AGC with AD8338

Hi MNK,

Good day!
For your inquiries, kindly see below response.

1) What details regarding attack time equation do you need?
2) For the schematic, you may check our users guide. Just connect the GAIN pin to the DETO pin and apply 0V to 3V on VAGC pin.
www.analog.com/.../UG-559.pdf
3) You can set the output by changing the voltage applied on VAGC pin. The RMS output voltage will be the absolute difference between VAGC pin and VREF. Example, if you have 2V on VAGC and 1.5V on VREF, then the output voltage will be 0.5V. Or if you have 1.2V on VAGC and 1.5V on VREF, your output will be 0.3V.
4) When AD8338 is set to AGC, you will only have control on the output thru voltage applied on VAGC. May I know why you want to control the gain?

Best regards,
Franz

Hi Franz,

Hope you are having a Good day as well .

Thank you for your prompt reply. I am a noob in regards to circuit design and These are the few issues in detail, I would appreciate your help and feedback on the following:

1) I needed more information on the following equation from data sheet which states T is approximately 1.75 millisecond.

T (sec) = 17,450 Ω × (285 pF + CDETO). but the units on RHS doesn't match with LHS.

2)  Down below is the Schematic for my circuit: The problem I am facing is I tried changing the voltage at VAGC form 0 to     3V. but Vrms remains at 1.51V.

3) The output of the AD8338 is out of phase with the input form the below figure.

4) For my application I need the Gain to be set within a range like from 30dB to 80dB for that I need the control on the output so that it lies within the corresponding range. Any additional suggestions regarding circuit are very much appreciated.

Best Regards,

MNK

Hi MNK,

Good day! I checked the spice model for AD8338 and I have seen that AGC functionality is not yet modeled.

I will update you once we have the updated model for this part.

Best regards,

Franz

Hi,

What is the noise coefficient of AD8338 in terms of dB?

Hi rookie1,

Does this older post answer your question about the noise factor of AD8338?

https://ez.analog.com/amplifiers/f/q-a/539393/ad8338

Regards,

Hooman

hi,Hooman

According to calculation, when G=0dB-->VGAIN =0.1V, the noise coefficient is very large. I request you to calculate and verify my result.

Regards,

Rookie1

Hi rookie1,

Yes, the AD8338 input referred noise voltage (G = 0dB) is 150nV/RtHz (a large number compared to generic op amps in the 10's of nV/RtHz range worst case) which would result in a rather large NF.

Here is what I calculate for NF assuming a 50ohm source impedance (0.91nV/RtHz noise density):

From the datasheet, with Vg = 0.1V --> Vout_noise = 1000nV/RtHz

NF = 10* log [1 + Na / (Ni * G)] = 10* log [ 1 + (1000nV/RtHz)^2 / ((0.91nV/Hz)^2 *1) ] = 60.8dB

Keep in mind that if you're dialing down the gain to the minimum 0dB, then you're dealing with a fairly large input signal to begin with. So, with such a large input signal, noise is not going to be a significant factor in most applications. The NF impact becomes a larger issue as gain is increased. At least, that's how I look at it.

Another way to look at this, with low AD8331 gain setting, your input SNR is already huge. So, the large NF of AD8331 would not impact anything. For example, assume a 1V_RMS input signal over a 10MHz bandwidth and 50ohm source impedance:

Input_SNR = 20 * log (1V_RMS / (0.91nV/RtHz * sqrt(10e6)) ) = 110.8dB

At the AD8338 output, the SNR will be:

Output_SNR = 110.8dB - 60.8dB = 50dB

Regards,

Hooman

Hi rookie1,

Yes, the AD8338 input referred noise voltage (G = 0dB) is 150nV/RtHz (a large number compared to generic op amps in the 10's of nV/RtHz range worst case) which would result in a rather large NF.

Here is what I calculate for NF assuming a 50ohm source impedance (0.91nV/RtHz noise density):

From the datasheet, with Vg = 0.1V --> Vout_noise = 1000nV/RtHz

NF = 10* log [1 + Na / (Ni * G)] = 10* log [ 1 + (1000nV/RtHz)^2 / ((0.91nV/Hz)^2 *1) ] = 60.8dB

Keep in mind that if you're dialing down the gain to the minimum 0dB, then you're dealing with a fairly large input signal to begin with. So, with such a large input signal, noise is not going to be a significant factor in most applications. The NF impact becomes a larger issue as gain is increased. At least, that's how I look at it.

Another way to look at this, with low AD8331 gain setting, your input SNR is already huge. So, the large NF of AD8331 would not impact anything. For example, assume a 1V_RMS input signal over a 10MHz bandwidth and 50ohm source impedance:

Input_SNR = 20 * log (1V_RMS / (0.91nV/RtHz * sqrt(10e6)) ) = 110.8dB

At the AD8338 output, the SNR will be:

Output_SNR = 110.8dB - 60.8dB = 50dB

Regards,

Hooman