I've done some scope plots for the problem on channel 3 and sent it to the support. It doesn't need to be a problem of the PSU in general, maybe my unit is just broken, but I doubt it. Maybe someone can verify it with his/her unit, if you have the same PSU?
I loaded channel 3 with an 8 ohm resistor, then I turned the PSU on with the power-switch (the PSU was turned off a few hours before). As you can see, there is a 1.5V spike, falling down over some milliseconds (1V and 1ms per division). The voltage selection for channel 3 was on 5V. If I turn it off for just a minute, and then on again, the spike can be even higher, I've measured up to 7V. There should be no voltage at all, because the channel is turned off initially when I turn on the PSU.
Same setup, this is what I see when I turn it off with the power-switch, and channel 3 still not enabled. A big 5V spike, falling over 100ms. And this is at 0,6A, because of the 8 ohm load. A sensitive DUT could be destroyed by it.
Same setup, channel 3 loaded with 8 ohm, the channel output was turned on with the CH3-ON/OFF switch, and then I shorted the output. As you can see, there are some serious overshoots, up to 12V! If I have a sensitive DUT attached and accidentally short the power supply outputs momentary, it could be destroyed by such overshooting. I don't trust this output at all anymore, because maybe there are some other overshoots, depending on the dynamic behavior of the DUT.
Just for comparing this with the other channels and to verify my measurements: same setup, but this time for channel 2 (8 ohm load, turned on). I shorted it momentary, as with the test for channel 3. As you can see, nice regulation without overshooting. Same for channel 1, no problems.