meter accuracy with half wave loads

[JohnR66]

I wonder how accurately my meter (rotating disk type) reads loads that are
half wave due to a diode. I have some heating equipment that derives its low
mode by incorporating a diode.
Thanks, John

 

[SQLit]

I wonder how accurately my meter (rotating disk type) reads loads that are
half wave due to a diode. I have some heating equipment that derives its low
mode by incorporating a diode.
Thanks, John

Your house "rotating disk meter'?

Try googling the manufacture and see for yourself.

In my experience with utility meters for 30 plus years, your barking up the
wrong tree.

 

[JohnR66]

They don't miss a beat. They will respond (properly) to even a
single half cycle loading.

Okay, thanks. It is a engine block heater. It has a high and low mode and
the I found the low mode is achived by using a diode. I wasn't sure how the
meter reacts.
John

 

[stevenal]

These meters truly truly integrate the product of the voltage and current
waveforms. Only when response drops at higher frequencies (harmonics), will
waveform error creep in into the energy measurement.

I have wondered, from time to time, how the dc flow created by rectifiers
will affect saturation in transformers. In the old days, it was best when
ac/dc radios were plugged in correctly, dc from the various sets would all
flow in the same direction. Polarized plugs would help insure that kind of
current flow. I have never heard of problems caused by such dc flow. Does
anyone know of situations in which such current flow was a problem?

Bill

-- Ferme le Bush

But a half wave signal is full of harmonics.??

 

[Don Kelly]

----------------------------

But a half wave signal is full of harmonics.??

 

[stevenal]

----------------------------

------------------

But the voltage at the metering point will be quite clean. There will be no
power involved due to the interaction of the fundamental of the voltage and
a harmonic of the current.

I agree that's what the result should be. But to be accurate, the meter must
either filter out the non-fundamental currents, or represent them properly.
I am not aware of any intentional filtering used to isolate the fundamental.
Here is what ANSI C12.1 says on the subject:

Frequency variations in a modem power system under normal operating
conditions are insignificant. Any inaccuracies that might result from
variations that occur are entirely negligible. However, the presence of
voltage harmonics or current harmonics created by nonlinear loads may cause
measurable inaccuracies. In the vast majority of metering installations, the
accuracy is still within ±2%. Cases of severe harmonic distortion must be
analyzed on an individual basis.

Westinghouse admits that disk driving torque is not neccessarily of the same
wave form as the current and voltage at harmonic frequencies (Distribution
Handbook), but also indicates the error is usually negligible. Newer
electronic meters might be better if the cost is right. No point in spending
dollars to save pennys.

 

[Don Kelly]

----------------------------

In terms of energy measurement, load current harmonics will only consume
power if there are harmonics at the same frequency on the line. Thus, if
the
power company supplies a pure sine wave voltage, harmonic currents will
not
lead to increased power consumption. If harmonic currents lead to the
introduction of harmonic voltages an the line side of the meter, then the
harmonics will drain power.

My question was not about harmonic power. I was asking, in effect, whether
the dc current component introduced by rectifiers would cause distribution
transformers to saturate.

Bill

---------------------
A significant DC component could cause saturation problems. Whether it
would, depends on the particular transformer and the load. Distribution
transformers have their maximum efficiency in the 25-50% of rated load and
this "usually" implies some generosity in design of the core.

I should think that any significant rectifier loads would be full wave
having no DC component.

 

[Paul Hovnanian P.E.]

[snip]

I agree that's what the result should be. But to be accurate, the meter must
either filter out the non-fundamental currents, or represent them properly.
I am not aware of any intentional filtering used to isolate the fundamental.
Here is what ANSI C12.1 says on the subject:

The theoretical situation, which Don mentioned, is one where the
fundamental plus the harmonic currents are multiplied (on an
instantaneous basis) by a pure fundamental voltage waveform. Each
harmonic current multiplied by the fundamental voltage will result in a
waveform that is sinusoidal and symmetrical about the instantaneous
power (y) axis. The average over one (or more) fundamental periods will
be zero. Only a fundamental current multiplied by a fundamental voltage
will produce a resulting waveform not symmetrical about the y axis. Only
this fundamental current will contribute to a net power flow integrated
over one or more cycles.

