This Time Self-Hosted
dark mode light mode Search

Saving energy?

Disclaimer: I’m not an expert, my physics knowledge is limited to high school classes, and I probably forgot most of it already; what I’m trying to do here is apply some low-grade math to a few numbers that got inferred out of reading about active PFC supplies around the net when I discovered I needed a new UPS. It might be accurate, in which case I’m lucky and happy, it might be totally bogus, in which case I’d like to know what is wrong, and what the truth would be. Don’t take my words from granted, and if you can give some more in-depth knowledge to either agree or disagree with this entry, I’ll see to update it.

This is a stream of consciousness entry, as I was thinking if it was worth the investment of replacing the three other PSUs I have in my home office with with active PFC, now that the UPS is replaced with the new one that supports them.

I’ve read that the improved efficiency of these PSUs is between 5 and 25… Alessio told me that it can reach 40% trying to avoid hyped values, I’ll try to make some calculations counting about a 10% saving over the size of the PSU, so that a 400W PSU replaced with an equivalent active PFC unit would save about 40W.

Now, I have three boxes in my home office, one of which is Farragut, which is turned on 24/7, meaning 8760 hours in an year; the PSU should be around 300W, so with a 30W saving, I would save 262.8 KWh in an year, which would then cost me (assuming the price for the second range: 1801-2640 KWh/year) €/cent 13.65 per KWh, which would be about €36 in an year. Considering the price of an active PFC unit to be about €80, in less than three years, it would have been paid back for good; if we count with the highest price, it would be about €59, so two years and I’m already in active.

The other two PSUs wouldn’t save me nearly as much, let’s say they would save me €10/year, it would take me a few years before paying them back, but then, they would help me with the noise (Farragut’s PSU’s is the most silent beside Enterprise, the other two are quite noisy) and the heat they generate.

It’s not bad after all, and saving energy is an important task nowadays anyway. But I wonder, how much can be saved for instance in my old ITIS (High School)?

I know of at least six or seven “servers” that are old PCs with ATX cases and PSUs, when I left the school three years ago; considering their age I’d suggest an average of 300W each too. That would be about 2000W and thus 200W saved by moving to active PFC equipment. They would also be active 8760 hours in an year, as most of them are active even when the school is closed during the summer; it would be already 1MWh/year saved.

Then there are about 9 laboratories with an average of 10 PCs turned on for about 9 months an average of 6 hours a day (and more than half of that time, someone is playing on them); they are usually newer so I’d suggest a 400W average PSU, saving 40W each. The save here would be 5.8 MHw/year.

For sure the price of energy for a school is lower than it is for an house, but it’s still a huge amount of energy, and money, even counting 10 eurocents per KWh, it would be about €600/year, for a single school; I don’t know how many schools like this are present in Italy, but seems like to me they can be quite a big amount of energy waste, considering the abundance of cheap hardware in their PCs, and servers.

Now that I’ve done a bit of math, I think I can say that the European Union directive that forces the use of active PFC power supplies does have a sense, and I’m glad it was issued, even if it forced me to buy a new (far from cheap) UPS.

Comments 5
  1. I’m sorry to say, but you are approaching it the wrong way around. The power used by a computer is determined by the power consumed by CPU/graphics-cards, etc and not by what the powersupply can (maximally) deliver.Have a look at this page from silentpcreview: http://www.silentpcreview.c…It shows you the typically power-consumption for a range of PC’s. The number under AC-input is the power that it actually draws from the wall-socket. So the first PC-configuration uses 71W when idle and 126W when running PCMark (so under maximum load). Bottomline is that you should not use the rating on the PSU to calculate the usage.Also interesting on that page is the graph that shows the efficiency of a PSU as a function of power output. There you can see that to reach the optimum efficiency of a PSU it needs to be carefully matched with the system. If you put a 600W PSU in a system that typically draws 60-120W efficiency drops an awfull lot. (80% efficiency is very good btw)So bottomline: If you worry about energy-consumption you need to approach it from multiple angles. The powersupply is only one of them, although putting a good one in clearly helps a lot. However most important for energy-savings is using energy-efficient CPU’s and none of these power-hungry graphics if you want to silence your pc the website I link to is a great resource. I specifically reviews components on their noise-production. Selecting a silent PSU combined with a good CPU-cooler makes a world of difference. 😉

  2. Just to make sure that I’m not misunderstood as it might be read that way:Putting energy-efficiency requirements on the PSU by the EU does indeed make a lot of sense. With the same configuration of your PC it allows a big reduction in energy-consumption without that people need to think about it. I’m afraid that efficiencies <50% are not uncommon for ‘cheap’ PSU’s as 50% was the minimum Intel design guide until 2003. And it is simply more expensive for a manufacturer to build efficient PSU’s as it requires better and more expensive parts.You might find it interesting to browse through this pdf titled: Power Supplies: A hidden opportunity for energy savings.…For televisions it shows how big the potential gain can be. Some manufacturers have 48% efficiency while others have 91% (see table 2). Nearly 40% lost as heat!! So clearly big savings can be made by forcing manufacturers to use efficient power supplies.

  3. moesaji, I know that 300W is not the drawn power usually, Farragut at the moment consumes about 170W or so (which is a freaking lot considering that it only has motherboard and company, and a single hard disk), the 10% was based upon the 25% declared and an average use of a PSU of 40-60%.And yes, this is only part of the solution, but if you see I’ve been working on the other parts of the solution already 🙂

  4. One should also consider the energy used to build the new and recycle the old PSUs, I’m sceptical that this really pays off, all things considered. Ok, perhaps it would save some money, but would it be good for the environment?

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.