Measure RMS current (non-sinusoidal)

[Hammy]

If I wanted to measure RMS current from the mains, as well as peak
inrush, is this setup valid. Using a scope (DSO).

Schematic
http://i15.tinypic.com/4puzr15.png

Taking a differential measurement across R1 (CH1 - CH2). My question
is should the ground clips on the probes be clipped to earth ground or
left floating? The scope calculates RMS magnitude so for RMS current
the calculation would simply be

Irms (in) = (CH1 - CH2) rms/R1 ----?

OR

Convert the waveform (E/R) to current and integrate over one period?

OR am I off in left field?

To answer any questions no I do not have a true RMS meter or true
power meter, or differential probe.

Thanks for any advice.

 

[Phil Allison]

"Hammy"

If I wanted to measure RMS current from the mains, as well as peak
inrush, is this setup valid. Using a scope (DSO).

Schematic
http://i15.tinypic.com/4puzr15.png

Taking a differential measurement across R1 (CH1 - CH2). My question
is should the ground clips on the probes be clipped to earth ground or
left floating?

** Is your scope grounded to the AC earth pin or not ???

Is it a hand held one ?

The scope calculates RMS magnitude so for RMS current
the calculation would simply be

Irms (in) = (CH1 - CH2) rms/R1 ----?

OR

Convert the waveform (E/R) to current and integrate over one period?

OR am I off in left field?

** The idea is workable - but quite unsafe done your way.

Use a small step down transformer ( say 12 volt to 120 volt ) in reverse to
isolate the voltage across the 0.1 ohms from the AC supply.

Thus will be OK for any steady AC current readings - once calibrated.

However, inrush surges drawn by iron core transformers are ( surprisingly )
not AC - but consist of a brief string of current pulses all of the same
polarity. The first one is by far the largest.

So you cannot transformer couple the event without distorting the waveform.




........ Phil

 

[Hammy]

** Is your scope grounded to the AC earth pin or not ???

Is it a hand held one ?

The scope is a bench top model (Earth grounded).

** The idea is workable - but quite unsafe done your way.

Use a small step down transformer ( say 12 volt to 120 volt ) in reverse to
isolate the voltage across the 0.1 ohms from the AC supply.

Yes I know. I'll exercise the necessary precautions. L & N PCB traces
will both be insulated and amply spaced, terminations heatshrinked;
the only exposed wiring at mains potential will be the resistors leads
..One goal is to measure inrush so my leads will be connected before
power is applied.

Thus will be OK for any steady AC current readings - once calibrated.

However, inrush surges drawn by iron core transformers are ( surprisingly )
not AC - but consist of a brief string of current pulses all of the same
polarity. The first one is by far the largest.

So you cannot transformer couple the event without distorting the waveform.

....... Phil

I considered using a transformer for isolation as you suggest but I
know it will affect the accuracy.

I've done similar setups in school; input testing of a flyback, motor
current, single three phase transformers etc. but I can't remember
where the ground lead was clipped or if at all (it's been two years).
We almost always used sampling resistors and a differential probe or
scope setup as I explained.

Thanks for your assistance.

 

[John Larkin]

If I wanted to measure RMS current from the mains, as well as peak
inrush, is this setup valid. Using a scope (DSO).

Schematic
http://i15.tinypic.com/4puzr15.png

Taking a differential measurement across R1 (CH1 - CH2). My question
is should the ground clips on the probes be clipped to earth ground or
left floating? The scope calculates RMS magnitude so for RMS current
the calculation would simply be

Irms (in) = (CH1 - CH2) rms/R1 ----?

That looks OK.

OR

Convert the waveform (E/R) to current and integrate over one period?

OR am I off in left field?

To answer any questions no I do not have a true RMS meter or true
power meter, or differential probe.

Thanks for any advice.

Move the sense resistor into the N side to reduce the common-mode
voltage. "Differential" scopes just subtract the channel signals and
won't be very good at extracting a small difference from two big
voltages. And/Or do like Phil suggests, and use a signal isolation
transformer.

John

 

[Paul E. Schoen]

If I wanted to measure RMS current from the mains, as well as peak
inrush, is this setup valid. Using a scope (DSO).

Schematic
http://i15.tinypic.com/4puzr15.png

Taking a differential measurement across R1 (CH1 - CH2). My question
is should the ground clips on the probes be clipped to earth ground or
left floating? The scope calculates RMS magnitude so for RMS current
the calculation would simply be

Irms (in) = (CH1 - CH2) rms/R1 ----?

OR

Convert the waveform (E/R) to current and integrate over one period?

OR am I off in left field?

To answer any questions no I do not have a true RMS meter or true
power meter, or differential probe.

Thanks for any advice.