Frequency variations in a modem power system under normal operating
conditions are insignificant. Any inaccuracies that might result from
variations that occur are entirely negligible. However, the presence of
voltage harmonics or current harmonics created by nonlinear loads may cause
measurable inaccuracies. In the vast majority of metering installations, the
accuracy is still within ±2%. Cases of severe harmonic distortion must be
analyzed on an individual basis.

In reality, all power systems will exhibit some impedance to higher
order current harmonics. Given a theoretical pure fundamental infinite
bus voltage source, the actual voltage at the metering point will
contain voltage harmonics due to the load current harmonics, plus any
other harmonic voltage drops upstream due to other loads. Now, if you
perform the same sort of instantaneous power calculation described
above, each current harmonic multiplied by the voltage of the same
harmonic will produce a net real power flow. In practice, the voltage
distortion (and subsequent voltage harmonics) is quite low and the
resulting i*e products are even smaller at higher frequencies. Keep in
mind that these products are real power flow and are doing real work at
the load, However, in spite of an electromechanical meter's poor
frequency response at these higher frequencies, there isn't much power
that gets missed.

 

[Don Kelly]

----------------------------

[snip]

I agree that's what the result should be. But to be accurate, the meter
must
either filter out the non-fundamental currents, or represent them
properly.
I am not aware of any intentional filtering used to isolate the
fundamental.
Here is what ANSI C12.1 says on the subject:

The theoretical situation, which Don mentioned, is one where the
fundamental plus the harmonic currents are multiplied (on an
instantaneous basis) by a pure fundamental voltage waveform. Each
harmonic current multiplied by the fundamental voltage will result in a
waveform that is sinusoidal and symmetrical about the instantaneous
power (y) axis. The average over one (or more) fundamental periods will
be zero. Only a fundamental current multiplied by a fundamental voltage
will produce a resulting waveform not symmetrical about the y axis. Only
this fundamental current will contribute to a net power flow integrated
over one or more cycles.

Frequency variations in a modem power system under normal operating
conditions are insignificant. Any inaccuracies that might result from
variations that occur are entirely negligible. However, the presence of
voltage harmonics or current harmonics created by nonlinear loads may
cause
measurable inaccuracies. In the vast majority of metering installations,
the
accuracy is still within ±2%. Cases of severe harmonic distortion must be
analyzed on an individual basis.

In reality, all power systems will exhibit some impedance to higher
order current harmonics. Given a theoretical pure fundamental infinite
bus voltage source, the actual voltage at the metering point will
contain voltage harmonics due to the load current harmonics, plus any
other harmonic voltage drops upstream due to other loads. Now, if you
perform the same sort of instantaneous power calculation described
above, each current harmonic multiplied by the voltage of the same
harmonic will produce a net real power flow. In practice, the voltage
distortion (and subsequent voltage harmonics) is quite low and the
resulting i*e products are even smaller at higher frequencies. Keep in
mind that these products are real power flow and are doing real work at
the load, However, in spite of an electromechanical meter's poor
frequency response at these higher frequencies, there isn't much power
that gets missed.

Westinghouse admits that disk driving torque is not neccessarily of the
same
wave form as the current and voltage at harmonic frequencies
(Distribution
Handbook), but also indicates the error is usually negligible. Newer
electronic meters might be better if the cost is right. No point in
spending
dollars to save pennys.

 

[Paul Hovnanian P.E.]

----------------------------

[snip]

I agree that's what the result should be. But to be accurate, the meter
must
either filter out the non-fundamental currents, or represent them
properly.
I am not aware of any intentional filtering used to isolate the
fundamental.
Here is what ANSI C12.1 says on the subject:

The theoretical situation, which Don mentioned, is one where the
fundamental plus the harmonic currents are multiplied (on an
instantaneous basis) by a pure fundamental voltage waveform. Each
harmonic current multiplied by the fundamental voltage will result in a
waveform that is sinusoidal and symmetrical about the instantaneous
power (y) axis. The average over one (or more) fundamental periods will
be zero. Only a fundamental current multiplied by a fundamental voltage
will produce a resulting waveform not symmetrical about the y axis. Only
this fundamental current will contribute to a net power flow integrated
over one or more cycles.

