16 replies to this topic

[cadcoke5]

I found a few threads on this, but they were quite old, and many of links were no longer valid.

 

I need to monitor power going to a heater, 120v AC at approx 100 watts.   It should be mostly resistive, but ideally I would allow for some non-resistive component.

 

Since this is for a temporary testing situation, ideally it would just have a household plug on one end, and a household socket on the to plug in the heating device. An example of this form factor might be the $25 "Kill AWatt" power monitor. But it is only a display, and some have commented about it reporting as much as 50% under the true value for items that are under 60 watts.  I have been unable to find published accuracy data for this or other similar devices.  If the sensor were not packaged as a unit, like the "Kill A Watt", I could make my own box to house the device and just wire in some power cords.

 

I suppose I could just obtain a simple 120v to 5v transformer, and wire it in, but I was hoping for something with more processing to be able to adjust for variations in the power supply. There was a thread that suggested using AD536 by Analog Devices. But, I am just not wanting to go to this level of electronics design, and really prefer a single component solution.

 

There is the $10 Non-Invasive Current Sensor from Spark Fun.  But, my electronics training is about 25 years old, and I would rather not try to design a circuit that will make this device work.

https://www.sparkfun.../products/11005

 

Perhaps there is something like this that has already built-in the components to provide a voltage range based on the current draw?

 

Any products to suggest?

 

-Joe

[LabJack Support]

1.  For a proper power/energy measurement that does not make any assumptions, you need to acquire voltage & current waveforms with appropriate sensors, multiply them to get the power waveform, and then do whatever math you want on the power waveform.  Or buy some sort of sensor/transducer that does that.

 

2.  The first simplification of that process would be to assume voltage is some particular value (120 VAC RMS, perfect sine, and in phase with current) and just acquire a current waveform or a True-RMS reading from a True-RMS current sensor.

 

3.  The further simplification would be to assume voltage as above, plus assume current is a perfect sine.  This allows you to use a current sensor that provides RMS output and True-RMS is not required.

 

 

To look for sensors I would search variations of "ac voltage/current/power sensor/transducer".  For example, I searched "ac power transducer" and found a $54 option at Amazon that might work for you:

 

http://www.amazon.co...e/dp/B006K3LAEG

 

Some other possible sources that came up are Yokogawa, Ohio Semitronics, ABB, and of course Omega:

 

http://www.omega.com...transducer.html

 

 

Related topics:

 

https://forums.labja...?showtopic=5519

 

https://forums.labja...?showtopic=4856

 

https://forums.labja...?showtopic=4263

[cadcoke5]

I am not sure if my post went through. I got a message that I had to log in, which I did, but I did not see any sort of confirmation that my post was accepted.  I know posts must go through a moderator first, so it is not possible to see if the post was successful my looking at the forum.  So, sorry if this is a repeat.

 

I am not sure if I am understanding the terminology on the Omega web site, describing their power transducers.  It says it does "true RMS", but is that the same as adjusting for "power correction" that is done for inductive or capacitive loads?

 

-Joe

[LabJack Support]

We just got the one post above from 10:16am.  No duplicate.

 

True-RMS means that it does not assume the waveform is a sine ... it gives the correct RMS for any shape waveform.  Post a link to something of interest and we will check it out.

[cadcoke5]

I just got the same error when I tried to reply. Apparently the board is letting me type a reply without being logged on.  Then when I hit "post" it takes me to a log-on screen, but after doing so it  takes me to an error screen. When I hit the back button on my browser, it only takes me to the forum, and not the message I was viewing before.  It looses what I typed.

 

Here is my reply;

Please pardon my poor understanding of non-resistive loads.  Does a "True RMS" power meter do the work necessary to measure the power in an inductive load?

 

-Joe

[cadcoke5]

I replied a few days ago, but have not seen the post, so I will try again.  It seems you must first log in, and even though there is a reply window that apparently works, anything you type there before you log in will just disappear.  Here is my question;

 

I am not clear if "True-RMS" means the device can measure true power in an inductive load? Does it?

