I am seeing a Rx QEC Init Calibration failure when re-running init calibrations at frequencies below ~150MHz
Out profile boils down to
Device Clock: 122.88MHz
RX1 Enabled, 200MHz @ 245.76 MHz
ORX1 Enabled, 200MHz @ 245.76 MHz
This profile has been tested on the EVK is is working fine. We have our custom hardware and seeing the calibration failure there.
The getInitCalStatus returns the following:calsSincePowerUp 0xc06ecalsLastRun 0x0calsMinimum 0x4e initErrCal 0xe initErrCode 0x3
RxQEC Init Cal: Error configuring Cal PLL
What could be causing the Cal PLL to fail during calibration that is hardware related?
After programming with the profile on our EVB, generate the initdata.c and then use that in your custom board. The profile configuration related issues will be eliminated. Then check the matching on your custom board and also make sure that there is no input at RX port when you are initializing. Check the power supply voltage levels when the calibration is running on your custom board.
Thanks for the reply. Using the Linux driver and DTS configuration, the same on both the EVB and our custom board.
We have checked and re-checked the matching on the board, and even terminated the RX paths completely with no success.
We have discovered and believed to have fixed a too low voltage on the 1.3V supply, with no difference to the abovementioned behavior. Any specific rail you could point out which might cause this?
Can you move the LO to 160MHz and then check if the init RX QEC cal is passing or not? If you bypass the cal, then is it programming successfully?
If with EVB, the cal is passing with the same configuration, then it points to matching issue at low frequency. Matching needs to be done taking into consideration the layout.
If I set the LO to 150MHz, the RX QEC passes. If I set it to 149MHz, it fails. Same configuration works perfectly for the EVB.
Yes, I can do full initialization and programming when at startup, and init cal for any frequency above 150MHz. When I re-run RX QEC init cals for frequencies below 150MHz it fails (currently using IIO oscilloscope to control the device).
We'll look at the matching again, thanks.If ORX2 and RX2 are disabled in the profile, could the matching on those paths still affect the RX QEC cal? Is there any indication in the RX QEC status from the ARM as to which channel is causing the failure?
email@example.com said:We'll look at the matching again, thanks.
Also, check if you are seeing this issue on multiple boards.