I have qty. 4 of the ADRF5716 Evaluation Boards (EVB). Using an HP E3631A power supply, and the +25V output, I have:
- The positive supply set to +3.30V and connected to EVB's VDD
- The negative supply set to -3.30V and connected to EVB's VSS
- Ground from the PSU connected to EVB's Ground.
I power these rails up using the "Output On/Off" button on the power supply.
The rails come up simultaneously. There are 1K pull-up resistors on the digital control inputs of the EVB. I have an in-line current probe on VDD. It's ≈200µA peak during power-up.
I have jumpers on both digital inputs (0 dB reference state). I measure with a PNA insertion loss of ≈1.5dB.
After several power cycles of this (maybe 10-15 or so), and/or multiple torque/untorque cycles of the connector, attenuation will randomly "drop" by some amount in the reference (0dB state). In some cases, the insertion loss is suddenly 8dB, 12dB, 30dB, etc. At first, I suspect the SMA connector is mechanically loose, so I have the RF pin soldered, and I re-torque the connector's mounting hardware. Remeasuring the insertion loss in the reference state, I still see the same loss I measured before.
Replacing the IC on the EVB will often fix this problem temporarily.
- Could this be a mechanical stress issue due to the footprint on the PCB?
- Is there a known issue with power sequencing that causes permanent damage?
- How critical is it that VDD power up before VSS? Can they come up together?
- If the current measured @VDD input is low during power-up, and the digital inputs are coming up at the same time as VDD, then is there any cause for concern if the VSS pin is floating as VDD comes up?
- Cables are long coming out of the Power supply. What happens if I get a transient due to inductive load?
Thanks