PC Power Draw?

[eccerr0r]

Re: PC Power Draw?

You were lucky to have an APFC PSU and you can compare results. Once you throw in a non PFC PSU you can't. Can't even compare with itself really, you can get a number but it's not meaningful, like if you got 3 when banging on the GPU and 1.5 when you're not, you can't necessarily say that it's double power in the first situation (it could be be more like 2.5x depending on the PF).

Of course this will be moot once you get the Kill-A-Watt which you can do the same as your PM157 and more, including comparing non PFC PSUs.

[Topcat]

Re: PC Power Draw?

It's a diagnostic equipment, not a way to compute how much you have to pay for electricity...

That;s not really what I was curious about.... Just curious as to what kinds of loads were being pulled from line under different kinds of tasks.

[eccerr0r]

Re: PC Power Draw?

Nonlinear loads predate SMPSs, so there have always been loads that tricked the PM157.
It's a diagnostic equipment, not a way to compute how much you have to pay for electricity...

[Topcat]

Re: PC Power Draw?

That is an APFC power supply according to Supermicro with a power factor of 0.99
So it should measure the same on the old and new meter.
But a good way to make sure the new (and old) meter works as intended is described in this old post:
https://www.badcaps.net/forum/showpo...postcount=2444

Here is the test report and specs of that PSU from what Google provided:
https://cdn.badcaps-static.com/pdfs/...7185a619d9.pdf
https://cdn.badcaps-static.com/pdfs/...87410c7421.pdf

Thanks, good info. Just for grins, I did order a P3 "Kill-A-Watt" model 4400, which I'll test against the old PM57 power monitor....which should read about the same....back when the PM57 was made, SMPS's weren't the standard; not sure if they had even been invented yet. I'm guessing that meter is circa ~1960's, maybe 1970's. I've had it since the early 90's, came from a defunct shop I bought out.

To everyone participating in this thread, good refresher!! I vaguely remember this from my first year in electronics....

[eccerr0r]

Re: PC Power Draw?

Ugh. I can understand running that 24/7 when it's cold in the winter and you need heating. But in the summer when it's hot, I keep nothing turned On when I'm not using it.

That's sort of the problem with PVRs and webservers, can't exactly tell when it needs to be on when it's getting "requests" off the air or online... so they need to be up waiting for them (And even if it's not actively recording, it still needs to pick up EPG data.)

The other thing that pisses me off is that one of the DVBs I have, it goes bonkers when suspended. Driver problem probably, else suspend/resume would have been a good solution, alas foiled...

[Per Hansson]

Re: PC Power Draw?

FWIW, this PSU is a Supermicro PWS-865-PQ 865w unit. I should have put that in my original post; it seems pertinent information.

That is an APFC power supply according to Supermicro with a power factor of 0.99
So it should measure the same on the old and new meter.
But a good way to make sure the new (and old) meter works as intended is described in this old post:
https://www.badcaps.net/forum/showpo...postcount=2444

Here is the test report and specs of that PSU from what Google provided:
https://cdn.badcaps-static.com/pdfs/...7185a619d9.pdf
https://cdn.badcaps-static.com/pdfs/...87410c7421.pdf

[Topcat]

Re: PC Power Draw?

^
That brought back some great memories!

[kaboom]

Re: PC Power Draw?

Hey, I love that show. Always nice looking women on it, and a ferrari model that I've actually driven (308GTB)!!

And they got over on poor Magnum every time!

Mix n match:

“Remember I told you I didn't like getting egg on my face? Well, right now, I'm wearing a whole omelet.”

Have you seen:
http://magnum-mania.com/Lists/Classic_Scenes.html
https://quotecatalog.com/communicator/thomas-magnum/

Work the lock, don't look at the dogs.
And you owe me $50!
"Recycled how?" "Through a cow."

[Topcat]

Re: PC Power Draw?

TU...

OT, but I shouldn't shorten your handle to "TC" too much, otherwise you could be mistaken for a helicopter tour guide who gives "free rides," LOL.

"Magnum, not again, dammit! Phooey!"

Hey, I love that show. Always nice looking women on it, and a ferrari model that I've actually driven (308GTB)!!

[kaboom]

Re: PC Power Draw?

