|
As this forum is rarely used anymore, we've locked it. Feel free to browse and read. Questions? Please reach out to us directly. Cheers! |
|
 |
|
|
| Author |
Message |
stefuel
Joined: 07 Mar 2006
Posts: 3353
Location: Green Harbor MA USA
|
| Posted: Fri Jun 28, 2013 10:41 am Post subject: |
|
|
Well, I'll be away from this project for probably a week. Starting today I'm working every day and on call every night.
I won't have much if any time until 7/5/13. That's the end of my Hell week and the beginning of the closest bartenders hell week......or two 
_________________
Chip
A Barco is only a AmPro with training wheels
Card carrying member of the AVS chain gang.
|
|
| Back to top |
|
 |
CasetheCorvetteman
Joined: 09 Nov 2008
Posts: 6326
Location: Australia
|
| Posted: Fri Jun 28, 2013 12:46 pm Post subject: |
|
|
Don't over do it mate 
|
|
| Back to top |
|
|
SisterOfMercy
Joined: 31 Oct 2007
Posts: 155
Location: Zwart Nazareth, The Netherlands
|
| Posted: Sat Jun 29, 2013 11:36 pm Post subject: |
|
|
| Jeremy112 wrote: | | Now, some might think the biggest con: Holy power hogging! - Not so. Almost any server in existence only uses ONE of the multiple hot swap power supplies that are installed in its chassis. The rest are there PURELY for redundancy, they DO NOT engage unless one fails or is pulled out while the machine is on. And even then if theres more than 2, only one more will power on in place of the removed/failed one. |
How would you know? The IPMI implementation I have seen on the one HP server I have dealt with is crap, and doesn't give you much information.
| Jeremy112 wrote: | | The biggest con of any server: They are LOUD, they cannot be quieted down, they are designed specifically like this to ensure they stay running at peak performance. I have read and heard of people modifying their machines to make them quieter, but in the end this is very harmful for any server due to their extremely packed together design. So if you intend to use this machine, either make a dedicated spot for it that it can be out of the way and not heard, or put up with the sound. |
They can be a bit more silent than is expected of them. My SM 836BA case has PWM fans that are all actively controlled. Compared to most servers it's quite silent.
Extremely packed together design?? what kind of servers do you buy?? 
| Jeremy112 wrote: | | Stefuel, I have to agree with most of the posts above regarding the age of your machine. It is fairly similar to the DL380 G4 I have here, which is a great machine, but really unless you are using it for processing and storage only, it is rather antiquated, and the performance of the machine would probably less than what you would want from it. However, since you are using it only for experimenting, it is a VERY good learning system. I learned how to work on servers for the first time with a pallet of Dell PE 2600 servers, even more outdated, but they taught me a lot of invaluable information for future work with servers. |
Well, those HP machines are quite sh*tty.
And in my opinion servers are for everyone. Just a small silent one near the internet connection which synchronises your backups with the data of someone else.
_________________
The kissing and the colour come crashing down
|
|
| Back to top |
|
|
stefuel
Joined: 07 Mar 2006
Posts: 3353
Location: Green Harbor MA USA
|
| Posted: Sun Jun 30, 2013 11:14 am Post subject: |
|
|
The PE 2800 is a power hog no doubt. When you first power it on, it's drawing more the 3.5 amps. After boot is complete, it settles to 2.5 amps. This is with only 2 SCSI drives inserted.
My desktop (tower) PC starts at 1.75 and settles to 1.5. As I'm only looking for a "file server" I don't need a turbocharged PC. Right now I am experimenting using a Dell Precision 390 with 4 HDDs installed to be used as a server. I have installed WD green HDDs. It boots at 1.5 but settles down to .7. If left alone long enough the vid card shuts down dropping the unit to .5 amps. If left longer, it goes into standby and drops to less then .1 amp but still fully booted up.
I have not had time to see what happens when there is a external call for info stored on it while in standby. I may have to send a WOL to bring it back to life when serving is required.
_________________
Chip
A Barco is only a AmPro with training wheels
Card carrying member of the AVS chain gang.
|
|
| Back to top |
|
|
CasetheCorvetteman
Joined: 09 Nov 2008
Posts: 6326
Location: Australia
|
| Posted: Sun Jun 30, 2013 12:50 pm Post subject: |
|
|
Current draw it not always a good indicator of power consumption, unless youre being charged by the kVA hour and not the kW hour, or youre running for your own power supply and not from mains supply.
