Any probs using different PSUs in a computer?
I'll explain more.
For my next computer, I plan using an ATX PSU for the motherboard, while the HDDs, optical drives and fans would run off a linear PSU. (With the same ground...)
Will it cause any problems? Current flowback?
I'm not sure that's what I'll be doing exactly tho.
OR
Instead of using an ATX PSU, are there not-too-noisy options that are still power-efficient?
I never worked with SMPSes, so I'll ask dumb questions.
What about a big toroid with switching regulators, and big capacitors to clean the noise? Would it still be efficient?
I'll explain more.
For my next computer, I plan using an ATX PSU for the motherboard, while the HDDs, optical drives and fans would run off a linear PSU. (With the same ground...)
Will it cause any problems? Current flowback?
I'm not sure that's what I'll be doing exactly tho.
OR
Instead of using an ATX PSU, are there not-too-noisy options that are still power-efficient?
I never worked with SMPSes, so I'll ask dumb questions.
What about a big toroid with switching regulators, and big capacitors to clean the noise? Would it still be efficient?
In fact, I first thought of making 1 big PSU for all the system, but, after I read the specs, I found it isn't as simple as green wire grounded = turn the thing on...
(There are things like overvoltage, overload, overheat protection, a very short delay before power on which isn't nice if you use a big toroid)
To power a 2nd ATX from the 1st ATX, all you need is a relay powered from the first to short green wire to ground.
I meant multiple, not different. Doh!
What I'm really looking for it to eliminate as much noise as possible, since my computer and sound system are right next to each other, and as I'm planning to DIY a case for my next PC, I told myself, why not DIY a PSU also.
I thought about a complete linear PSU, but linear regulators aren't really efficient... (And would make a 300W heater)
But, what's noisy in SMPSes, the 400Hz transformer, the switching regulator or all these answers? If I take a toroid and switching regulator, I probably put less junk on the power line, but what about RFI / noise / interference? Is the regulator making some itself or can just putting big caps on the power line fix things?
(There are things like overvoltage, overload, overheat protection, a very short delay before power on which isn't nice if you use a big toroid)
To power a 2nd ATX from the 1st ATX, all you need is a relay powered from the first to short green wire to ground.
I meant multiple, not different. Doh!
Maybe...
What I'm really looking for it to eliminate as much noise as possible, since my computer and sound system are right next to each other, and as I'm planning to DIY a case for my next PC, I told myself, why not DIY a PSU also.
I thought about a complete linear PSU, but linear regulators aren't really efficient... (And would make a 300W heater)
But, what's noisy in SMPSes, the 400Hz transformer, the switching regulator or all these answers? If I take a toroid and switching regulator, I probably put less junk on the power line, but what about RFI / noise / interference? Is the regulator making some itself or can just putting big caps on the power line fix things?
Umm... multiple mice on separate treadmills? I feel sure they would be fine as long as you could coordinate turn on.
Noise from the power supply is only one source from a computer. It does not necessarily upset the operation of the computer. Even so, adding your own chokes and caps at the outputs is one way to achieve clean rails.
I guess you could also mount your computer far from anything critical, and consider power conditioning (even perhaps to stop noise getting back into the power mains?)
Noise from the power supply is only one source from a computer. It does not necessarily upset the operation of the computer. Even so, adding your own chokes and caps at the outputs is one way to achieve clean rails.
I guess you could also mount your computer far from anything critical, and consider power conditioning (even perhaps to stop noise getting back into the power mains?)
It seems someone finally woke up to this with PFC. Maybe it was the power suppliers, or maybe it's just hard to live in this world if you are an appliance that is susceptible to the noise, and your neighbours are noisy. 
Just a thought, would anyone consider running the ground connection (that the computer's case is connected to), outside to a dedicated ground stake?
Just a thought, would anyone consider running the ground connection (that the computer's case is connected to), outside to a dedicated ground stake?
By reading an Antec Phantom review I think it's to improve the efficiency percentage that needs to be better or equal to 80% in the ATX 2.2 specs.
I know it's only one of them, but even a powered off ATX computer is too much for AM radio. ;-) (Just powering +5VSB is making lots of noise. I need to unplug it)
So if I filter the rails and use a toroid transformer(Avel Lindberg states 93% efficiency -- at full load) switching regulators are OK? The IC isn't making lots of RFI?
Just thinking at it, if I want clean power I also have to invest in ferrite clamps.
And the screen and everything? Since I have the soundcard out to my amp, I would say no (ground loops probably). But at the same time, I don't have any grounded audio equipment - Double insulated. Would it even change something at all?
DragonMaster said:
This can only be good.
I see. Probably not. So your problem is either airborne, or power line borne. Since your psu is in a grounded iron case, it has electrical and magnetic containment. Maybe this is not complete but it must live up to FCC/equivalent regs. It seems the larger problem must be the power line.
I cant help but think the ferrite beads plus a clean psu (careful PFC etc.) must be most of the issue.
Would you consider plugging that AM radio into an extension cord and riding it around the room to confirm which of these two issues this the larger?
Not my 30lbs harman/kardon...
http://dmsaudiocorp.deep-ice.com/system/sys2.html
With the sound card and amp unplugged and a 7.8mH power filter, it was almost worse : The noise turned in a whine.
BTW, it's a -very- cheapo generic PSU with the transformer starting to whine in the whole room. I guess I should try my old Fortron in there(I have one in an other unused computer, but the fan bearing is almost dead)?
About the Fortron, maybe I'll choose something else wih -active- PFC rather than passive.
IMO a shunt capacitor is a better filter that a series choke, but the choke easily works at the higher frequencies. Both may be ideal. (perhaps the choke was somehow containing the noise or reacting with strays or reradiating it?)
Anyway, I have found a huge difference between cheap and expensive power supplies. Fan noise (which is my pet hate, I use series resistors in winter), case quality etc. An expensive psu will have Elna or Rubycon electros for example, and a whole range of niceties that a cheapskate would consider unnecessary. Basically, evidence that the designers cared about what was giong on.
Buying an overrated supply would probably reduce radiation due to the components being run without overload (I wouldn't pay heed to the ratings on cheap psu's). Running a second psu would also 'lighten the load'.
Soundcard choice may also have an effect. Power supply rejection ratio's and such. You may also consider physically shielding your soundcard.
Anyway, I have found a huge difference between cheap and expensive power supplies. Fan noise (which is my pet hate, I use series resistors in winter), case quality etc. An expensive psu will have Elna or Rubycon electros for example, and a whole range of niceties that a cheapskate would consider unnecessary. Basically, evidence that the designers cared about what was giong on.
Buying an overrated supply would probably reduce radiation due to the components being run without overload (I wouldn't pay heed to the ratings on cheap psu's). Running a second psu would also 'lighten the load'.
Soundcard choice may also have an effect. Power supply rejection ratio's and such. You may also consider physically shielding your soundcard.
If a 300A ATX PSU is running on one fan (not mainly used by the own supply, but used to move air through the computer case and cool drives, cards and the own motherboard) and having well over 80% efficiency, how much noisy fans do yo expect to require for a 300W linear PSU with below 50% efficiency? Well, you can also use a 50cm by 50cm by 4cm heatsink plate like in serious class A amplifiers
BTW: An active PFC PSU always produces more EMI than a non-PFC or a passive-PFC PSU. PFC stages are inherently noisy unlike the push-pull converter that produces the low output voltages that is inherently quiet.
BTW: An active PFC PSU always produces more EMI than a non-PFC or a passive-PFC PSU. PFC stages are inherently noisy unlike the push-pull converter that produces the low output voltages that is inherently quiet.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.