The conversation on this page has been archived and is no longer active.

Thirty-five watts

Message #1 - Posted 2009/03/08 - Michael H. Phillips

I've been playing with a plug power-meter and found the following for my G5:

On and idle: 108W
Sleep mode: 43W
Shut down (but plugged in): 34W

Both wake options in Energy Saver are turned off.

How can the G5 be using 34W when shut down?

I've done a simple calibration of the metre against bulbs of known wattage and it seems, if anything, to be under-reading.

Michael

mhphillips at gmail dot com

Message #2 - Posted 2009/03/08 - Don Bruder

Previously, Michael H. Phillips wrote:

I've been playing with a plug power-meter and found the following for my G5:

On and idle: 108W
Sleep mode: 43W
Shut down (but plugged in): 34W

Both wake options in Energy Saver are turned off.

How can the G5 be using 34W when shut down?

I've done a simple calibration of the metre against bulbs of known wattage and it seems, if anything, to be under-reading.

Unless it's unplugged, there's juice running through the transformer at all times. The power supply has to do its thing to let you push the power key and get the machine to turn on, the clock gets fed, the power management unit draws its bit, USB devices that are bus-powered are slurping up whatever they need (despite the computer being powered down) etc, etc, etc...

You're just seeing the "phantom draw" that's incredibly common these days - TV sets, TiVos, stereo gear, fax machines, cordless phones - There's a *HUGE* number of things in common use today that draw power at all times. Even wall-warts that aren't connected to whatever they're supposed to power draw some. Add 'em all up, and there's a pretty amazing amount of power being burned in the average home even when everything is "turned off".

Don Bruder - dakidd@sonic.net - If your "From:" address isn't on my whitelist, or the subject of the message doesn't contain the exact text "PopperAndShadow" somewhere, any message sent to this address will go in the garbage without my ever knowing it arrived. Sorry... <http://www.sonic.net/~dakidd> for more info

Message #3 - Posted 2009/03/08 - Salmon Egg

Previously, Don Bruder wrote:

Previously, Michael H. Phillips wrote:

I've been playing with a plug power-meter and found the following for my G5:

On and idle: 108W
Sleep mode: 43W
Shut down (but plugged in): 34W

Both wake options in Energy Saver are turned off.

How can the G5 be using 34W when shut down?

I've done a simple calibration of the metre against bulbs of known wattage and it seems, if anything, to be under-reading.

Unless it's unplugged, there's juice running through the transformer at all times. The power supply has to do its thing to let you push the power key and get the machine to turn on, the clock gets fed, the power management unit draws its bit, USB devices that are bus-powered are slurping up whatever they need (despite the computer being powered down) etc, etc, etc...

You're just seeing the "phantom draw" that's incredibly common these days - TV sets, TiVos, stereo gear, fax machines, cordless phones - There's a *HUGE* number of things in common use today that draw power at all times. Even wall-warts that aren't connected to whatever they're supposed to power draw some. Add 'em all up, and there's a pretty amazing amount of power being burned in the average home even when everything is "turned off".

Without knowing the details, it is difficult to understand what is happening. Of the top of my head, 35W for an off computer souynds high.

I would not trust the watt meter unless I knew it read true power irrespective of waveform and power factor. It is very easy to design digital equipment including voltmeters, ammeters, and wattmeters that do not take that into account. Modern power switching supplies can produce current pulses and harmonics that can be read as consuming power when they do not.

Old fashioned instrumentation without electronics was pretty good at measuring true power and rms values. In particular, watt=hour meters were could at rejecting reactive power and harmonic power measurement for truly sine-wave supply voltage.

Once you convince yourself that your power meter can handle such situations, then you can move on to other considerations.

Bill

Private Profit; Public Poop! Avoid collateral windfall!

Message #4 - Posted 2009/03/09 - Michael H. Phillips

On Sun, 8 Mar 2009 14:38:19 +0000, Don Bruder wrote:

Unless it's unplugged, there's juice running through the transformer at all times. The power supply has to do its thing to let you push the power key and get the machine to turn on, the clock gets fed, the power management unit draws its bit, USB devices that are bus-powered are slurping up whatever they need (despite the computer being powered down) etc, etc, etc...

With all cables (USB, Ethernet, Firewire) disconnected from the back of the G5, it's still using 34W when shut down. It seems very high.

Michael

mhphillips at gmail dot com

Message #5 - Posted 2009/03/09 - Paul Russell

Michael H. Phillips wrote:

On Sun, 8 Mar 2009 14:38:19 +0000, Don Bruder wrote:

Unless it's unplugged, there's juice running through the transformer at all times. The power supply has to do its thing to let you push the power key and get the machine to turn on, the clock gets fed, the power management unit draws its bit, USB devices that are bus-powered are slurping up whatever they need (despite the computer being powered down) etc, etc, etc...

