Go Down

Topic: Energy monitor values (Read 19656 times) previous topic - next topic


I dont want to kill myself, or worse, my oscope.
Good one. :))

Now I see the waveform of the 0.5VA looks exactly like my 0.35VA's waveform. The smallest had me fill a complaint about the first transformer and send it to TME, the company I ordered it from, because of the waveform and the high output voltage. They sent me a "replacement" and the results are the same. Sorry TME! I had no idea.

Looks to me the TV19's waveform is identical to the 5VA's. Based on what I've seen above (many thanks!!!) I'll order both TV19 and 6VA.


Apr 29, 2015, 03:11 am Last Edit: Apr 29, 2015, 03:21 am by AppCrash
Yes the 5VA is nearly as good as the TV19. Output is shifted by less than -200 µs (analogRead takes 112µs). This is equivalent to about 3°

The TV19 is far better for phase : not graphically measurable with my oscope, less than 0.5° (6x better) according to the manufacturer data.

Linearity is good (very little distorsion)

A larger image of the 5VA, it will show you the difference :


A 1000W heater is *NOT* a reference load
What would you suggest I use?


Excuse my ignorance. Been reading bits and pieces of these posts about the voltage transformer, and I don't get everything.
AFAIK, mains transformers are designed to have some voltage regulation.
e.g. if you feed 110% of the mains voltage in the primary, you don't get 10% more from the secondary. Bigger transformers have a "better" regulation, and AFAIK small toroids can be the worst.
And if you scope  the output of a transformer via an isolation transformer, you are doubling the problem.
Did you try two small transformers with the primaries in series and the secondaries in parallel?
Half the voltage on the primary should not saturate the core.
Measure without isolation transformer ofcourse, and mind the phase.


Apr 30, 2015, 01:32 am Last Edit: Apr 30, 2015, 01:50 am by AppCrash
"AFAIK small toroids can be the worst"

The only small toroid is a *measurement* transformer, with 20-20kHz BW, and therefore comparable to an audio transformer.

"if you feed 110% of the mains voltage in the primary, you don't get 10% more from the secondary"

It is intended for !

"Measure without isolation transformer of course"

This has already be done and shown !

>>> answered post #39 !

YHDC 0.5VA 9V directly connected to mains, attachment : pic_20_1.png. You can see it is exactly the same waveform, with or without isolation transformer. No difference. I hotlink it so you can't miss it :

YHDC 0.5VA 9V, direct to mains

", and mind the phase."

>>> answered post #53, attachment : pic_11_1-1.png.

YHDC 0.5VA 9V vs YHDC TV19 1000:1000, both connected directly to mains (remember : TV19 is *designed for voltage measurements*). TV19 is phase correct better than 30' (0.5° = 1/720 period = 28µs) according to datasheet, with 20Hz 20KHz BW. See manufacturer website. I also hotlink the pic so you can't miss it :

YHDC 0.5VA 9V vs YHDC TV19 1000:1000, direct to mains :

You can see that the results are exactly the same, without and with isolation transformer. The only difference with isolation transformer is that 230V mains is not connected to the scope and L cannot be shorted to earth through the scope. I had to show the "input" signal vs the output with direct input.

Just a question : how many transformers between the nuclear plant, the high voltage lines (63000 to 40000V), and the 230V outlet ? ;) A 150VA isolation transformer for non loaded and less than 5VA transformer won't make any difference : there is *no load to nearly no load* (only milliamps on the secondary side), so regulation and saturation are not a concern. Don't be confused by the clipped aspect of the sine. The reason is *not* the transformers. It has been explained to me that this is caused by millions electronics devices (rectifiers + capacitors), pumping current only on top of waves.

"Did you try two small transformers with the primaries in series and the secondaries in parallel?
Half the voltage on the primary should not saturate the core."

Why would I do this ? I compare a transfomer designed for voltage measurements with ordinary transformers. I am *not* trying to use a mains transformer as voltage sensor. I am *avoiding* it after I observed the crappy signal coming from a small transformer.

I just spent more than 3 weeks for testing and coding, I will not do more. The bluetooth power meter is ready for PCB, with proper probes : a current transformer as current sensor, and a 1000:1000 current transformer *designed for voltage measurements* as voltage sensor. These components are designed for, why use anything else ?

These tests were only intended for waveforms, not for tweaking inadapted transformers (in fact, I tried filtering waves a couple of days...). And guess what ? I get correct results, and waves from the two sensors are in phase on a resistive load ! Just had to programmatically shift one because of the ADC latency.


They have to be used with a resistor in series with the primary ; for testing : 3x 270k 0.25W // = 90k 0.75W (max dissipated power = 0.7 W). For the final version, it will be 39k 1W + 39k 1W in series (=78k, 2 watts).
Can you please detail why can't I go directly with the 78k variant and what's with the testing variant?


I have plenty of 270k 0.25W, did'nt care about frying them, and had to order for 2x 1 watt ... So I tested with 3x // 270k, but don't do this ! it is just because I had them by hand !

Rule of thumb : at least twice the max dissipated power : carbon resistors have a bad thermal coef ; 2 resistors in series in order to divide voltage by a factor two (max voltage / resistor, see some carbon film resistor datasheet).

78k or 90k ? 10% more Vpp at ADC input, E12 series only, minimum order 10, same value.


May 01, 2015, 04:56 am Last Edit: May 01, 2015, 05:04 am by TomGeorge
I have found the problem with using the small and PCB mount transformers is that they run hot due to their construction.
The manufacturers do not expect great efficiency, and to keep lamination noise to a minimum they weld the laminations together down two sides.

