 |
Search Sci.Electronics.Basics |
|
|
|
|
|
|
 |
|
|
Sci.Electronics.Basics -> Questions about Current Transformers:
There are 11 messages in this thread.
You are currently looking at messages 1 to 11.
|
Author: theCaseDate: 20:32 27-08-06
|
|
I've got a tentative plan to try and log power usage in my home, here's
a summary of the plan:
1) Buy two current transformers like these:
http://www.crmagnetics.com/newprod/ProductView.asp?ProdName=CR3110
2) Place one on each "hot" leg of the 240 volts coming into the house,
leaving the ground alone.
3) Place an appropriately sized burden resistor in parallel with each
transformer (this would be done prior to placing the transformers in
place as I understand the possibility of developing high voltages if I
ever have an "open" across the transformer).
4) Sense the voltage using a cheap A/D converter like this;
http://www.dataq.com/194.htm
5) Write some VBA code inside Excel for the conversion of the voltage
to power, data analysis, and logging.
My questions are:
a) Do I need an external amplifier / power supply for any of this, or
would the voltage developed be sufficient to sense through the
transformers directly?
b) Would I need a capacitor/diode across the transformer to rectify the
AC and provide an "average" voltage as opposed to the 60Hz ripple? If
so, any guesses appreciated as to the size needed. I'm thinking the
sampling rate may be around 50/sec.
c) Could I expect reasonable results without getting into lead/lag
power factors and all that crazy math stuff? This is just for an
ordinary home circuit; I'm not running welders or 60 hp pumps.
d) Does George Bush have a clue about what to do in Irag? (Just kidding
on this one, I already know the answer)
This is just a fun for-my-own-benefit type of project. One of those
where the hacking/ doing is more fun than the end results. I also know
enough about electricity/basic electronics to have a healthy respect
for working with the voltages involved and understand that messing
around in the breaker box can be dangerous. And finally I plan on
doing the calibration/testing outside the breaker box using maybe a
toaster to help calibrate it.
|
|
|
|
Author: John PopelishDate: 21:02 27-08-06
|
|
theCase wrote:
> I've got a tentative plan to try and log power usage in my home, here's
> a summary of the plan:
>
> 1) Buy two current transformers like these:
> http://www.crmagnetics.com/newprod/ProductView.asp?ProdName=CR3110
>
> 2) Place one on each "hot" leg of the 240 volts coming into the house,
> leaving the ground alone.
>
> 3) Place an appropriately sized burden resistor in parallel with each
> transformer (this would be done prior to placing the transformers in
> place as I understand the possibility of developing high voltages if I
> ever have an "open" across the transformer).
>
> 4) Sense the voltage using a cheap A/D converter like this;
> http://www.dataq.com/194.htm
>
> 5) Write some VBA code inside Excel for the conversion of the voltage
> to power, data analysis, and logging.
>
> My questions are:
>
> a) Do I need an external amplifier / power supply for any of this, or
> would the voltage developed be sufficient to sense through the
> transformers directly?
You may not need an amplifier, since the transformer is capable of up
to 10 volts output across the burden resistor. Unfortunately, the
output is an AC signal, which will be hard to digitize with the 194.
> b) Would I need a capacitor/diode across the transformer to rectify the
> AC and provide an "average" voltage as opposed to the 60Hz ripple? If
> so, any guesses appreciated as to the size needed. I'm thinking the
> sampling rate may be around 50/sec.
You can connect a bridge rectifier directly to the transformer, and
add a burden resistor and capacitor across its output. This will
produce a proportional DC output that is much better suited for
digitization.
> c) Could I expect reasonable results without getting into lead/lag
> power factors and all that crazy math stuff? This is just for an
> ordinary home circuit; I'm not running welders or 60 hp pumps.
The signal from this device tells you the current in the conductor.
It does not tell you the power being consumed, without being processed
in combination with voltage signals.
> d) Does George Bush have a clue about what to do in Irag? (Just kidding
> on this one, I already know the answer)
Everyone knows the answer except for George.
|
|
|
|
Author: John O'FlahertyDate: 01:05 28-08-06
|
|
theCase wrote:
> I've got a tentative plan to try and log power usage in my home, here's
> a summary of the plan:
>
> 1) Buy two current transformers like these:
> http://www.crmagnetics.com/newprod/ProductView.asp?ProdName=CR3110
>
> 2) Place one on each "hot" leg of the 240 volts coming into the house,
> leaving the ground alone.
>
> 3) Place an appropriately sized burden resistor in parallel with each
> transformer (this would be done prior to placing the transformers in
> place as I understand the possibility of developing high voltages if I
> ever have an "open" across the transformer).
>
> 4) Sense the voltage using a cheap A/D converter like this;
> http://www.dataq.com/194.htm
>
> 5) Write some VBA code inside Excel for the conversion of the voltage
> to power, data analysis, and logging.