Measuring current flow in an AC circuit is all well and good, but you need to know the power factor in order to calculate true power consumption. This is not like DC where power in watts is equal to current multiplied by the voltage, with AC that calculation will get you VA and not watts. To convert VA into watts, you will need to multiply by the power factor.
VA=VxI
P=(VxI)xPF
Assuming voltage of 110v AC nominal and putting in a mythical power factor value cause chances are you have nothing there to accurately measure it:
110vx3.5A=385VA
If power factor is 0.8:
385VAx0.8PF=308 watts.
All that said, without measuring power factor, you have no way to know how much power in watts is being consumed by either machine and the only benefit to knowing apparent power ( VA ) is for sizing a UPS to suit the required duty.
|
|
| Back to top |
|
|
SisterOfMercy
Joined: 31 Oct 2007
Posts: 155
Location: Zwart Nazareth, The Netherlands
|
| Posted: Sun Jun 30, 2013 8:21 pm Post subject: |
|
|
And the power factor has also to do with the reactive power, and cos (phi), it's all the same. 
With the server board/case I currently use I can see the power going in the supplies, and the power coming out of the supplies. Losses are in the PSU fan and of course the conversion.
If you know the price per kWh, you can easily calculate that a high efficiency PSU will pay for itself very soon!
_________________
The kissing and the colour come crashing down
|
|
| Back to top |
|
|
stefuel
Joined: 07 Mar 2006
Posts: 3353
Location: Green Harbor MA USA
|
| Posted: Sun Jun 30, 2013 10:04 pm Post subject: |
|
|
If you are a electrician I am using a tool you know well. I'm using a clamp over type amp meter with a 6 amp scale.
I made a test cord that allows me to clamp over one line.
_________________
Chip
A Barco is only a AmPro with training wheels
Card carrying member of the AVS chain gang.
|
|
| Back to top |
|
|
stefuel
Joined: 07 Mar 2006
Posts: 3353
Location: Green Harbor MA USA
|
| Posted: Sun Jun 30, 2013 10:19 pm Post subject: |
|
|
I just looked up the specs for both machines. The precision 390 has one 375 watt power supply. The Poweredge 2800 has two 930 watt supplies (one being redundant) so way more then double and it doesn't go into standby. You will have a hard time convincing me that I won't see significant energy savings with the 390 over the 2800
_________________
Chip
A Barco is only a AmPro with training wheels
Card carrying member of the AVS chain gang.
|
|
| Back to top |
|
|
CasetheCorvetteman
Joined: 09 Nov 2008
Posts: 6326
Location: Australia
|
| Posted: Mon Jul 01, 2013 6:04 am Post subject: |
|
|
| stefuel wrote: | If you are a electrician I am using a tool you know well. I'm using a clamp over type amp meter with a 6 amp scale.
I made a test cord that allows me to clamp over one line. |
Yep, i know that tool very well indeed
But it still only measures current flow, which is not all you need to see measure power consumption.
Another thing you are not understanding here is power supply ratings mean sweet FA to actual consumption, that rating is the maximum load capacity of the power supply, not the input power and certainly not the consumption.
If you take a 930w power supply and a 375w power supply and connect exactly the same load to both, lets say 200w load, both will supply 200w, no more, and no less. The efficiency of the power supply will dictate which one draws more power on the line side. Im not saying it doesnt use more power here, cause it probably does, but what i am saying is your method of measurement is of very little benefit and somewhat inaccurate.
Without seeing ( and understanding ) a spec sheet or schematic diagram on both supplies, we have very little chance of determining if either supply has active power factor correction, or even any power factor correction at all.
The ONLY way you can effectively measure power consumption is with a watt-hour meter, which will tell you true power ( watts ), apparent power ( volt-amps ), and reactive power ( volt-amps reactive ).
So while you may or may not think im going to have a hard time convincing you of anything, there is a hell of alot more to the measurement of AC power consumption than just measuring current flow, and whether youre convinced or not is irrelevant, cause its a fact.
