Output of transformer tied to earth

Other Hardware General Electronics
1

I am attempting to monitor the presence of a 28VAC output from a transformer. I planned on using a bridge rectifier and a 33uF capacitor to convert it to ~38-40VDC then a 10:1 resistor divider to reduce the 38-40VDC to a safe value for the Arduino analog input. The problem is that one side of the transformer is connected to earth ground.

I built the circuit shown in the attached diagram. If I use the meter to look at the OUT node or the top side of R2, and the ground lead is attached to the CM node, I get exactly what I expect. However my scope does not have a floating ground (it is tied to earth as well). So obviously I can't connect the ground lead to "CM" or it would make D4 and D2 useless and put quite a bit of current through D3 and likely burn up my rectifier.

So my questions are:

  1. How can I look at those nodes on the scope? I tried using two probes (one to the node of interest and the other to CM) and then using MATH to look at CH1-CH2, but that yielded some very strange results. It showed a voltage about what I expected with periodic very high voltage that is off the display regardless of the vertical scale setting. I expect something could be wrong with the MATH setting or the scope.

  2. How do I connect my Arudino to this OUT node? I assume a computer GND would be tied to earth GND at the outlet it is plugged in to. So if I plugged a USB cable from that computer to the Arduino, it would connect the CM node to Earth GND which would damage the rectifier. If I use a wall-wart power supply to power the Arduino (one that doesn't have a GND terminal where it plugs in the outlet) would that work? Or what about a laptop which isn't even connected to the earth ground at all?

2

Consider an optocoupler. No galvanic connection between the sensing circuit and Arduino, light only.

3

Much better and much safer idea: connect the transformer output to an AC optocoupler and a 5K series input resistor.

4

I have an optocoupler but it just provides a digital output and I was wanting to monitor the approximate value of the voltage. I guess I could try to build something that will pass analog signals. Thank for the input.

BTW, I found the problem with the Math on the scope. One of the two signals had the scale set incorrectly and that was messing up the Math function.

5

A 3-resistor voltage divider can be used to measure AC signals referred to ground.

Connect transformer ground to Arduino ground (assuming you can do safely).
Resistor from AC hot to Arduino analogue pin.
Resistor from analogue pin to ground.
Resistor from analogue pin to 5volt.

Resistor values must be calculated so that 28volt AC results in 5volt peak/peak with VCC/2 bias.
Then simply sample the AC wave.

Note that 28volt AC is almost 40volt peak (80volt peak/peak).
And that can be more if the transformer is unloaded.
Leo..

6

lemmy999:
I am attempting to monitor the presence of a 28VAC output from a transformer.

lemmy999:
I was wanting to monitor the approximate value of the voltage.

It always helps if you state what your really want... this are two different things!

You can indeed measure the complete AC waveform, find the peak, calculate the voltage. Lots of sampling to do.

Alternative approach: rectify (half wave is enough - single diode for just the positive half, a MOSFET should also do if you don't like the diode drop interfering with your signal, then it should be safe to also directly connect Arduino GND to earth ground on the other side of the transformer), a capacitor between diode and GND and two resistors in series - as voltage divider - in parallel to the capacitor. Other than a little ripple this will always give you the peak value.

7

Hi,
What is the 28Vac used for?
Does it already got to a rectifier.

Have you really built this circuit in real life?
Is the 28Vac winding really grounded at one end?
Usually they aren't and it is after the bridge do you connect the negative terminal to gnd.

What is the REAL WORLD application.
Use a DMM to measure the voltages.

Tom... :slight_smile:

8

wvmarle:
It always helps if you state what your really want... this are two different things!

You can indeed measure the complete AC waveform, find the peak, calculate the voltage. Lots of sampling to do.

Alternative approach: rectify (half wave is enough - single diode for just the positive half, a MOSFET should also do if you don't like the diode drop interfering with your signal, then it should be safe to also directly connect Arduino GND to earth ground on the other side of the transformer), a capacitor between diode and GND and two resistors in series - as voltage divider - in parallel to the capacitor. Other than a little ripple this will always give you the peak value.