>
> My questions are:
>
> a) Do I need an external amplifier / power supply for any of this, or
> would the voltage developed be sufficient to sense through the
> transformers directly?
>
> b) Would I need a capacitor/diode across the transformer to rectify the
> AC and provide an "average" voltage as opposed to the 60Hz ripple? If
> so, any guesses appreciated as to the size needed. I'm thinking the
> sampling rate may be around 50/sec.
>
> c) Could I expect reasonable results without getting into lead/lag
> power factors and all that crazy math stuff? This is just for an
> ordinary home circuit; I'm not running welders or 60 hp pumps.
>
> d) Does George Bush have a clue about what to do in Irag? (Just kidding
> on this one, I already know the answer)
>
> This is just a fun for-my-own-benefit type of project. One of those
> where the hacking/ doing is more fun than the end results. I also know
> enough about electricity/basic electronics to have a healthy respect
> for working with the voltages involved and understand that messing
> around in the breaker box can be dangerous. And finally I plan on
> doing the calibration/testing outside the breaker box using maybe a
> toaster to help calibrate it.
In the 1980s I did something very like this, but digitized each
circuit leaving the fuse box separately. I found some surplus toroidal
8mH chokes with a hole in the middle that would just allow a 12 gauge
wire through. I put one on the load side of each breaker, and wired all
the toroids to a couple of 16-channel multiplexers.The voltage out of
these was rather low, and it had a lot of harmonic content, but it was
still pretty well proportional to the current in the circuit.
After being digitized by the 6502 system at the fusebox, the output
went through a serial link into an Apple II computer, which mulched the
data into a form that fed an MX-80 printer, with the output simulating
a multi-channel strip chart recorder that updated at a controllable
interval.
One of the problems I had was to digitize all the channels quickly
enough, and there would have been a problem with an absolute value
circuit and LPF after the multiplexer, because of the time needed to
settle to the value of each channel. What I finally did was to simply
digitize the 60Hz AC directly, and find the average value of each
channel in software, which could be done in a maximum of 1.5 cycles of
the line, or about 10 ms. All it needed was to find a zero crossing,
and sum the absolute values until the next crossing in the same
direction. I wasn't looking for phase or power factor information
either, just current.
--
John
|
|
|
|
Author: BobGDate: 13:00 28-08-06
|
|
> theCase wrote:
> > I've got a tentative plan to try and log power usage in my home, here's
> > a summary of the plan:
==============================
Plan B might be to get The Energy Detective... current transformers
clip on to the wires in the breaker box, attractive LCD display inside
reads projected monthly total....
|
|
|
|
Author: John O'FlahertyDate: 13:29 28-08-06
|
|
John O'Flaherty wrote:
{snipped}
>What I finally did was to simply
> digitize the 60Hz AC directly, and find the average value of each
> channel in software, which could be done in a maximum of 1.5 cycles of
> the line, or about 10 ms.
Whoops! 1.5 cycles or 25 ms.
--
John
|
|
|
|
Author: theCaseDate: 21:03 28-08-06
|
|
John Popelish wrote:
> You can connect a bridge rectifier directly to the transformer, and
> add a burden resistor and capacitor across its output. This will
> produce a proportional DC output that is much better suited for
> digitization.
Any idea what a ballpark range would be for the capacitor?? I could
muddle through it I'm sure, my guess would be between 2 & 10 mF?
> The signal from this device tells you the current in the conductor.
> It does not tell you the power being consumed, without being processed
> in combination with voltage signal.
Yea, I just planned on P=I*E as an approximation. I doubt if I'd be
able to correlate voltage and current at the same time, I'll probably
use 115 volts RMS x my current sample.
Thanks for your comments!
|
|
|
|
Author: theCaseDate: 21:10 28-08-06
|
|
BobG wrote:
> ==============================
> Plan B might be to get The Energy Detective... current transformers
> clip on to the wires in the breaker box, attractive LCD display inside
> reads projected monthly total....
Yes! The Energy Detective looks like a really neat device, I
wouldn't mind getting one, but the big drawback for me is that one
can't connect to a PC to log data. Plus! there's the "hacking is fun"
part of this that I'd miss out with an out-of-box product.
I still might buy one, especially if they do open it up to logging. I
guess one could rig up a webcam to look at the display, use some image
recognition software to convert to text, and then write some code to
write that out to a text file....
Anyway, thanks for your response.
|
|
|
|
Author: John PopelishDate: 21:30 28-08-06
|
|
theCase wrote:
> John Popelish wrote:
>
>
>>You can connect a bridge rectifier directly to the transformer, and
>>add a burden resistor and capacitor across its output. This will
>>produce a proportional DC output that is much better suited for
>>digitization.
