Measuring Amps with CAI webcontrol

I'm trying to monitor the amps consumed by an electric motor using my cai board.
 
I've found these pretty cheap devices on ebay
 
http://www.ebay.com/itm/New-2pcs-design-20A-range-Current-Sensor-Module-ACS712-Module-Arduino-module-/321111721218?pt=LH_DefaultDomain_0&hash=item4ac3c00502
 
I am not sure how well this will work with cai.  It looks like it takes a 5v source and outputs 2.5v at 0 amps increasing by .1v with each amp.  So at its 20 amp max it would output 4.5v for a working range of 2.5-4.5v.
 
Any experience with this?  Is there a better way to do it?  It would certainly be nice if it used more of a voltage range on the output.

These do look like very easy to use parts and ok price.  If the motor is AC then the output will go above and below 2.5V for a range of .5 to 4.5.  Also need to consider RMS vs peak so you don't exceed the sensor range.  If DC then, yes, the range could be spread out and you should be able to find similar sensors for DC.  Sorry I don't have any model numbers handy.  If it is AC you might look up the "open energy monitor" which is arduino based.  You wouldn't have to use an arduino though and there is a fair amount of sensor information (but it is all current transformer based and only good for AC).  The open energy monitor also gives power factor and watts which can give more precise loading information.

JimS said:

If the motor is AC then the output will go above and below 2.5V for a range of .5 to 4.5.  

Click to expand...

 
Hmm, this part is confusing me.  I understand that with DC voltage if the current flows one way it goes .5 to 2.5, and if it flows the other the range is 2.5 to 4.5.  But I am not understanding how this works with an AC source.  The spec sheets say that it is for both AC and DC.  The specs sheets do not show output vs amps for AC, only DC.  I sort of assumed that the chip filtered out the one component of the AC signal so that it would show like a DC amp.  http://www.robotshop.com/content/PDF/datasheet-ebs011.pdf  But if the device shows 2.5v when no amps are flowing, how is it going to use the .5. to 4.5 range as you go through the current spectrum?
 
It is an AC blower motor that I am monitoring.  The output airflow resistance is dynamic and i am trying to control the various components of that outflow resistance to keep the fan running at its rated amps (about 10 amps).  I am feeling that a 20amp monitor should survive startup on a 10amp motor.  It is only a $3 part so I won't be too upset if it fails.  I am not trying to measure wattage so power factor isn't a concern.

Lou Apo said:

Hmm, this part is confusing me.  I understand that with DC voltage if the current flows one way it goes .5 to 2.5, and if it flows the other the range is 2.5 to 4.5.  But I am not understanding how this works with an AC source.

Click to expand...

 
I've used these for years. They are convenient but they have quite a few downsides too. Most of these sensors will operate well into the KHz range, so the output is proportional to the current at that instant. You would need to sample very quickly and do many calculations to get your AC current. I think the webcontrol board would not be able to do sufficient A->D conversions in the time required to do it accurately.
 
You would need a precision rectifier (opamp probably easiest) and level-shifting to make it useful in AC applications with the WC.

rossw said:

I've used these for years. They are convenient but they have quite a few downsides too. Most of these sensors will operate well into the KHz range, so the output is proportional to the current at that instant. You would need to sample very quickly and do many calculations to get your AC current. I think the webcontrol board would not be able to do sufficient A->D conversions in the time required to do it accurately.
 
You would need a precision rectifier (opamp probably easiest) and level-shifting to make it useful in AC applications with the WC.

Click to expand...

 
 
So your saying the value of the output will follow a sign wave up and down as the current moves back and forth?

Lou Apo said:

So your saying the value of the output will follow a sign wave up and down as the current moves back and forth?

Click to expand...

 
If your power is sinusoidal, yes. If your supply is distorted, or say from an MSW inverter, or if your load is nonlinear, then the output will be some other (complex) waveform, hence requiring many samples to calculate RMS.

Any other suggestions on how to get AC current measurement into a CAI?  I'm pretty sure CAI can't follow a 60hz analog signal and make any sense of it.

