| General > General Technical Chat |
| What is the motivation of intel to want PSUs that are 12V only? |
| << < (12/14) > >> |
| DimitriP:
--- Quote ---What is the motivation of intel to want PSUs that are 12V only? --- End quote --- Let's see...google has been doing it since at least 2009 https://www.cnet.com/news/google-uncloaks-once-secret-server-10209580/ So either someone from google got hired by Intel , or Intel independently decided to catch up. They might also have decided to stop relying on varying quality third party power supplies. |
| coppice:
--- Quote from: DimitriP on March 09, 2020, 10:59:26 am --- --- Quote ---What is the motivation of intel to want PSUs that are 12V only? --- End quote --- Let's see...google has been doing it since at least 2009 https://www.cnet.com/news/google-uncloaks-once-secret-server-10209580/ So either someone from google got hired by Intel , or Intel independently decided to catch up. They might also have decided to stop relying on varying quality third party power supplies. --- End quote --- Rack mount servers moved to 12V only power supplies long ago. The present discussion is about the desktop catching up. |
| Psi:
Yep, i'm all for single voltage PC power supplies. Although having a separate low current standby voltage in addition to the high current feed does make sense. It seems pretty wasteful to design all PC PSU to do +3.3 at x Amps, +5 at y Amps and +12 at z Amps etc.. and then have all the motherboard manufactures decide to use some of that available power on some of the rails. And it would be super cool if they had the spec allow a voltage range that matched up to a number of lipo cells between flat and charged. |
| SiliconWizard:
--- Quote from: Psi on March 09, 2020, 11:37:03 am ---Although having a separate low current standby voltage in addition to the high current feed does make sense. --- End quote --- Given that it's hard to design a switching PS that is both very efficient at high and low currents, that could make sense. Unfortunately, I don't know how many people put their desktop PCs in standby instead of just leaving them on idle (or switching them off altogether when they are done), so I don't know how much impact that would have in practice. I for one never used to put my desltop PCs in standby (and it was often buggy in the past depending on your hardware and OS), but I've been doing so for the past year or so and it works very well - saves a lot of power indeed. Either way, the low current rail could still be 12V. I haven't read Intel's spec carefully, but I don't think anything would prevent to do this with a single rail with a little more "intelligence" in the PSU. It could switch automatically to different blocks in the PSU depending on the load - that would fit the bill without having to make the spec more complex or having to deal with more voltages/or do the switch-over on the motherboards if it can be done in the main PSU. --- Quote from: Psi on March 09, 2020, 11:37:03 am ---It seems pretty wasteful to design all PC PSU to do +3.3 at x Amps, +5 at y Amps and +12 at z Amps etc.. and then have all the motherboard manufactures decide to use some of that available power on some of the rails. --- End quote --- Given that on modern motherboards, much of the current is already taken from the +12V rail (CPU, GPU...), this all looked like a natural move. --- Quote from: Psi on March 09, 2020, 11:37:03 am ---And it would be super cool if they had the spec allow a voltage range that matched up to a number of lipo cells between flat and charged. --- End quote --- I'm not sure what purpose there would really be, I think they are mostly targetting desktop computer. Why would you want to power them with LiPo batteries? |
| mariush:
ATX was invented 20+ years ago... and it's still going. I think it would be somewhat shortsighted to standardize some voltage based on the number of cells of a particular battery technology. Who knows, 5-10 years from now we could have totally different battery cells based on nanotubes or hydrogen cells or graphene (ultracapacitors) 12v seems reasonable and I mentioned several times that 20v (chosen in the usb power delivery) sounds like a good compromise (would allow use of solid polymer capacitors rated for 25v or ceramic capacitors, mosfets should still work relatively fine in vrms and so on. I feel like a low voltage should remain to make low power devices efficient.. we now have Intel and Ryzen laptop processors on 7nm-14nm that consume 10-15 watts while doing actual work, so it wouldn't surprise me if a few years from now we'd have processors consuming less than 5w on average. Such tablets and laptops could probably work better with lower voltages like 7.4v (2 lithium cells in series) or whatever new battery technology will be (what if future battery cells will be 1.8v ..2.4v each...) I don't see why we could not have the standby 12v from an offline switcher (let's say 1-2A max and 90+ efficient) and 3-5 seconds after motherboard orders start up, you could have a mechanical relay or something in the power supply switch the 12vSB wire to the actual 12v dc-dc converter in the power supply, which is designed for 30-50A or even more, and which can do maybe upwards of 95% efficiency. |
| Navigation |
| Message Index |
| Next page |
| Previous page |