Antminer C1 melted power cables after settings change, re-powering

I have been using an Antminer C1 to keep a fishtank from freezing. I had it underclocked due to a mild winter, but during a recent cold snap it wasn't keeping up so I restored it to what I thought were defaults (possibly 225M/0700 or 250M/0725 depending on whether the one screenshot I took was before or after).

I made the change through the web interface but realized I had something on the stove so I neglected to power cycle the hardware itself until maybe 5 minutes later. But when I got there the power cabling (which was four 6-pin plugs harnessed off a single 8-pin "GPU" plug on the power supply) had melted, and the GPU plug on the PSU was carbonizing.

So now I'm wondering:

1. Did my failure to power cycle the unit immediately after changing the frequency+voltage cause this? I have run this miner for years at both normal and underclocked settings and never noticed the cabling anywhere close to this hot before.

2. Will powering the unit with four PCI-E male-male cables be a bit better than through the other/remaining GPU plug on the PSU? I understand 8 cables is ideal, but my supply only has six 6-pin sockets and one remaining 8-pin socket (I desoldered and cleaned up the remains of the other plug). Not to mention every $10 invested in this rig is not really coming back.

Obviously this isn't efficient hardware, but — so long as I don't have to keep buying new parts — it beats a straight electric heater and (for the fishes' sake) it's nice to have notifications from the mining pool when the power goes out to it.