Relaying data from the "Kill A Watt" to a PC

[electron]

These guys have hacked the popular Kill A Watt unit, so it can transmit the power usage to a PC using the Xbee chip Schematics and source will be shared as well:

Click for more info

Thanks to TonyNo for mentioning this project in the chatroom.

 

[drozwood90]

We've been discussing something like this over at the Homeseer Forums.

http://forums.homeseer.com/showthread.php?...ght=kill-a-watt

I didn't want to really hook up the way they did, until I can confirm that the output from the Op-Amps are calibrated.

I think it might be, but the issue is that I gave my only Kill-a-Watt away for X-mas...so, until I get another...I'm at a standstill.

I know that the Serial RF modules I was going to make...well, I found a supplier that can sell them to me in a custom frequency for $3.99 and $2.99.

So, whip a PC side box, and then assign an ID number to each module and the PC can query each KAW module for it's current (haha) voltage and current / power readings.

--Dan

 

[hollisin]

Holy Heathkit I need to build one of these. I haven't seen the "will be available" stuff anywhere. I'll bet he will get 100 emails a day until he publishes something.

 

[project_x]

We've been discussing something like this over at the Homeseer Forums.

http://forums.homeseer.com/showthread.php?...ght=kill-a-watt

I didn't want to really hook up the way they did, until I can confirm that the output from the Op-Amps are calibrated.

I think it might be, but the issue is that I gave my only Kill-a-Watt away for X-mas...so, until I get another...I'm at a standstill.

I know that the Serial RF modules I was going to make...well, I found a supplier that can sell them to me in a custom frequency for $3.99 and $2.99.

So, whip a PC side box, and then assign an ID number to each module and the PC can query each KAW module for it's current (haha) voltage and current / power readings.

--Dan

I'm interested in the Serial RF modules, are they your design, or off the shelf at the supplier (in particular frequencies)? Would you mind passing on some more information?

Thanks,
Rob

 

[drozwood90]

We've been discussing something like this over at the Homeseer Forums.

http://forums.homeseer.com/showthread.php?...ght=kill-a-watt

I didn't want to really hook up the way they did, until I can confirm that the output from the Op-Amps are calibrated.

I think it might be, but the issue is that I gave my only Kill-a-Watt away for X-mas...so, until I get another...I'm at a standstill.

I know that the Serial RF modules I was going to make...well, I found a supplier that can sell them to me in a custom frequency for $3.99 and $2.99.

So, whip a PC side box, and then assign an ID number to each module and the PC can query each KAW module for it's current (haha) voltage and current / power readings.

--Dan

I'm interested in the Serial RF modules, are they your design, or off the shelf at the supplier (in particular frequencies)? Would you mind passing on some more information?

Thanks,
Rob

http://www.ananiahelectronics.com/PCR1A.htm
$3.99 for receiver

http://www.ananiahelectronics.com/FS100A.htm
$2.99 for transmitter

-------------

http://www.blueatomonline.com/products.php...&category=5



If you buy from the top guy, tell him that drozwood90 referred you.

Just building up the brownie points!

--Dan

 

[politics123]

Update: They've posted a parts list, schematic, how-to, and code:

http://www.ladyada.net/make/wattcher/

Seems like a great way to monitor how thirsty your computer is

 

[AnthonyZ]

Update: They've posted a parts list, schematic, how-to, and code:

http://www.ladyada.net/make/wattcher/

Seems like a great way to monitor how thirsty your computer is

I hope to head to Sparkfun tomorrow for some "bits and pieces". I'm also a Twitter user so this is a natural fit for me...

 

[damage]

they are now selling kits

http://www.adafruit.com/blog/2009/03/26/tw...-now-available/

 

[mustangcoupe]

I have not seen any press release on this yet but I did see a Kill A Watt sensor unit recently received (rf transmitter) FCC approval (according to the FCC database). Hope they used a open protocol and not something we cant access.

I hope to see more information soon.

 

[vc1234]

I have not seen any press release on this yet but I did see a Kill A Watt sensor unit recently received (rf transmitter) FCC approval (according to the FCC database). Hope they used a open protocol and not something we cant access.

I hope to see more information soon.

I just stumbled over the KaW+Xbee solution. I took a brief look at the Python code and they appear to be measuring apparent power (V*A) rather than actual power (watts). It means (if true) that they do not take into account the power factor which can be around 0.6-0.8 for motor loads such as dishwashers/fans/air-conditioners etc. I'll try post a question on their forum to get clarification.

