Mains wire connection issue in some PSUs

[LongRunner]

Take a good look at this:

Notice the thick black wire in the middle?

Incorrect wire-to-board connection in HP-D3057F3H.jpg (97.67 KiB) Viewed 19375 times

The wire in question is one of the voltage selector leads (the other of which goes to the main PCB). It's the reinforced type of wire — UL AWM 1617, which has an outer jacket on top of an inner insulated wire resembling UL AWM 1015 (and both types are 600V rated). This is used in many units to provide double-insulation between the input leads and the secondary side, but as the photo shows, the wire-to-board connectors used aren't designed to accommodate the larger outer diameter. This is the unit pictured, but Hipro is not the only one to have made that mistake — In Win also did it in some of their units (e.g. the IP-P410Q3-2), and Casing Macron did a similar thing, only with a pluggable connector, for the input leads in the original MPT series (but at least fixed it in the second generation, although that's small consolation considering that they continued using the same crappy caps).

Hipro's newer HP-D3537F3H has reinforced wiring for all mains wires but the earth, and it turns out that the wire-to-board connectors used aren't that strong (one of those in my unit has broken).

The most disappointing part is that this issue is completely avoidable — the outer jacket of UL AWM 1617 can simply be stripped off near the connections if needed, and of course the other alternative is to use UL AWM 1015, with additional tubing to provide supplementary insulation where needed.

But for what it's worth, this issue could eventually become a thing of the past. There's a new type of wire insulation that's much thinner (yet also much more durable) for the same voltage as traditional PVC, and it is my hope that this (claimed) breakthrough will eventually extend to mains cordsets, among other things (including tubing, at least of the non-shrinkable type). Well, to be honest PVC is an old dog of a material, but I can handle a few more years with it.

[c_hegge]

I can see the issue, but I doubt if it is going to cause high failure rated for most users. I mean, if you don't take the PCB out, you won't stress the wires much and it probably won't be a problem.

[LongRunner]

Mind you, the wire that broke its connector (though most of the wire strands are still intact) in my HP-D3537F3H is the neutral from the EMI filter to the bridge rectifier on the main PCB. That pair of wires shouldn't even need supplementary insulation in the first place, as they're nowhere near the secondary, and it does seem to me that this oversight will result in a good deal more connection failures, than there would have been insulation failures otherwise.

I suppose the reason the wire-to-board connectors aren't stronger is because that would require larger through-holes in the circuit boards. But that's an expense they should be willing to bear.

Also, when I get the opportunity to repair that unit to working condition, I'm considering using Chemi-con's new KYB series for the 2200µF 16V 10mm caps on +12V. Comparing the specifications against Panasonic FK, it looks like KYB is the better of the two, overall (with an endurance rating of 10K vs. 4K hours at 105°C). 3300µF 6.3V 10mm is also included. Now we'll just have to wait another decade for 3300µF 16V 10mm to become a standard size…

One other problem with the unit besides the caps (both primary and secondary) and wire connections, though — a Zener diode used in the snubber on the primary side of the standby supply has gotten very hot . I just hope that's a consequence of the capacitor failure and not a design error.

[Behemot]

I have those 2200/16 KYBs in the process and I already have custom 3300/16 D10×40 mm if you want some Long. It is still possible friend of my GF will take some caps to Sydney, or I'll be sending caps to GaryR to Brisbane so they may ship from there to you.