 

-Joe

[LabJack Support]

Because your system is mostly a resistive system a True-RMS type sensor should work just fine.  You will get some error if you plug in several inductive load devices.  However depending on how large the resistive load is versus the devices applying inductive loads you should be fine.  Any setup you use will have some error associated with it and this difference may alter your readings by a relatively small amount.  You may be interested in reading through the below page,

http://openenergymon...er-introduction

 

True-RMS sensors calculate "Apparent Power".  Some of these sensors will attempt to compensate for the addition of inductive load devices however not all will.

 

To calculate Real Power, using something like Sparkfun's current sensor and the referenced circuit for calculating voltage or a transformer is the best way to go.  This will compensate for any resistive and/or inductive devices you want to plug into your system.  It will require some extra circuitry.  I don't know of any sensors that just do this for you.  Sparkfun and the open source electronics community have a plethora of example material to help you get started making a circuit with that sensor.  LabJack devices are great alternatives to the Arduino as they will remove one extra step of unknown's in your system as each device has provides calibrated voltage readings to help reduce errors in your calculations.

[LabJack Support]

I updated some forum settings so the "Reply to this topic" window is only available when logged in. This should help prevent a lost post when not logged in. Thanks for bringing this up.

 

As for your post not displaying immediately, that was due to moderation and a delay in approval. Posts made late Friday and on the weekend may not be approved/visible until Monday.

[cadcoke5]

Thank you for the link to the true power explanation. The challenge remains about how to measure true power on an AC circuit.  The "Open Energy Monitor" is a great design, but it seems to be designed just for communications with an Arduino via a radio module.

 

If anyone knows of an easy-to-use sensor for reporting true power to the LabJack, please recommend it here.

 

-Joe

[LabJack Support]

One thing I notice on the open energy monitor is that it seems to use a wall-wart to step down the ac voltage for measurement.  I am not sure if it is a special wall-wart, but in general such wall-transformers are not particularly accurate ... a problem that is often worse the more non-sinusoidal the waveform is.

 

Does a "True RMS" power meter do the work necessary to measure the power in an inductive load?

 

 

What you often see is a True-RMS current sensor.  Such a current sensor can be used to properly calculate power if you know/measure the magnitude of the voltage and the voltage waveform meets 2 conditions:  1. It is a pure sine, and 2. It is in phase with the current.

 

If someone is claiming a True-RMS power sensor, I would expect that it provides actual power regardless of the shapes & phase of voltage & current.

 

For the power sensor from Amazon linked above, we posted a question there but not seeing a response yet it sounds like that sensor is for sines only.

 

 

If anyone knows of an easy-to-use sensor for reporting true power to the LabJack, please recommend it here.

 

 

Not necessarily easy-to-use, but the correct method is #1 from post #2 above.

 

I will look around a little and see if I can find a COTS sensor.

[LabJack Support]

Contacted a bunch of potential sources I found including Yokogawa, Ohio Semitronics, TransData, Measurement Technologies (mtech.biz), and CR Magnetics.  So far CR Magnetics is the only one that responded, and they do not seem to have a power sensor that will provide a DC output correspond to proper power for AC current & voltage waveforms with any shape and phase difference.

 

CR Magnetics does have the simple CR9580 series of current sensors available on Amazon for $30 to $35.  They appear to provide a simple 0-5V DC output that corresponds to 0 Amps to Max Amps (10A, 20A, or 50A):

 

http://smile.amazon....y/dp/B005CWOOPS

 

http://smile.amazon....y/dp/B005CWOPCU

 

http://smile.amazon....y/dp/B005CWOQV0

 

They just measure current, and are average-based RMS, not True-RMS, but look like a good choice for applications where accuracy requirements are not so stringent or where it is known that the current is sinusoidal and in-phase with a known sinusoidal voltage.

[LabJack Support]

Amazingly we have not gotten any response from any of those companies, except for CR Magnetics who did respond right away.

[Paul.L.]

I spent some time looking for CTs for measuring 120V AC and eliminated a couple CTs that had DC outputs but also had load resistance greater than allowed for a U3, App A. 

