EcoDog vs Arduino

Re: EcoDog vs Arduino

I've been looking into Brultech ECM-1240. Have you looked into that one?

Got any details on how you're building your Arduino?

Re: EcoDog vs Arduino

OK, I've looked into that Arduino project before. Just haven't had time to mess with it.

The Brultech starts at about $200 - more if you want automatic data logging and ethernet or wireless capability will double the price. But seems to be a better value than things like the TED if you want to monitor multiple circuits. And if you have a computer to do the logging via serial it's a pretty good price.

I've also looked at EcoDog - but can't find pricing. How much does it cost?

Re: EcoDog vs Arduino

I just got a price quote this morning, a bit over $900 for a system with 16 CTs.

Yup, I'm going with the Brultech 1240

Re: EcoDog vs Arduino

sub3marathonman wrote: »

I was wondering what the price of the Arduino was, and where the CTs were purchased from. To me, the Brultech prices seem a bit high, and their shipping charges are extremely high.

I'm also wondering why the calculations are being done on the Arduino, and if anything could be set up to just get the values from the CTs to a full-size computer that has far more computing power.

I'm guessing that you haven't tried actually pricing any of this stuff yet.

Before I came across the Brultech system I bought a couple CTs off ebay. I paid $8 each for them plus another $6 shipping (ea)
Brultech only charges $5 for their 40 amp CTs, a bit more for the 80 amp ones.
For less than $180 you can get the Brultech computer, 2 100 amp split CTs (look up pricing for them), 2x 80 amp and 10x 40 amp CTs.

If you were going to duplicate this setup with an Arduino you'd probably end up spending more (even if you buy the CTs from Brultech), and THEN you'd have to write the custom software to calculate power factor, etc. from the voltage and current measurements.
Plus you'd probably need several Arduinos, or add a couple extra analog to digital converters to match the 8 inputs from the Brultech (7 inputs for CTs and one to measure voltage). I'm not sure but I think the Brultech has higher resolution on it's A-to-D converters.

There might be something available that is cheaper than the Brultech, but everything I've found is either more expensive, less capable, or both.
Look into "The Energy Detective" less capability and higher price, or the EcoDog WAY higher price.

You might be able to do better with an Arduino if you're really good at programming. Just trying to figure out the coding for calculating power factor was more than I wanted to try.

Re: EcoDog vs Arduino

sub3marathonman wrote: »

That would obviously, based on my post, be incorrect. I stated, "To me, the Brultech prices seem a bit high, and their shipping charges are extremely high." If I hadn't researched the price I could not have made that statement, which I stand by..

My apologies then. Since you've researched this, can you please post a link to similar equipment with cheaper prices?

Best of luck with the Arduino, please keep us posted on your progress.

Re: EcoDog vs Arduino

mr.radon wrote: »

I have a couple of Arduino boards, the one I'm playing with is the Max 1280.
http://www.bizoner.com/arduino-atmega1280-mega-board-with-usb-cable-and-9v-power-cable-p-180.html
It has 16 ADC inputs. Right now I've got a few CT's, when I figure it out, I'm going to use these to measure the current: http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=ACS750LCA-050-ND

I can get the CT's hooked to the Arduino to output raw data. I can use excel to crunch numbers, but i want to just output data sets every second, save it and not have to crunch numbers.

The Ecodog is working real well. No faults, but not worth the money...
It does allow access to your data.

Are you planning on sampling the current once a second, or are you going to take the hundreds(thousands?) of samples per second needed to accurately measure the current waveform and then output those hundreds of samples once per second?
You won't be able to measure current with only one sample per second, worse case example would be sampling the current every second exactly when it happens to be at the zero crossing point and you'll always see zero current.

I'm guestimating that in order to get an reasonably accurate idea of the waveform shape you'll need to sample the waveform at least 20 times per cycle, or 1200 times per second. Without knowing the exact shape of the waveform, you won't be able to calculate power factor which means you won't be able to accurately measure true power. A lot of the homemade systems I've seen just guess at the power factor and voltage which means there measurements can be off by up to 20%.