Frequency variations in a modem power system under normal operating
conditions are insignificant. Any inaccuracies that might result from
variations that occur are entirely negligible. However, the presence of
voltage harmonics or current harmonics created by nonlinear loads may
cause
measurable inaccuracies. In the vast majority of metering installations,
the
accuracy is still within ±2%. Cases of severe harmonic distortion must be
analyzed on an individual basis.

In reality, all power systems will exhibit some impedance to higher
order current harmonics. Given a theoretical pure fundamental infinite
bus voltage source, the actual voltage at the metering point will
contain voltage harmonics due to the load current harmonics, plus any
other harmonic voltage drops upstream due to other loads. Now, if you
perform the same sort of instantaneous power calculation described
above, each current harmonic multiplied by the voltage of the same
harmonic will produce a net real power flow. In practice, the voltage
distortion (and subsequent voltage harmonics) is quite low and the
resulting i*e products are even smaller at higher frequencies. Keep in
mind that these products are real power flow and are doing real work at
the load, However, in spite of an electromechanical meter's poor
frequency response at these higher frequencies, there isn't much power
that gets missed.

Westinghouse admits that disk driving torque is not neccessarily of the
same
wave form as the current and voltage at harmonic frequencies
(Distribution
Handbook), but also indicates the error is usually negligible. Newer
electronic meters might be better if the cost is right. No point in
spending
dollars to save pennys.

----------------
Not to mention that with harmonics in both current and voltage, some of the
harmonic torques may be retarding torques -to the benefit of the customer!
I note that the utilities are not overly concerned about this effect on
their billing.

Only if the power is flowing back out to the utility (this may in fact
be the case). Except for the frequency response problem, the meter will
measure the power correctly.

 

[stevenal]

----------------------------

stevenal wrote:


[snip]

I agree that's what the result should be. But to be accurate, the meter
must
either filter out the non-fundamental currents, or represent them
properly.
I am not aware of any intentional filtering used to isolate the
fundamental.
Here is what ANSI C12.1 says on the subject:

The theoretical situation, which Don mentioned, is one where the
fundamental plus the harmonic currents are multiplied (on an
instantaneous basis) by a pure fundamental voltage waveform. Each
harmonic current multiplied by the fundamental voltage will result in a
waveform that is sinusoidal and symmetrical about the instantaneous
power (y) axis. The average over one (or more) fundamental periods will
be zero. Only a fundamental current multiplied by a fundamental voltage
will produce a resulting waveform not symmetrical about the y axis. Only
this fundamental current will contribute to a net power flow integrated
over one or more cycles.

Frequency variations in a modem power system under normal operating
conditions are insignificant. Any inaccuracies that might result from
variations that occur are entirely negligible. However, the presence of
voltage harmonics or current harmonics created by nonlinear loads may
cause
measurable inaccuracies. In the vast majority of metering installations,
the
accuracy is still within ±2%. Cases of severe harmonic distortion must be
analyzed on an individual basis.

In reality, all power systems will exhibit some impedance to higher
order current harmonics. Given a theoretical pure fundamental infinite
bus voltage source, the actual voltage at the metering point will
contain voltage harmonics due to the load current harmonics, plus any
other harmonic voltage drops upstream due to other loads. Now, if you
perform the same sort of instantaneous power calculation described
above, each current harmonic multiplied by the voltage of the same
harmonic will produce a net real power flow. In practice, the voltage
distortion (and subsequent voltage harmonics) is quite low and the
resulting i*e products are even smaller at higher frequencies. Keep in
mind that these products are real power flow and are doing real work at
the load, However, in spite of an electromechanical meter's poor
frequency response at these higher frequencies, there isn't much power
that gets missed.

Westinghouse admits that disk driving torque is not neccessarily of the
same
wave form as the current and voltage at harmonic frequencies
(Distribution
Handbook), but also indicates the error is usually negligible. Newer
electronic meters might be better if the cost is right. No point in
spending
dollars to save pennys.

----------------
Not to mention that with harmonics in both current and voltage, some of the
harmonic torques may be retarding torques -to the benefit of the customer!
I note that the utilities are not overly concerned about this effect on
their billing.

Only if the power is flowing back out to the utility (this may in fact
be the case). Except for the frequency response problem, the meter will
measure the power correctly.

Paul,

Ideally, yes. But we were talking of the meter error here. If the torques at
the higher frequencies are not in phase with their respective voltages and
currents, the resultant torque at any given frequency might be negative for
a positive power flow at that frequency. An ANSI meter only needs to be
accurate within + or - 2% of fundamental.