Always nice to see you.... Very informative thread as well.

TU...

OT, but I shouldn't shorten your handle to "TC" too much, otherwise you could be mistaken for a helicopter tour guide who gives "free rides," LOL.

"Magnum, not again, dammit! Phooey!"

[Topcat]

Re: PC Power Draw?

Been "reading the mail" sporadically, but finally dropped in to post.

With any luck, more to come...

Always nice to see you.... Very informative thread as well.

[kaboom]

Re: PC Power Draw?

(snip)Eventually, it lost the WiFi, which I suspect is due to the nVidia NB loosing it again. Oh well (snip)

This made me LOL... It's practically meme-worthy considering NV's "reputation," both past and current.

For this reason, even the Kill-A-Watt is not super-accurate when it comes to non-APFC SMPS... but it's better than a simple ammeter, though.

As the KAW needs to remain accurate for low current levels too, and given there's a fixed number of bits for its ADC, does it run out of bits and "clip" or truncate above a given current level? If they use the ATMega's (or whatever uC) ADC, then maybe so, especially considering they'd make the ADC's "granularity" finer for more accuracy at lower current levels & given the fixed (low?) number of bits available. Favoring low-moderate current loads that operate over long periods of time vs short-time, high current loads.

So the KAW's ammeter may (does?) lose resolution as current goes much over 20A. KAW flashes the display and beeps over ~1800W@120V. Remember, to be accurate for non-APFC cap charging current, it should be accurate to something like 30A. But it probably truncates far below this, almost like trying to drive beyond 0dB in a WAV/MP3 file; anything over the ADC limit is truncated to the limit, just like anything over 0dB limits, "digitally clips," to 0dB- and sounds horrible.

You funny!

But am I reactive?

Anyways, good to see you back here! Been wondering where/how you've been.

Just killing some spare time...
Been "reading the mail" sporadically, but finally dropped in to post.

With any luck, more to come...

[momaka]

Re: PC Power Draw?

Yes, I understand the concept. I think perhaps my question was rather how would a 'modern' say "Kill-A-Watt" read this versus the trusty old Sencore?

Not quite sure... but I think that Sencore (or just any AC ammeter) likely just measure the voltage drop over a current shut and then amplifies the magnitude of that to give a reading on the scale.

Whereas with the Kill-A-Watt, it measures not only that but also the AC voltages before and after the current shunt, and probably some other parameters about the voltage waveform. From these, it's likely what enables it to calculate the real power.

Uhggg....I could have gone the rest of the day without you mentioning that. I barely even remember the class....let alone the subject matter. It's been way too long.

Oh, don't worry, I'm rusty AF too when it comes to those.
For one of my uni classes, though - industrial motors and machinery - we did have to do a ton of W, VA, and VARS calculations... so just the very basics of trig knowledge I still retain from that (and it was probably one of the most fun labs/classes I took.)

Which is kind of a rare bird, as most opted for the nvidia "performance" model vs the "boring" intel one. The irony, as you already know, is that the supposedly low-performance intel ones are still going. The former NV "high performance," or "gamer" models are now no performance, lol...

Yep, lol.
In fact, I also have a Compaq V6000, which is nearly the same thing as this laptop, but with AMD CPU and nVidia chipset... and in the case of that one, it was a no performance machine, indeed. But with a reflow, I was able to get it working just to mess around with it. (The thread is here somewhere on BCN.) Eventually, it lost the WiFi, which I suspect is due to the nVidia NB loosing it again. Oh well At least I'll have a spare keyboard and display for the DV6000. I'm already using the adapter and HDD caddy from it.

There's literally another layer to consider. Those caps are behind a rectifier; you've now got a nonlinear load. Simply put, those caps cannot recharge on each half-cycle until the instantaneous line voltage (less rectifier drop) exceeds the voltage across the caps. As this occurs over a few degrees, the currents are large, with the caps "topping off" over a brief period.
...
So while there are inductive & capacitive loads, a non APFC SMPS is neither. It's a non linear load. The former two simply shift or displace the sinewave of their current draw; the latter distorts it.

Yeah, that too.
For this reason, even the Kill-A-Watt is not super-accurate when it comes to non-APFC SMPS... but it's better than a simple ammeter, though.

