Hi Analog Devices,
I'm seeing issues with the AD8309 Limiter outputs duty cycle as I approach smaller signal (and still a bit away from the noise floor) and can see significant component-to-component AD8309 variation.
In my application I am exclusively using the Limiter Outputs and have selected the AD8309 for it's significant dynamic range. It exists in a 50-ohm environment and is driven by a single-ended buffer.
I am driving single-frequency 10MHz with a precision input source. I can probe the Limiter outputs with a FET probe without disturbing the AD8309 detector, or effecting the load the Limiter is driving.
In the attached waveforms you will see two DIFFERENT AD8309 chips in the same circuit, both driven at the same signal input amplitude which is about -36dbm. You will notice that the Duty Cycle of the waveform goes to ~52/48 instead of 50/50. (*Note: This is just a single Limiter-output pin... probing the complementary Limiter-output since it is differential will show 48/52).
NOTE: these results are true for several AD8309's that I've now tested. Around QTY=10, not just QTY=2.
I've noticed that some AD8309's will track with a correct 50% Duty Cycle for all the way down to -40dbm, whereas others will start to have non-50% duty cycle as early as -24dbm... I can swap different AD8309's to see the difference in performance (literally soldering in/out different AD8309's).
Note that I've estimated my application noise floor to be -53dbm so I'm surprised I'm having issues as early as -24dbm.
I've attached several snipits:
1)SchematicForAD.jpg - Simplified schematic showing relevant information / test setup
2) LIMHI_GOOD.png - Showing AD8309 output for chip#1; note the 50% duty cycle
3) LIMHI_Bad.png - showing AD8309 output for chip#2; note the non-50% duty cycle
@Larger signal input, the duty cycle will return to 50%.