I NEED to monitor the presence of the voltage. I WANT to measure more than just if it is there or not to see if the voltage is just getting too low or if it is completely going away.

I was going to use the single diode approach, but I knew the ripple would be much higher using this method. It isn't really a problem I guess and that would definitely be the simplest approach. I definitely fulfills the NEED and would let me know if the 28VAC is changing significantly. Thanks.

9

TomGeorge:
Hi,
What is the 28Vac used for?
Does it already got to a rectifier.

Have you really built this circuit in real life?
Is the 28Vac winding really grounded at one end?
Usually they aren't and it is after the bridge do you connect the negative terminal to gnd.

What is the REAL WORLD application.
Use a DMM to measure the voltages.

Tom... :slight_smile:

The 28 VAC is part of an HVAC system

No, it doesn't go to rectifier. I added the bridge rectifier to convert it to DC.

Yes I built that circuit and it works great. In the original post when I said I "used a meter" it was a Fluke DMM. It gave the exact result I expected. I just wanted to look at it with a scope. The CH1-CH2 method on the scope works as well, but since the CH1 and CH2 signals are fairly large, the resulting MATH signal doesn't look great. I assume due to limited number of bits for CH1 and CH2.

It really is connected to earth. I put one DMM probe in to the earth ground connection of a wall plug and the other probe on one of the transformer secondary leads and it was 0.8 ohms. I also measured from the transformer secondary to the scope problem clip and it was about 0.8 ohms as well.

Because of this I can't connect the negative of the rectifier output (CM in my drawing) to earth ground, or it would destroy the rectifier. On the negative part of the cycle, nearly 40V would be placed across the diode D3. (I actually have a bridge rectifier, not 4 individual diodes).

The real world application is I am having issues with a LUX GEO smart thermostat and the company says it is losing power and resetting. So I want to monitor the 28VAC that comes to the thermostat to see if and when it is dropping out in an attempt to diagnose my problem and have a functioning thermostat.

I have used the DMM to measure the voltages and it has worked well. I just wanted to look at it with my scope as well. Also want to connect the output to an Arduino and don't want there to be a grounding issue there.

Thanks.

10

Google shows this: "This universal system compatible thermostat offers 3-power options: 2 AA Lithium batteries, a micro-USB, or the C-wire.". Have you replaced the batteries recently? Have you tried a new thermostat?

Paul

11

" If I use a wall-wart power supply to power the Arduino (one that doesn't have a GND terminal where it plugs in the outlet) would that work? "

In the U.S. the neutral wire is connected to Earth ground at the breaker box. But the wallwart should have a transformer giving you the isolation you want.

12

lemmy999:
The 28 VAC is part of an HVAC system

It really is connected to earth. I put one DMM probe in to the earth ground connection of a wall plug and the other probe on one of the transformer secondary leads and it was 0.8 ohms. I also measured from the transformer secondary to the scope problem clip and it was about 0.8 ohms as well.

Thanks.

Why a full wave rectifier, a half-wave rectifier would let you use the transformer grounded end of the winding as your gnd reference.
Tom.... :slight_smile:

13

Paul_KD7HB:
Google shows this: "This universal system compatible thermostat offers 3-power options: 2 AA Lithium batteries, a micro-USB, or the C-wire.". Have you replaced the batteries recently? Have you tried a new thermostat?

Paul

It runs off of EITHER batteries or the 24VAC. If it runs off of the batteries the update times vary greatly and can be 10-15 min. If it runs off of 24VAC changes you make via the app will sync with the thermostat in a few seconds. So for that reason I want it to run off of 24VAC if possible. The issue I am seeing doesn't seem to happen when using batteries. Also, the company sent me a second one and it is doing the same thing. The one upstairs on a different HVAC system does not have this issue. So it really does point to a power problem.

14

TomGeorge:
Why a full wave rectifier, a half-wave rectifier would let you use the transformer grounded end of the winding as your gnd reference.
Tom.... :slight_smile:

You are correct it would. I had a bridge handy and also figured it would give me less ripple. But yes a half wave would solve all of my issues.