Lou Apo said:

Any other suggestions on how to get AC current measurement into a CAI?  I'm pretty sure CAI can't follow a 60hz analog signal and make any sense of it.

Click to expand...

 
Depending on how accurate you want, and how quickly you want it to follow the load, what I've done in the past has been to use a simple coil around the active lead (a few turns helps increase the output if it's low current), then fed the coil output to an opamp as an amplifier, then a precision rectifier then a filter. This can be a single quad-opamp and a few components.

How about a CR Magnetics 4310-20.  It will take the 120VAC amps (max 20) and convert it to a proportional 0-5VDC output.  
 
Otherwise you can build your own.  Diagram I saved.

sda said:

How about a CR Magnetics 4310-20.  It will take the 120VAC amps (max 20) and convert it to a proportional 0-5VDC output. 

Click to expand...

 
The problem with this circuit is its VERY poor performance at low currents.
You need around 1.2V from the coil before you get ANY output, and at low currents that can be a real problem.
 
This (below) is about the simplest, precision full-wave rectifier I know of - and its unusual input configuration means that you can actually have gain without needing an additional amplifier stage.

 
Follow this with  a sample-and-hold or even just a filter with a suitable time-constant, and you'll get quite good results.
(If you use an opamp lowpass filter, it gives you the benefit of adjusting gain for a perfect input span matching your WC to the source - eg, if you are measuring a nominal 0-1A, or 0-10A, you can set the whole system for 10.24V full scale, which makes the ADC bits fall perfectly on milliamp (or tens-of-milliamps) transitions.

ACS712 output is 0.5XVcc + Vout. For 5V supply power, when 0A, its output is 2.5V. I think you could directly output that to a 4.7K resistor, then have a sample and hold amp like this one to keep the output
http://www.linear.com/product/LF398N
http://www.ebay.com/itm/221137608627
 
This IC is very simple to wire and Ch determine how long you want the output hold time.
 
Ross's full bridge rectify circuit will have better accuracy.

A sample and hold would work IF you knew when to sample - I don't see any easy way to sync that to the AC line.  The precision rectifier would work followed by an RC filter.  But I think the solution by sda is the simplest and would give reasonable accuracy.  Even if the accuracy at low currents if poor, 10A does not seem like a low current - it is half the full scale value.  It can be calibrated with an accurate meter to determine the output voltage that corresponds to 10A. 

Just to make it clear, which it doesn't seem to be ...
 
The CR4310/CR4311 series of CT's has built in DC conversion - run the wire through the hole and connect to the +/- on the CT.  I don't know what's inside the CT.
http://www.crmagnetics.com/4310_a.pdf
 
The circuit diagram was for a roll your own using a CR3110.  I think this was more for determining if something was on/off rather than measurement.

sda said:

Just to make it clear, which it doesn't seem to be ...
 
The CR4310/CR4311 series of CT's has built in DC conversion - run the wire through the hole and connect to the +/- on the CT.  I don't know what's inside the CT.
http://www.crmagnetics.com/4310_a.pdf
 
The circuit diagram was for a roll your own using a CR3110.  I think this was more for determining if something was on/off rather than measurement.

Click to expand...

 
I gather then that the 4310 is a finished solution.  Just plug it in, correct?
 
I can't seem to find it for sale anywhere though.  Google search came up with almost nothing.  Digikey and Neward don't have it.  
 
What if I took this http://www.newark.com/triad-magnetics/cst-1020/current-sense-transformer/dp/89M4938?in_merch=Popular%20Products  and added the bridge rectifier, resistor, and capacitor per your drawings of the 3110?  

Rectifier has 0.6V barrier voltage. If the AC signal is less than that, you will not read anything. The ACS712 is designed to have zero volt at 0.5 * Vcc to overcome this problem.  Its output can be bipolar or unipolar.
 
I think could add rectifier and  large enough filter capacitor to try, get a calibrated curve by measuring few know current.