My friend "j" said the same thing!

You funny!

Anyways, good to see you back here! Been wondering where/how you've been.

[kaboom]

Re: PC Power Draw?

Same here.
Currently sitting in front of a HP Pavilion DV6000 laptop (the C2D ones with Intel chipset, luckily, and not nVidia shitset)

Which is kind of a rare bird, as most opted for the nvidia "performance" model vs the "boring" intel one. The irony, as you already know, is that the supposedly low-performance intel ones are still going. The former NV "high performance," or "gamer" models are now no performance, lol...

Essentially, real power is how much power the device really is using (and what we measure in Watts)... or more precisely, the power draw of the circuit associated with any resistive parts.
But when inductors and capacitors get involved, their impedance becomes a function of the frequency. So any inductive and capacitative components in the circuit will make the current draw from the line to be a bit "funky" and not exactly following the shape of the voltage wave. This "funky" component, in math terms, has an imaginary number value. And when it comes to power, there is also a corresponding imaginary component, called reactive power. It is measured in Volt-Amps (so it's like Watts, but not quite ) In simpler terms, the reactive power is the amount of power the circuit "borrows" from the line temporarily but doesn't end up using it, and then returns it back.
Finally, there is apparent power (also measured in VA.) Perhaps now would be a good time to brush up on your Trigonometry here. Apparent power is essentially the combined component of real and reactive power. Using vectors, the real component is for a horizontal vector and the reactive component is for a vertical vector. If these two vectors are put together as the two sides of a right-angle triangle, then the apparent power is the hypotenuse (the combined vector of these two vectors.) And this is why the apparent power can appear as a much higher value that the real power. The power factor, PF, is the ratio of real power to apparent power. Knowing any of these two values (PF, real power, apparent power, or reactive power) will allow you to determine the other two.

My friend "j" said the same thing!


Also:

In the case of a non-PFC (or PPFC) PSU, the input bulk caps make the input circuit appear quite capacitative to the AC line. Thus, the apparent power will read a higher value than the real power due to non-zero reactive power component. So if you try to use a simple current meter on the AC line to measure current on a device that is not purely resistive, you will see higher power draw than what is actually being drawn.

There's literally another layer to consider. Those caps are behind a rectifier; you've now got a nonlinear load. Simply put, those caps cannot recharge on each half-cycle until the instantaneous line voltage (less rectifier drop) exceeds the voltage across the caps. As this occurs over a few degrees, the currents are large, with the caps "topping off" over a brief period. You can see this if you scope the line with the appropriate divider probe. The "flat topping" (more of a slope) of each half cycle is thousands of cap-input rectifiers.

PPFC is only somewhat useful, since it attempts to "stretch" the recharge over more degrees of each half cycle vs its contemporary. But APFC is best in that it makes the current draw co-incide with the voltage waveform. IOW, input current just about follows input voltage; the very definition of an easy load.

So while there are inductive & capacitive loads, a non APFC SMPS is neither. It's a non linear load. The former two simply shift or displace the sinewave of their current draw; the latter distorts it.


Yes, TC. If you connect a typical 35uF motor run cap directly across 240VAC, you'll have 3.2A or 768VA flowing (or "bouncing" to/from the POCO transformer), but the only power consumed will be the I2R losses in the wiring and a minimal loss in the cap. More to the point: the same I2R losses occur in the POCO transformers, which, of course, is on their side of the meter. Now consider what APFC does, vs a few degrees of huge currents for non PFC SMPS topping off. Also reduces losses for the POCO. And keeps triplens from adding on the neutral in 3ph Y secondary dist systems.

DO NOT take this to mean "makes the SMPS more efficient."
It does not! As previously "discussed" on bcn, the APFC front end reduces overall SMPS efficiency- a power supply feeding a power supply. It's thru control schemes & better MOSFETs & sync rectification where efficiency is gained.

It makes distribution more efficient because of lower transformer/conductor/gen losses, and makes distribution more effective since there's less overhead of the (k/M)VARS; more conductor ampacity is available for delivering billable power.

Because resistive losses are I squared R.

