Has anyone seen this DIY electricity monitor?

[mpooley]

http://www.billporter.info/not-so-tiny-power-meter/


Happy new year everybody

Mike

[wookey]

No. Looks interesting. Well-found.

I did find this one a year or so ago: http://www.delorie.com/electronics/powermeter/
and now (after much faffing) have only one component left to solder on (need to find where they keep the binocular microscope at the new job) before I can program it and try it . It has the advantage of being multi-channel, but is also a much more expensive PCB and project. I've probably spent 12 hours just soldering it up.

[SimonHobson]

I did find this one a year or so ago: http://www.delorie.com/electronics/powermeter/
and now (after much faffing) have only one component left to solder on ...

Hmm, details seem rather 'scant' !
Before I start delving in, do you know if it measures voltage and phase angle (ie reporting real power and VA), or is it yet another "measure current and assume the rest" job ? If it does real power and VA then it could be just the thing I've been looking for.

[Dyslexicbloke]

Hi folks …
Just wondering if anyone has any insight into using power meters with a modified sine wave inverter.

OK I know but I can’t afford a true sine inverter and won’t be able to for some time.

Anyway I have tried a couple of the plug in type energy meters and they just read rubbish when on the inverter. One of the things I need to do is collect a load of data and use it to manage my CHP system but I am not anticipating that being easy.

If anyone has any insight into weather these, or other circuits and units, will cope with a poor waveform I for one would be very interested.

Alistair

[SimonHobson]

Just wondering if anyone has any insight into using power meters with a modified sine wave inverter.

OK I know but I can’t afford a true sine inverter and won’t be able to for some time.

Anyway I have tried a couple of the plug in type energy meters and they just read rubbish when on the inverter. ...

I'm not in the least surprised.

I've come across (yet another) power monitoring project here. He describes how he measures voltage and current - and he is just measuring peak and dividing to get RMS, assuming of course that both are sine waves. That assumption is reasonable for voltage when running from mains, but not for current, nor for voltage when running from anything but a 'solid' supply.

Modern equipment is much better since regs now restrict the current waveform and power factor - large switch mode power supplies now generally have another switching circuit in front of them to form the current to be close to a sine wave and unity power factor. But older equipment can be very poor with a peaky current waveform and consequent modification of the voltage waveform when running on a 'soft' supply like a generator (even if it's sine wave).

Now, back to the question ...
Two things come to mind. One is to ask the vendor if their unit measures true power with a non-sine voltage/power waveform. Don't expect an answer, or at least a truthful answer, from most vendors as they'll have no idea and either ignore you or just make something up :roll: I suspect you'll have to go to something quite pricey to get true RMS reading units - if you think about it, it won't be enough to work out voltage and current over a waveform, they'll actually have to do lots and lots of small measurements, multiply VxA (Volts time Amps) to get W, and integrate that over a waveform. Without looking it up, I suspect a ferraris disk electric meter (ie the old fashioned spinning wheel type) will work fairly well since these are effectively an integrator of VxA. You'd also be back to manual readings unles you use a pickup to count wheel rotations like this project.

Alternatively, you could measure the DC voltage and current going into the inverter - but then you would need to somehow calculate the efficiency of the inverter on various loads to get a reasonable approximation of AC power out.

[wookey]

Simon - how is the complete schematics, BOM, and the PIC code used to run it 'scant'? There _are_ no more details that could be provided :-)

I do agree that it doesn't have a nice user-friendly 'spec'. You have to read the schematics or the gEDA list thread where it was designed for that sort of info: http://www.mail-archive.com/geda-user@moria.seul.org/msg14029.html

where it says "The chip measures the current many times per 60Hz cycle and integrates the real power used". i.e it's a 'real' power meter, not a cheap power meter. More details are given on the datasheet for the ADE7753.

[SimonHobson]

Simon - how is the complete schematics, BOM, and the PIC code used to run it 'scant'? There _are_ no more details that could be provided :-)

I do agree that it doesn't have a nice user-friendly 'spec'. You have to read the schematics or the gEDA list thread where it was designed for that sort of info: http://www.mail-archive.com/geda-user@moria.seul.org/msg14029.html

where it says "The chip measures the current many times per 60Hz cycle and integrates the real power used". i.e it's a 'real' power meter, not a cheap power meter. More details are given on the datasheet for the ADE7753.

