Unable to obtain Log output at the Limiter Output pin

The limiter output is intended to be a fixed amplitude ac signal that maintains the zero crossings of the original signal. So you would generally expect to see a square wave with a fixed peak to peak level. The limiter path has tons of gain and also has an output clamping mechanism. So a 10mVpp input ac signal or a 1Vpp input ac signal will come out of the limiter output at the same level. For this reason, some folks use the limiter output as part of an FM or phase modulation circuit where all of the modulation information is in the phase of the signal and none is in the amplitude of the input signal. For your testing I suggest that you start with a sine wave input. Observe the limiter output. Then start varying the pp input level and you will notice that the output level does not change. 

Thanks for the reply, We have performed some experiments, as per your direction : 

(a) Only Sine Wave:

     The input parameters are as follows: Sinusoid of 1MHz, with the p-p varying from 100mV to 2V

Output observed at the limiter port is a square wave as follows & doesn't change in amplitude. This is in line with the description that you have provided.  

(b) Sine Wave With a triangular modulating waveform: 

The input parameters are as follows: Carrier of 5 MHz, with the p-p varying from 100mV to 2V and the message signal frequency of 3KHz

The log of the message signal was observed in the Vlog output as below. There was a deformed square wave that was observed in the limiter port

Also, when we varied the signal power (P-P amplitude) of the input, the output was seen as shown in the graph below. This is the average DC output when measured from the multimeter:

Based on the above experiments we had the following questions:

1. Our application requires us to feed the log output of the input signal to the ADC for signal compression and peak signal detection. For this we were anticipating to use the limiter output. But , from the experiments and you suggestion, we intend to go with the Vlog output. Could you please provide your inputs for the same?

2.  Suppose an application requires to measure the input signal strength and also input signal compression, can both the operations be performed simultaneously ? Or does it require to use two different devices ?

Regards,

ARK

Hello,

Could you please let us know if this analysis is correct?

With Regards,

Avinash

The Vlog output is a signal strength indicator. Typically, you low pass filter the output signal to remove the ac (at 2x the input frequency) and give you a dc voltage for stead state CW input (or a log x type waveform if you apply a signal with a triangular envelope).  I'm not sure quite what you mean by "signal compression and peak signal detection" but can say that the limiter output will give you no information about the instantaneous or average level of the original input signal.  If you don't filter the log output, you will get an instantaneous output which is roughly equal to log(abs(vin)). If you filter the log output, you will get Avg(log(abs(Vin))).

I'm not sure how to answer question 2 because I don't know exactly what you mean by "input signal compression". If you are trying to detect signs that the input signal is somehow clipping, then you should probably do as little filtering of the signal as possible (so that you can see the instantaneous log(abs(vin)). You could digitize this and do the averaging in the digital domain to give you signal strength. 

Thank You. 

(a) Our application requires us to detect peaks in a time varying signal. The signal itself is centered around frequencies of 60MHz to 66MHz. A typical time representation of the signal is as below: 

The dynamic range of the signal can vary over a large range for which we intend to use the log amplifier. We expect this input signal to be "compressed" with the log amplifier to a range that is suitable for the ADC. Thereby  we wanted to know the suitability of the AD8309 & verified the same with a sinusoid signal modulated with a  triangular waveform which gave us the log(triangular w/f).  We primarily want to know if we will be able to retain the peaks information for our application, by processing the output of the log amplifier.

(b) Please see (a) for our intension of signal compression  

With Regards,

ARK