GearDraxon

If you look back through the thread, you'll see a couple of posts that detail rewiring the fan to allow for slower speeds, or using an Nvidia reference fan in the HG10. You're correct, though: many of us were looking to this product to reduce the overall noise level, and the fan they chose made it (significantly) louder in many installations.

An Oculus Rift is on its way to my home, so I gave up and ordered a replacement HG10 N980 from Amazon. If you're keeping score, this means Amazon got me in 2 days what Corsair couldn't in a month. Installed with the revised standoffs. Idle temps at ~32c, going to 75c under 3dmark testing, with the Corsair blower fan sounding like an airplane takeoff. Adjusted screws/thumbscrews and reinstalled about 7 times. Temps never got better, and frequently got worse. Gave up and reinstalled the EVGA ACX cooler. Idles at ~29c, quiet under load. Now I've got 2 HG10s and an H100i GTX with no use. What an amazing waste of time and money.

Last timeframe that I was told: Late March. In short, those of us that bought HG10 980s when they were released are stuck waiting, because they shipped out the fixed version to retail first, instead of us. Ridiculous.

So, Amazon has the fixed part in stock. That means I can spend another $40 and order a complete replacement for the non-working part I've had since November, and get it Wednesday. Or, I can wait until Corsair gets more of the standoffs for those of us with RMA tickets in (apparently late March.) This is getting silly, guys.

I've posted a comment to my support ticket - I'm a bit concerned that Graybeard phrases it as "there are a few left in stock," as I submitted my ticket right as we were advised. I've not received an update since then, so if they are suddenly out of the part needed to make this ~3-month-old issue right, I'll be pretty displeased. Edit: Yup. Ticket just updated with "Unfortunately, we do not yet have any of these available." I'm hoping that this is just a miscommunication.

Any word on replacement units shipping out to those of us who have had the non-working version for months?

Dustin, big thanks for responding. A quick question: - You mention only finger-tightening the thumbscrews. Does that mean that the other screws (mounting the N980 to the PCB) should be fully tightened? I think that myself and others perceived that tightening the mounting screws was leading to much of the bending.

That's actually what I've been trying to do. The problem is that, when installed, it's quite hard to get to the screws in question. Two of the thumbscrews are quite tight up against the HG10 shroud, and of course the whole thing is 90° to where it'd be convenient to adjust screw tightness. Oh, and I haven't seen anyone with idle temps in the 20s. Low 30s are the lowest I've seen.

Guys, I'm getting pretty good at wiping thermal paste off of GPUs and water blocks. Current setup: - Trimmed thermal pads (@mattlach, you will see what they're supposed to be covering when you look at them: a bunch of very similar chips that all of the pads line up with) - AMD bracket being used on H100i cooler - Thumbscrews on cooler fairly tight: more than finger-tight, but not going nuts with a screwdriver - Screws on PCIe side of the PCB tight - Screws on opposite side of PCB (with the SLI connector, etc.) not tight: holding the HG10 to the PCB, but not snug by any means - Freshly-applied thermal paste Result: less apparent bending, and idle temps of ~32-33c. This indicates to me that there's reasonably-good contact between the GPU and the water block, as previous setups idled at ~40c and crashed upon running 3dmark. Current setup runs 3dmark, but sends temps into the 90s, and causes the HG10 blower fan to go nuts, attempting to help. Then, because I care about you, I decided to run 3dmark whilst pressing firmly on the water block, near the PCIe connector. Temps went only to ~75c, and the blower fan stayed reasonable. As soon as I let up? Temps to the 90s. This thing is *just* slightly off of making good contact, and the culprit appears to be the lack of a securing screw where you'd expect one: above and to the right of the PCIe connector. There's a hole on both the PCB and HG10 there, but no integrated standoff (as there are for the other 6 holes on the HG10), and it appears to not be perfectly aligned - which may explain why they didn't do this. Either way, it seems pretty clear that if there were a way to tighten the connection at this spot, it'd help the water block -> GPU connection.

Hey look, all of us ACX SC+ers are in the same spot! I've now switched to using the AMD bracket with my H100i. It appears as though there is *far* less bending. That being said, the thermals have not improved: idling at ~40c, running 3dmark got me an immediate thermal shutdown.

That's good to know - the instructions say that you can you use either the Intel or the AMD ones. I used the Intel bracket, as it's what was pre-installed on my H100i, but perhaps there's less of an issue with the AMD one?

Right - they use a 970 for their demonstration, which has none of the bending / thermal pad issues. I'm curious if they'll release some official info on the "right" way to install this thing.

So, to fix the issue caused by poor instructions / weirdly non-precise installation needs, I should drill new, exciting holes into this thing? Nah, I'm good.

Correct. I'm attempting to install an HG10 N980 on my EVGA ACX 2.0 980ti, and the issue is two-fold: 1. The HG10 attaches to the PCB with 6 screws. On the HG10, there are small posts to receive each screw. There's a hole on the PCB (directly above the right-hand side of the PCI contacts) that could be used with a 7th screw, but there's no receiving post. Tightening the 6 screws to the point that you'd expect (so that the HG10 is tight to the PCB) bends the PCB. 2. The thermal padding on my unit was incorrectly placed/sized: it covered most of the VRM chips, but also covered a bit of some taller chips, which (combined with the lack of a 7th screw post) causes a fair amount of the bending. I trimmed the thermal padding, and made sure to not tighten the 6 screws completely (as another poster was advised by support.) I'm no longer getting the immediate-on-startup overheating/shutdown on my 980ti that I was before, but it's still not cooling it correctly - it idles at ~31c, but when running 3dmark goes directly to 90c, causes the HG10's blower fan to go nuts, etc. This is while hooked up to a new H100i GTX. Any thoughts?