The problems around this connection do not stop, but exact numbers about real cases of damage are always rare. Even lists in social networks, forums, or collective threads on Reddit do not accurately reflect the real damage that has occurred. The number of people who do not speak out publicly could be much higher. Unfortunately, retailers and board partners are also extremely tight-lipped, making it very difficult for outsiders to publish statistically useful figures and remain objective.
Statistics and possible conclusions
Of course, one can ask providers of third-party solutions who have also implemented statistically evaluable numbers of units to get an approximate state of affairs. That’s exactly what I did and simply asked CableMod, since I have a very good relationship with the manufacturer. The fact that they were so transparent with me that I didn’t just get a few numbers has to be credited, because other sources are unfortunately less willing to talk and provide information.
Now let’s get to the facts. I think that a number of about 80,000 adapters sold (as of last week) alone speaks for a real need for action with regard to the screwed-up plug-in solution, and at the same time it is large enough to conduct statistical surveys. This includes all damages that were reported directly by the customer to CableMod as well as the feedback from the board partners to CableMod (RMA). Not included are possible dealer services where neither the AIC nor CableMod were billed or communicated with. However, this number is probably negligible due to the high costs, which no dealer likes to bear himself. Let’s now look at the recorded damage cases (GeForce RTX 4090 only) and notice a few interesting things:
| Manufacturer | Installation direction 12VHPWR | Damages | Percent |
| Asus | 180° rotated | 18 | 0,02250 % |
| Gigabyte | normal | 5 | 0,00625 % |
| MSI | normal | 5 | 0,00625 % |
| Zotac | normal | 2 | 0,00250 % |
| PNY | normal | 1 | 0.00125 % |
| Total | – | 31 | 0.03875 % |
We can see that 58% of the whole boards, i.e. far more than half, have a turned 12VHPWR connector and come from Asus. I have illustrated the difference with two board pictures of mine. Since CableMod also sells native, relatively flexible cables and no damage was recorded there, you have to ask yourself whether the protected story of the allegedly too stupid users is really still so generally tenable. The fact is, as I have proven several times, that any adapter solutions only cause problems when additional lateral forces act on the plugged-in connector (e.g. as a result of cable management).
Currently I’m trying to get one of the damaged cards with a repaired (inverted) connector to my lab, in order to do exactly the same heat flow analyses with this air-cooled card in a closed setup, which I could already do with a water-cooled card. Thematically, I will then also concentrate on the initial heating of the connectors by the board (waste heat of the VRM tracks and the shunts). With one card I could already observe a kind of “avalanche effect”, where the heated pins were the initial trigger for further heating in the connector (spring contacts, terminal). But this has yet to be proven for sure, so please be patient.
Summary and conclusion
The statically so low percentage of real failures is pleasing, but actually nothing should happen there. The fact that Asus is disproportionately represented here could also be due to the alignment and the opposite bending forces of the cable connections. Especially since the inverted socket also doesn’t have six ground connections as a thermal buffer between tracks/shunts and the contact pins in the socket. But this, too, still needs to be checked in detail by means of precise heat flow analysis.
However, we can also clearly see that one must be quite careful with simple assignments of blame to the users (or manufacturers of the cards). I would never have put on this shoe in my life, because the matter is really too complex for that and in the meantime it has been proven that even cards that were installed by professional assemblers are affected. And now? Do we start again from scratch? My self-built bridge, whose circuit board was not heated up by anything, also endures 900 watts with adapter or native cables. Gladly also slightly twisted or with lateral pressure, as long as the PCB is hot under 60 °C. That’s already a clue for a new story….
165 Antworten
Kommentar
Lade neue Kommentare
Urgestein
Urgestein
Urgestein
1
Neuling
1
Urgestein
Veteran
Urgestein
Urgestein
Veteran
1
Urgestein
Veteran
Urgestein
Neuling
Urgestein
Urgestein
Urgestein
Alle Kommentare lesen unter igor´sLAB Community →