Here is something for you to test that will prove the theory to you, if you have extra low voltage lighting in your home and have a spare iron core transformer there, connect it through your clamp meter with no light bulb and see how much current flows, then test again with a halogen lamp connected. If you could do the same with a watt-hour meter, youd instantly see where it all ties in. Abit of a quick calculation using P=VxI formula will show you that the numbers dont add up to what you expect to see.
|
|
| Back to top |
|
|
stefuel
Joined: 07 Mar 2006
Posts: 3353
Location: Green Harbor MA USA
|
| Posted: Mon Jul 01, 2013 10:31 pm Post subject: |
|
|
A computer power supply is not a iron core transformer. I know both use power with no "draw" on them So lets agree to dis-agree (line from Men in Black 3) For the sake of arguement my voltage available at the plug is 120.0 vac and I know a device plugged into that 120.0 vac is drawing .5 amps. The formula for a simple amps to watts conversion is
W=A X V
Dell Precision 390 120.0 X .5 = 60 watts
Home brewed tower 120.0 X 1.5 = 180 watts
Dell PE 2800 120.0 X 2.5 = 300 watts
If I have a hundred bucks in my pocket and every hour I have to give someone 50 cents, how long will my money last?
Use the same example but fork over a dollar and fifty cents a a hour and then re-calculate for two dollars and fifty cents
a hour
Obviously this crude example does not take into account the rate but the theory is correct. So with this can you actually tell me that using the 390 won't be more energy efficient then the 2800?
If you still insist that I'm wrong, then let me be wrong and say "I told you so". I'm happy. 
_________________
Chip
A Barco is only a AmPro with training wheels
Card carrying member of the AVS chain gang.
|
|
| Back to top |
|
|
stefuel
Joined: 07 Mar 2006
Posts: 3353
Location: Green Harbor MA USA
|
| Posted: Tue Jul 02, 2013 12:24 am Post subject: |
|
|
So then I converted my watt hours to Kwatt hrs and best I can figure based on my most recent bill where I'm paying
18 and a half cents per KWH that it costs me $23.98 per month to leave this PC running full time. When I switch to the precision 390 the cost will drop to $7.99 per month. If I stuck with my original plan to use a commercial server like this
PowerEdge 2800 the cost to leave it running would be $39.96.
_________________
Chip
A Barco is only a AmPro with training wheels
Card carrying member of the AVS chain gang.
|
|
| Back to top |
|
|
CasetheCorvetteman
Joined: 09 Nov 2008
Posts: 6326
Location: Australia
|
| Posted: Tue Jul 02, 2013 2:45 am Post subject: |
|
|
What qualifications are you basing this theory on? Cause you still are not listening. Youre still going about it wrong, and the theory is NOT correct.
A power supply, regardless of design, will affect power factor. Power factor affects current flow. Current flow DOES NOT dictate true power consumption in AC circuits. I dont know how many times i will have to repeat it for you.
A simple power calculation of P=VxI is only of any use in the cause of straight resistive loads, which a switch mode power supply is not. The formula is NOT that simple, agreed or not, youre still incorrect and your measurements mean very little in the real world.
The example given is not only crude, its incorrect. For one youre only mesuring current flow in an AC circuit, youre taking the voltage value at a nominal level instead of actual, and youre not calculating anything you can relate to a cost per unit.
|
|
| Back to top |
|
|
Tom.W
Joined: 09 Mar 2006
Posts: 6635
|
| Posted: Tue Jul 02, 2013 4:28 am Post subject: |
|
|
How about getting back on the topic ?
Chip what WD drives did you get and have you considered a Dune ? I have no need for a server I just plug in the external drives I want to watch and power them up individually if needed...
Power issue solved
These look nice.....
http://www.duneplayer.com/base-3d.html
|
|
| Back to top |
|
|
gjaky
Joined: 05 Jun 2010
Posts: 2802
Location: Budapest, Hungary
|
| Posted: Tue Jul 02, 2013 6:11 am Post subject: |
|
|
| stefuel wrote: | A computer power supply is not a iron core transformer. I know both use power with no "draw" on them So lets agree to dis-agree (line from Men in Black 3) For the sake of arguement my voltage available at the plug is 120.0 vac and I know a device plugged into that 120.0 vac is drawing .5 amps. The formula for a simple amps to watts conversion is
W=A X V
Dell Precision 390 120.0 X .5 = 60 watts
Home brewed tower 120.0 X 1.5 = 180 watts
Dell PE 2800 120.0 X 2.5 = 300 watts
If I have a hundred bucks in my pocket and every hour I have to give someone 50 cents, how long will my money last?
Use the same example but fork over a dollar and fifty cents a a hour and then re-calculate for two dollars and fifty cents
a hour
Obviously this crude example does not take into account the rate but the theory is correct. So with this can you actually tell me that using the 390 won't be more energy efficient then the 2800?
