Archived Forum

Measuring the flow of energy at  20+ points would not be practical with one emonTx.  There's only processing power to monitor two or three circuits at once.  I think you'd be best to measure just at the points where power enters or leaves the premises.

I don't think there's any maximum on the amount of AC current that can be measured.  It's the CT along with its input circuit which determines the signal which goes to the processor.  If the output signal from the CT is too great, you could maybe use only a fraction of it as is generally done with the AC voltage feed.  Or use a smaller burden resistor.

With such a high maximum current to be measured, you could not expect to get good accuracy at low levels of power. 

Maybe you could switch off all your equipment for a while and see whether the numbers keep on counting or not.

As Robin says, there is no limit to the current that the emonTx can handle, it's the current transformer you use that determines the maximum current. If you want 200 A, then I suggest a 600 A Magnelab, the -000 version with an external burden resistor (600 A because Magnelab c.t's are designed for 0.333 V output, whereas you need about 1 V for the emonTx. You change the 18 Ω burden on the emonTx PCB to 37.5 Ω - 36 Ω is the best standard value in practice. You also need to alter one line in the code to change the calibration.)

With the emonTx, you can expect reasonable accuracy down to about 1% of the maximum, below that the accuracy deteriorates.

My own Mk2 rig uses the smaller version of this CT which claims to go up to 200A. 

http://www.magnelab.com/products/current-voltage-sensors/split-core-ac-current-sensor-sct-0750

(I'm using the SCT-0750-000 version, with a burden resistor of 680R.  This CT is much less sensitive that the YHDC one)

Thanks guys for replying!!!

I have double checked the main's capacity, it is 125A not 200A!

I understand the accuracy issue with sensors having a high capacity when reading low values....  Really when you think about it, if the main is sized for 125A, its very unlikely that I would use more than 100A at the same time...

Assuming the WORST case scenario:  the default emonTX sensor rated at 100A represents 80% of my main's capacity (125A).  80% is a LOT, probably way too much for any scenario to be realistic.

I guess the default sensors at 100A is sufficient.

Now on to the questions:

In the future, lets say I opt to "upgrade" the CT sensor with a 200A or more, would I need to purchase another emontx or can I retrofit it easily?

The current shop carries only 433 & 868MHz emontx nodes.  I am in canada and my Raspberry Pi uses a 915MHz RFM12Pi antenna.  How can I retrofit the emontx with a 915MHz RF module??

Finally, the shop no longer carries emontx nodes.  What are my options?  I am comfortable building my own provided there is an online guide of some sort...  I am very strong in soldering but I need guidance not to solder the components in the wrong locations...

EDIT:  I have noticed that there are emontx shields in the shop.  I already own a Arduino UNO.  Perhaps this is a viable option?

Well I think I have answered most of my questions!

Looking at the CT sensors , the product page says they have an opening iof only 13mmx13mm... I am pretty sure the main cable coming from the utility is much bigger than that (top of my head, I'd say around 20mm including the insulation and jacket).  That would mean to buy a CT sensor capable of such wire size.  I suspect most sensors that are capable of clamping around such wire would have a capacity rated >> 100A... 

I will confirm the wire size as soon as possible.

Other than that, I was thinking about this strategy:

EmonTX shield (no RF)

Standalone RFM12B module at 915MHz then soldered onto the emontx shield

Arduino UNO

3X 100A CT sensors from this shop (to measure power from 3 larger circuits (hot water heater, living room heaters, stove))

1X 200A (ideally if possible 150A) CT sensor to be capable of large wire (18-20mm?)

Would that work?

The short answer - yes.

The long answer - if you go for a c.t. other than the shop standard one, you need to consider the output voltage that the c.t. is capable of. If you go for the larger Magnelab (with a large aperture), remember what I wrote above about the VA rating.

You can swap c.t. and burden resistor at any time in the future (and of course you'll need to re-calibrate the sketch).

But just a word of warning: don't consider changing the RF module. Once you've soldered the RFM12B module, you will find it very hard if not impossible to remove, as it is surface-mount with 14 joints that need to be melted simultaneously.

The emonTxShield_CT1234_voltage.ino sketch can only deal with one input at a time.  It will measure each input (CT & voltage) for a short period (0.2 seconds I think) before moving on to the next one.  Then it will dispatch the relevant data via RF and wait for 2 seconds before starting the sequence again. 

To measure power at all four inputs simultaneously would require a new interrupt-based sketch.  Each ADC conversion takes 104 uS, so each set of 5 need could take as little as 520 uS.  At 50 Hz, that's 38 sets of samples per 20000 uS period. 

Just to put calypso_rae's comment into context and to add a bit of explanation, exactly the same applies to the emonTx, but with only 3 current inputs slightly more samples can be taken each cycle.

The "basic" sketch mentioned - primarily designed for battery operation - measures current and voltage for 10 cycles (0.2 s for us, not for you, but this can be set in the sketch) and this measurement is repeated for each current input in use. The result is then transmitted by radio, and the processor put to sleep for 5 s. Then the sequence is repeated. In this mode, around 46 sample pairs are read each mains cycle (with 60 Hz mains), giving a good measure of the true rms voltage and current and the average real and apparent powers with a reasonably steady load, but not so good with rapidly fluctuating or switched loads.

In the "interrupt-based" sketch, the analogue to digital converter free runs in the background and feeds values to the main sketch. Thus it can run continuously and faster, whilst still performing the same calculations and feeding the results by radio. Unfortunately the downside is a mains power supply is required as the battery life would be too short. With 60 Hz mains, you're still looking at maybe 32 sets of 5 samples per cycle - much faster.

With 60 Hz mains, you're still looking at maybe 32 sets of 5 samples per cycle - much faster.

Indeed ... and someone to write the new code!