Dick Cappels' project pages http://www.projects.cappels.org
Return to HOME (more projects)
A Simple Attenuator And Preamp For DC Measurements
A battery powered attenuator and preamp to extend the range of a voltage recorder.  Gains cover 40db, 20db, and 0 db
and can run for thousands of hours on a single coin cell.

The To keep things simple, a terminal block is used to connect
(from left to right) INPUT, COMMON, and OUTPUT.


After building the voltage recorder with playback, the first use I had planned for it was to record the activity of the water pump that supplies my household needs. This was going to require detection of signals in the tens of millivolts. The philosophy behind the voltage recorder was to make is very straight-forward and as simple as possible, and to make adapters as necessary for particular measurement applications. It was clear that I could take care of many of these applications by making an attenuator and preamp.

I had a few New Japan Radio Co. LTD NJU7051 single opamps that I had bought years ago. They have adequate offset voltage, offset input bias current, and input bias current specifications and -this is the good part -are rail-to-rail devices that can operate with a single power supply as low as 1 volt drawing less than 20 uA. This particular part may be obsolete now, but I fond similar opamps from New Japan Radio Co. at Digikey in January, 2013. Texas Instruments and probably others also make suitable opamps.

The current drain is so low when driving high resistance loads, I sometimes wonder whether an off position in the rotary switch is really warranted. Whether needed or not, its like the OFF button on a solar powered calculator; it gives the use peace of mind.

The gains are -40 db, -20 db, 0 db, +20 db, and +40 db, which correspond to gains of X 1/100, X 1/10, X 1, X10, and X100 respectively.

The Circuit

There is a single common ("GROUND") connection shared by input and output signals.              

In the -40, -20, and 0 db modes, the opamp is connected as a high input resistance voltage follower. The attenuation is accomplished with resistors, with the input resistance being between 1.1 Meg and 1.11 Meg Ohms.

In the +20 and +40 db modes, the opamp is configured as a non-inverting amplifier.

The input attenuator is made with a 1 Meg Ohm input resistor in series with the input signal, which is followed by a 10k resistor in the case of 40 db attenuation, a 111k resistor in the case of 20 db attenuation, and no resistor in the case of 0 db attenuation. To reduce stress on the opamp's inputs in the case of excessive voltage being applied to the input, two diode-connected JFETs are used in conjunction with the 1 Meg input resistor to limit current. One JFET clamps the incoming signal to the positive power supply and the other JFET clamps the signal to common (or "GROUND").

I started with Schottky diodes, hoping I would get lucky and get reverse leakage current a low lower than the specificed typical, but no such luck. The resulting offset at the input was comparatively huge at 15 millivolts at room temperature, which would double every 10 C. The 2N5485 JFETs have a very low reverse gate current of only 1 nA. I can easily live with that.

A Zener diode limits the power supply voltage just in case the input is exposed to a high voltage when the battery is not installed.

Connected at the same point in the circuit as the protection diodes, is a 0.022 uf capacitor which is intended to limit bandwidth to reduce noise some.  The intended -3db bandwidths when driven by a low resistance source are intended to be (I did not check each setting) as follows:

-40 db                            720 Hz
-20 db                            72  Hz
0 db, +20 db, +40 db    7.2 Hz

A slightly regrettable effect of having only a two pole rotary switch avalable for mode selection is that the resistances seen by the inverting and non-inverting inputs are quite different from each other in the -40 and +40 db modes. The resulting difference results in a small DC offset at the output.


No printed circuit board was needed, but it would have made it nicer to look at.

The opamp, the 1 Meg Ohm feedback resistor, the battery holder, the Zener diode and bypass capacitor are mounted on a hand-made circuit board made of a predrilled prototyping board with one pad per hole on 2.54 mm centers. Mounting the SO packaged opamp took a little time, but it was done successfully not once, but twice because I had destroyed the first opamp and had to replace it. I found that I could accommodate two IC leads on one pad by cutting the pad in half.