Help Get Low Side of Voltage Output

--- You seem to be saying that you have a (say) 10 volt signal which varies (say) 10 millivolts on either side of 10 volts, and what you want to do is find a way to measure the tiny fluctuations in the presence of the huge 10V signal.

Right?

Tell us some more about your setup, OK?

What's the high voltage?

Is it AC? DC?

What kinds of impedances are we talking about? That is, where's the high voltage coming from and what does its load look like?

At what rate do the fluctuation occur?

What kind of measurement accuracy do you need?

-- John Fields Professional Circuit Designer

Two cases: ELF receiver and Fluxgate Magnetometer which I built from schematic on web. ELF max V is 5 volts which is stepped down across a resistive voltage divider to output 1.96v, 0.96v, and 0.5v. The voltage varies on the low side of millivolts and microvolts, of course. The voltage is varying DC and varies slowly. I do not believe this to be circuit related as a result of my testing. The magnetometer also has been given a voltage divider with a similar voltage range. As to how fast they change, well, I can watch it on my digital volt meter but the output is destined for a sound card which sees just the 0.5 volts so I would like to truncate the 0.5x then I can push the result through a voltage multiplier.

If the small voltages you want to monitor are higher in frequency than the near-DC baseline, then AC coupling will work. This can be either a big series capacitor feeding a hi-Z input gain stage, or a servo stage that subtracts out the average value of the input.

If the small voltages are also very low in frequency, you need an "expanded scale voltmeter" circuit, that basically just subtracts off a constant amount and amplifies the difference. Here you can use the heart of the above servo circuit, namely a large-TC low-pass filter to determine a baseline DC value, only implement it as a sample/hold. This is easy: Put a (CMOS) switch in series with the input resistor of the low-pass, and let the cap hold the charge. A hi-Z buffer looks at the cap and feeds that to a diff-amp that subtracts that value from the "live" input. When the output drifts out of range, close the switch to allow the RC to function again until it's caught up.

Best regards,

Bob Masta dqatechATdaqartaDOTcom D A Q A R T A Data AcQuisition And Real-Time Analysis

You won't believe! My tattered copy of Forrest Mims fell on the floor. I can use his inverting comparator for what I need. Thanks

Hi WSTORM. If you're still there, a non-inverting amplifier might not be the best idea. It has an input impedance equal to the value of the input resistor, like this (view in fixed font or M$ Notepad):

| ___ | .--|___|----. | C | Rf | | | | | Vin || ___ | |\\| | | o---||---|___|-o--|-\\ | Vout | || Ri | >-----o------o | .--|+/ | | |/| | === | GND (created by AACircuit v1.28.6 beta 04/19/05

For high frequencies, Vout = - ( Rf / Ri ). But for lower frequencies, the combination of C and Ri will result in a rolloff of lower frequency signal.

Better might be to use a non-inverting amplifier. You can have a single supply setup with an LM358 that can interface well with your sound card like this:

| | +5V | + | | | .-. | 3K3| | | | | +5VC .------------------. | '-' + | | | | |\\| | | | o-----|+\\ 2V| .-. | | | >----o 33K| | | .-. .-|-/ | | | | 2K2| | | |/| | '-' | | | | GND | | | '-' | | .----------o | | '---------o | | | | | | .-. | === .-. | 1K| | | GND 1M| | | | | Vout | | | | |\\ '-' 1uF To Sound | 1uF '-' '---|-\\ | || Card | Vin || | | >---o---||----o | o--------||----o-----------|+/ || | || |/ | | (created by AACircuit v1.28.6 beta 04/19/05

You will get a gain of 1 + (Rf / Ri), or +34 with this circuit. This should easily give you fair amplification of signals up to 100Hz, with a maximum output amplitude of up to 3Vp.p. The biggest limitation of the LM358 here is some crossover distortion. If you want very low distortion, or want more gain or a higher maximum frequency, you'll have to replace the LM358 with a better single-supply op amp.

Good luck Chris

Hi Chris Tnx 4 that. You took time and interest. I had decided to probe the direction of non-invert but WOW this was real help. TNX AGAIN.