Autosensing PSUs - Do They Work?
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee, Devonavar
Autosensing PSUs - Do They Work?
Do these new PSUs that increase or decrease their fan-speed depending on load really work, or is it just a gimmick?
If they are worth it, whats a decent one available around the 300-Watt mark, for under about £40... $80ish...?
Cheers
Rich
If they are worth it, whats a decent one available around the 300-Watt mark, for under about £40... $80ish...?
Cheers
Rich
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Please look in the Recommended Section.
BTW, virtually all modern PSUs have thermal fan speed control They all work in varying degrees -- the question is how well (meaning how aggressively does the fan speed get ramped up), and how good is the fan to begin with.
BTW, virtually all modern PSUs have thermal fan speed control They all work in varying degrees -- the question is how well (meaning how aggressively does the fan speed get ramped up), and how good is the fan to begin with.
-
- Posts: 170
- Joined: Tue Dec 03, 2002 11:29 pm
-
- Posts: 380
- Joined: Thu Jul 15, 2004 12:44 pm
- Location: UK
I think he means "why bother monitoring electrical load when monitoring thermal load will do", and he has a point.
Why bother speeding the fan up on high power load if the PSU is coping ok and isn't getting hotter?
It does seem a bit like a marketing gimmick to me. Thermal fan regulation is fair enough and has a use, but load monitoring seems like an inferior way of acheiving a similar result.
Why bother speeding the fan up on high power load if the PSU is coping ok and isn't getting hotter?
It does seem a bit like a marketing gimmick to me. Thermal fan regulation is fair enough and has a use, but load monitoring seems like an inferior way of acheiving a similar result.
MikeC: I think the original poster and larrymoencurly distinguish between fan regulation based on PSU temperature alone, and regulation also taking into account the instantaneous load drawn by the system.
larrymoencurly: As to why one would also regulate based on load, I can only speculate. By definition, if regulation is based on temperature alone, it will be impossible to achieve a constant temperature inside the PSU if load is not constant. If load goes up, temperature has to go up to make the fan spin faster.
On the other hand, measuring instantaneous load and PSU temperature, it would in theory be possible to devise a controller that keeps the PSU temperature completely stable.
Imagining two PSUs, the first controlling based on temp alone, the second also using load to keep temp constant under all conditions, both set to run at a given desired temperature for max load, the fan of the second will spin slower at sub-max loads, since it allows the PSU to run hotter under those conditions.
I'm pretty sure that that won't work perfectly in reality, and to know how useful this is one would need to know how much stability of temperature can be gained. Using a flat response curve at low temperatures and a steep curve above some critical (or desired) temperature with some smooth transition between the two, as Seasonic does, should get temperatures pretty closely regulated. The problem with regulation based on load is that fan speed (and hence noise) will change more rapidly, which in itself is more annoying than a higher but slowly changing noise.
Actually, my SuperTornado Rev2 does regulate based on load - in a very non-intuitive manner
EDIT: meglamaniac: You beat me to it, plus you said it better than me in about a quarter of the space
larrymoencurly: As to why one would also regulate based on load, I can only speculate. By definition, if regulation is based on temperature alone, it will be impossible to achieve a constant temperature inside the PSU if load is not constant. If load goes up, temperature has to go up to make the fan spin faster.
On the other hand, measuring instantaneous load and PSU temperature, it would in theory be possible to devise a controller that keeps the PSU temperature completely stable.
Imagining two PSUs, the first controlling based on temp alone, the second also using load to keep temp constant under all conditions, both set to run at a given desired temperature for max load, the fan of the second will spin slower at sub-max loads, since it allows the PSU to run hotter under those conditions.
I'm pretty sure that that won't work perfectly in reality, and to know how useful this is one would need to know how much stability of temperature can be gained. Using a flat response curve at low temperatures and a steep curve above some critical (or desired) temperature with some smooth transition between the two, as Seasonic does, should get temperatures pretty closely regulated. The problem with regulation based on load is that fan speed (and hence noise) will change more rapidly, which in itself is more annoying than a higher but slowly changing noise.
Actually, my SuperTornado Rev2 does regulate based on load - in a very non-intuitive manner
EDIT: meglamaniac: You beat me to it, plus you said it better than me in about a quarter of the space
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Ah -- 99.9% of fan control advertised or said to be based on "load" is actually thermal control. "At this operating ambient temp, with that load, the thermistor will reach such temp and the voltage to the fan will be so, so that the fan speed will be this. This means we can say the fan speed is controlled by load."
Real load based fan control may be found on industrial PC PSUs -- serious stuff for "mission critical" servers, etc. Very rare in retail PSUs. I have not seen it.
Real load based fan control may be found on industrial PC PSUs -- serious stuff for "mission critical" servers, etc. Very rare in retail PSUs. I have not seen it.
Re: Autosensing PSUs - Do They Work?
I've seen fortron supplies for sale at microdirect.co.uk, seasonic supplies are a bit more common in the uk.Blade wrote:Do these new PSUs that increase or decrease their fan-speed depending on load really work, or is it just a gimmick?
If they are worth it, whats a decent one available around the 300-Watt mark, for under about £40... $80ish...?
Cheers
Rich
griff
Seasonic Super Tornado and Super Silencer are pretty good - the SS 300W is £51.99+P&P at Kool'n'Quiet, and the Nexus 4090 400W is another excellent option, £61.10+P&P at Kustom PCs and Novatech. The ST 350W is £62.99 at Kool'n'Quiet.
Finally, the Nexus 3500 is £37.32 at Novatech, and looks like it contains the same fan as the 4090 now. I can't speak for the electronics inside, but it's probably based on a Fortron, as the 4090 is.
Finally, the Nexus 3500 is £37.32 at Novatech, and looks like it contains the same fan as the 4090 now. I can't speak for the electronics inside, but it's probably based on a Fortron, as the 4090 is.
Re: Autosensing PSUs - Do They Work?
the nexus nx-3000 is available for about 25 quid at novatech.co.ukBlade wrote:Do these new PSUs that increase or decrease their fan-speed depending on load really work, or is it just a gimmick?
If they are worth it, whats a decent one available around the 300-Watt mark, for under about £40... $80ish...?
Cheers
Rich
There was a 320W enermax noisetaker for 38 quid or 370w for 48 quid at aria.co.uk
edit: I have no experience of either these PSUs but the nexus makes the SPCR recommended list as does the more powerful enermax noisetaker
Its hard to get a powerful silent psu that is cheap as I'm finding out