What methods exist for determining energy usage?
What methods exist for determining energy usage?
With the T junction at 105C, a CPU operating at 92C remains within safe limits. It seems Intel or AsRock might have optimized it for such temperatures, similar to how the 13th gen i7 and i9 function at around 95C.
Consider using a PC case that incorporates a large heatsink into its design. These are often referred to as "passive PC cases."
Searching online will provide detailed information about these options.
For example, GamersNexus recently posted a video (uploaded an hour ago):
View: https://www.youtube.com/watch?v=ZVpO4Owmgko
From that video, you can access:
https://shop.streacom.com/collections/fanless-cases
Yeah, I've seen some from Streacom and different industrial, fanless designs before.
This project was meant to be an upgrade from a micro-server I ran on a Raspberry Pi.
I didn't want a large, heavy, costly case with fins everywhere.
Those cases were built for standard mini-ITX boards, not this one.
I'll check how much improvement I can get with a better CPU heatsink and case, then I'm ready to accept whatever comes next.
But I still want to understand how effective my cooling solution is compared to the one it came with.
The AsRock MoBo IS is indeed a standard mini-ITX design.
Isn't this too compact?
Details:
https://shop.streacom.com/collectio...-f...9753255121
It features fewer fins yet a larger size,
specifications:
https://shop.streacom.com/collectio...-f...6662813905
Regarding cost, the DIY option is affordable. Otherwise, the unique design and production of these specialized cases tend to increase the price.
Of course, you're welcome to build it yourself and add a more robust heatsink on top of the Pentium processor.
Consider an open bench PC case.
Alternatively, if aesthetics matter, Thermaltake has several choices:
https://www.thermaltake.com/products/cha...t_limit=30
The phrase "at the wall" is not acceptable because it implies an unrealistic or extreme condition.
I just thought about it. I've seen some review sites say they measure CPU power use, but I don't recall which ones or how they do it. I imagined there might be power meters between your PSU and motherboard, but I couldn't find one like that. I also considered checking the efficiency curve of the power supply and adjusting accordingly. I'm open to a more straightforward method. Thanks for the response.
Techpowerup clarifies this information on their efficiency page for all CPU reviews. They explain that all power readings are derived from a direct physical measurement of voltage, current, and power through the 12-pin CPU power connector, ensuring accuracy for the CPU alone—not the entire system. You can use a thin multimeter lead to touch the metal cuff and obtain the reading. In several videos I've watched, multimeter probes have also been placed directly on the CPU pins on the back of the motherboard, though this is risky unless you have a proper setup and know the pinout. A recent example with a timestamp is available here: https://youtu.be/kiTngvvD5dI?t=732
Thanks for the input. If I were setting up a regular measurement routine, I’d likely purchase an ATX power extension cable and connect probes to an ammeter. Have you seen any PSUs with I2C or USB ports for software-based reporting? It wouldn’t change much for me, but it does seem like a useful feature. Surprisingly rare in top-tier PSUs as of 2023.
So I purchased an AC power meter that claims to be more accurate than my previous Kill-A-Watt or the LCD display on my UPS.
This appears to be the most precise option available, at least in terms of cost:
Amazon.com
It also looked promising because the product description and Q&A section emphasized accuracy:
Amazon.com
Another model seemed interesting until a customer review mentioned incorrect readings when the auto-ranging shifted the device to another scale. This could significantly reduce its reliability if your load exceeds those limits.
Other models only provided kWh measurements with low precision (such as 0.1 or 0.01), which might seem acceptable at first, but it becomes problematic when dealing with small devices. For instance, running a 10 W device for ten hours would be necessary to get an accurate reading on the first meter, yet even then, the result could be off by up to 100%. I don’t need to use kWh mode, but I’d appreciate having that option.
Once it arrives, I’ll share my findings. If no better alternatives exist (like simple and safe ways to measure current directly across the ATX power connection), I think @Aeacus deserves the top answer for being first and offering the most options.
I've explored straightforward and safe methods such as hwinfo64, wattmeter, and UPS. Additional scientific approaches were also considered, like using probes in the +12V EPS power connector or at the CPU socket, which offer greater accuracy but come with significantly higher risks.
I’m contemplating another possibility, but the financial cost is prohibitively high, making it not worthwhile.
The concept involves a digital PSU, for example Corsair AXi, and monitoring its power output through iCUE software. This allows checking wattage ratings and efficiency differences (e.g., 64W from 51W wall to PC). However, simply purchasing the Corsair AXi just to track watts is a major waste of money.