Allison]

"John Fields"
"Suzy the Self Appointed Net Kop Mad Bitch "


** Fraulein " Suzy the Net Kop " is likely the Austrian one.

Uncle Adolph would be proud of the dumb a dog shit NAZI slut.





....... Phil

 

[Daniel Mandic]

http://www.yates.com.au/Products/PestControl/RatsMiceHouseholdPests/Ra
tsakDoubleStrengthRTURatBait.asp

yar..... you can take it with you when going into... )))

 

[Daniel Mandic]

Aus(trian) or Aus(tralian)?

AUS (Australia), AUT (Austria).



Best Regards,

Daniel Mandic