 

[Don Kelly]

----------------------------

It is possible that harmonics are putting power back onto the line! That
is
what parametric oscillators can do. The nonlinearity of the load may
actually convert power at the fundamental frequency into power at a
harmonic. The conditions would have to be unusual, but possible.

Bill

-- Ferme le Bush

 

[Don Kelly]

----------------------------

I thought they only cared about third order and below.

Not really, regulatory standards nowadays are concerned with the first 80
harmonics.

Prior to the existence of large non-linear loads, even harmonics were not a
problem and triplen harmonics (3rd, 9th, 15th, etc) could be eliminated
without filters. 5th and 7th harmonics were virtually eliminated through
winding design of generators and motors, leaving small 11th and 13
harmonics, etc. In fact, if you use your body as an antenna and touch the
input of a scope, you will see these (emphasised) where they will not show
up with a direct 60 Hz input. Fifty years ago, harmonics and the need for
suppression was not a big issue. Now it is.

 

[daestrom]

----------------------------
stevenal wrote:


[snip]

I agree that's what the result should be. But to be accurate, the meter
must
either filter out the non-fundamental currents, or represent them
properly.
I am not aware of any intentional filtering used to isolate the
fundamental.
Here is what ANSI C12.1 says on the subject:

The theoretical situation, which Don mentioned, is one where the
fundamental plus the harmonic currents are multiplied (on an
instantaneous basis) by a pure fundamental voltage waveform. Each
harmonic current multiplied by the fundamental voltage will result in a
waveform that is sinusoidal and symmetrical about the instantaneous
power (y) axis. The average over one (or more) fundamental periods will
be zero. Only a fundamental current multiplied by a fundamental voltage
will produce a resulting waveform not symmetrical about the y axis. Only
this fundamental current will contribute to a net power flow integrated
over one or more cycles.

Frequency variations in a modem power system under normal operating
conditions are insignificant. Any inaccuracies that might result
from
variations that occur are entirely negligible. However, the presence of
voltage harmonics or current harmonics created by nonlinear loads
may
cause
measurable inaccuracies. In the vast majority of metering installations,
the
accuracy is still within ±2%. Cases of severe harmonic distortion must be
analyzed on an individual basis.

In reality, all power systems will exhibit some impedance to higher
order current harmonics. Given a theoretical pure fundamental
infinite
bus voltage source, the actual voltage at the metering point will
contain voltage harmonics due to the load current harmonics, plus any
other harmonic voltage drops upstream due to other loads. Now, if you
perform the same sort of instantaneous power calculation described
above, each current harmonic multiplied by the voltage of the same
harmonic will produce a net real power flow. In practice, the voltage
distortion (and subsequent voltage harmonics) is quite low and the
resulting i*e products are even smaller at higher frequencies. Keep
in
mind that these products are real power flow and are doing real work at
the load, However, in spite of an electromechanical meter's poor
frequency response at these higher frequencies, there isn't much
power
that gets missed.

Westinghouse admits that disk driving torque is not neccessarily of the
same
wave form as the current and voltage at harmonic frequencies
(Distribution
Handbook), but also indicates the error is usually negligible. Newer
electronic meters might be better if the cost is right. No point in
spending
dollars to save pennys.

--
Paul Hovnanian mailto[email protected]
------------------------------------------------------------------
Misery loves company, especially this one.
----------------
Not to mention that with harmonics in both current and voltage, some of the
harmonic torques may be retarding torques -to the benefit of the customer!
I note that the utilities are not overly concerned about this effect on
their billing.

Only if the power is flowing back out to the utility (this may in fact
be the case). Except for the frequency response problem, the meter will
measure the power correctly.

Paul,

Ideally, yes. But we were talking of the meter error here. If the torques
at
the higher frequencies are not in phase with their respective voltages and
currents, the resultant torque at any given frequency might be negative
for
a positive power flow at that frequency.