With all cables (USB, Ethernet, Firewire) disconnected from the back of the G5, it's still using 34W when shut down. It seems very high.

It may well just be a bogus reading from your power meter, due to the load being reactive and/or non-linear. One method you can try to see whether it really is taking 35W is the following: get an extension board with at least two sockets; put the power meter in line with the extension board; plug a moderate well-behaved load, such as a 100W lamp, into one socket of the extension board and note the reading (1); plug the G5 into a second socket in the extension board and note the new reading (2); then finally subtract (2) from (1) to get the power draw of the G5.

Paul

Message #6 - Posted 2009/03/10 - David Empson

Don Bruder wrote:

Previously, Michael H. Phillips wrote:

I've been playing with a plug power-meter and found the following for my G5:

On and idle: 108W
Sleep mode: 43W
Shut down (but plugged in): 34W

Both wake options in Energy Saver are turned off.

How can the G5 be using 34W when shut down?

I've done a simple calibration of the metre against bulbs of known wattage and it seems, if anything, to be under-reading.

Unless it's unplugged, there's juice running through the transformer at all times. The power supply has to do its thing to let you push the power key and get the machine to turn on, the clock gets fed, the power management unit draws its bit, USB devices that are bus-powered are slurping up whatever they need (despite the computer being powered down) etc, etc, etc...

You're just seeing the "phantom draw" that's incredibly common these days - TV sets, TiVos, stereo gear, fax machines, cordless phones - There's a *HUGE* number of things in common use today that draw power at all times. Even wall-warts that aren't connected to whatever they're supposed to power draw some. Add 'em all up, and there's a pretty amazing amount of power being burned in the average home even when everything is "turned off".

I've done power measurements on a few Mac models I have access to, with a reasonably accurate power meter.

Two late model PowerMac G4s drew about 6W while shut down, and about 13W while sleeping. An iMac Core 2 Duo drew less than 1W while shut down and less than 2W while sleeping.

I haven't measured a PowerMac G5 yet, but a friend has one and is going to give me some measurements when he gets a chance (using the same meter). We expect it to be similar to my figures for the PowerMac G4, or lower.

35W is certainly not a normal amount of power consumption for any Mac while shut down. Either the meter is giving a spurious reading, or the computer is faulty.

As Salmon Egg noted, a cheap meter might produce a meaningless reading due to power factor issues - if it is designed to measure resistive loads only (e.g. light bulbs), then it might not be able to measure a reactive load (with capacitance and/or inductance) with any accuracy.

I'd try another (better) meter before assuming there was anything wrong with the computer.

David Empson
dempson@actrix.gen.nz

Message #7 - Posted 2009/03/10 - Michael H. Phillips

On Mon, 9 Mar 2009 11:35:25 +0000, Paul Russell wrote:

It may well just be a bogus reading from your power meter, due to the load being reactive and/or non-linear. One method you can try to see whether it really is taking 35W is the following: get an extension board with at least two sockets; put the power meter in line with the extension board; plug a moderate well-behaved load, such as a 100W lamp, into one socket of the extension board and note the reading (1); plug the G5 into a second socket in the extension board and note the new reading (2); then finally subtract (2) from (1) to get the power draw of the G5.

The subtraction gives me 34W - no change.

This is probably of no relevance, but I measured a low-end Dell running Windows 98 and got 2W for sleep mode and 0W when shut down.

Michael

mhphillips at gmail dot com

Message #8 - Posted 2009/03/10 - Paul Russell

Michael H. Phillips wrote:

On Mon, 9 Mar 2009 11:35:25 +0000, Paul Russell wrote:

It may well just be a bogus reading from your power meter, due to the load being reactive and/or non-linear. One method you can try to see whether it really is taking 35W is the following: get an extension board with at least two sockets; put the power meter in line with the extension board; plug a moderate well-behaved load, such as a 100W lamp, into one socket of the extension board and note the reading (1); plug the G5 into a second socket in the extension board and note the new reading (2); then finally subtract (2) from (1) to get the power draw of the G5.

The subtraction gives me 34W - no change.

OK - that's a useful data point. So it sounds like the G5 really is taking 35W. I'd expect the power supply area to be noticeably warm in that case. I have an old G5 downstairs that I can check - I'll see whether it's warm and if I can find my power meter I'll measure it too.

Paul

Message #9 - Posted 2009/03/10 - Paul Russell

Paul Russell wrote:

Michael H. Phillips wrote:

On Mon, 9 Mar 2009 11:35:25 +0000, Paul Russell wrote:

It may well just be a bogus reading from your power meter, due to the load being reactive and/or non-linear. One method you can try to see whether it really is taking 35W is the following: get an extension board with at least two sockets; put the power meter in line with the extension board; plug a moderate well-behaved load, such as a 100W lamp, into one socket of the extension board and note the reading (1); plug the G5 into a second socket in the extension board and note the new reading (2); then finally subtract (2) from (1) to get the power draw of the G5.

