Question Is it bad to use a power supply that is very over-powered for the system it is powering?

[UltronXavier]

https://pcpartpicker.com/list/2jwZ4q

This is my setup. I'm just curious whether it is bad for any of the components to use a 1000w PSU to power a system that draws around 400w with peripherals?

 

[PC Tailor]

Welcome to the forums my friend!

No it is not, it's perfectly fine. The components will only draw what they need from the PSU, all you have effectively is a PSU that's barely touching what it CAN supply. The PSU regulates how much power is distributed to the components, and doesn't just push all of the power it has. (Some might argue poor quality power supplies this is questionable!).

Ultimately the RMi is also an excellent quality PSU, so it will be more efficient too.

 

[UltronXavier]

Thanks for the welcome!
I'm glad to hear that! When I was buying parts I was simply looking for something 80+ Gold certified and found a deal on the RMi1000 which was cheaper than all the other 500-600w PSUs I'd been considering. I figured why not, it gives me a lot of options for upgrades in the future👍

iCue tells me the efficiency and I've not once seen it drop below 80% (not sure how if calculates that, but still...)

 

[PC Tailor]

It's an excellent unit with great efficiency, excellent ripple suppression and load regulation.
Solid unit all round.

7/10 warranty on the RMi I believe depending on when bought.

 

[BFG-9000]

PSUs are most efficient around 50% load so @ 400w is fine. Unfortunately the 80plus certification does not measure below 20% load (unless it's 80+ Titanium. Which your Gold PSU is not) so it may well be quite inefficient down there, and your system will certainly idle well under 200w.

So it's probably fine but may be less efficient than you'd expect.

 

[PC Tailor]

Actually the RMi is pretty efficient at low loads. (If I'm understand the correct RMi model). Even below 200W is showed around 80-90% in most tests, especially being as it had long hold up times. But that being said, many older PSUs do struggle to maintain that efficiency at lower loads, thus why so many go through low load efficiency tests aftermarket.

But not as efficient as the previous RMi versions, which is a little disappointing in some respects. probably falls below the 80% threshold at around 50W.

 

[Deleted member 2731765]

No, it is not bad to use a higher wattage PSU. The system will still use/pull only the maximum wattage required, regardless of whether you have 500 or 850 Watts PSU.

Wattage number is not always important. The quality and the make matters. The main concern is the "quality" of the power, the quality of the components used/CAPS, as well as the total AMP drawn on the +12V RAIL (output), the efficiency under load, "ripple suppression", among other factors.

The total wattage number of any PSU is not always really the most important deciding factor, primary concern is the 'quality' of power it produces, and the total capacity of the 12V source etc.

And no, the extra power won't go wasted, it just won't get fully utilized.

Let me explain------------> If suppose you have a 500W PSU installed on any RIG, then that does not mean that it will always draw the full wattage/500 W.

Nope....... The amount of power drawn is determined by the number of components on that rig/computer, and how much they actually require. The amount of power drawn will only be equal to what is required, and not more.

Suppose, if all the components of a rig require 400 W to run, then the "load" on the PSU will be 400 W, and hence the power draw of that particular PC will always be 400 W (depending on the efficiency obviously), regardless of whether a 500 Watt or a 1000 W PSU is installed. PSU efficiency is a different matter though, as how well the PSU converts the AC power it receives from the outlet, to DC.

Any electricity which is not converted from AC to DC, is given off as heat. A PLATINUM certified PSU might help you save a little on your electricity bill, over a year, though this depends on many other factors as well.

 

[UltronXavier]

Thanks for all the responses! I figured my rig would only draw the power it needed, but I wasn't sure if the power supply would somehow provide too much power and cook the components... Thanks all for clearing it up for me

 

[digitalgriffin]

https://pcpartpicker.com/list/2jwZ4q

This is my setup. I'm just curious whether it is bad for any of the components to use a 1000w PSU to power a system that draws around 400w with peripherals?

Just throwing money down the drain. That said, I like my power supplies a little over-spec'd. Buy a quality one for @$125, use it for 10 years, instead of Mediocre $60 every 3->5 years.

For example, if my PC is spec'd for 450 Watts, I divide that number by .64 (20% loss for capacitor aging, and 20% below max capacity = .8 * .8 = .64). That leaves me with about a ~700 Watt PSU.

I bought a Platinum Seasonic 850Watt with a 10 year warranty for $115. My last seasonic rebadge (Corsair AX850 Gold) is still running good 7 years later. (A little noisier on the +12V/+5V/+3.3V voltage lines, but good just the same.) Until recently I pulled up to 550 watts on a regular basis. I'm pulling closer to 400 now with graphics cards power improvements.