If you still insist that I'm wrong, then let me be wrong and say "I told you so". I'm happy. |
The best way to tell how much power does your equipment draws to measure the RMS value of current and voltage and their angle. Most multimeters are only estimating AC voltage from the measured voltage's peak value.
Now the power factor in general is a thing that worth a look. The main effect of the reactive power is voltage drop accross the power line, usually the consumer power meters don't measure ractive power meter at all, they measure its effect at second hand.
So if you have a 120VAC nominal line voltage, if a heavy reactive load present in the circuit your line voltage may fall, let's say by 10V, so if you have an incadescent bulb it will draw less power (less amps) because of the lower voltage (this is true for all resistive loads), but what if you have a switching power supply in that same circuit? Let's assume the switching PSU eats 120W(=120VACx1A) but with the lower line voltage it still wants to eat the 120W so it needs to draw more current (110VACx1.1A=120W) more current -> more voltage drop, you see? And not only that equipment affected that has bad power factor.
_________________
projectors in the past : NEC 6-9PG xtra, Electrohome Marquee 6-7500, NEC XG 1351 LC ( with super modified Electrohome VNB neckboard !!!)
current: VDC Marquee 9500LC
The MOD: VNB-DB, VIM-DB
|
|
| Back to top |
|
|
CasetheCorvetteman
Joined: 09 Nov 2008
Posts: 6326
Location: Australia
|
| Posted: Tue Jul 02, 2013 7:29 am Post subject: |
|
|
| gjaky wrote: |
The best way to tell how much power does your equipment draws to measure the RMS value of current and voltage and their angle. Most multimeters are only estimating AC voltage from the measured voltage's peak value.
Now the power factor in general is a thing that worth a look. The main effect of the reactive power is voltage drop accross the power line, usually the consumer power meters don't measure ractive power meter at all, they measure its effect at second hand.
So if you have a 120VAC nominal line voltage, if a heavy reactive load present in the circuit your line voltage may fall, let's say by 10V, so if you have an incadescent bulb it will draw less power (less amps) because of the lower voltage (this is true for all resistive loads), but what if you have a switching power supply in that same circuit? Let's assume the switching PSU eats 120W(=120VACx1A) but with the lower line voltage it still wants to eat the 120W so it needs to draw more current (110VACx1.1A=120W) more current -> more voltage drop, you see? And not only that equipment affected that has bad power factor. |
Very much the point im trying to make here. The switch mode power supply is filled with inductors and capacitors, both of which have an effect on phase angles, which in turn affects current flow in an AC circuit. The principal behind it is the same regardless of the components. Inductive loads cause lagging power factor, and capacitive loads cause a leading power factor, and can also cause out of control voltages if the lead angle gets too large. Both will cause increased current flow due to reactive power.
Without the correct equipment to measure ALL aspects of the equation, calculation is inaccurate, and as near as pointless as it can get.
|
|
| Back to top |
|
|
CasetheCorvetteman
Joined: 09 Nov 2008
Posts: 6326
Location: Australia
|
| Posted: Tue Jul 02, 2013 8:23 am Post subject: |
|
|
| Tom.W wrote: | How about getting back on the topic ?
|
We are on topic, this is an off topic thread anyway!!
|
|
| Back to top |
|
|
stefuel
Joined: 07 Mar 2006
Posts: 3353
Location: Green Harbor MA USA
|
| Posted: Tue Jul 02, 2013 9:58 am Post subject: |
|
|
Are we having fun yet
Lets throw away all the amps, watts and kilowatts for a second and look at the two devices at both ends of the scale.
The PowerEdge 2800 server runs 24/7. It has like 10 cooling fans that run full time. It sounds like a jet engine spooling up for take-off. It remains in that state 24/7 and you can't slow it down. It's designed that way for reliability. The fans in this beast are thermostatically controlled to "ramp up" when needed. It holds up to 10 SCSI drives. I have three 15,000 rpm drives installed and when called on, waste no time in causing the fans to "ramp up".
The Precision 390 work station has 3 cooling fans. When not in use for a few minutes the video card shuts down. A few minutes later of non use it goes into "away mode" Note: not sure if it has to complete any processing before "away mode" is activated but, in that mode just about everything is shut down including cooling fans BUT it is still able to stream or record to and from the hard drives. This is where I expect to see the most energy savings as I estimate the 390 will be in "away mode" probably 90% of the time.
Tom,
I am using Western Digital Green drives. It is my understanding that those drives park and shut down when not in use.