If the Sencore simply has an ammeter whose scale also read "watts," you're likely missing the mark with modern equipment. Actually, (amps@120V)=watts persisted for quite some time. I have an early 90's fan whose nameplate claims .8A @ 120V, supposedly 96 "watts." No, 96VA, maybe 60W on the KAW.

Because with AC, V * A =/= watts.
Again, see motor runcap example above- hardly "new math," which was kind of a stupid comment.
Try an oil cap with the Sencore. Bet you get a "watt" reading even though it's just VA...
If it's an actual dual coil wattmeter (like a mechanical kWH meter), it should read zero with just a runcap.

[Topcat]

Re: PC Power Draw?

Nope, no new math. Real power vs. apparent power theory is probably about as old as AC power lines. Kaboom pretty much called it all out in Post #2 above.

Yes, I understand the concept. I think perhaps my question was rather how would a 'modern' say "Kill-A-Watt" read this versus the trusty old Sencore? I also have a 90's era Sencore PR570 Variac/isolation transformer; which also measures this kind of thing (Amps/Watts), but its limited to 4A, so I don't usually use it for heavy loads (like testing a SLI system shown above)...I never even questioned it when using it to read current & wattage pulls for whatever it is I was working on....

Perhaps now would be a good time to brush up on your Trigonometry here.

Uhggg....I could have gone the rest of the day without you mentioning that. I barely even remember the class....let alone the subject matter. It's been way too long.

Kill-A-Watt meter.
And, uhm, I got mine from Harbor Freight. The company that makes the Kill-A-Watt meter (P3 International) isn't specific to Harbor Freight, though, so you can get these from any other place. Only reason I got mine from HF is because they used to send me 25% off coupons a while back (so the meter cost me a little under $20 with taxes instead of $25.

I bought one for $22. We'll find out how off it is versus the old Sencore PM57. Probably a handy thing to have around anyway....but it would be interesting to see, just for giggles.

[momaka]

Re: PC Power Draw?

I have too many PCs.

Same here.
Currently sitting in front of a HP Pavilion DV6000 laptop (the C2D ones with Intel chipset, luckily, and not nVidia shitset) that averages about 20-25W at the wall (real power, that is), measured with a Kill-A-Watt meter. I don't remember what its Power Factor (PF) was or if the adapter had APFC (it probably does, as many original OEM adapters from that era and onwards do), but the Kill-A-Watt can differentiate between real power (Watts) and reactive power (Volt-Amps.) So it's rare that I get to fool it.

Otherwise, my daily-use desktop at home is that old Pentium 4 Optiplex 170L dinosaur. IIRC, it draws somewhere around 70 or 80 Watts idle (I can't even recall when was the last time I measured it) and peaks at about 120-130 Watts under full CPU and (onboard) GPU load. As for the CRT monitor (21" Sony) attached to that desktop - it varies based on what I'm viewing on the screen. Dark images/backgrounds make it use a lot less power (about 60-ish Watts, IIRC), and light/bright images/background make it use closer to 100-ish Watts. So total for that whole setup is about 120-ish Watts idle and could go as high as 200 Watts peak.

one of which is my PVR that I run 24/7 and it's probably like 80W or so idle, and I am suspecting near 130W or so if I'm submitting batch jobs to it.

Ugh. I can understand running that 24/7 when it's cold in the winter and you need heating. But in the summer when it's hot, I keep nothing turned On when I'm not using it.

Also, recently I fixed up a mini SFF desktop with a Pentium G3260 and onboard everything. This thing sits at 18 Watts in idle. Watching Youtube in 1080p doesn't make it go past 25 Watts. I can make it peak up to 70 Watts with synthetic benchmarks, or about 50-60W with games that push both the CPU and the IGP to the max... but that's about it. Despite being an SFF desktop, it actually has a normal mATX-sized mobo in it with normal desktop RAM. According to HP, the mobo can take up to a mid-range i5-3xxx CPU (I think rated around 70W TDP.)

Is this like the 'new math'? Raw amperage pulled from line, when did this change?

Nope, no new math. Real power vs. apparent power theory is probably about as old as AC power lines. Kaboom pretty much called it all out in Post #2 above.

