Australian AC Voltage Sensor Adapter:


Just wondering if there is the possiblity of purchasing an AC Voltage Sensor Adapter with an Australian plug from either the shop or elsewhere?

As far as I’m aware, the shop stocks only UK, US and EU style adapters. You can however use any a.c. output adapter, or a 9 V transformer (if you put it in a box). The output should be a nominal 9 V, in reality the calibration of the standard (factory-loaded) sketches is set for 11.6 V, which is the open-circuit voltage we expect. If you do go that route, you probably need a programmer to adjust the calibration (though, with the exception of phase error correction, it can be done in emonHub).

A local one was discussed here:

The link to the Jaycar product page was posted by dBC in post #3 just above it.

Curses! I knew I’d seen a mention of one somewhere, but couldn’t find it on my crib notes.

Did you ever get round to testing that Jaycar one for phase shift?

There’s no mention of the no-load output voltage, so calibration will be essential. (The no-load voltage must not exceed 14.3 V at the maximum mains voltage you experience, 260 V if dBC’s mains is anything to go by - which relates to about 13.2 V at the nominal 240 V. It’s unlikely that a nominal 9 V output will reach that on no-load.)

I didn’t, but I haven’t forgotten either. The paying job seems to be consuming significant time this year, its already May and I’m sure it was only just Christmas!

That’s the same VT I’ve been using in all my stm32 experiments with the emontxshield demo stuff (STM32 Development). So I can add some data, at least for my specimen. My no-load voltage ratio is almost exactly 1:23, i.e. when I put in 230V I get out 10V. And my overall VT+emontxshield voltage ratio is almost exactly 300. When I put in 230V I get out 766mV. The transformer divides by 23 and the shield divides by 13 (nominally), for a total division of 299 (measured at closer to 300).

My phase measurements have all been relative to the SCT013 phase error, not absolute. At 230V, 10A, I found I had a net phase error of ~4.8 degrees. I then calibrated that away using the simple ADC lag technique discussed in the stm32 thread, so now at 230V and 10A I have no phase error at all.

[Section about distortion at higher voltages retracted as I’ve now discovered much of the distortion at the higher voltages is actually coming from the calibrator, which is limited to driving just 20mA on it’s V output].

In theory I should be able to bypass the VT altogether, set the calibrator to ~10V and feed that straight into the AC jack on the shield. Then the phase error will be entirely down to the CT, so I can measure that at 10A (or any other current for that matter) and then turn those relative VT phase errors into actual VT phase errors. I’ll add that to the list of things to try when I get a chance.


I might need to set it quite a bit lower than 10V with that 10uF cap loading it up.

The 10 μF is tying the bias mid-rail voltage to GND a.c-wise, it’s not loading the c.t. (i.e. the calibrator) output.

If the calibrator output is tied to the same GND as the emonTx GND (and you can’t apply a d.c. offset), then you need to a.c. couple the output to the biased input, and in that case you’ll need a load (>10 kΩ or so) between the ADC input and the bias mid-rail to establish the d.c. conditions.

Yeh, thanks… not sure what I was thinking there… but I was clearly mis-remembering where that cap goes.

Actually, the calibrator outputs are isolated (with some common-mode restrictions). In theory, I think I ought to be able to:

  1. replace the VT with the calibrator’s V output (suitably scaled down to ~10V or so) and fully characterise the CT (that’s more like my typical use of it, since my energy monitor doesn’t involve a VT but instead uses shunts).

  2. replace the CT with the calibrator’s I output (suitably scaled down to some tens of mAs) and somewhat characterise the VT - limited by the restriction of only being able to pump 20mA into it before the V signal starts to distort.

That way there’ll only be one source of phase errors at a time as there’s only one transformer (CT or VT) involved at a time.

Sounds good to me. The common-mode voltage should only be 2.5 V maximum (5 V Arduino Shield).

Can i ask a really dumb question in regards to the original question… Is there anything wrong with buying the UK AC sensor and using an UK to AU adapter? Don’t both countries run on the same voltage?

In theory, no. In practice, it’s better to not have adapters.

Yes, and frequency.

The Jaycar adapter is 9VAC 1A Unregulated Power Supply 7DC Plugs | Jaycar Electronics also 9VAC 1A Unregulated Power Supply 7DC Plugs | Jaycar Electronics New Zealand
Voltage calibration coefficient - 300.0 (No-load = 10.4 V)
Unfortunately, we don’t have details for its phase error.

I have a sample.

Powertech MP-3027 0.92° at 230V/50Hz. Rated 1000mA 445g

There is also

DCSS AC910 2.17° Rated 1000mA 327g.

As usual, the heavier transformer exhibits lower phase shift.

These were both measured across a 13k load.

I am running with an adapter and it works fine, but obviously depends just how deep in to calibration and accuracy you want to get. I have updated my P1&P2 scales slightly and emon monitoring is now very close to what my smart meter is gathering in terms of totals.

nodename = emonpi
    names = power1,power2,power1pluspower2,vrms,t1,t2,t3,t4,t5,t6,pulsecount
    datacodes = h, h, h, h, h, h, h, h, h, h, L
    scales = 1.07,1.16,1,0.01,0.1,0.1,0.1,0.1,0.1,0.1,1
    units = W,W,W,V,C,C,C,C,C,C,p

I am running with three adapters. They were pretty much on the measured voltage via a multimeter, whereas the jaycar unit was considerably off. I don’t have the exact measurements, but all three are behaving quite consistently and reliably.

All a.c. adapters will have a similar way of specifying the output: There will be the nominal voltage at full load, which will have a manufacturing tolerance, usually a few percent. The there might be two ways of specifying the output at no-load - which is a good approximation to our usage: they might specify the no-load voltage, or they might specify the regulation as a percentage, i.e. the percentage rise in voltage above the voltage at full load.

If you use the UK adapter, that’s nominally a 9 V unit: the output is 9 V ± 5% at full load, but 11.6 V ± 3% at no-load. So the regulation is nearly 30% - not unusual for a very small transformer.