Precision rectifier/clamp works near 0V for single-supply circuits


EiW-voltage-clamp-classic And here it is, modelled using LTSpice.

It is the same circuit as the classic simple precision rectifier (set to pass the negative half-sine), but with the non-inverting input of the op-amp connected to the voltage-to-clamp-to rather than 0V as it would be in a rectifier.

In this case, the input is blue, the output (the diode anode), is yellow. It is a 2Vp-p sine-wave and the clamp voltage is set at 1V.

All is working beautifully (although output resistance is 1k, so buffering mught be required downstream, and none of the circuits in this blog get over that limitation).

However, because it is a single-supply design, the op-amp is powered between 0V and a positive voltage (5V in this case), so the diode presents a problem if the clamp is required to be below 0.7V.

EiW-voltage-clamp-stressedSet to clamp at 0.5V, it all starts to look a little soggy, even with a rail-to-rail op-amp like the LT6015 specified, which pulls the diode cathode very close to 0V.

Needing to clamp to less than 0.5V, I had a little think, and came up with this arrangement, replacing the diode with an npn transistor and a 1k resistor, and swapping op-amp inputs to keep the negative feedback working (because the transistor adds an inversion in the loop).

EiW-voltage-clamp-unstressedNow it all starts to work again (output taken from transistor collector).

That said, the transistor has put a lot of extra gain inside the loop, so, to avoid oscillation, I would put a 100pF capacitor between the op-amp’s output and its inverting input – compensation experts might have a more elegant solution.

EiW-voltage-clamp-comfortable-at-0.2VThis scheme works right down to close to 0V rail – here it is clamping to 200mV.

Other restrictions

As well as the 1kΩ output impedance and the need for anti-oscillation compensation, this circuit (like all ‘simple’ precision rectifiers) leaves the op-amp output to fly-off into one rail or the other when it is not actively clamping (because the feedback loop is temporarily open).

Because the output flies off, not only will the amplifier draw more current (as the output structures are saturated) but the output also has to slew all the way back from the rail to the appropriate operating point when it is time to clamp again – meaning that clamping only starts after a delay – not an issue with this 100Hz waveform.

Luckily, in the transistor case, the output slews to the nearby 0V rail, so it does not have far (0.6V) to get the transistor operating.

In the simple ‘precision rectifier diode’ case, this situation is generally solved using the ‘improved precision rectifier’ (google if need be) circuit, but that would definitely need a negative rail to clamp near 0V (unless some kind soul can figure a way to do it with two transistors?)


Source link