Another energy saving feature
_________________
Chip
A Barco is only a AmPro with training wheels
Card carrying member of the AVS chain gang.
|
|
| Back to top |
|
|
CasetheCorvetteman
Joined: 09 Nov 2008
Posts: 6326
Location: Australia
|
| Posted: Tue Jul 02, 2013 11:48 am Post subject: |
|
|
We are having a ball
But youre skirting the issue, i asked you on what qualifications you base your theory on AC power consumption measurement.
Never at any point did i even suggest the old server will use less power than this new machine. Quote me stating that if you can find it, but you keep coming back with that with each reply so at some point we seem to have gotten confused with that. What i did say and am still saying is your theory on power consumption measurement is flawed and youre not grasping the concept behind AC power circuits with regard to inductive and capacitive loads and the effects they create, which you seem to be passing off as non existant and using methods useful only for resistive loads or pure ripple-free DC loads, neither of which apply in this case.
Any hard disk can be set to go to sleep or standby mode at any time, after a given amount of time, or when a specific power saving mode is set. Windows controls that, not Western Digital, and not the specific model of hard disk. Caviar Green drives save power by using a lower performance drive motor, and are spun at 5,400RPM, making them as slow as a wet week. That is the only way they save energy. Any other theory you may have on it is false, nothing is special about them. And they are slow as f***, ive got two here, what you save in instant power consumption you lose again in the amount of extra time taken to complete a task.
No video card shuts down. Operating system controls that, and its easy to set in Windows. Again you gained nothing that any other Windows user hasnt had the option to use for 12 years. Did you know you can also slow down the CPU and disable cores when they arent needed? You can save even more power yet!!
Youre doing nothing there on that machine that cant be done on any decent PC running Windows. The later the Windows the easier it is to set those features up. Its as simple as ticking boxes and selecting options from drop down menus.
But that doesnt matter, cause its not which machine uses more power for longer that we are talking about, we are talking about your methods of power consumption measurement and your qualifications and experience to back up your theory on AC power characteristics.
|
|
| Back to top |
|
|
CIR Engineering
Joined: 25 Aug 2008
Posts: 4269
Location: Chicago USA & Berlin Germany
|
| Posted: Tue Jul 02, 2013 3:03 pm Post subject: |
|
|
You can look up the power usage on most computer components from their spec sheet or by contacting the manufacturer. For example, a WD green drive uses 6 watts while spinning up, 2 watts while idle spinning, and 0.5 watts in spun down mode. A typical home grade motherboard uses about 40 watts peak. My M1015 controller uses 10 watts...
A 10k RPM hard drive uses about 27 watts spun up and a 15k RPM drive even more.
I don't care what equations we want to discuss, but the idea that a commercial server is going to be an economical choice over a low power home PC is kind of silly.
craigr
_________________
JETI 1501-HiRes 2nm Spectroradiometer
JETI 1211 Spectroradiometer
Photo Research PR-650 Spectroradiometer
Klein K10-A Colorimeter
Murideo Fresco SIX-G HDMI 2.x Multimedia Generator
Murideo Fresco SIX-A HDMI 2.x Analyzer
Light Illusion ColourSpace XPT Color Calibration Software
Light Illusion LightSpace XPT Pro Version 10.x Color Calibration Software
OMARDRIS JVC Software Patch to use K10-A and Jeti with JVC OEM AutoCal Software!
Sencore CR7000 CRT Tube Analyzer / Rejuvenater
Authorized Dealer for Lumagen & just about everything worth buying
www.CIR-Engineering.com - craigr@cir-engineering.com
Phone: 865-405-6892
|
|
| Back to top |
|
|
ecrabb
Forum Moderator
Joined: 13 Mar 2006
Posts: 15909
Location: Utah
TV/Projector: JVC RS40, Epson 5010
|
| Posted: Tue Jul 02, 2013 3:08 pm Post subject: |
|
|
| CIR Engineering wrote: | | I don't care what equations we want to discuss, but the idea that a commercial server is going to be an economical choice over a low power home PC is kind of silly. |
Exactly. Power factor correction aside, and the fact that no PC actually runs at the stated PSU power consumption rating aside, the commercial rack server is overkill for a media server. The real point is that older commercial rack servers often offer little, if any usable benefit for a home media server. The downsides will probably outweigh any of the positives. And yes, the desktop PC will very, very likely consume less, or much less, power than the server, regardless of how you configure the video card, hard drives, and CPU.
SC
|
|
| Back to top |
|
|
|
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum
|
Forum powered by phpBB © phpBB Group
|
|