 

[drozwood90]

Thanks! That was one thing I was trying to understand.

What I am trying to get an understanding of is how to actually just read the display. This way I can get a reading of actual power use...with any power factor included.

Good catch...glad I'm not the only one out there thinking about this!!

--Dan

 

[vc1234]

Thanks! That was one thing I was trying to understand.

What I am trying to get an understanding of is how to actually just read the display. This way I can get a reading of actual power use...with any power factor included.

Good catch...glad I'm not the only one out there thinking about this!!

--Dan

I did not get an answer yet but after looking a bit more at the python code, here's what I posted on another forum:

"
I took another look at the python code, and it seems that calculations are OK. I did not notice the first time that you take the average of 17 samples which is in fact active power given suitable sampling interval.
"
What they do is take 17 one millisecond samples (about one AC sine wave period) and get an average of instantaneous I and V products which is a pretty good approximation to the theoretical average power value. There still may be a substantial error due to AD conversion and possible uncompensated phase shift, but it's hard to say without measuring...

Do you happen to know what power metering chip Kill-a-Watt uses ?

 

[drozwood90]

They don't use one.

They use a piece of wire that's used to generate the current reading. This value as well as a scaled Voltage reading is put into a Op-Amp (the same one that is pictured that people tie into with the Zig-bee stuff). This is then read using PWM into the chip, which I had a link to at one time (after people started to pull the things apart, the MFTR. started to hide what the chip was). I have a listing of it somewhere, but it is a micro. that was SPECIFICALLY designed to do power stuff...although without an ADC built in...

The value of the wire and other stuff doesn't matter to the chip, as they are calibrated out. The values from the calibration that is performed (again, I had links to the procedure in how to get into the cal. mode...but have since lost it) are stored in the EEPROM that is next to the microchip.

That is one reason that I'm trying to capture the digital value...

BUT, if I find that the offset or power factor can be nulled or is so small it doesn't matter for what I'm doing, then the value really is getting something simple, as opposed to accurate (diminishing returns). Is it worth it to get the digital value if it costs 10x more, or can I suffer 1% tolerance? 2%? When does the tolerance really start to impact when I would rather go to read the post calibrated value?

Your thoughts?

--Dan

 

[hobbes487]

I have jumped in on this and actually built a couple of these. It seems accurate enough for my uses. I use them to determine if certain appliances attached to them are turned on or not. For example, my home entertainment equipment (HTPC, projector, receiver, speakers) use about 27 watts even when they are all turned off. So I have set it up like this:

Wall -> Kill-a-Watt -> Z-Wave Appliance module -> power strip with all equipment

Initially I would just have the zwave module to switch off during the wee hours of the morning, but now I can be more intelligent about it. What if I am in the middle of a 24hr movie marathon? Based on my schedule the zwave module should be turned off, but because the KAW shows much higher usage than 27 watts, it won't turn off. I also use occupancy, but this just adds to the intelligence of the overall system.

I don't use the python code that was given in the tutorial. I made my own application in C# that interfaces with vCrib so I can use the readings in any sort of intelligence that I want.

 

[vc1234]

They don't use one.

They use a piece of wire that's used to generate the current reading. This value as well as a scaled Voltage reading is put into a Op-Amp (the same one that is pictured that people tie into with the Zig-bee stuff). This is then read using PWM into the chip, which I had a link to at one time (after people started to pull the things apart, the MFTR. started to hide what the chip was). I have a listing of it somewhere, but it is a micro. that was SPECIFICALLY designed to do power stuff...although without an ADC built in...

I found this picture: http://cafeelectric.com/killawatt/index.php

The chip is called "PRODIGIT 53920005". I could not find any datasheet info for it which is odd as it looks like an ATMEGA micro controller or similar clone. A datasheet would be interesting to see ...

The value of the wire and other stuff doesn't matter to the chip, as they are calibrated out. The values from the calibration that is performed (again, I had links to the procedure in how to get into the cal. mode...but have since lost it) are stored in the EEPROM that is next to the microchip.

If you do not care about precision, what they do with the XBee ADCs should probably be fine although I do not like their waveforms much -- they look too distorted, especially the voltage one (http://www.ladyada.net/make/tweetawatt/software.html). The dimmer voltage is unreal -- there is no way the 120V mains sine wave can be distorted so much ! Something is wrong. The bulb waveforms are not very pretty either. For comparison, I've attached electrical bulb V/I waveforms obtained from a metering chip ADCs.