I'm not all that familiar with the capabilities of the Arduino, but doing a rough estimate exporting 1200 samples per second it's going to require something like 19,200 bps per channel. For 8 channels that's a bit over 1.5 Mbps.

Can the Arduino handle that high of an output data stream while simultaneously sampling 8 channels 1200 times per second?

FWIW I don't see any way you can output raw data once per second unless you're only looking at one channel of data. Seems to me like it's going to have to be continuously outputting data just to keep up with the sample rate.

Re: EcoDog vs Arduino

mr.radon wrote: »

I'm not feeding the output straight into the A-D, it goes to a circuit that outputs a 3 second moving average current reading based on a 60hz system. As far as power factor, I'm working on that.
The Arduino board work is mainly to satisfy my curiosity and help me remember things from engineering school.
I've been monitoring my Enphase data verses utility production meter data verses, TED 1000 verses Ecodog.
Enphase to Utility production meter, Enphase reports 2.5% more power.
TED 1000 to Utility production meter reports 11% more power.
Ecodog to Utility production meter reports 2.3% more power.
So when I finish the Arduino board I'll be able to see how it reports compared my other meters. I'd like to do an instructables write up.
I'm not happy with the TED1000.

About the amount of data.... 10 bits per A-D channel and 8bits/byte
@ 1 sample/sec: 16 channels = 160 bits/sec, 20 bytes/sec
Over a day that is only 1,728,000 bytes of data. Its well within the Arduino's capability, I've gotten it to record all 16 channels 100X a second.

And I think your calculations are a bit off.
@ 1,200 sample/sec: 16 channels = 19.2K bits/sec, 24K bytes/sec

So you plan on multiplexing the data together before transmitting and then separate it out again on the other side?

At low data rates that should work fine, though you'll probably want to include some framing and sync bits/bytes, possibly include some CRC checks for data integrity. Have you ever heard of Generic Framing Procedure (GFP)? It has some rather elegant solutions to framing and errors checks that combine the two and result in very low overheard.

I didn't think the Arduino would have enough processing power left if it was running thousands of samples per second to mux the data include framing/sync/error checks, which is why I just assumed you'd use two bytes per sample on the output, you could do simple parity checks and maybe throw in an extra byte for framing/sync every time you restart the sequence.

I don't see how you can measure power factor without getting enough samples to see the actual shape of the current waveform.
I wonder if that is the problem with the TED, they might be guessing at the power factor.

FWIW I think your math is off
1,200 samples per second of 10 bits per sample = 12,000 bits per second x 16 channels = 192,000 bits per second, etc.
Though to be fair I also screwed up the math, it should have read 153.6 Kbs instead of 1.5 Mbs.

Like I said, my original statement assumed keeping each channels data in separate bytes (instead of muxing it together), so 2 bytes per sample.

Re: EcoDog vs Arduino

mr.radon wrote: »

Here is a snapshot of the Ecodog pie chart...

Cool, ...what the heck do you have plugged in in the master bath that draws that much energy?

Re: EcoDog vs Arduino

sub3marathonman wrote: »

From http://bangaloreinteractive.wordpress.com/common-arduino-commands/ it states that the maximum reading rate is about 10000 times/second (just for the reading command), which would work out to at most 625 times/second on each of the 16 channels (just for the reading command).

I hope people will forgive the faux-pas of quoting yourself, but I was hoping somebody would be able to add some additional information.

If you measure 600 times/second, that is 10 times/cycle. I'm wondering if that is enough data to get an accurate measurement of power usage. Also, is it possible to measure fewer times if it is a circuit where the power factor is an unchanging value of 1.0?

I thought I saw somewhere that there was a mathematical method to accurately determine a function with just four data points, but I can't find that now.

It would be a bit ironic if the Arduino with 16 inputs would not really be able to be used because it is too slow to measure all 16 lines accurately.