>
>
> Any idea what a ballpark range would be for the capacitor?? I could
> muddle through it I'm sure, my guess would be between 2 & 10 mF?
(snip)
Well, the transformer has a 1 to 3000 turns ratio, so a peak current,
say 20 amps, would output a current of 6.6 mA. If you choose a burden
to produce something like 10 volts at that current, that would be 1.5
k. You would want the capacitor to have a time constant with that
burden resistor that is longer than a half cycle, to keep the ripple
reasonable. For instance, a time constant of .1 second would require
a capacitor of 68 uF. 47 uF or 100 uF would probably work fine.
Since the voltage is always unipolar, an electrolytic is appropriate.
|
|
|
|
Author: theCaseDate: 21:51 28-08-06
|
|
John Popelish wrote:
> theCase wrote:
> > John Popelish wrote:
> >
> >
> >>You can connect a bridge rectifier directly to the transformer, and
> >>add a burden resistor and capacitor across its output. This will
> >>produce a proportional DC output that is much better suited for
> >>digitization.
> >
> >
> > Any idea what a ballpark range would be for the capacitor?? I could
> > muddle through it I'm sure, my guess would be between 2 & 10 mF?
> (snip)
>
> Well, the transformer has a 1 to 3000 turns ratio, so a peak current,
> say 20 amps, would output a current of 6.6 mA. If you choose a burden
> to produce something like 10 volts at that current, that would be 1.5
> k. You would want the capacitor to have a time constant with that
> burden resistor that is longer than a half cycle, to keep the ripple
> reasonable. For instance, a time constant of .1 second would require
> a capacitor of 68 uF. 47 uF or 100 uF would probably work fine.
> Since the voltage is always unipolar, an electrolytic is appropriate.
Hey thanks! I was screwing around with some online calculator to get my
2/10 mF, but I now realize the resistance value I was using was way
off, your numbers make a lot more sense.
|
|
|
|
Author: Ingvar EskDate: 08:25 29-08-06
|
|
"theCase" <TomCase@gmail.com> wrote in message
news:1156816306.850348.152190@h48g2000cwc.googlegroups.com...
>
> John Popelish wrote:
> > theCase wrote:
> > > John Popelish wrote:
> > >
> > >
> > >>You can connect a bridge rectifier directly to the transformer, and
> > >>add a burden resistor and capacitor across its output. This will
> > >>produce a proportional DC output that is much better suited for
> > >>digitization.
> > >
> > >
> > > Any idea what a ballpark range would be for the capacitor?? I could
> > > muddle through it I'm sure, my guess would be between 2 & 10 mF?
> > (snip)
> >
> > Well, the transformer has a 1 to 3000 turns ratio, so a peak current,
> > say 20 amps, would output a current of 6.6 mA. If you choose a burden
> > to produce something like 10 volts at that current, that would be 1.5
> > k. You would want the capacitor to have a time constant with that
> > burden resistor that is longer than a half cycle, to keep the ripple
> > reasonable. For instance, a time constant of .1 second would require
> > a capacitor of 68 uF. 47 uF or 100 uF would probably work fine.
> > Since the voltage is always unipolar, an electrolytic is appropriate.
>
>
> Hey thanks! I was screwing around with some online calculator to get my
> 2/10 mF, but I now realize the resistance value I was using was way
> off, your numbers make a lot more sense.
>
Even if this approch is correct you will have quite big error due to the
voltage drop over the diods. If 20A gives you 10V then you loose 1.2V in the
brigde (12% error). 5A gives you 2.5V and almost 50% error. For small
currents you will not get any reading at all.
Ingvar Esk
|
|
|
|
|
|
Ingvar Esk wrote:
> > > > John Popelish wrote:
(snip)
> > > >>You can connect a bridge rectifier directly to the transformer,
and
> > > >>add a burden resistor and capacitor across its output. This will
> > > >>produce a proportional DC output that is much better suited for
> > > >>digitization.
(snip)
> Even if this approch is correct you will have quite big error due to the
> voltage drop over the diods. If 20A gives you 10V then you loose 1.2V in the
> brigde (12% error). 5A gives you 2.5V and almost 50% error. For small
> currents you will not get any reading at all.
The current out of the transformer is proportional to the current
throgh its primary wire, as long as the output voltage is low enough
that the core does not saturate. The voltage drops of the diodes have
to be included in this caveat. If the core does not saturate, the
current through the burden resistor is almost independent of the diode
drops.
Your analysis assumes the current trransformer is a voltage transformer
(voltage across its primary is independent of the load on its
secondary). It is not (it has a current through its primary that is
independent of the load across its secondary).
|
|
|
|
1 |
|
|
|
|
Contact | Electronic Portal
|
|
|
