Monitoring ac power

[rwfurness]

I am wanting to monitor the power consumption of a domestic fridge at the same time as monitoring ambient temperature, internal temperature and condensing coil temperature. I plan to build a datalogging system based around a PIC and digital temperature sensors. I need some help with how to sense the power being used by the fridge, and how to get this information into the PIC based datalogger.

Robert

 

[MBrej]

Hi, as far as measuring the AC current goes, look at something like the CSLA2CD, which is a hall effect sensor and will give an output voltage proportional to the current, but the mains and the electronics are isolated. As the output from the sensor will be a AC voltage, rectify the voltage to keep it positive, and put a capacitor in parallel (perhaps with a large resistor) after this to smooth the signal so the voltage will be the peak voltage. This can be read by an ADC on the PIC, and then can then be converted to RMS voltage/current quite easily if you assume the waveform is a sine wave (which it should be).

The output of the hall effect sensor may need feeding into an opamp to change the output impedance (the sensor probably cant supply alot of output current to drive the electronics on the output to get an accurate reading)

As for the rest of the data logging circuitry there is plenty of information on the forum already.

Matt

 

[manuka]

Do you need true W power, reactive VA or just a casual insight into what's being used? If just the latter then a simple inductive "clamp" around the phase lead will deliver quite a few mV AC. Well before PICAXE era READADC etc I used a spark plug lead clamp to great effect, calibrating it against a 100W lamp & 1000 W heaters etc via a small trim pot, with the output fed to a DMM. It had good dynamic range too.

Some dirt cheap AC clamp meters are now around - here in NZ the "red shed" sells a nifty digital one for ~US$10!

Temp. reading via DS18B20s is of course pretty turn key now. You may want to use the Rev.Ed data logger for the overall task. What exactly are you trying to achieve with these fridge insights?

 

[krypton_john]

I am wanting to monitor the power consumption of a domestic fridge at the same time as monitoring ambient temperature, internal temperature and condensing coil temperature. I plan to build a datalogging system based around a PIC and digital temperature sensors. I need some help with how to sense the power being used by the fridge, and how to get this information into the PIC based datalogger.

Robert

Hey, if you are going to monitor the fridge don't forget to add a monitor to answer a question of the ages - 'does the light go off when the door closes' - a simple LDR might suffice!

But seriously - would be interesting to log the amount of time the door is open when considering the efficiency of the fridge.

 

[fernando_g]

Hey, if you are going to monitor the fridge don't forget to add a monitor to answer a question of the ages - 'does the light go off when the door closes' - a simple LDR might suffice!

Oh Boy, don't even get me started on that one.
Some years ago I noticed that indeed the compressor at home would run longer than usual. Could not explain it why.

One night, arriving at a completely darkened home, I noticed the faintest sliver of light coming from the fridge's door. I immediately figured it out, the light was on.

It turns out that my daughter had spilled some chocolate milk. My wife cleaned the mess, but unknown to her, some of the spilled milk had seeped into the door activated light switch, which was located at the bottom. The milk hardened and glued the contacts shut. So although the switch's lever would still move freely, the light would remain on forever.

So indeed, any fridge efficiency monitor should include a light sensor.

Now my two yen on the subject:
REAL Power is the product of voltage, current AND power factor. I've built a circuit to do this, let me dig it out and I'll post it

 

[fernando_g]

power meter schematic

I found it...

On the schematic, the areas shaded in pink are no longer required when using a PICAXE.
When I developed this circuit, I was not able to write microprocessor assembly code, and therefore, I used a F/V converter to convert the AD7755's frequency output (pin 22) into a DC voltage which was read with a multimeter.

This frequency is proportional to the real power, and with a PICAXE one could use a "count" command to measure the frequency and scale it to watts.
The AD7755 also has another output (pin 24) which can be used to integrate watt-hours.

You can read about this versatile device in Analog Device's web site.

Please note, this was designed for a 120V powerline; for 240V applications, you need to scale the voltage divider R4, R9, R10, R11 accordingly.

 

[rwfurness]

Hi, as far as measuring the AC current goes, look at something like the CSLA2CD, which is a hall effect sensor and will give an output voltage proportional to the current, but the mains and the electronics are isolated. As the output from the sensor will be a AC voltage, rectify the voltage to keep it positive, and put a capacitor in parallel (perhaps with a large resistor) after this to smooth the signal so the voltage will be the peak voltage. This can be read by an ADC on the PIC, and then can then be converted to RMS voltage/current quite easily if you assume the waveform is a sine wave (which it should be).

The output of the hall effect sensor may need feeding into an opamp to change the output impedance (the sensor probably cant supply alot of output current to drive the electronics on the output to get an accurate reading)

As for the rest of the data logging circuitry there is plenty of information on the forum already.

Matt

Matt

Thanks for this suggestion. I have looked up this chip, and it looks a good option.

