[PIC] Power factor measurement using PIC18f4520

[betwixt]

Be careful though, the switch will have to be 'closed' throughout the entire AC cycle so if a solid state relay is used it must not have zero crossing detection. If it switches at zero crossing there is a risk that the capacitor will connect and disconnect in phase with the voltage waveform and you might get strange results.

Brian.

 

[c_mitra]

Be careful though, the switch will have to be 'closed' throughout the entire AC cycle so if a solid state relay is used it must not have zero crossing detection.

This is *very* important and the time constant for connection and disconnection must be at least 10 periods or cycles of the mains frequency. The software must keep track (of the turning on and off of the relay) and ignore power factor computations after 2-3 cycles of turning on the relay (the system must be given time of at least 1 cycle to stabilize) and must take action only after the power factor stays low for 2-3 periods or cycle.

 

[asimov_18]

You can do it many ways. All of the above are valid methods but I would suggest there is a better way of using the optocoupler that avoids the high power rated resistor and is also more efficient. What isn't clear is those schematics show a 24V AC input but I'm guessing you really have 220V AC and they do not show how the voltage is dropped to 24V. If you use a reactive circuit, for example another transformer, it will introduce a phase shift of it's own which only complicated things further.

I suggest you do it this way:
Connect the 220V AC input to two 47K 0.5W resistors, one in each wire (for safety, they are electrically in series anyway) then to a bridge rectifier with it's output wired to the LED side of the optocoupler. So its similar to the final schematic in your last post but the bridge rectifier is on the LED side of the resistors. By moving the bridge you no longer have to use high voltage diodes, you can use small signal diodes (1N4148 for example) because the LED clamps the voltage across them to no more than about 2V.

The optocoupler method has almost zero phase shift and is entirely voltage operated so it gives a good reference to reset the delay counter.

Brian.

Hi Betwixt,
Since this topic related to zero crossing is being discussed I was wondering if it is possible
to use a transformerless capacitive power supply for driving the optocoupler LED after removing the
filter capacitor at the output stage. This design is very common
for example: https://ww1.microchip.com/downloads/en/AppNotes/00954A.pdf

I know there would be a phasing shifting involved in this but if this phase shifting can be accounted for
then its not a problem and accuracy can be achieved. Your opinion about the same would be highly
appreciated.

Regards
asimov

 

[c_mitra]

I was wondering if it is possible
to use a transformerless capacitive power supply for driving the optocoupler LED after removing the
filter capacitor at the output stage.

It is certainly a good suggestion to replace the transformer with a capacitor diode and a zener but for phase calculation, but we need two voltages that are proportional to the voltage and current respectively. I fail to see how we can get voltage and current readings from a constant voltage power supply that can of course drive the optocoupler LED.

We need not bother about the zero crossing if we can capture about 10 points within a complete period (any 20 ms time interval would be enough for a integration).

 

[betwixt]

Since this topic related to zero crossing is being discussed I was wondering if it is possible
to use a transformerless capacitive power supply for driving the optocoupler LED after removing the
filter capacitor at the output stage.

It's easier than that, to generate zero crossing pulses just take the AC, pass it through a bridge rectifier, a series resistor and then to the LED. Any voltage above LED Vf will light it, only the brief period when the voltage is close to zero crossing will turn it off.

To generate a transition from high to low and low to high at alternate zero crossings, use a series resistor then a reversed ordinary silicon diode across the LED, it will light on one half cycle and turn off on the other.

Both these methods do not intoduce any phase shift but there is a short time either side of zero crossing where the voltage will not be high enough to operate the LED. The time will be fairly constant though so it is easy to compensate for it in measurements.

Brian.

 

[c_mitra]

If adding parallel capacitors would not work, shall I order to make a 10,000/50VAC transformer, and use its secondary as the choke, by assuming that its first inductance is of 8H, then the secondary inductance is of few mH? Or at 10KV the inductance is far from 8H?

Are there any possible solution to use same secondary as the choke? From seeings, it is clear that its inductance is not in the milli Henries range. What to do in such case?

Finally, If this was used, and resultant ampere was in milli-amps. When branching a parallel capacitor, I would see less milli-amps, would that be an effective demonstration of controlling power factor, or I should keep thinking about several amps drawing such that its reduction would be more profound ?

How are the things coming out?

 

[KhaledOsmani]

How are the things coming out?

Hello c_mitra,

First of all, there is an error in the RC filtering circuit for phase calculation: Large electrolytic capacitors within are taking too much time for averaging, what results in an output that is not in real-time. i.e the RC circuit takes an image of the proportional to voltage output way beyond its supposed output. Delayed image in other words.

Second, I couldn't find/make a choke of 50mH to test what BradtheRad made as simulation. I did made many chokes/autotransformers, but all their impedances were greater beyond than 50mH. This results in a meaningless compensation.

Meaningful power factor correction, takes place for big loads, and as you proposed all motor-pumps rated at >1HP are best to seek this correction.

Other than this, project is working fine. :lol: :lol:

Thank you

 

[c_mitra]

I often used to tell my students that wild goose chases are very interesting from a scientific point: in research, catching the goose is not the ultimate aim; the objective is to study and learn their behaviours under different conditions. These knowledge will come useful and handy in some other conditions, perhaps unexpectedly.

 

[BradtheRad]

Second, I couldn't find/make a choke of 50mH to test what BradtheRad made as simulation. I did made many chokes/autotransformers, but all their impedances were greater beyond than 50mH. This results in a meaningless compensation.

A plain inductor is the obvious load to use, in order to see waveforms which have to do with power factor.

If you can be flexible with the frequency, then your range of options is broader.
One experiment worth trying:

(1) Make an LC tank circuit. (The capacitor should be non-polarized.)

(2) Apply a sine sweep, while looking for the resonant frequency.

(3) When you find it, detach the capacitor.

(4) This leaves only the inductive load. This produces the same situation which you are designing your project to detect, namely power factor error.

(5) Observe waveforms. You'll probably see current lagging voltage by 90 deg (or almost).

(6) Reconnect the capacitor, and observe that it restores power factor (current waveform coincides with voltage waveform).

(7) Try different C values, and observe the different amounts of power factor correction.

 

[KlausST]

Hi,

I'm sorry to read this.

there is an error in the RC filtering circuit for phase calculation: Large electrolytic capacitors within are taking too much time for averaging

The RC filters for U, I and pf should be the same. All with the same time to settle. A raw estimation is about 200ms.
Calculated to meet a typical update rate of a measurement device display.

Klaus