Help me pick out a power transformer.

I want a transformer to use as a voltage reference in a power monitoring application. But, I guess I just don't understand the specs on transformers well enough to actually choose one. I went to digikey and actually got two of
Nice little transformers, but they weight a lot and all I want is a voltage reference so I don't need much power. What I'm hoping to find is a transformer that is maybe 3/4 inch square and only needs to provide enough power to split across a voltage divider and monitor on an analog input pin on an arduino. The idea is to sample the AC waveform a few thousand times a second to calculate power usage, so I need the AC. A voltage around 6V or so would be nice, but it's not real important as long as it's below skin resistance to lessen the risk of shock.

Input US mains level 110 (220 would be nice though) output 6 up to maybe 20VAC. Power level, real low.

I've seen this kind of transformer in various devices, but I don't know what they are called or how to search for them.

Actually, I'm trying to put this into a device. I never thought of using a capacitor to drop the voltage, that could work, but the details might drive me nuts. The idea is to feed 110VAC into a transformer primary (with some kind of over current protection) then take the voltage out of the secondary across a voltage divider that will give me a mid point on the arduino analog input pin. That way I can measure the entire sine wave. Say I reduce it to +- .5V and have it offset from ground by a volt. See, I have the 110V reduced to .5 to 1.5 or near that anyway. This way I can use the instantaneous voltage and (separately measured) the instantaneous current to compute power usage.

The problem is a transformer that is small and I can cram into a small space. I think 3/4 inch would do it. I'll get power for this device from a dismantled wall wart most likely since those things are so darn cheap now. I'd actually love to be able to buy the guts of a wall wart, but haven't found a supplier yet.

The idea is to sample the AC waveform a few thousand times a second to calculate power usage, so I need the AC.

Assuming you find the correct transformer, how are you going to read this AC voltage with an Arduino. Analog input pins can only utilize positive voltages in the 0-+5vdc range, no negative voltages allowed. Assuming you rectify it with a diode, you know that the the wave form will to a constantly changing voltage, so reading it a thousand times a second is bound to give you a mess of numbers to deal with. Assuming you then decide to filter the rectified AC voltage and use a voltage divider if needed to maintain the 0-+5vdc measurement range of the Arduino, then there is not much sense in reading it a 1,000 a second as the time constant of the filter capacitor and the load resistance of the divider will determine how fast that sampling system can respond to actual line AC voltage variations.

You may have some more basic research to do, and carefully define your measurement objective and then determine what you will need in external components to meet that objective.

A true AC power analysis requires AC voltage, AC current, and phase angle between voltage and current to determine power factor. This is not a trivial pursuit.


You need what's called a Low Phase Shift Potential Transformer. For accurate power measurements this is essential. Otherwise your calculations could be all over the map.

OK, here's how I plan to do it. I'll divide the mains voltage down to a 2V range using a transformer and a voltage divider. Then I'll offset the -1 to +1 voltage up to vary around 1V. That will give me 0-2V peak to peak. Then I'll measure the voltage using an arduino analog pin. To measure the current I'll use a CT for something around 20A which will give me 3V to mess with for current. Then, I'll measure both of them based on a timed interrupt and do the calcs from there. Yes, there will be a phase difference between the current and voltage and I'll measure that using a purely resistive load so there is no induced phase difference giving me a correction factor to mix in with the calculations. Phase difference is only a problem if you don't allow for it.

This is essentially what the kill-a-watt does, but I want to build my own. The reason I'm doing it myself instead of just buying a kill a watt is because it would be fun. Additionally, I can take the readings and do things with them using the computing power of an arduino and my in house XBee network. And, I'm tired of limits on the kill a watt.

Oh, I guess I should give an example. Suppose I want to implement load controls for power usage. If the A/C is on and someone turns on the electric dryer, there could easily be devices that are measuring power on each of them and when it sees this situation, it shuts down one of them based on the priority of the task. I want to take load shedding to the next level and put a little autonomous intelligence in it. Especially intelligence I'm controlling, not the power company or government.