PowerMeter for aquarium

[Elysion]

Hi,

I have a 650lt. aquarium at home. There 2x 300 watt heaters, several pumps and several lights which mimics Sun's movement.
I want to follow the spent electrical energy. I'm wanting my data logger client to call a restfull service of mine in my host pc.
My data logger will be a 8 bit PIC. I have experience with calling restfull services from PIC over http. My question is about power sensing.
I looked around and find that they are generally

shunt based
http://www.microchip.com/DevelopmentTools/ProductDetails.aspx?PartNO=mcp3905ev

or hall effect based
http://www.infineon.com/cms/en/product/sensor-ics/current-sensor/TLI4970-D050T4/productType.html?productType=db3a30444449c62b01444e11450f0508

What I need is the watt/hour. No active/passive issues.

What do you suggest for sensing the current/power?

Thanks,
Oguz
 

[Signal]

What I need is the watt/hour. No active/passive issues.

I'd say "Watt*hour" and "active/reactive".

I want to follow the spent electrical energy.

Do you want to make a device yourself or just to solve the problem. For latter did you consider consumer products like this one http://www.amazon.com/P3-P4400-Electricity-Usage-Monitor/dp/B00009MDBU ?

[Gallymimu]

I'd consider a split core transformer as well.

not this one per se, but it's an example.
http://www.amazon.com/0-01-120A-0-5M-Split-core-Current-Transformer/dp/B005FIFT4E

[dezso]

I built one 4-5 years ago based on Linear's LT2940 with homemade fully isolated current transformer.

[Elysion]

I'd say "Watt*hour" and "active/reactive".
Do you want to make a device yourself or just to solve the problem. For latter did you consider consumer products like this one http://www.amazon.com/P3-P4400-Electricity-Usage-Monitor/dp/B00009MDBU ?

I want to do it myself. Otherwise a simple smartplug-plugmeter will make the job.

[Kombinator]

Alone is difficult to make a powermeter.
Will likely have a large reactive ingredient(ballast for fluorescent lamps) and big error.
Try to find an calibrated ready electronic energy meter
with AD7755 or similar.
Use time between pulses the out of the chip to calculate the momentary active power.
It will be the average power in time between 2 pulses. Will be very accurately measured.
Forgive me for bad English and incorrect terminology.
 Regards.

[Vineyards]

I would suggest something along the lines of the circuit posted by dezso with a special emphasis on isolation. Even the slightest amount of stray AC current would make your IC and analog measurement stages (including the onchip ADC) unstable. Isolation comes at a price. If you need an analog isolation barrier you will need something like ISO124 etc and a dedicated DC/DC converter. You may also have to think about clamp circuits to protect your analog interface. There are more options in the digital isolation realm like the ADUM series. If you don't want absolute precision, you can simply reckon the voltage as 220V (or whatever in your region) and just measure current and make the calculation based on this value vs timer value adding the result at the end of say one second to the power consumption counter and displaying the temporary consumption every second.

I haven't done anything like that before but I would normally proceed like this. Perhaps others users may have better ideas.

[dezso]

Also keep in mind the AC power P = VI formula will not work since the voltage and current constantly changes.
Pavg = VI cosφ a lot harder to calculate without a designated custom IC like Kombinator suggested AD7755 (obsolete) use AD71056

[Gallymimu]

Also keep in mind the AC power P = VI formula will not work since the voltage and current constantly changes.
Pavg = VI cosφ a lot harder to calculate without a designated custom IC like Kombinator suggested AD7755 (obsolete) use AD71056

you can get close enough looking at average voltage and current, or simply sampling voltage and current at about 1KHz and doing a point by point power calculation.  Even sampling voltage and current sequentially instead of simultaneously produces awfully good results.

[Signal]

<...> Even sampling voltage and current sequentially instead of simultaneously produces awfully good results.

I imagine simple linear interpolation would be a useful improvement for that case.

[Gallymimu]

I imagine simple linear interpolation would be a useful improvement for that case.

if you look at the relative error between linear interpolation and simple rectangular integration it doesn't make much difference compared to the rest of the errors (A/D linearity, reference voltage, etc.)   "GOOD" accuracy is of course relative, but I'm talking on the order of 1% for the measurements.

Heck, these days oversampling is so "cheap" it really reduces the level of analog filtering and post processing needed for applications like this.

[Elysion]

Hi to all,

I ordered some evaluation boards from DigiKey which use MCP39F501 and it's other family members. Especially MCP39F501 calculates
"Active, Reactive and Apparent Power, RMS Current and Voltage, Power Factor, Line Frequency".

When I test them, I'll share the results.

I bought TLI4970 based solutions, too but due to active/reactive/apparent power issues, I think, I'll use these current measurement devices only for my high power laser diode linear driver, not in this.

[Vineyards]

I don't think precision is very critical for this application. What will change if you have %10 error? If precision of any sort is needed you should consider an RMS/DC converter like AD736, AD536 or LTC1966 or LTC1967. 536  input can accept up to 7.5v, 736 is good for 0-200mV and LTC's will accept up to 1V (for the sake of linearity). Check the pdfs though these are from my memory. I worked with all of these IC's. Sampling cap is important. Check out the application notes for concepts like crest factor etc. My favourite is Ltc1966 but it is SMD only.

[David_1]

that home made current meter is just what I need, Thanks for the idea I am so pleased....

[jumulab]

Dear all,
Heve you seen the new TI device and application note / evaluation kit for MSP430 MCU ( MSP430i2040 Sub metering EVM). ?
I think it is avery cool device and information for your project.
TI have a very wide portfolio in energy meters  with their  loe power MSP430 family.
Yoy can see some info, aplication notes, data sheets and  all to start in :
http://www.ti.com/tool/TIDM-SERVER-PWR-MON.   Follow the  pages in this sibject.
Regards

[iot]

I have experience with calling restfull services from PIC over http.

Check for openhab for controlling your devices, you can save a lot of work and have a simple interface to control your aquarium.