Domain: doubleed.com
Stories and comments across the archive that link to doubleed.com.
Comments · 8
-
I use a "Watt's Up"
1800 watt (~15 amp) maximum; LCD display; ampere, watt & power-factor modes; built-in tracking totals; possible PC integration: http://www.doubleed.com/
-
Re:cheaply measure a single device
I love my kill-a-watt but I've been thinking of picking up a Watt's Up? for the datalogging capabilities. But the price is silly, I should just build one.
Anyway, a clamp-on ammeter should be in your toolkit. (Get a DC-capable model and watch motherboard/peripheral power draw inside your PC!) Instead of slicing open an extension cord, consider an AC line splitter to make your measurements with. The 10-winding side makes small measurements more accurate, and it looks more professional if you end up using it on the job. -
Re:I will do one better!
I've never seen one personally, but this seems to have what you're looking for.
-
Note From Autor:Wow, that was fast. I only posted the page on Sunday evening...
To address questions asked here on Slashdot. The main purpose of this project was to build a wearable debian box. We had a 2 square inch VGA LCD heads up display that we were using. Also we had toyed around with Festival to read output to the user. The application did not require any user input, so no keyboard was nessasary. If someone wanted user input one could use bluetooth and a Linux PDA to run ssh or VNC. Also I have seen some IR remote controls that have a full keyboard (Do a Google search on CarPC's and remote controls), so user input is not out of the question.
Also several people were talking negativly about the battery life. My run time test was conducted under heavy load so the Mac Mini was pulling around 14-20W at the time. The device I used for measuring the power was a Watts Up Pro which was specificly designed for measuring the power draw on consumer electronics, so the readings are accurate. Bear in mind that these batteries are preproduction prototypes that were sent to me for evaluation, and SKC Films informed me after I recieved them that they would not perform optimally do to a mistake in that batch. They also offered to make me another batch, which I declined as my application only needed about 45min runtime and the ones I already had would do fine for proof of concept. The point here is that with further work the battery life can be inproved. Simply using CPUFreq in the Linux Kernel will help strech out the battery life a good deal. To address concerns someone brought up about charging the Li-ion batteries, A power supply with a current limiting knob _is_ a safe method of charging the batteries. The chargers that are designed for charging Li-ion batteries, say for instance in cell phones, do just that. I have recharged my battery pack just fine using a Topward power supply set to 20VDC and limited the current to about 50mA and the batteries didn't even get warm let alone explode in my face
;0).
Granted my application was very specific, but this could be used for lots of things. Slashdot has already mentioned quite a few. How about this one:
A portable compute brick. In a lab setting one might need to take a part of a Mosics cluster from one lab bench to another to collect and process data. Having a built in UPS on the Mac Mini with a Wifi network interface allows you to move the compute node physicly without having to first remove the node from the cluster and migrate all of it's processes off the node. Please note that Wireless comunications with the Mac Mini need to use a USB or Firewire Wifi card because the Mac airport card uses the broadcom chipset, which Linux users have learned to hate with a passion. Before someone mentions NDISWrappers I would like to state the obvious Mac is non-X86 and the binary drivers that are used with NDIS are compiled for X86.
"Why do this?" was also asked. My response...
I am a Geek, and it was fun!
Cheers,
Silas Bennett
P.S. My uname should read Fuzzy_The_Quantum_Duck but Slashdot didn't like the last 2 characters... -
measurements
I've been trying to minimize my power usage for some time and have measured several different systems using a watts up meter. (Unfortunately filtered through my memory)
- Sony Vaio Picturebook C1VN ~12 watts
- Desktop without monitor using fanless VIA C3 ~25 watts
- DWI 7550 SBC w/ standard HD ~15 watts
I think the newer eden boards are a little bit better than the C3 in a ordinary mother board.
I think my G4 powerbook averages about ~15 watts as well. (Charging is closer to 30 watts.)
The important note is that the laptops include the LCD monitor whereas I was running the desktops headless.
Also to cut down on energy lossage with either the small desktop or the laptop try to get a DC to DC power supply. From what I've read an inverter will sap another 10 watts.
-
Re:Turning monitor off
Many household devices today use more power when their state is changed than they would if they were just left on constantly.
Go get yourself a meter like the WattsUp and gather some stats. In my experience, for computers there is an initial power surge that is maybe 2-3x normal draw, and lasts anywhere from 20 seconds to a couple of minutes. So worst case, if you will be turning your computer on again within 6 minutes, you probably should leave it on. Otherwise, turning it off will save power.
For screens, it's more complicated; CRTs draw more while warming up, but only for 15 seconds or so. But they draw less when showing a fully black screen, and different resolutions draw different amounts. Also, modern CRTs have assorted sleep modes. So whether or not to turn off a CRT depends on your usage patterns. For LCDs, I have no data.
I was also interested to note that high CPU usage produces measurable extra power usage. On a dual-processor P3/733 system, each processor pulls an extra 10 watts under load. So it turns out that things like SetiAtHome aren't free.
My mother, for instance, wastes more power than she saves by going around turning off all her fluorescent lights.
A common misconception, but not true. See this article for the details. They recommend leaving them on if you'll need them again within 15 minutes. But this isn't because of power usage; it's because turning them off and on too often reduces bulb life.
Aren't LCD monitors designed in the same way? They only use power when they are updated?
You'll note that in a dark room, you can see the LCD just fine. This means that it's producing light, which takes power. According to this week's Economist, "an LCD is only 10-15% efficient at converting energy from its power source into a readable image." -
Re:Speaking of things like this...I'm amazed that the Kill-a-watt has been mentioned numerous times in this discussion, but I've seen no reference to the far superior (if a bit more expensive) Brand power meter ($150 - $350) or the Watts Up? meter ($100 - $130). Both are capable of measuring a single 120V AC load both instantaniously and cumulatively over time. They can also show the Voltage and power factor and such (check out individual specs, as I forget which does what).
The high-end Brand meters can also connect to a PC via RS-232 serial (Win 95/98 only) and log data that way, and there is even a model with non-volitile memory and a datalogger onboard, so that you can plug in your fridge and come back in a month and download its power consumption history. I highly recommend these meters, since they don't simply measure Watts as Volts * Amps, but calculate an implicit power factor by sampling instantanious Volts (this is a sine wave in AC, remember) & Amps at 4 kHz in order to measure your power consumption just like your utility does.
For more & larger loads, check out the Brand "One Meter" which can handle 120 or 240 V AC loads, DC loads, and 64 channels simultaniously. Sensors are networked to the datalogger via Cat-5 (not over TCP-IP, though). There is a small LCD for instantanious readout, and it can download data through any com program and a serial cable (I use Win 2k, never got it to work with my Mac).
One of the coolest devices I've seen is the WebDAQ which is a 32-channel datalogger with built-in server; just plug in some sensors and an ethernet network, and you can view data live over the web! A little pricey for me at $1k, but cheaper than putting together the pieces. -
Re:Power usageWell thanks to a nifty meter know as the watts up. I can athoritatively say that my computer eats as much as my refridgerator. 'fridge 5.03 KwH used over 43.4 hours, yielding an average watts used of 115.9. AMD athlon based system with ultra 160 scsi drive on 24/7; 2.5 KwH used over 26.2 hours, yielding an average wattage of 95.4. Tossing in my 21 inch monitor, 2.2 KwH is 150 hours yields 14.7 watts.
Of course I haven't gotten around to enabling power management because I want remote access to my system.
Also I haven't gotten around to measuring my ether switch or cable modem yet. So the computer equiment should be using a bit more. Also those measurements were for a particular usage pattern which might have changed when I moved. diane