Essentially, real power is how much power the device really is using (and what we measure in Watts)... or more precisely, the power draw of the circuit associated with any resistive parts.
But when inductors and capacitors get involved, their impedance becomes a function of the frequency. So any inductive and capacitative components in the circuit will make the current draw from the line to be a bit "funky" and not exactly following the shape of the voltage wave. This "funky" component, in math terms, has an imaginary number value. And when it comes to power, there is also a corresponding imaginary component, called reactive power. It is measured in Volt-Amps (so it's like Watts, but not quite ) In simpler terms, the reactive power is the amount of power the circuit "borrows" from the line temporarily but doesn't end up using it, and then returns it back.
Finally, there is apparent power (also measured in VA.) Perhaps now would be a good time to brush up on your Trigonometry here. Apparent power is essentially the combined component of real and reactive power. Using vectors, the real component is for a horizontal vector and the reactive component is for a vertical vector. If these two vectors are put together as the two sides of a right-angle triangle, then the apparent power is the hypotenuse (the combined vector of these two vectors.) And this is why the apparent power can appear as a much higher value that the real power. The power factor, PF, is the ratio of real power to apparent power. Knowing any of these two values (PF, real power, apparent power, or reactive power) will allow you to determine the other two.

In the case of a non-PFC (or PPFC) PSU, the input bulk caps make the input circuit appear quite capacitative to the AC line. Thus, the apparent power will read a higher value than the real power due to non-zero reactive power component. So if you try to use a simple current meter on the AC line to measure current on a device that is not purely resistive, you will see higher power draw than what is actually being drawn.

But really, the main reason why all of this matters is that reactive power, in big amounts, can be bad news for the power lines, as it still puts a toll on the power lines when the circuit is "borrowing" extra power (that it is then not using and returning back.)

Unless someone has something better than a clamp meter or inline from harbor freight, feel free to share it.

Kill-A-Watt meter.
And, uhm, I got mine from Harbor Freight. The company that makes the Kill-A-Watt meter (P3 International) isn't specific to Harbor Freight, though, so you can get these from any other place. Only reason I got mine from HF is because they used to send me 25% off coupons a while back (so the meter cost me a little under $20 with taxes instead of $25.

[eccerr0r]

Re: PC Power Draw?

BTW might be worth to archive this with PM157 data if you don't have the owners manual around:

https://cdn.badcaps-static.com/pdfs/...22b67bf8a9.pdf

Seems the BAMA doesn't have the manual, which is unfortunate, this pdf might be the only manufacturer data available...

[Topcat]

Re: PC Power Draw?

Ok, just warning you based on the photographs, there doesn't seem to be a way for it to discern the difference between a PF=1 and PF << 1 target device, and you'll get misleading numbers from the latter. Just mentioning it may be "obsolete" because there are devices out there that can tell the difference.

I suspect you probably use a APFC PSU so the numbers shouldn't be that far off from real power consumption. Just that a non PFC PSU will give "wattage" numbers larger than real power consumption - if you have an APFC and a non PFC PSU giving the same numbers, the APFC unit is drawing more money from your wallet at the power company.

(And as an addendum, this is not a "bad" thing. It's just that you'd actually be using that power, versus wasting it or unable to utilize it in the case of the non-PFC PSU!)

It's always nice learning something on my own forum. Thanks for the information. FWIW, this PSU is a Supermicro PWS-865-PQ 865w unit. I should have put that in my original post; it seems pertinent information.

only real way of knowing what matters is watching the meter that clocks up the units you have to pay for .

That is indeed true...

[petehall347]

Re: PC Power Draw?

only real way of knowing what matters is watching the meter that clocks up the units you have to pay for .

[eccerr0r]

Re: PC Power Draw?

Ok, just warning you based on the photographs, there doesn't seem to be a way for it to discern the difference between a PF=1 and PF << 1 target device, and you'll get misleading numbers from the latter. Just mentioning it may be "obsolete" because there are devices out there that can tell the difference.

I suspect you probably use a APFC PSU so the numbers shouldn't be that far off from real power consumption. Just that a non PFC PSU will give "wattage" numbers larger than real power consumption - if you have an APFC and a non PFC PSU giving the same numbers, the APFC unit is drawing more money from your wallet at the power company.

(And as an addendum, this is not a "bad" thing. It's just that you'd actually be using that power, versus wasting it or unable to utilize it in the case of the non-PFC PSU!)