If you have a good DSO, then the measurement is probably good, but it must
be the type that calculates the RMS between two cursors using a
mathematical root-mean-square algorithm. I do this exact measurement for
sub-cycle and multi-cycle pulses in my circuit breaker testing equipment.
We use a Rogowski coil (air-core inductor as magnetic field sensor), which
can be calibrated to show the current waveform to an accuracy of better
than 1%, if properly integrated. You can make a crude coil by wrapping one
of your mains leads around an air-core choke. You'll probably get about 0.1
mV/ampere, so you need a good preamp. This method is more suited to very
high current measurement. Also look into Hall Effect transducers, or a DC
clamp-on current probe.

I have software that can show the waveform of a powerline current pulse,
and give readings of true RMS and peak current and time between two
cursors. The hardware is proprietary (but I could sell you an old prototype
board). If you can save the waveform in the format I use for my system, you
can use my software for analysis.

More information on my website www.pstech-inc.com, and the software can be
downloaded at www.Ortmaster.com.

Paul

 

[Phil Allison]

The scope is a bench top model (Earth grounded).

** For God's sake - buy yourself a Hall effect sensor.

Made by Lem Heme, Honeywell and others.

Just add a small +/- 15 volt DC supply and pass one of the AC wires through
the hole and you are away for AC and DC.

DC to 100 kHz bandwidth - very accurate - very safe.

100 amp peak current capability even in the cheapest models.



........ Phil

 

[Chris]

** For God's sake - buy yourself a Hall effect sensor.

Made by Lem Heme, Honeywell and others.

Just add a small +/- 15 volt DC supply and pass one of the AC wires through
the hole and you are away for AC and DC.

DC to 100 kHz bandwidth - very accurate - very safe.

100 amp peak current capability even in the cheapest models.

....... Phil

Absolutely. You're going to have a very difficult time separating out
the line neutral voltage from the measurement, and it really isn't a
good idea or safe to put a scope probe on the line. It usually
results in very large ground loop currents, and the scope won't like
it (if it survives).

In addition to Mr. Allison's sources for current sensors, try Allegro
for an inexpensive solution that has total isolation and an adequate
50KHz bandwidth:

http://www.allegromicro.com/en/Products/Categories/Sensors/currentsensor.asp
http://www.allegromicro.com/en/Products/Part_Numbers/0750/index.asp

The ACS750 operates off a single +5V supply. It's typically accurate
to within several percent, more than adequate for most applications.
If you AC couple the sensor output, you don't have to worry about zero
drift, and can easily get an accurate measurement of inrush current.
And by the way, inrush current is usually measured in peak amps. I'm
not sure why you'd want an RMS reading for a one shot event at turn-
on.

Good luck
Chris

 

[Hammy]

Thanks for the advice.

In addition to Mr. Allison's sources for current sensors, try Allegro
for an inexpensive solution that has total isolation and an adequate
50KHz bandwidth:

http://www.allegromicro.com/en/Products/Categories/Sensors/currentsensor.asp
http://www.allegromicro.com/en/Products/Part_Numbers/0750/index.asp

The ACS750 operates off a single +5V supply. It's typically accurate
to within several percent, more than adequate for most applications.
If you AC couple the sensor output, you don't have to worry about zero
drift, and can easily get an accurate measurement of inrush current.
And by the way, inrush current is usually measured in peak amps. I'm
not sure why you'd want an RMS reading for a one shot event at turn-
on.

Good luck
Chris

I know what inrush is.You didnt read my original post.I also want RMS
to determine whether I can use a $4.00 relay or a $15.00 dollar
relay,but thanks for the suggestion.I'll look into it.

 

[Chris]

I know what inrush is.You didnt read my original post.I also want RMS
to determine whether I can use a $4.00 relay or a $15.00 dollar
relay,but thanks for the suggestion.I'll look into it.- Hide quoted text -

You might be going the long way around here, as you're using a relay
to switch a transformer. Absent arcing, relay contacts are rated
according to the surface heating that will occur as current is passing
through the contacts.

Since you're using full-wave rectification, you should size your
transformer at about 1.2 times the DC current you're going to be
using. So, let's say you've got a dual 6VAC secondary, and are
shooting for an unregulated 8VDC at 1 amp at the output. That would
mean you should size your 12VDC transformer so it can supply 1.2 amps
AC. For this application, a 3A relay contact should do, because the
relay contacts have a thermal mass which should keep surface
temperature well below melting point for a turn-on surge -- except for
one thing.

Assuming you're switching the transformer primary, you're switching an
inductive load here. And when you interrupt current to the
transformer primary, you're going to be getting a good arc, which will
pit the contacts. When that happens, they don't fit together as well,
so you will probably end up with very excessive heating over a small
fraction of the original contact surface after a few close/open
cycles, which will eventually weld the contacts together. This is
about the worst thing that can happen.

Look for a relay which has a motor load rating. That usually means
the contacts are farther apart, with a heavier spring to pull the
contacts apart quicker. This also means you're going to need a lot
more coil current here.