Robert

 

[rwfurness]

Hey, if you are going to monitor the fridge don't forget to add a monitor to answer a question of the ages - 'does the light go off when the door closes' - a simple LDR might suffice!

But seriously - would be interesting to log the amount of time the door is open when considering the efficiency of the fridge.

Krypton John

Yes, I am plannng to log the time the door is open. I plan to use a real time clock, and log the times the door is opened and closed. I had thought of using an LDR, but I think a door switch is simpler.

Robert

 

[rwfurness]

I found it...

On the schematic, the areas shaded in pink are no longer required when using a PICAXE.
When I developed this circuit, I was not able to write microprocessor assembly code, and therefore, I used a F/V converter to convert the AD7755's frequency output (pin 22) into a DC voltage which was read with a multimeter.

This frequency is proportional to the real power, and with a PICAXE one could use a "count" command to measure the frequency and scale it to watts.
The AD7755 also has another output (pin 24) which can be used to integrate watt-hours.

You can read about this versatile device in Analog Device's web site.

Please note, this was designed for a 120V powerline; for 240V applications, you need to scale the voltage divider R4, R9, R10, R11 accordingly.

Fernando

This looks an interesting option. I have looked up the Datasheet. It seems there is now an updated chip - the ADE7757. Thanks for the suggestion. I've a bit of reading to do now.

Robert

 

[krypton_john]

Krypton John

Yes, I am plannng to log the time the door is open. I plan to use a real time clock, and log the times the door is opened and closed. I had thought of using an LDR, but I think a door switch is simpler.

Robert

And of course the door switch is already part of the fridge!

You'll need to be careful about getting your signal wires out of the fridge without affecting the integrity of the door seal or cabinet.

 

[rmeldo]

Hi,

I thought I would resume this thread since I am trying to do a similar job as kripton_john.

I want to monitor and log the energy used by a night storage heater.

See message 20 the thread below for a full description of my project:

http://www.picaxeforum.co.uk/showthread.php?t=11145

The part that is relevant to this thread is the one regarding the metering of the AC power adsorbed by the heater overnight.

I liked the idea of implementing the circuit that Fernando G posted here so I went ahead and ordered a couple of samples of the ADE7769 (16 pins), which to me seems quite similar to the 7757 mentioned above.

The chip could be made to measure voltage directly and current indirectly, throught the use of a shunt.

The problem that I would like to explore is the one about isolation of the Picaxe circuitry from main voltage.

It is my understanding that the voltage reading could be done indirectly by putting a transformer in between the mains and the Picaxe. However transformers are noisy and being the heater in the bedroom, night noise is highly undersirable.

Also what could be done with respect to measuring the current?

Also there is the problem of keeping the relative phase of voltage and current measurements consistent before and after the isolation.

What is people's opinion on the matter?

Thanks

Riccardo

 

[hippy]

For the ADE7769, Application Note AN-679 could be a good place to start.

http://www.analog.com/static/imported-files/application_notes/81462932AN679_a.pdf

Figure 1 and Figure 33 ( full circuit ) shows the basics including a transformerless power supply. Providing all that is safe and away from prying fingers it would likely be okay for use ( you'll always get the ultimate call and have to take responsibility for that decision ).

For electrical isolation with the PICAXE, an opto-isolator as shown on outputs K7 and K8 would do that job. No other connections needed or wanted to the live-side circuit from whatever else you connect.

If you want to isolate the ADE7769 itself from live mains, I'm not sure how you'd go about that as it seems to be implicit that it is connected directly to the live system and its internal calculations may depend on correct phasing of signals.

 

[hippy]

A thought ... Are there cheap products which already use the ADE7769 available ? If so they should ( caveat emporer with dodgy kit from dodgy suppliers ) be certified and therefore safe for use. Taking one of those and adding an opto-isolator could be a much simpler job than having to create the entire circuit for yourself.

 

[rmeldo]

Thanks Hippy. I'll start reading. That opto-isolator seems a good idea from th e outside.


Regarding your suggestion of finding something ready made, I had thought exactly the same. I have on my desk a plug-in plug-socket to plug loads on, which I bought with that intention exactly. I wanted:

1) see what was inside.

2) find a way to hijack the device and extract a signal for my own use.

Sadly there is no name on the chip so I could not see how to do it.

I might have another go after reading the application note.

Thanks

Riccardo

 

[BeanieBots]

The phase of the signals is critical, however, using transformer isolation should not introduce much in the way of phase shift providing a good quality low loss transformer is used on BOTH the voltage AND the current sense.

A transformer should NOT be noisy. If it is, something is wrong (eg loose laminations) and it will be lossy and hence not suitable. Even a cheap wall-wart transformer should be OK.

Getting "ready made" kit as suggested by Hippy would be the best option and would only require a simple opto isolator for complete safety. Probably a very expensive option but the current domestic consumer units come with a digital output. Just needs an LDR to read the 'pulses'.