Power Supply Dell 330W AC-DC constantly drawing ~121W minimum?

[zovc]

Hey!

This is likely a non-issue for a lot of people, especially when compared to running a full desktop, but I noticed this today.

Before switching over to the 330W Dell AC-DC unit (bought on Amazon last week, ~$115) my system was hooked up with a HDPLEX AC-DC 160W going to a HDPLEX DC-ATX 160W. The system (when on) was drawing between ~20W and ~30W, as it's an Atom-based system and it didn't have a GPU or any other PCIe device installed.

Upon switching to the Dell 330W (still using the HDPLEX DC-ATX 160W), I noticed the system (when on) is drawing ~121W, fluctuating mostly between 121.1W and 121.3W... is this the lowest amount of power the Dell 330W can draw from the wall?

 

[jeshikat]

How are you measuring?

Because other than the fans, computers do no mechanical work. So all the energy consumed is converted to heat and either your brick would be very toasty or your motherboard would probably be fried.

 

[zovc]

My Kill-a-watt plug meter. It's going into a 'power strip', the Dell AC-DC power brick is going into the Kill-a-watt, and the Dell brick's plug is going into the connector it fits into that came with the 160W HDPLEX.

I did notice that since I've made this switch, my Supermicro board is blinking on LED8 which indicates "Overheat/PWR Fail/Fan Fail LED" or, specifically when it's blinking means "Blinking: PWR Fail or Fan Fail..."

Now, I had connected a Noctua PWM fan before noticing the light so I thought that might be the problem. But with the fan connected and everything shut off and restart, we're still getting that blinking. No part of the equation is getting hot, though. Not the heatsink on the atom chip, not the HDPLEX, and not the RAM or the drives I have in the system.

 

[zovc]

Okay I'm a dingus. Apparently, this motherboard just doesn't like Noctua PWM fans?

First red herring, the ~121 "Watts" was Voltage. I didn't consider that the power had flickered while we were getting some really bad storms Wednesday, Volts is the first mode the Kill-a-watt defaults to. This system is drawing its normal ~20W.

The indicator LED is definitely only because of the Noctua fan I'm connecting (NF-A9 PWM), which is strange, but immediately the light comes back (when I'm in Windows 10?) when I connect the fan. Whether I connect it while the system is on or off, regardless of which header I use.

 

[jeshikat]

That's strange with the fan, but not as strange as defying the laws of conservation of energy

 

[zovc]

That's strange with the fan, but not as strange as defying the laws of conservation of energy

A lot of GOOD science gets done in my lab.

 

[malcky]

Hey!

This is likely a non-issue for a lot of people, especially when compared to running a full desktop, but I noticed this today.

Before switching over to the 330W Dell AC-DC unit (bought on Amazon last week, ~$115) my system was hooked up with a HDPLEX AC-DC 160W going to a HDPLEX DC-ATX 160W. The system (when on) was drawing between ~20W and ~30W, as it's an Atom-based system and it didn't have a GPU or any other PCIe device installed.

Upon switching to the Dell 330W (still using the HDPLEX DC-ATX 160W), I noticed the system (when on) is drawing ~121W, fluctuating mostly between 121.1W and 121.3W... is this the lowest amount of power the Dell 330W can draw from the wall?

Hi there, I'm just wondering why did you buy the 330w adapter? Looks like your system is a low powered set up....unless I'm missing something? Was the original 160w HDPlex AC-DC broken or something?

Thanks

 

[zovc]

Hi there, I'm just wondering why did you buy the 330w adapter? Looks like your system is a low powered set up....unless I'm missing something? Was the original 160w HDPlex AC-DC broken or something?

Thanks

A system I'm working on building right now is probably going to draw around the 330W under load. I'm hoping I can manage to use an internal AC-DC brick, but it may not end up fitting into my case without a lot more modding than I'm already planning.

So, the brick wasn't for this system.

 

[stree]

I was led to believe that underworking a PSU was just as bad for efficiency as overloading, so not a good long term match.

 

[zovc]

I was led to believe that underworking a PSU was just as bad for efficiency as overloading, so not a good long term match.

That's not necessarily true, each PSU has its own efficiency curve. Nevertheless, this isn't a permanent pairing.

 

[malcky]

I was led to believe that underworking a PSU was just as bad for efficiency as overloading, so not a good long term match.

That's not necessarily true, each PSU has its own efficiency curve. Nevertheless, this isn't a permanent pairing.

Surely a PSU is supposed to work on a large power range/curve, from idle range of say 20w to a heavier load range of say 100w, so in this scenario would a 230w PSU be classed as under worked and be causing problems for the PSU?

I am in the process of putting together a small form computer and reckon it will be pulling between 20w at idle and upto 100w max....this is without a graphics card for now, which is why I bought a HP 230w psu as I will one day add that too perhaps.

 

[jeshikat]

I was led to believe that underworking a PSU was just as bad for efficiency as overloading, so not a good long term match.

I see that idea often, but for similar efficiency curves it really doesn't matter much.

80-Plus Platinum rated PSU:

• 90% efficient at 20% load
• 92% efficient at 50% load

So say you had two Platinum-rated power supplies, a 1000W and a 400W.

Running the numbers at a 200W DC load (thus the 1000W is running at 90% efficiency and the 400W at 92% efficiency):

• 217W AC draw for the 400W PSU
• 222W AC for the 1000W PSU

So it's just a 5W difference. You could definitely argue that it's silly to pay more for a PSU that will be less efficiency, but it's only slightly less in this example.

Of course there are other comparisons that would be more substantial, but you'd have to crunch the numbers for that specific situation.