OK, I'll rephrase my original comment - there's scant information given on the web page "The code is the documentation" attitude is one of those things that does put people off open source in many cases, I'm not exactly non-technical myself, but I haven't the time (at the moment) to go reading the code to see if there's any comments giving clues as to whether the project merits a better look. There isn't a reference (that I've noticed) on the page to the mail thread.

Now I've seen what it's doing, it's definitely something I'll be looking into further - but for me, I'd need a different board (I'd need the current sensors remote from the rest) and I've never got to that level. I have gEDA 'sort of' installed on my Mac, but I don't know how to use it

[SimonHobson]

Without looking it up, I suspect a ferraris disk electric meter (ie the old fashioned spinning wheel type) will work fairly well since these are effectively an integrator of VxA.

Chatting with Dad over breakfast, he seems to think it would work as well.

[SimonHobson]

Simon - how is the complete schematics, BOM, and the PIC code used to run it 'scant'? There _are_ no more details that could be provided :-)

I'm installing (via Fink) gEDA as I type so I can take a look at the full schematics - I can see that the PDF on the page is incomplete (a few "hanging" net flags which don't connect to anywhere else in the document). There is enough to show me what the hardware is though. The code is, not unsurprisingly, "sparsely" commented and C isn't a language I've ever worked in.

However, it does look like just the job I've been looking for. I think I'll be needing to polish up my soldering skills before long ! If the first works, the boss will have me making five of them - one for the UPS loads in the server room, and one each for the mains distribution boards in each office And then I can start on the house, by which time there might be a few parts 'surplus' at work

I guess boards will be a matter of sending the files to one of the outfits advertising in the electronics mags ?

[Pat_]

Just been dipping in to this thread. But the documentation mentions current detectors whilst not mentioning voltage measurement. So presumably power factor is not dealt with, which seemed to be the critical question.

[SimonHobson]

Just been dipping in to this thread. But the documentation mentions current detectors whilst not mentioning voltage measurement. So presumably power factor is not dealt with, which seemed to be the critical question.

Yes, voltage is measured. The power supply is via a small transformer, and the AC signal is picked off upstream of the rectifier to provide the voltage input - the assumption for this design being that each channel is on the same supply (no multiphase support), although that would be possible if you create your own board. The chips used do a proper power calculation - instantaneous volts*amps to get instantaneous power and integrate it. They also do a 90˚ phase shift on one of the signals and calculate true reactive power as well.

Info on the chips is available on the Analog Devices website.

[KenB]

You could also consider the open source energy monitor being developed by Openenergymonitor.org a fresh new start-up company based in Snowdonia. They have developed energy monitoring solutions for the Centre for Alternative Technology (CAT) and are working on solutions for larger organisations including the National Trust.

They too have a voltage input, provided from a small transformer, and the hardware is based on readily available Arduino kit.

They have a real time graphing output to a pc running an open source programme (java based so works on MAC, Linux, PC) called KST and are working on ethernet hardware to allow energy monitoring to be remotely accessed from a web based application.

Their developments are all based on open source, licence free and there are no proprietory protocols involved.



Ken

[Pat_]

Yes, voltage is measured. The power supply is via a small transformer...

Ah, thanks. When I have time, I'll take a further look.

[Dyslexicbloke]

Hay guys .... and Gals, havnt checked the names:-)

Thanks for input ... I will be building, can't afford commertial offerings.
I just bought 4 of these (below) on the asumption that I was looking at a uP project of some sort.

http://www.devicecraft.com/haefcuse1.html (The guy gives details on how to modify the board, how cool is that)

It remains to be seen how stable they are but you can't knock the price, $5.50 each and $4.00 shipping,  thats circa £17.00 for the lot, delivered!

I am going with Pic, because I have some and a programmer, which will run at 20Khz.
If I use an averaged input, standard analouge read strategy, and do it on a zero crossing triggered interupt I should be able to get 20 samples of V and A within each target cycle.

That said .... I plan to look at an old style meter as well, I like the simplicity of the solution, I am all for simplicity where it works and counting revolutions is hardly a chalange. I suppose its like a heat based power measurment but without the need for temperature compensation.

All good stuff folks thanks.
Al

[ericw]

If you are not interested in ultimate accuracy I have done a Halfbee which measures/calculates the RMS current, using a current transformer, and can take in a phase reference to give information on direction. It doesn't check the magnitude of the voltage or power factor however.

If you get a modern meter with a flashing led it makes measuring power so much easier.