Thanks for the replies. They help me understand what is being asked. Some points:
Using the HP 4-port USB 3.0 card and an Adapter, like the akasa, is an alternative. The USB3 card's internal connector matches the Intel specification, and is compatible with the adapter. It is good data that SDH has success with it. However, you discuss header #27 with power button, LED, etc. On a Z600, P27 is a 1x5 (with missing pin key) header for internal USB 2.0. P5 is the 2x9 front panel header with these signals. The front panel audio signals are on header P28. Some earlier workstations name headers differently. FYI.
Some cautions about signal integrity and cables. Poor signal integrity can cause all sorts of problems, like dropping devices, incomplete transfers, etc. There are many items to consider, including the motherboard/card USB3 printed circuit board signal trace lengths, cable length, and extremely important, the quality of the USB 3.0 cable from the motherboard or expansion card to the USB 3.0 connectors. The USB3 card has short PCB trace lengths, and the akasa cable seems to be shorter than 18 in. Shortening PCB trace lengths and cable lengths, even by an inch or less, can make a significant difference in performance. Cable construction is critical, especially how the wires are separated from the bundle and then soldered to the connectors. I have seen some poorly made cables that have an error rates up to a billion times higher than those that are well made. I can't go into details, but trust me on this. Intel's 18 in. spec assumes an (unrealistic?) best case for cards, motherboards, and cable construction. My $0.02.
BTW, the USB3 card does not have any USB signal redrivers. The PCB trace lengths are very short, so that is why I recommend using well-made cables that are as short as possible. FYI.
As SDH wrote, make sure that the USB3 card is plugged into a PCIe Gen2 slot. It will work in a Gen1 slot, but at half the bandwidth.
The Z620 motherboard has longer USB3 PCB signal trace lengths. With the longer cable to the front IO connectors, HP determined redrivers, located in the cable, were necessary to give reliable operation. HP testing requires exceeding the USB standard bit error rate by a significant amount, and to not have any USB errors, caused by signal integrity, when running constant data traffic for days at a time. HP spent a lot of time making this work. Trust me on this.
Back to the Z620 front panel. The panel's blue 20 pin cable has the same pinout as the Intel standard. However, since the cable has the redrivers "between" the motherboard and the USB connector, if these redrivers are not powered up by the 3.3V wires, they will be "off" and the USB 3.0 signals will not work.
The Z600 power supply does not provide 3.3V, only 5V, 12V, and -12V. The 3.3V power is created by circuitry on the motherboard. There is no easy way to tap into the 3.3V power, and I cannot recommend soldering wires to the motherboard (I am an HP employee, and must warn that mods can damage the system)
On a Z620, the redriver 3.3V power is on during normal operation, and also during sleep. Connecting 3.3V power that is active only when the system is "on", and not in sleep, will cause any USB devices to drop off during sleep. Wake, such as by a keyboard, wil not work. If 3.3V power is supplied all the time, it might work (again, cannot recommend), but system power consumption will be higher.
The Z620 front panel has 2 USB 3.0 and 1 USB 2.0 port. The USB 2.0 signals are routed from the "odd" pins on the connector going to P27. I absolutely cannot recommend removing the 2 orange 3.3V wires and ground wires, and plugging the connector with the remaining wires into blue 2x5 internal USB header like P25 and P26, to make the USB 2.0 port on the front panel work. (cough) Making USB 3.0 ports work is not as simple.