The 20Amp CT linked above seems to have a load resistance of 1 mega Ohm. Will they, the 3 CTs above, work with a U3-HV ? If so I'm misunderstanding something about Input Impedance and Source Impedance. Please elucidate. 

 

Product description from Amazon:

The CR9580 Series Current Sensors provides a cost effective method for monitoring electrical current. The sensor generates a 0-5 VDC signal proportional to the input AC current. The output signal is average sensing, calibrated to RMS. The sensor is used with process control and industrial instrumentation equipment. Especially suited for OEM applications that require a low cost solution for numerous monitoring locations. The DC output can be connected directly to an analog input connection without additional signal conditioning. Care must be taken to ensure the burden impedance of the instrumentation is greater than 1.0 megohm. The unit will operate with lower burden impedance but at reduced accuracy. Accuracy: +/-0.5 percent full scale (FS). Ripple: 1 percent Max. Signal out: 0-5 VDC. Max. Signal out: 12 VDC. Frequency: 50 to 400 Hz. Insulation class: 600 V. Operating temperature: -30 degree C to + 60 degree C. Storage temperature: -55 degree C to + 85 degree C. Dielectric withstand: 2500 Vrms. Response time: 250 ms. max. 10-90 percent FS. Calibration: Avg. sensing, RMS calibrated. Output load: 1.0 Megohm or greater for rated accuracy.

[LabJack Support]

My interpretation is that this CT has a passive circuit consisting of a rectifier and filter to convert its AC current signal to a DC voltage signal.  That circuit must present a load resistance of about 5 kohms.  If you add a load resistor in parallel of 1 Mohm, that load resistor will create a ~0.5% error.  If the load resistor was 50 kohms, you would create a 50% error.  If the load resistor was 10 Mohm, the error would be ~0.05%.

 

The input impedance of a high-voltage analog input on the U3-HV is about 1.3Mohm, so you should be good:

 

http://labjack.com/s...uide/appendix-a

[Paul.L.]

It looks like your describing 3 circuits with 2 resistors each:

 

resistor 1; 5k ohms + resistor 2; 1M ohms :: error of ~0.5%  ;;; this is what you think is in the CR9580 ?

 

resistor 1; 50k ohms + resistor 2; 1M ohms :: error of ~ 50%

 

resister 1: 5k ohms + resistor 2: 10M ohms :: error of ~0.05%

 

From App A; what is the difference between Input Impedance and Source Impedance ?

[LabJack Support]

From the frame of reference of a U3 analog input:

 

Input Impedance:  The impedance some source sees as it drives a U3 input.  The impedance looking into the U3 input.

 

Source Impedance:  The impedance of some source connected to a U3 input.

 

 

My examples all assume the CR9580 acts like a 5 kohm source impedance.  Assume it consists simply of a 5k resistor.  The current from the CT goes through that resistor to produce a voltage.  To measure that voltage you connect either side of the resistor to AIN and GND.

 

If the impedance from AIN to GND is infinite, then all current flows through the 5k resistor as designed and you get the expected voltage.

 

If the input impedance from AIN to GND is 5k, then half the current flows through the sensor 5k and half the current flows through the AIN 5k.  Since only half the current is flowing through the sensor 5k, you only get half the voltage, and thus there is 50% error.

[cadcoke5]

I eventually did come across a plug-in current sensor for a standard US household outlet.

 

$69.73  CURRENT SENSING MODULE Sends a 12VDC signal when power is

turned on or off ITEM #: 81822

http://www.smarthome...ing-module.html

 

It is way overpriced for what it is, which I think is typical for most of their items.

 

For another project I was hoping to get a similar item that I can plug in a 240v plug used on clothes dryers in the U.S.  The idea was to make a system that would report to the user that the dryer has stopped running.  All the sensors I have come across require the user to mess with the electrical connections, and many users will be nervous about that.  A simple plug-in unit is what I am seeking.  Though, it will be complicated by the fact that there are actually a few different standards for the 240v appliance plugs.

 

Thanks for any leads on a plug-in 240v sensor.

 

-Joe