I have made a design to use the "MicroZed 7020 SOM - Industrial Grade" where the PL is used for digital processing of filters and the Zynq is loaded with Ubuntu, during the design in "Vivado" the power report gave a value of "Total on-Chip Power = 2.461 W "and Juntion Temperature of 53.4 ° C. Which were nice values to our design.
In the design, the XADC has been instantiated using AXI4Lite and an application was made to read the values of power rails and temperature. The values are as follows:
PL rails:
VCC INT (1.0 V) VCC AUX (1.8 V) VCC bRAM (1.0 V)
--------------- --------------- ----------------
now: 0.963 V now: 1.789 V now: 0.967 V
max: 0.984 V max: 1.812 V max: 0.979 V
min: 0.948 V min: 1.783 V min: 0.962 V
PS rails:
VCC INT (1.0 V) VCC AUX (1.8 V) VCC DDR (1.5 V)
--------------- --------------- ---------------
now: 0.967 V now: 1.789 V now: 1.480 V
max: 0.988 V max: 1.813 V max: 1.488 V
min: 0.960 V min: 1.782 V min: 1.471 V
[Temperature]
now: 84.72 C
max: 91.67 C
min: 67.18 C
I am calculating the temp as per the formula in the XADC user's guide:
(code >> 4) * 503.975 / 2^12 - 273.15
It can be seen that the reading is correct for the power rails, but the temperature value is higher than expected.
We have reviewed the formula and the temperature tables to relate if there is a bad reading or calculation, but apparently the reading is correct.
This test has been carried out on three platforms, generating similar values on all three.
So, I am a bit worried about the die temperature. At an ambient temp of 24.4 C, XADC is reporting 85 C under nominal operating conditions.
Do you know why the discrepancy is so high between the estimated and the actual temperature value?
Is it possible that the design of the ZYNQ is not suitable for the processes carried out or the power report is not well calculated?
I found something similar in:
https://microzed.org/content/microzed-temperature-readings
Thanks