PCI Express* Add-in Card Considerations
613768 23
NOTES:
Rest of Platform power here will not apply to all systems.
a. If Rest of Platform power is higher than what is in the table, increase the PSU size
respectively.
b. If Rest of Platform power is lower than what is listed in the table the PSU size can
also be reduced, but Peak Power requirements are also expected to increase beyond
what is listed in Table 3-3
.
CPU Power and Rest of Platform Power can vary from this table which would result in
custom Peak Power requirements
3.1.2 PSU Power Excursion
Based on the power budgets in Table 3-2 and peak power of both the Processor
detailed in Table 2-1 and the PCIe* Add-In-Cards in Section 3.1, the following Peak
Power Requirements are defined for the Power Supply. The peak time for each power
level is defined in
Table 3-3. The PCIe* Add-in card power excursions are only
designed for the new 12VHPWR connector. Therefore, the table is split into two
different power levels where it is expected that power supplies greater than 450 Watts
will have at least one 12VHPWR connector.
Table 3-3: PCIe* AIC and PSU Power Budget used for Peak Power Excursion
Power Excursion
% of PSU Rated Size
PSU ≤ 450 Watts &
PSUs without
12VHPWR Connector
Power Excursion
% of PSU Rated Size
PSU > 450 Watts &
12VHPWR Connector
present
Time for Power
Excursion (TE)
Testing
Duty Cycle
150% 200% 100 µs 5%
145% 180% 1 ms 8%
135% 160% 10 ms 12.5%
110% 120% 100 ms 25%
100% 100% Infinite --
NOTES:
Peak Power defined in this table correspond to Platform Level Power Budgets described
in
Table 3-2, if CPU or Rest of Platform power is reduced from Table 3-2 to reduce PSU
size, then custom Power Excursion % of PSU rated size must also be calculated.
This Table is only applicable if a power supply does not provide the I_PSU% signal. As
I_PSU% is required for ATX12VO power supplies refer to Section
4.3.7, refer Table 4-2
for an updated table.
The Testing Duty Cycle is not defined in the PCIe* CEM Gen 5.0 ECN, but it must be
defined to create a test criteria specific to a power supply. The Testing Duty Cycle
defines the percentage of time the Power Excursion value peaks with the remainder of
the time defined as a Time_Constant (T
C
). The Power Level defined is to have the RMS
value during this Dynamic Load test to be the Rated Wattage of the Power Supply.
Table 3-4
shows an example of how the testing criteria for power excursions of 1000
Watt Power Supply using an RMS value to average the rated power of a power supply.
This calculation is the same as what is described in the PCIe* CEM Gen 5 Power
Excursion ECN. For all power supplies with a rated wattage different than 1000 watts
a similar RMS calculation needs to be performed.