Since you didn't mention how much current you're switching, it would
be difficult to size up a relay. But just offhand, if you're driving
the coil with a DC voltage, a Potter&Brumfield T73-series relay or a
less expensive equivalent might be a good, relatively inexpensive
place to start. I've used them successfully to switch 60Hz power
supply transformer primaries successfully several times with no
problems. Coil power is about 1/2 watt.

Good luck
Chris

 

[Hammy]

snip

The ACS750 operates off a single +5V supply. It's typically accurate
to within several percent, more than adequate for most applications.
If you AC couple the sensor output, you don't have to worry about zero
drift, and can easily get an accurate measurement of inrush current.
And by the way, inrush current is usually measured in peak amps. I'm
not sure why you'd want an RMS reading for a one shot event at turn-
on.

Good luck
Chris

Hi Chris your last post wasn't picked up by my news server, but I read
it in an alternative site. But here's an explanation of what I'm
doing. Hopefully they don't drop this, my news server has been screwy
the past couple of days.

I'm building a LM317 12Vdc @ 0.150mA PSU. My transformer is a 14VAC
4.4VA model. This is to be used to power three different devices that
are separated from the PSU by 20 to 100ft.For reasons I won't bore you
with its desirable to keep the PSU in a remote location, hence the
need for the relay, to switch the PSU on using a control signal
generated by a battery operated cct.

I'll be using a "latching" type relay. I will also have TVS from L to
N and from L and N to ground. Would this not be adequate? The other
reason I want the measurement is for sizing of the "power entry
module" which includes an EMI filter, and fusing. I know there are
rules of thumbs and semi-accurate methods for estimation "schade
curves", simulation. But I like to measure and be as sure as I can be
with what I have available, before I leave something plugged into an
outlet for extended periods of time.

I've been reading application literature from the relay manufactures,
and the TVS I'm using is what they recommend. I would have this
irregardless of the relay. Are there additional measures that I can
use?

 

[Phil Allison]

"Hammy

I'll be using a "latching" type relay. I will also have TVS from L to
N and from L and N to ground.

** It not safe or AFAIK legal to connect a TVS to the local safety ground
or exposed metal within an appliance - if the TVS fails it may send the
safety ground live.

BTW:

You have seriously wasted our time with your misleading Q that turns out to
be about a miniscule 4.4 VA tranny.

Strewth !! What inrush surge ???

How the HELL did you expect to measure * 360 uV * across that 0.01 ohms
resistor ??



........ Phil

 

[Hammy]

"Hammy


** It not safe or AFAIK legal to connect a TVS to the local safety ground
or exposed metal within an appliance - if the TVS fails it may send the
safety ground live.

BTW:

You have seriously wasted our time with your misleading Q that turns out to
be about a miniscule 4.4 VA tranny.

Strewth !! What inrush surge ???

....... Phil

You idiot the fuse would blow if it fails short!

Straight from OMRON

http://i10.tinypic.com/6bw4viw.png

So this is illegal is it?

I never stated anywhere that I was building a MW supply.I still think
its good to measure inrush no matter the supply power.

How the HELL did you expect to measure * 360 uV * across that 0.01 ohms
resistor ??

Correct I should have thought it through more,but it was the general
idea,I didnt run through a detailed calculati of component values.

 

[Hammy]

You idiot the fuse would blow if it fails short!

Straight from OMRON

http://i10.tinypic.com/6bw4viw.png

So this is illegal is it?

I never stated anywhere that I was building a MW supply.I still think
its good to measure inrush no matter the supply power.


Correct I should have thought it through more,but it was the general
idea,I didnt run through a detailed calculati of component values.

Diferential Y type Caps are also connected to ground is this also
illegal.Better recall all EMI/RFI filters.

 

[Hammy]

Diferential Y type Caps are also connected to ground is this also
illegal.Better recall all EMI/RFI filters.

Scratch differential off my above post.

 

[Phil Allison]

"Hammy"
"Phil Allison"

You idiot the fuse would blow if it fails short!

Straight from OMRON

http://i10.tinypic.com/6bw4viw.png

So this is illegal is it?

** Yep - it is very unsafe.

One cannot connect devices that are prone to fail short circuit from either
AC line to the LOCAL safety ground in an appliance. MOVs are not permitted.

The only component that is generally allowed is a class Y cap.

I never stated anywhere that I was building a MW supply.I still think
its good to measure inrush no matter the supply power.

** Then you are an even bigger IDIOT.

Correct I should have thought it through more,but it was the general
idea,I didnt run through a detailed calculati of component values.

** Was MORE than an tad misleading to the readers of this NG.

Invalidates much of the advice given.



........ Phil

 

[Phil Allison]

"Hammy"

Diferential Y type Caps are also connected to ground is this also
illegal.

** No - class Y caps are the exception as they can withstand very
high pulse voltages and are not likely to ever fail short circuit.

However, the value used must not allow more than a very small current to
flow to ground - usually less than 1 mA AC.



........ Phil