Hello everybody,
I have a problem with precision rectifier circuit.I want to rectify a 80mV signal of 1Mhz.I also want a excellent linearity.
Can anybody help me.
Are there any other methods to accomplish the process at high frequencies.

Regards,
Vdd

 

The usual implementation of a precision rectifier uses an operational amplifier. For low frequencies this works fine. The problem at 1 MHz is that you need an opamp that has both gain and linearity at that frequency. I think it will be both difficult and expensive to find. I think that making a stable circuit will also be tricky.

If you could tell us more about the application it might help.

 

Vdd,

We use a CLC4208BJP (comlinear) for a precision rectifier, this device has a fast response and good slew rate/linearity. I use 2 op amp rectifiers into a buffer then combine the outputs for full wave. Put the diode in the feedback path to overcome linearity problems. The operation will suffer from a high junction capacitance on the diode and some frequency compensation may be needed. I use an MBD701 Schottky diode, less than 1pF at 20V.

Here is the textbook op-amp rectifier. I don't know if this will work to 1MHz, I would try 2 half waves and sum them.

 

The usual implementation of a precision rectifier uses an operational amplifier. For low frequencies this works fine. The problem at 1 MHz is that you need an opamp that has both gain and linearity at that frequency. I think it will be both difficult and expensive to find. I think that making a stable circuit will also be tricky.

If you could tell us more about the application it might help.

Thank you for your kind response.
Here is my application,
I am working on inductive analog sensor.I have an oscillator whose frequency is around 1Mhz.As the target approaches the sensor the amplitude of these oscillations decreases.I want to convert this into equivalent DC linearly.Is it feasible to use precision rectifier or some other method will do the trick.
Please suggest some method where the rectification and linearity can be achieved simultaneously.

Regards,
Vdd

 

You can make a surprisingly accurate peak detector with a discrete circuit. Below is a schematic and simulated output waveforms. The 80mv p-p, 1MHz sine wave is modulated down to 20mv p-p. Feedback is what provides the accuracy. This is analogous to the op amp version.

 

An alternative approach is to apply your input signal to one input of a comparator, the other input comes from a DC , lets call this Vx. Now if the peak input exceeds Vx, then allow a pulse to be fired, integrate this pulse via a CR network and use this as Vx. What will happen is that on application of the input, the Vx will pump up until it reaches Vin- the comp input differential (<10mV). Its linearity is superb, its accuracy is the input differential and the ratio between the Dv/Dt on Vx and the time to reach or change from the input voltage. It can catch the amplitude of a 200nS pulse repeated at 25Hz to +=.1%. As used in TV ITS measurements.
Frank

 

Here is the textbook op-amp rectifier. I don't know if this will work to 1MHz, I would try 2 half waves and sum them.

It should work for 1MHz with Schottky diodes and sufficiently fast OpAmps (such as AD8039, which has 425 V/us slew rate).