Initial thoughts

I have heard so many great things about Noctua products throughout the years. Customers claim their fans to be the greatest and their quality control to be superb. However, the cooler I received was significantly bent so much so that the case this cooler was specifically designed for didn’t even close.

To remedy this, I had to unmount the cooler, remove the fan, and bend the heat pipes into place. It seems like others online are also facing quality control issues with this product. It’s a shame, because this product took forever to come to market, so I was hoping it would be perfect.

Benchmarks

I’m testing the NH-L12 Ghost Edition against the Ryzen Wraith Stealth, as the main question to answer here is if this new cooler is actually worth the extra $55 over the $0 Wraith Stealth stock cooler. To be totally honest, the stock Ryzen Wraith Stealth cooler is very good already. It’s not too loud, and it’s fairly compact fitting into the Ghost S1 with plenty of clearance. However, there are a few reasons why I wanted to replace it. The main issue being that it lets the max CPU temperature goes above 90˚C. Additionally, Ryzen 3000 CPUs have a weird tendency to intermittently fluctuate the fan speed, and the Wraith Stealth makes this change very audible via an odd groaning noise.

Tested on a Ryzen 5 3600 in a Louqe Ghost S1 with side panels on. Temperature, RPM, and clock speed recorded using OpenHardwareMonitor. Noise levels recorded using Decibel X iOS app. Each benchmark was repeated three times and results are an average of the three runs.

Cinebench R20

Cinebench multicore benchmark

Noise

• Wraith Stealth noise: 55.23 db
• Noctua noise: 58.75 db

Thermals

Fan Speed

CPU Boost Frequency

Blender

Render benchmark done using BMW demo by Mike Pan.

Noise

• Wraith Stealth noise: 56.90 db
• Noctua noise: 59.15 db

Thermals

Fan Speed

CPU Boost Frequency

Conclusion

While the Noctua is technically louder, it does keep the Ryzen 5 3600 cooler and has a subjectively prettier sound signature. The Ryzen 3000 boost speeds are highly dependent on how cool the chip runs (along with voltage and other factors). We observe a higher sustained boost frequency with the NH-L12, but it’s certainly nothing that would make a noticeable difference in day-to-day usage. After taking a look at the numbers, I have to say I’m fairly disappointed in the NH-L12 Ghost Edition. This is pretty much the best air cooler that will fit inside the Louqe Ghost S1, and I was hoping it would provide a lot more thermal headroom to push CPUs in the case of a future upgrade.

There is a chance that having to bend the heat pipes into place to get the cooler to fit may have caused some change in the thermal performance, but this feels like it would be very unlikely. Nevertheless, it will be something to follow up with Noctua on.

I wonder how this would fare in a more sustained workload like a 2 hour session of a demanding game. Maybe the Noctua would pull ahead against the Wraith Stealth and maintain a much lower max temperature. More so, these are just synthetic benchmarks where the GPU is not being utilized. In a real example where both the GPU and CPU are at full blast, there may be different results, as the CPU cooler also has to deal with heat from the GPU. There’s definitely more testing to be done. So far, the Noctua NH-L12 Ghost Edition is not bad, but it definitely fell short of expectations.