15

Wawa:
A 3-resistor voltage divider can be used to measure AC signals referred to ground.

Connect transformer ground to Arduino ground (assuming you can do safely).
Resistor from AC hot to Arduino analogue pin.
Resistor from analogue pin to ground.
Resistor from analogue pin to 5volt.

Resistor values must be calculated so that 28volt AC results in 5volt peak/peak with VCC/2 bias.
Then simply sample the AC wave.

Note that 28volt AC is almost 40volt peak (80volt peak/peak).
And that can be more if the transformer is unloaded.
Leo..

When I first read your post I wasn't sure and was thinking I would be better off to AC couple the signal and bias at VCC/2 with a resistor divider, but then I drew it out and calculated the values and it works great. 10k for the AC to analog pin, 1428.6 (1.43k) for the analog pin to ground and 1250 (1.24k) yields a 0-5V AC signal. I think I would tweak the resistors a bit to allow more room between 5V and 0V and also add some schottky clamp diodes.

The only issue is that I want to sample this over an entire day so if I wanted to sample the AC signal at a high right I would have quite a bit of data. I guess I could sample quickly and every time the signal went from above mid scale (512) to below mid scale, I could save the peak value sampled since the last time the samples went from above mid scale to below mid scale. So it would save the peak sample from every cycle. The code is a bit more complex, but the circuitry is much simpler and my results are very precise.

Thanks for the suggestion!

16

Pick the resistor values at least 10x higher, so you won't need clamping diodes.

Could occasionally measure peak/peak value of the AC, like this..

lowPeak = 1023; // reset to opposite extreme
highPeak = 0; // reset
startMillis = millis(); // mark
while (millis() - startMillis < 15) { // 3/4 of 50Hz wave
  sample = analogRead(A0); // read
  if (sample < lowPeak) lowPeak = sample; // update
  if (sample > highPeak) highpeak = sample;
}
sample = highPeak - lowPeak; // peak/peak value

Leo..

17

The ultimate solution to your problem is to buy two small signal transformers and wire them back to back so they isolate the 24 volts of the #2 transformer from ground. I have some removed from circuit boards that are very small, physically.

Paul

18

I think you are making too much of this situation.
The thermostat is just a switch that responds to
temperature. If I remember correctly, when the
temp is below the set-point, it closes the circuit
that activates a relay and calls for heat. So, it
is simple on or off. Go with jremington's reply #2.
Herb
P.S. I think there is no good reason for that circuit
to be earthed. It just may be a pinched wire
that causes the ground connection.

19

Half wave works wonders, too, when it comes to finding the peak value.

Make your voltage divider to get down to within the 0-1.1V range, and you can use the more stable internal ADC reference.

20

herbschwarz:
I think you are making too much of this situation.
The thermostat is just a switch that responds to
temperature. If I remember correctly, when the
temp is below the set-point, it closes the circuit
that activates a relay and calls for heat. So, it
is simple on or off. Go with jremington's reply #2.
Herb
P.S. I think there is no good reason for that circuit
to be earthed. It just may be a pinched wire
that causes the ground connection.

This is a smart/wifi thermostat. Yes it just connects the 24VAC (R) to the W to call for heat, but it gets power from the 24VAC and C wire (common). I am unable to manually change the temperature because it appears to be losing power momentarily and reverting back to the schedule. So it isn't like I am trying to do this on a basic/simple thermostat. I just wanted to see if it was indeed losing power or if the thermostat was just screwing up (this has been a rather buggy thermostat). I ended up using the circuit I posted but solved the ground issue using a 9V wallwart power supply to power the Arduino and it worked great.

This helped me track things down to a faulty blower time delay relay in the furnace. It gets the call for heat, and 25 seconds later the blower should come on, but sometimes it isn't. So the heat exchanger gets too hot and the primary limit switch correctly shuts everything down (that is why I was getting the temporary loss of power). Thanks everyone for the suggestions.