AD8036: need clarification on "fast rectifier" use.

First a disclaimer, I'm only a mechanical engineer, doing this out of necessity as a grad student.

I'm hoping to use either AD8036 or AD8037 as a peak detector for a nearly sinusoidal waveform at 100Hz-5MHz. The datasheet gives this handy schematic on page 19 and claims good performance to 20MHz. I am willing to tolerate error of a couple % at high frequency.

However, the same datasheet also says on page 16 (bottom of right column) that "The AD8036's and AD8037's CLAMPIN input clamp architecture works only for noninverting or follower applications..."

Doesn't the schematic show an inverting application?  Is this just an "oops" or is it the correct layout?

Now just a couple of basic questions. My intent is to use a simple RC filter on the output. I am sampling the output at 10Hz so a large time constant is desired. Can I assume that in this case, a relatively large C (0.1uF) does not constitute a problematic capacitive load?


Top Replies

  • 0
    •  Analog Employees 
    on Sep 15, 2014 11:16 PM

    Hi Ilya,

    1. The circuit is configured as inverting of gain unity and this is not an error in the datasheet.

    The CLAMPIN circuit only affects the A1's non inverting input. The CLAMPIN is comprised of comparators CH and CL, which drive S1 through a decoder. Considering the circuit, VH left open  and VL is referenced to 2Vin (due to noise gain of 2) and +Vin is grounded, S1 connects positive input of A1 to VL. Please refer on the datasheet page 19 ad 20 further explanation on how the circuit will operate.

    2. With your RC filter on the output, can you provide schematic?

  • Hi JinoL,

    thanks for the reply.

    Re 2: I realized after posting this question that a simple RC filter probably will not work. The amplifier will discharge the capacitor as fast as it will charge it. I may have to use a more classic 2 diode + op-amp peak detector circuit. There does not seem to be an advantage to using AD8036 here except perhaps using a 1 diode set-up and clamping the output to ground to avoid saturating at negative rail. Is that correct?

    Re 1: I understand the circuit explanation but then I don't understand why page 16 says that the "CLAMPIN architecture works only for non-inverting or follower applications". Does follower in this case mean specifically unity gain inverting?


  • 0
    •  Analog Employees 
    on Sep 20, 2014 11:32 PM

    Hi Ilya,

    Sorry for the delayed response.

         1. First, the datasheet theory of operation (pg. 15, paragraph 2) says that AD8036 and AD8037 can be used in either inverting or non inverting configuration but the clamp in function will only work at non inverting mode. in conjunction to this, Application Section (pg 17) says that the clamping feature of AD8036 and AD8037 only works when the amplifier is configured in non inverting mode, but to clamp in an inverting circuit an additional inverting gain stage is needed. With the circuit in Figure 12, these restrictions are met.

          With your concern 'does follower in this case means unity gain inverting?', Yes you're right but for the case when the input is in negative cycle only. For the positive input cycle, the operation is different (the operation explained on page 19).

         2. Can you share to me the circuit you have in mind? I want to have a closer look to assist you better. Thanks.

    Best Regards,


  • Regarding the peak detector function: integrating the rectified (out of the AD8036) wave will provide an image of the signal's amplitude, when this amplitude is constant.

    Detecting peaks is always a tradeoff, you have to think about the rise and decay times of the input signal.

  • 0
    •  Analog Employees 
    on Aug 2, 2018 2:46 PM
    This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin