PreAmp for Voltmeter and Oscilloscope
Ultralowdrift preamp gives a regular multimeter 100nV resolution and a scope 10uV/div sensitivity (dc-100kHz). The PCB serves as front-panel and holds all components. Simple one-opamp circuit, runs on 2x AA batteries.
PreAmp for Voltmeter and Oscilloscope:
For low voltage applications it could be used floating in combination with a Voltmeter. When used as an oscilloscope pre-amplifier, the amp is grounded by connecting it's BNC output to the oscilloscope input.
The gain can be switched from 1-10-100 V/V, the input can be switched between dc/ac coupling.
Adding the preamp (set to 100V/V) to my scope lowered the equivalent input noise of my measurements up to a factor of 14.
The document "PreAmp-VoltmeterScope.pdf" shows the circuit, measurements, calculations and build-data. In addition, it points out the limitations that I found in using a "zero-drift" chopper opamp.
Below is some basic information in text form (also to be found in the document):
Voltmeter preamp application, measured noise:
Measured noise (readout fluctuations) as a voltmeter preamp 100V/V
(shorted input):
Connecting to a voltmeter with 10uV resolution I get effectively 100nV resolution with <100nVpp fluctuations. Voltmeter example for this: I used a Aneng AN8002 multimeter (16Euro) having 10uV resolution.
To really see and quantify the fluctuations, I connected the output of the preamp to a high-end 6.5digit Keithley 6500 DMM. Measured preamp output (shorted input) fluctuations were 2.8uVpp at gain 100V/V over a 6 minutes timespan.
This means 28nVpp for the preamp input noise. The measured noisetrace can be found in the uploaded document.
The DMM was set to a moving average filter resulting in approx. 0.5Hz bandwidth.
Scope application:
Using 100V/V gain and the scope set to 1mV/div results in an increased input sensitivity of 10uV/div.
For a digital scope with FFT function the preamp lowers the equivalent input noise, helps to get a flat noisefloor and prevent seeing the peaks from internal scope interference (spurs) in the spectrum. See "PreAmp-VoltmeterScope.pdf" for further data.
Scope application, measured noise:
I first checked the scope noise itself (no amplifier, shorted input) that measures 50uVrms and the spectrum shows steep increasing noise at frequencies below 50kHz combined with peaks (spurs).
With the (shorted input) preamp on 100V/V connected to my scope I measured 350uVrms, a flat noise spectrum and no peaks.
Going from 50uV noise to 350uV noise with 100V/V additional gain means that you get 100 times more signal but only 7 times more noise. As the bandwidths are not the same, it is more correct to compare the noise spectra (FFT).
The FFT spectra that I took indicate improvement of the input noise floor up to a factor of 14 for frequencies below 50kHz.
PreAmp general data:
For dc input mode the input resistance is 10Mohm
For ac input mode the input resistance is 1Mohm, ac-coupling is from 1Hz up.
Supply current, including powerLED: 2mA
Battery supply is 2x AA, lasting for some 1500 hours of operation
Maximum output swing is 3Vpp in a high impedance load.
Some notes on the circuit:
-Blue powerLED also acts as batterystatus indicator. The LED goes off when the 2xAA total voltage get's below 2,7V.
-The ac/dc switch discharges the coupling capacitor when switched to dc. This prevents previous dc voltages being
imposed on the device under test.
-On power-off the supply nodes and the input protection diodes are switched to ground for a reason:
This prevents that input overloads during power-off are able to charge the supply lines above opamp damage levels.
Notes on the opamp:
Chopper opamps usually have higher input bias and noise currents compared to non-chopper CMOS opamps.
For this chopper opamp (OPA388) the input offset voltage is only 0.25uV but the input bias current is typically +/- 30pA. Connected to a 1Mohm resistor this results in 30uV input offset. The noise current is also relative high (100fA/sqrtHz).
A non-chopper CMOS opamp like the MCP601 specs 1pA bias current and 1 fA/sqrtHz noise current but has much higher offset and suffers from 1/F noise. If your application is only aiming at oscilloscope use you might consider using a non-chopper opamp.