 

[jonnyguru]

PSUs are most efficient around 50% load so @ 400w is fine. Unfortunately the 80plus certification does not measure below 20% load (unless it's 80+ Titanium. Which your Gold PSU is not) so it may well be quite inefficient down there, and your system will certainly idle well under 200w.

So it's probably fine but may be less efficient than you'd expect.

The myth that PSUs are typically most efficient at 80% load is one that coincides with the older double forward topologies. But the newer switching technologies like LLC resonant mode actually allow for a pretty flat efficiency "curve" because the switching frequency can change with the load and when you add to that a PWM controller that supports "burst" mode, you can get good efficiency all the way down to 2% load. (http://en.kediman.com/attaches/2017/04/916-wi0e4n.pdf )

Don't bother with 80 PLUS's website anymore. If you want to know the efficiency of a PSU across MANY loads, you really should be looking at Cybenetics' website instead:

https://cybenetics.com/

 

[UltronXavier]

Just throwing money down the drain. That said, I like my power supplies a little over-spec'd. Buy a quality one for @$125, use it for 10 years, instead of Mediocre $60 every 3->5 years.

For example, if my PC is spec'd for 450 Watts, I divide that number by .64 (20% loss for capacitor aging, and 20% below max capacity = .8 * .8 = .64). That leaves me with about a ~700 Watt PSU.

I bought a Platinum Seasonic 850Watt with a 10 year warranty for $115. My last seasonic rebadge (Corsair AX850 Gold) is still running good 7 years later. (A little noisier on the +12V/+5V/+3.3V voltage lines, but good just the same.) Until recently I pulled up to 550 watts on a regular basis. I'm pulling closer to 400 now with graphics cards power improvements.

It may have been throwing money down the drain if it wasn't the same price as the 550w RMi at the time of purchase. I thought I may as well buy more than I needed and a high quality PSU so it gives me upgrade headroom and will last a long time.
What do you use to track your power draw? Is there a program that gives real-time readings?

 

[refillable]

https://www.techpowerup.com/review/corsair-rm1000i/6.html

"The RM1000i performed very well at light loads;"

Even at 10% the load (100W) its efficiency is almost 89%.

What do you use to track your power draw? Is there a program that gives real-time readings?

You need a wattmeter, and to analyze the quality of the output the power supply makes, you need an oscilloscope and a variable DC load.

Also, try Corsair Link. It might show something useful. Never used it before, though.

 

[digitalgriffin]

It may have been throwing money down the drain if it wasn't the same price as the 550w RMi at the time of purchase. I thought I may as well buy more than I needed and a high quality PSU so it gives me upgrade headroom and will last a long time.
What do you use to track your power draw? Is there a program that gives real-time readings?

My Asus board has a built in monitor and recorder for voltage lines. You can look for sag or overshoot when you get near the limits. Staying within 5% is ideal. But for noise, I have a scope for a detailed look.

 

[jonnyguru]

My Asus board has a built in monitor and recorder for voltage lines. You can look for sag or overshoot when you get near the limits. Staying within 5% is ideal.

It's getting those numbers from an IC on ther motherboard and it's only looking at voltages and not actual power consumption. He asked what people used to determine power DRAW.

That monitoring software is about as useful as the random number generators used in power supply calculators.

He has an RM1000i which has an MCU inside the PSU that captures that data. If the PSU doesn't have an internal device for measuring voltages, current, etc. then you're not getting actual readings from thh PSU.

If you don't have a "digital" PSU with this capability, the best way to measure your power consumption is with a Kill-A-Watt or Watt's-Up (which is better because it has logging capability).

 

[digitalgriffin]

It's getting those numbers from an IC on ther motherboard and it's only looking at voltages and not actual power consumption. He asked what people used to determine power DRAW.

That monitoring software is about as useful as the random number generators used in power supply calculators.

He has an RM1000i which has an MCU inside the PSU that captures that data. If the PSU doesn't have an internal device for measuring voltages, current, etc. then you're not getting actual readings from thh PSU.

If you don't have a "digital" PSU with this capability, the best way to measure your power consumption is with a Kill-A-Watt or Watt's-Up (which is better because it has logging capability).

My bad. I missed the intent of the question. I thought he was asking about the quality of power in reference to my previous post.

But yes a Kill-A-Watt is the easiest way. And using his RM1000i internal sensors would also work. Otherwise an open loop current sensor, or closed loop sensor is another way to measure power draw on individual rails. I have each for the scope along with the voltage probes.