ad8334 cascading issue

Hi all

I'm building a circuit to detect the envelope of a very small signal (some 10s of uV @250kHz, an ultrasound echo. I calculated that I'd need approx 90dB of gain before sending it to a leaky peak detector (and then to ADC: ltc2334-16).

Now I'm having strange or incorrect measurement results:

  1. I'm cascading first the LNA, then 3 times a VGA channel. According to the datasheet I'd expect (all in Hi mode) 19+3*36.5 dB = 128.5dB. 
    I'm measuring something more like 46dB of total gain, which is insufficient as I have a voltage drop in the leaky peak detector (due to the diode). 
    1. I'm wondering: is it possible to cascade 3 times a VGA channel (with each time 100nF) as a connection?
  2. In the signal there's at first a 'big' signal coming from driving the speaker, this is not possible to remove, yet of no interest to me. 
    1. Could it be that this is somehow blocking the next VGA (overloading it so quick recover is not quick enough?) (Not sure how quick the X-AMP can recover, so perhaps it is limiting all signals afterwards?)

concerning the picture, apparently scaling did drop of. input signal is scaled at 10mV/div, blue output is single-ended (via RF transformer as in the AD8331 eval board) output, 1V/div. 

based on the mechanical return signal I get 20log(4V/20mV) = 46dB



corrected the gain calc
[edited by: wouter.devriese at 2:15 PM (GMT 0) on 2 Aug 2019]
Parents
  • 0
    •  Analog Employees 
    on Aug 15, 2019 6:12 PM

    Hello wouter.devriese,

    Would you mind sharing the AD8334 portion of your schematic so I can see all of the connections?  If possible, can you label the supplies, where you are applying the input, and where you are measuring the output?

    If the large mechanical return signal is over-driving the part, calculating the gain based off of that won't be very accurate. Do you have a function generator available to replicate your signal of interest to apply to the inputs instead?  If so, the function generator likely won't be able to make a clean 10-100uV signal, so you may need to make a voltage divider to the input of the AD8334 and use a larger signal from the function generator (sine wave should be fine).  

    Thanks,

    Jordyn

Reply
  • 0
    •  Analog Employees 
    on Aug 15, 2019 6:12 PM

    Hello wouter.devriese,

    Would you mind sharing the AD8334 portion of your schematic so I can see all of the connections?  If possible, can you label the supplies, where you are applying the input, and where you are measuring the output?

    If the large mechanical return signal is over-driving the part, calculating the gain based off of that won't be very accurate. Do you have a function generator available to replicate your signal of interest to apply to the inputs instead?  If so, the function generator likely won't be able to make a clean 10-100uV signal, so you may need to make a voltage divider to the input of the AD8334 and use a larger signal from the function generator (sine wave should be fine).  

    Thanks,

    Jordyn

Children