Other opamps that run on 2.7V supply could also be used here as the PCB footprint is a standard pin-out SOIC package. Example opamp:MCP601
For low voltage applications it could be used floating in combination with a Voltmeter. When used as an oscilloscope pre-amplifier, the amp is grounded by connecting it's BNC output to the oscilloscope input.
The gain can be switched from 1-10-100 V/V, the input can be switched between dc/ac coupling.
Adding the preamp (set to 100V/V) to my scope lowered the equivalent input noise of my measurements up to a factor of 14.
The document "PreAmp-VoltmeterScope.pdf" shows the circuit, measurements, calculations and build-data. In addition, it points out the limitations that I found in using a "zero-drift" chopper opamp.
Below is some basic information in text form (also to be found in the document):
Voltmeter preamp application, measured noise:
Measured noise (readout fluctuations) as a voltmeter preamp 100V/V
(shorted input):
Connecting to a voltmeter with 10uV resolution I get effectively 100nV resolution with <100nVpp fluctuations. Voltmeter example for this: I used a Aneng AN8002 multimeter (16Euro) having 10uV resolution.
To really see and quantify the fluctuations, I connected the output of the preamp to a high-end 6.5digit Keithley 6500 DMM. Measured preamp output (shorted input) fluctuations were 2.8uVpp at gain 100V/V over a 6 minutes timespan.
This means 28nVpp for the preamp input noise. The measured noisetrace can be found in the uploaded document.
The DMM was set to a moving average filter resulting in approx. 0.5Hz bandwidth.
Scope application:
Using 100V/V gain and the scope set to 1mV/div results in an increased input sensitivity of 10uV/div.
For a digital scope with FFT function the preamp lowers the equivalent input noise, helps to get a flat noisefloor and prevent seeing the peaks from internal scope interference (spurs) in the spectrum. See "PreAmp-VoltmeterScope.pdf" for further data.
Scope application, measured noise:
I first checked the scope noise itself (no amplifier, shorted input) that measures 50uVrms and the spectrum shows steep increasing noise at frequencies below 50kHz combined with peaks (spurs).
With the (shorted input) preamp on 100V/V connected to my scope I measured 350uVrms, a flat noise spectrum and no peaks.
Going from 50uV noise to 350uV noise with 100V/V additional gain means that you get 100 times more signal but only 7 times more noise. As the bandwidths are not the same, it is more correct to compare the noise spectra (FFT).
The FFT spectra that I took indicate improvement of the input noise floor up to a factor of 14 for frequencies below 50kHz.
PreAmp general data:
For dc input mode the input resistance is 10Mohm
For ac input mode the input resistance is 1Mohm, ac-coupling is from 1Hz up.
Supply current, including powerLED: 2mA
Battery supply is 2x AA, lasting for some 1500 hours of operation
Maximum output swing is 3Vpp in a high impedance load.
Some notes on the circuit:
-Blue powerLED also acts as batterystatus indicator. The LED goes off when the 2xAA total voltage get's below 2,7V.
-The ac/dc switch discharges the coupling capacitor when switched to dc. This prevents previous dc voltages being
imposed on the device under test.
-On power-off the supply nodes and the input protection diodes are switched to ground for a reason:
This prevents that input overloads during power-off are able to charge the supply lines above opamp damage levels.
Notes on the opamp:
Chopper opamps usually have higher input bias and noise currents compared to non-chopper CMOS opamps.
For this chopper opamp (OPA388) the input offset voltage is only 0.25uV but the input bias current is typically +/- 30pA. Connected to a 1Mohm resistor this results in 30uV input offset. The noise current is also relative high (100fA/sqrtHz).
A non-chopper CMOS opamp like the MCP601 specs 1pA bias current and 1 fA/sqrtHz noise current but has much higher offset and suffers from 1/F noise. If your application is only aiming at oscilloscope use you might consider using a non-chopper opamp.
Other opamps that run on 2.7V supply could also be used here as the PCB footprint is a standard pin-out SOIC package. Example opamp:MCP601
Diskussion (0 Kommentare)