If the voltage and current are of opposite phase for some harmonic, doesn't
that mean power is flowing the other direction? How can the current develop
a counter torque and not be also carrying power in the opposite direction?

daestrom

 

[stevenal]

Don Kelly wrote:

----------------------------
stevenal wrote:


[snip]

I agree that's what the result should be. But to be accurate, the meter
must
either filter out the non-fundamental currents, or represent them
properly.
I am not aware of any intentional filtering used to isolate the
fundamental.
Here is what ANSI C12.1 says on the subject:

The theoretical situation, which Don mentioned, is one where the
fundamental plus the harmonic currents are multiplied (on an
instantaneous basis) by a pure fundamental voltage waveform. Each
harmonic current multiplied by the fundamental voltage will result

in
a

waveform that is sinusoidal and symmetrical about the instantaneous
power (y) axis. The average over one (or more) fundamental periods will
be zero. Only a fundamental current multiplied by a fundamental voltage
will produce a resulting waveform not symmetrical about the y axis. Only
this fundamental current will contribute to a net power flow integrated
over one or more cycles.

Frequency variations in a modem power system under normal operating
conditions are insignificant. Any inaccuracies that might result
from
variations that occur are entirely negligible. However, the

presence
of

voltage harmonics or current harmonics created by nonlinear loads
may
cause
measurable inaccuracies. In the vast majority of metering installations,
the
accuracy is still within ±2%. Cases of severe harmonic distortion must be
analyzed on an individual basis.

In reality, all power systems will exhibit some impedance to higher
order current harmonics. Given a theoretical pure fundamental
infinite
bus voltage source, the actual voltage at the metering point will
contain voltage harmonics due to the load current harmonics, plus any
other harmonic voltage drops upstream due to other loads. Now, if you
perform the same sort of instantaneous power calculation described
above, each current harmonic multiplied by the voltage of the same
harmonic will produce a net real power flow. In practice, the voltage
distortion (and subsequent voltage harmonics) is quite low and the
resulting i*e products are even smaller at higher frequencies. Keep
in
mind that these products are real power flow and are doing real

work
at

the load, However, in spite of an electromechanical meter's poor
frequency response at these higher frequencies, there isn't much
power
that gets missed.

Westinghouse admits that disk driving torque is not neccessarily

of
the

same
wave form as the current and voltage at harmonic frequencies
(Distribution
Handbook), but also indicates the error is usually negligible. Newer
electronic meters might be better if the cost is right. No point in
spending
dollars to save pennys.

of
the

harmonic torques may be retarding torques -to the benefit of the customer!
I note that the utilities are not overly concerned about this effect on
their billing.

Only if the power is flowing back out to the utility (this may in fact
be the case). Except for the frequency response problem, the meter will
measure the power correctly.

Paul,

Ideally, yes. But we were talking of the meter error here. If the torques
at
the higher frequencies are not in phase with their respective voltages and
currents, the resultant torque at any given frequency might be negative
for
a positive power flow at that frequency.

If the voltage and current are of opposite phase for some harmonic, doesn't
that mean power is flowing the other direction? How can the current develop
a counter torque and not be also carrying power in the opposite direction?

daestrom

Yes to the first question, assuming you mean power at the harmonic in
question and not total power. I'm afraid I don't know enough meter theory to
answer why to the second. Anyone else? I have referenced two sources above
that explain the phase shift and error exists, but neither details why.

 

[Paul Hovnanian P.E.]

Yes. But since a part of the line impedance which causes the harmonic
voltage drop is common to multiple customers, they will see the voltage
harmonics produced by other customers' nonlinear loads.

 

[Don Kelly]

----------------------------

Yes. But since a part of the line impedance which causes the harmonic
voltage drop is common to multiple customers, they will see the voltage
harmonics produced by other customers' nonlinear loads.

 

[Don Kelly]

----------------------------

Is this not why the equipment used by businesses have power factor
correction requirements?

I mean... a poor power factor will cost ya... then it will cost ya
again, if the power company finds out that it is you causing the
anomaly.

Poor power factor problems and power factor correction preceded harmonic
producing loads by many many years. Most power factor correction is
intended to compensate for inductive loads such as induction motors, rather
than harmonic suppression.
However, the latter may well improve the apparent (depending on definition)
power factor of loads which do have appreciable harmonics.

 

[Paul Hovnanian P.E.]

Partially.

Typical non-linear loads (rectified power supplies, etc.) tend to draw a
big spike of current near the top of each voltage sine wave. The
resulting system voltage drop flattens the tops of the sine waves, so
the peak voltage is less than sqrt(2) * Vrms. This doesn't bother
resistive or motor loads too much, but it can mess up other electronic
loads, which are more sensitive to the peak voltage.