This changes the magnetic characteristics as well as causing eddy currents, this will distort the output waveform as the magnetic field is just about saturated in the core. The flat topping of the voltage waveform could be field saturation.

If you have available to check if this is the problem, try a larger transformer that hasn't got its laminations welded together.

Also is the CT inline so it is not measuring the voltage sense transformer current?

The thread is getting so long its hard to follow, so if its already been mentioned, then I stand corrected.

Tom...... :)
Everything runs on smoke, let the smoke out, it stops running....


May 01, 2015, 12:50 pm Last Edit: May 01, 2015, 01:04 pm by AppCrash
Interesting !

A few weeks ago, on EEVblog, while I was suggesting that these miniature transformers where saturating even unloaded, its been told to me that miniature cores were left slightly open to avoid this problem... Did'nt go further, as the component is molded, and it had been told they did'nt saturate when unloaded. But examining in/out in XY mode showed a nasty hysteresis curve.

A couple of years ago, while I needed some wire in order to add a winding to another (big) transformer, I dismantled a tiny outlet transformer, a classical EI. Laminations were not soldered but seemed having not been oxidized (no isolation, verified with an ohm meter). Laminations where glued or impregnated, but with large resin free surfaces from place to place. I told myself that Foucault would be happy. I played a lot with transformers a long time ago (rewinding CRT TV transformers for audio amps), and never seen such a crap !

I hesitate to dismantle one of my transformers (0.5VA is fine for a Pro Mini !). But I have some fryed wall transformers, it is worth taking them apart !

Testing all the small transformers I have, it seemed that over 3VA, there's less or no distorsion (unloaded secondary). 5 VA seems just fine. And measurement transformer is cheaper and far better ! (so why not use them ?)

Current sensor inline or not with voltage sensor is not really a problem here, as these very basic power meters are intended for high power measurements only (> 100 or 200 watts). Current signal becomes extremely noisy as power decreases, and gives only low bits on the ADC. Voltage sensor (mains transformer or measurement transformer) takes milliamps, and currents to be measured are two to three orders of magnitude greater. On my side, for test circuit, current sensor is not measuring voltage sensor current. But it is on PCB, because of easier component implementation.

P.S. : Eddy currents are a myth. There's only Foucault currents.


They have to be used with a resistor in series with the primary ; for testing : 3x 270k 0.25W // = 90k 0.75W (max dissipated power = 0.7 W). For the final version, it will be 39k 1W + 39k 1W in series (=78k, 2 watts).

Burden at secondary : recommanded value, 500 ohm.
The output level is adjusted by varying the primary series resistor. With the given values, it can be connected directly to a 5V Arduino (the same way, with 2.5V middle point). Just now : 3.32V peak to peak for 230V input (0.84V to 4.14V, so there is room for mains peaks)
So all I have to do is add the primary resistors and the TV19, connect a 500 Ohm resistor on the output and then connect it to the rest of the schematic like in the attached image?


May 17, 2015, 08:16 pm Last Edit: May 17, 2015, 08:51 pm by CrossRoads


The resistor is in series with the primary, and creates the relation between voltage and current flowing through the primary. This current is replicated by the secondary and builds up a voltage when flowing through the burden.

>>> I suspect you don't know or understand the Ohm's law. <<<

RTFD (read the f****** datasheet)

If you don't understand such basic things, immediately stop playing with mains !!!

Yes, I am rude, but I have to be. First, learn electricity if you want to build such circuits. IT IS A PREREQUISITE !!!
[edit - language please - you can be rude without being crude. Moderator]


So what did you mean when you said "directly with the 2.5V middle point"? This transformer's datasheet is the "driest" I've ever seen. Only info found about it is the description from their website:



May 18, 2015, 03:47 am Last Edit: May 18, 2015, 03:59 am by AppCrash
*ALL* information needed *IS* on this page, with the schematics : a resistor in series with primary, a burden in parallel with the secondary, the value of the current, the isolation voltage, the bandwidth, the phase distorsion, the max output before distorsion. ALL. Oh yes ! Something is missing : the weight ! There's even the formula that gives the voltage knowing the current and the burden resistance.

Only three informations are needed : the ratio (1000:1000), the rated current (5mA), and the isolation voltage (2kV). And only basic knowledge for calculation : Ohm's law. They even wrote the Ohm's law on the datasheet ! 

A datasheet is not a tutorial, and a datasheet is not a curse.

Here's a circuit that will make openenergymonitor happy, as is :

The prerequisites are :

- electricity : at least Ohm's law, a minimalist understanding of transformers, and RMS/peak current/voltage (pk to pk must fit within the MC power supply)

- safety : what are neutral, live, and earth, how electricity is wired in a house and in devices (earthed ones and double insulated ones), and why.

If you are not fluent with this, you *MUST NOT* build such a circuit. Ask someone else to do it for you, or buy a commercial device. It is (extremely) simple, but it has *NEVER* been intended for beginners.


Sorry. Now I've seen the mistake. I've put the 78k in parallel with the mains.

Still, the openenergymonitor.org states that the 5V supply should go with 470k resistors. Well, that's nothing that can't be changed.


The assembly is finished with the exact recommended parts. This assembly has an output of 1.47V at the 500 ohm ends. Instead of a 500 ohm resistor I've used 2x 1K in parallel.

Why is this happening? Is the voltage value important for the power factor measurement?

Go Up