I have a design with a tuner providing a 650KHz to 750KHz IF which is then sampled by an MCU at 1.6MSPS. The MCU has single end ADC input and the tuner has a differential output of up to 2.0V P-P amplitude.
In the first iteration of this design the tuner IF feeds a 4:1 impedance transformer used as balun, the unbalanced output of the transformer goes to an OPA365 for scaling and adding1.65V DC bias then it goes to the ADC input.
This solution is working but the transformer is rather large and the room is sparse on this design (USB dongle form factor), also the OPA365 gets easily saturated and introduces asymetrical distortion (the bottom tends to clip and the top tends to fold back and make unwanted product), also there is the problem that the IF amplitude varies too much which decrease the effective range sine the signal processing chain can only handle 3 bit of resolution loss (on 12bit).
Also later on i discovered that if i grab the IF signal on one side of the differential output before the transformer there is alreadya 1.65V DC bias, the amplitude is pretty good and the effective range is the same as if i use the IF at the output of the OPA365, with the advantage that the clipping is nice (it is the tuner clipping, which is clean and symetrical) and the noise floor is lower.
So i want to get rid of the transformer and OPA and add an AGC on the IF to keep it level constant as long as possible, in order to stay in the best range of my ADC. My idea is to grab the signal where i currently take it, directly on one of the differential lines of the tuner, and feed it to the AD8367 directly, which would be configured as an AGC and would hoepfully keep my ADC with high signal level even when the IF at the tuner output gets lower.
HEre are my question in this regard:
-will the AD8367 accept an IF that is already DC biased at 1.65V?
-does the AD8367 offer any option to introduce a DC bias at the output?
-will the AD8367 provide sufficient output level for my ADC (current range 0 to 3.3V, if needed i could possibly lower the VCC and thus the VDDA, which must be kept close on this MCU)
Thank you in advance.
[edited by: Hoel at 5:41 PM (GMT 0) on 9 Apr 2020]