AC Voltage Meter - Non Transformer Solution

A

Thread Starter

Arg

I'm measuring the power consumption of an AC device. I'm using Arduino/ an ACS712 cs. However, I had guessed the voltage was 230VAC as it's nominally stated.

But, I'm seeing some anomalies in the voltage values. I'm thinking I'll need something more accurate to measure the consumption. hoping to go with something transformerless. Can be too expensive and space-consuming.

Thoughts? Reccomednations? Ideas?
 
Do you want to measure AC (measure each part of the waveform) or will Average or RMS work?
I am trying to find out if you are measuring every couple of seconds and Voltage x Current. Or are you measuring V x I (100 times 60hz).
 
What about E-metering AFE (similar IC that is used in E-Meters) ? It can use shunt as current sensor, resistive divider as voltage sensor and provide your host MCU with full information on consumption - active/reactive/apparent energy, RMS voltage & current and so on... Such ICs are manufactures by ADI, ST, Silergy and many other vendors from China... The only issue to take into account is that it will require isolated power isolated communication interface as connected to Live voltage..
 
I'm measuring the power consumption of an AC device. I'm using Arduino/ an ACS712 cs. However, I had guessed the voltage was 230VAC as it's nominally stated.

But, I'm seeing some anomalies in the voltage values. I'm thinking I'll need something more accurate to measure the consumption. hoping to go with something transformerless. Can be too expensive and space-consuming.

Thoughts? Reccomednations? Ideas?
Most electrical safety standards I am familiar with allow a direct connection to mains via a couple of series connected resistors in each and every leg of the circuit separating the mains from your non mains circuit. The resistors should be rated for the peak voltage, be fusible types and of sufficient value that even if one of the two series resistors were to go short circuit the current in the remaining resistor would still be below the limit (which right now I cannot remember Grrr).
I used this technique several times and feed the resistors into a differential amplifier circuit (with a DC offset to keep the waveform within the ADCs input range) which then drove the ADC input. Some quick sampling over several waveform periods and an ISR that accumulates the square of the readings and a scheduled task to average those and get the square root (quick and easy in integer maths) and hey presto I have true RMS results and my uC circuits are still classed SELV so no other isolation required.
Just be sure that you treat the mains connected conductors appropriately, keep the creepage and clearances required and use the correct resistor types and resistor values and all will be well. From memory I used 4 x 510k resistors, 2 for Line and 2 for Neutral but if you have more than one Line or Neutral connection made this way you may need higher values to keep below the current limit (which I still can't remember).
If you don't have access to the standard I could try to find mine and look it up for you. Best if you have a copy though so you can check this for yourself.
Hope that helps.
S.
 
I'm measuring the power consumption of an AC device. I'm using Arduino/ an ACS712 cs. However, I had guessed the voltage was 230VAC as it's nominally stated.

But, I'm seeing some anomalies in the voltage values. I'm thinking I'll need something more accurate to measure the consumption. hoping to go with something transformerless. Can be too expensive and space-consuming.

Thoughts? Reccomednations? Ideas?
If you were to use a small transformer but not an instrumentation type the waveform can be so distorted by the transformer it is useless for anything more than a rough guess or the typical, dodgy and usually flawed assumption that peak and RMS voltages have a fixed and incontrovertible correlation in the raw mains waveform and that the small power transformer does nothing to upset that fact.
Unless you use a transformer designed for instrumentation I would suggest you use a different technique (see my previous response) because it sounds like you need reasonable accuracy and small mains transformers are not likely to give you that.
S.
 
Thank you.
You are welcome. Let us know how you get on.
I didn't tell you: the standard to use if you look it up, will be one of the 950 variants (EN60950, IEC950 etc etc) and that
standard will tell you the isolation barrier you need must have as a minimum a creepage distance of 6.4mm and a clearance of 3.2mm between anything mains side to everything non mains side of the isolation barrier. I still can't remember the exact limit for the current through the resistors but I am fairly sure it was in the order of 1mA or 1.5mA RMS so for the 510k resistors (with a single fault; one resistor is short cct) the current is less than 500uA. The resistors will obviously have to span the barrier and you must maintain the creepage and clearances at the same time. So no copper immediately under the resistors or within 6.4mm of the body of either resistor except for the connections at each end of each resistor. Through hole resistors are easiest to get and use in this application but make sure they are fusible types and rated for the full mains voltage or more; 230 + 10% = 253V RMS.
Hope I wasn't teaching grandma to suck eggs. I thought it better not to assume you or another reader would know the numbers etc.
 
Top