The subtraction gives me 34W - no change.

OK - that's a useful data point. So it sounds like the G5 really is taking 35W. I'd expect the power supply area to be noticeably warm in that case. I have an old G5 downstairs that I can check - I'll see whether it's warm and if I can find my power meter I'll measure it too.

OK - curiosity got the better of me so I tried it on an old G5 (single CPU, 1.6 or 1.8 GHz). On its own with my cheap power meter it reads around 24W when shut down or sleeping and around 150W when running. However with the light bulb trick it seems it's taking pretty much 0W when shut down, so for me at least it looks like just the power meter's inability to deal with the unusual load. Also I couldn't feel any warmth at all from the PSU area before I powered it up (it had been plugged in and shut down for a long time).

So I can only guess you have a faulty PSU ?

Paul

Message #10 - Posted 2009/03/10 - Salmon Egg

Previously, Michael H. Phillips wrote:

On Mon, 9 Mar 2009 11:35:25 +0000, Paul Russell wrote:

It may well just be a bogus reading from your power meter, due to the load being reactive and/or non-linear. One method you can try to see whether it really is taking 35W is the following: get an extension board with at least two sockets; put the power meter in line with the extension board; plug a moderate well-behaved load, such as a 100W lamp, into one socket of the extension board and note the reading (1); plug the G5 into a second socket in the extension board and note the new reading (2); then finally subtract (2) from (1) to get the power draw of the G5.

The subtraction gives me 34W - no change.

This is probably of no relevance, but I measured a low-end Dell running Windows 98 and got 2W for sleep mode and 0W when shut down.

While a variation on the way things are usually done, I see no logical reason to believe that the measurement will improve in accuracy. I was thinking that you could put a diode in series with a lamp to test the meter, but the resistance of the lamp would be lower than the wattage indicates.

Bill

Private Profit; Public Poop! Avoid collateral windfall!

Message #11 - Posted 2009/03/10 - Salmon Egg

Previously, Paul Russell wrote:

Michael H. Phillips wrote:

On Mon, 9 Mar 2009 11:35:25 +0000, Paul Russell wrote:

It may well just be a bogus reading from your power meter, due to the load being reactive and/or non-linear. One method you can try to see whether it really is taking 35W is the following: get an extension board with at least two sockets; put the power meter in line with the extension board; plug a moderate well-behaved load, such as a 100W lamp, into one socket of the extension board and note the reading (1); plug the G5 into a second socket in the extension board and note the new reading (2); then finally subtract (2) from (1) to get the power draw of the G5.

The subtraction gives me 34W - no change.

OK - that's a useful data point. So it sounds like the G5 really is taking 35W. I'd expect the power supply area to be noticeably warm in that case. I have an old G5 downstairs that I can check - I'll see whether it's warm and if I can find my power meter I'll measure it too.

Paul

If you have an oscilloscope available, you can make a current transformer also called a Rogowski coil to look at the current waveform will isolating the oscilloscope from line voltage. It will not measure the dc component of the current. Make sure that the secondary load is not opened. That is, know what you are doing.

Look for current spikes that might come from charging a capacitor in the power supply from a rectifier. Also see if you can get true rms voltmeters and ammeters and/or a wattmeter that actually multiplies instant current by instant voltage and displays an average overf several cycles.

Bill

Private Profit; Public Poop! Avoid collateral windfall!

Message #12 - Posted 2009/03/11 - Michael H. Phillips

On Tue, 10 Mar 2009 11:14:42 +0000, Paul Russell wrote:

OK - that's a useful data point. So it sounds like the G5 really is taking 35W. I'd expect the power supply area to be noticeably warm in that case. I have an old G5 downstairs that I can check - I'll see whether it's warm and if I can find my power meter I'll measure it too.

Having left the G5 Shut Down overnight, there is a slight warmth around the back. So it looks like I do have a problem.
Thanks all for the advice and testing.

Michael

mhphillips at gmail dot com

Message #13 - Posted 2009/03/11 - Paul Russell

Michael H. Phillips wrote:

On Tue, 10 Mar 2009 11:14:42 +0000, Paul Russell wrote:

OK - that's a useful data point. So it sounds like the G5 really is taking 35W. I'd expect the power supply area to be noticeably warm in that case. I have an old G5 downstairs that I can check - I'll see whether it's warm and if I can find my power meter I'll measure it too.

Having left the G5 Shut Down overnight, there is a slight warmth around the back. So it looks like I do have a problem.
Thanks all for the advice and testing.

You're welcome - with any luck you might be able to pick up a replacement PSU reasonably cheaply (maybe keep an eye on eBay for broken G5s that you can scavenge a PSU from ?).

Paul

Need Help? Have a Question?

Looking for more help, comments, and answers?

Ask your questions on Ask Different. Ask Different is a community of Apple users ready to help.