Another problem is circulating harmonic currents in distribution
systems. Strange things can happen if there is some capacitance on the
system (pf correction banks or a lot of UG cable), some transformer
inductance and they happen to form a tuned circuit near one of the
harmonics produced by non-linear loads. Weird stuff, like high voltages
can occur.

 

[Paul Hovnanian P.E.]

Don Kelly wrote:

----------------------------
stevenal wrote:


[snip]

I agree that's what the result should be. But to be accurate, the
meter
must
either filter out the non-fundamental currents, or represent them
properly.
I am not aware of any intentional filtering used to isolate the
fundamental.
Here is what ANSI C12.1 says on the subject:

The theoretical situation, which Don mentioned, is one where the
fundamental plus the harmonic currents are multiplied (on an
instantaneous basis) by a pure fundamental voltage waveform. Each
harmonic current multiplied by the fundamental voltage will result in
a
waveform that is sinusoidal and symmetrical about the instantaneous
power (y) axis. The average over one (or more) fundamental periods
will
be zero. Only a fundamental current multiplied by a fundamental
voltage
will produce a resulting waveform not symmetrical about the y axis.
Only
this fundamental current will contribute to a net power flow
integrated
over one or more cycles.

Frequency variations in a modem power system under normal operating
conditions are insignificant. Any inaccuracies that might result
from
variations that occur are entirely negligible. However, the presence
of
voltage harmonics or current harmonics created by nonlinear loads
may
cause
measurable inaccuracies. In the vast majority of metering
installations,
the
accuracy is still within ±2%. Cases of severe harmonic distortion
must be
analyzed on an individual basis.

In reality, all power systems will exhibit some impedance to higher
order current harmonics. Given a theoretical pure fundamental
infinite
bus voltage source, the actual voltage at the metering point will
contain voltage harmonics due to the load current harmonics, plus any
other harmonic voltage drops upstream due to other loads. Now, if you
perform the same sort of instantaneous power calculation described
above, each current harmonic multiplied by the voltage of the same
harmonic will produce a net real power flow. In practice, the voltage
distortion (and subsequent voltage harmonics) is quite low and the
resulting i*e products are even smaller at higher frequencies. Keep
in
mind that these products are real power flow and are doing real work
at
the load, However, in spite of an electromechanical meter's poor
frequency response at these higher frequencies, there isn't much
power
that gets missed.

Westinghouse admits that disk driving torque is not neccessarily of
the
same
wave form as the current and voltage at harmonic frequencies
(Distribution
Handbook), but also indicates the error is usually negligible. Newer
electronic meters might be better if the cost is right. No point in
spending
dollars to save pennys.

--
Paul Hovnanian mailto[email protected]
------------------------------------------------------------------
Misery loves company, especially this one.
----------------
Not to mention that with harmonics in both current and voltage, some of
the
harmonic torques may be retarding torques -to the benefit of the
customer!
I note that the utilities are not overly concerned about this effect on
their billing.

Only if the power is flowing back out to the utility (this may in fact
be the case). Except for the frequency response problem, the meter will
measure the power correctly.

--

Don Kelly @shawcross.ca
remove the X to answer

--
Paul Hovnanian mailto[email protected]
--------
If the first attempt at making a drawing board had been a failure,
what would they go back to?

Paul,

Ideally, yes. But we were talking of the meter error here. If the torques
at
the higher frequencies are not in phase with their respective voltages and
currents, the resultant torque at any given frequency might be negative
for
a positive power flow at that frequency.

If the voltage and current are of opposite phase for some harmonic, doesn't
that mean power is flowing the other direction? How can the current develop
a counter torque and not be also carrying power in the opposite direction?

daestrom

Yes to the first question, assuming you mean power at the harmonic in
question and not total power. I'm afraid I don't know enough meter theory to
answer why to the second. Anyone else? I have referenced two sources above
that explain the phase shift and error exists, but neither details why.

At high frequencies, the voltage drop across the leakage inductance of
the V and I meter windings could be significant and will have some phase
lag between the terminal voltage or current and the flux which is
actually producing the torque. Not only is leakage inductance a factor,
but at higher frequencies, the distributed capacitance between the
windings may be a factor as well.