Current Transducer Problems

[Instrument_Rookie]

Hey All,
I am new to using forums so please bear with me. I am currently working for a researcher on a Solar Irrigation pivot project. I am monitoring the voltages and currents with CR Magnetics transducers. The measurements made are voltage and current coming in from the solar panels (96V, 16.7A at maximum), current and voltage charging the batteries (four 6V batteries in Series) and the current going to the pumps. There is a Outback Flexmax 80 charge controller between the solar panels and the battery and a Lorentz PS100 pump controller between the batteries and the pump. The Voltage transducers are behaving properly but the current transducers are giving step outputs of 20A, 40A, 60A and then back down 40A, 20A 0A, when minimal current is flow occurring in the system (the pumps are not running and the batteries are fully charged). The current transducers measuring the current from the solar panels seem to give reasonable values most of the time but spike up to 60A sometimes. There isn't anything in the system that would draw 60A. The following link is a graph generated from raw data collected by the datalogger (Unidata Starlogger) so that my explanations are more clear https://obrazki.elektroda.pl/68_1311616218.png. Any suggestions of how to remedy this problem would be greatly appreciated. Thank you for your time.

 

[chuckey]

Which current transducer are you using, I can only find AC ones on their website.
Frank

 

[Instrument_Rookie]

I am using the CR5210-30 transducer.

---------- Post added at 13:04 ---------- Previous post was at 12:42 ----------

The transducers can also be found at DIGI Key

 

[FvM]

How do you measure 40 and 60 A with a 30 A sensor range? Also the discrete current steps suggest a trivial software bug.

 

[Instrument_Rookie]

That is what I was wondering. The sensor has a 0-5V output range so has a sensitivity of 6 if I am not mistaken. I don't know why it would generate a amperage output greater than what the sensor can sense. The data logger software is pretty simple in that you pick a Voltage input sensor and apply a multiplier to it. I don't know how I would mess it up. It is not like campbell scientific loggers where you write the program. I am sorry that I am not much help with understanding the problem. I guess that is why I am here.

---------- Post added at 15:18 ---------- Previous post was at 15:18 ----------

The system also has 20A breakers that would blow before the 30 A range.

 

[FvM]

What's the input voltage range of the data logger? Did you check, how it reacts on voltages exceeding the range? Did you check, what's the actual saturation limit of the current transducer output?

 

[Instrument_Rookie]

The input on the channel is -5V to 5V. I have checked the transducers in the lab before using them and they agreed with what current I was putting through them (used a multimeter in series) but it was low current I never tested at the FSO. It also responded with a proper negative value when current direction was reversed. Good idea on checking higher outputs to see how it reacts. I will test that. How do I find the actual saturation limit of the current transducer output. Sadly I have tried to find information from CR Magnetics on the transducer but they do not have much for a manual.

---------- Post added at 15:42 ---------- Previous post was at 15:39 ----------

Do you think that field remenance would be an issue. They were left with current running through them without supply voltage to them for a couple days during installation.

 

[FvM]

What is the meaning of "current pump" being stuck at about 60 A? What's the corresponding logger input voltage. According to your information (+/- 30A current give +/- 5V sensor output) this value could never occur. Bu it's displayed permanently. Even if the sensors would be defective, this numbers shouldn't be seen.

 

[Instrument_Rookie]

It is the current sensed between the battery bus and the pump controller. I am not sure on the corresponding logger voltage. So are you thinking its somewhere in the logger/program?

 

[Orson Cart]

This is not something silly like the current sensors are AC only and you are trying to measure DC currents?

Or you have an ouput that is a real time replica of the through current? instead of a DC value corresponding to the RMs through?

Better post the data sheet or web link to the transducer...

 

[nsaspook]

You problem might be that you're reading the current pulses from the Solar controllers PWM output in a random manner. To get the true current you need to read the current avg over at least one entire PWM cycle. When the controller is sending max current the sensor can read the DC current with only one sample because the output is almost always on but if the controller is limiting voltage by reducing the duty cycle to a small fraction of the entire PWM window you will sometime sample the small on pulse instead of the larger off time of the cycle causing errors.

Example:

long sensor_data;
int i;

sensor_data=0;
for (i=0;i<SAMPLE_TIMES;i++) {
   sensor_data+= read_adc_current_sensor(); // sample a complete pwm cycle, could be several milliseconds
}
sensor_data/=SAMPLE_TIMES; // divide total value to get avg current

 

[FvM]

It is the current sensed between the battery bus and the pump controller. I am not sure on the corresponding logger voltage. So are you thinking its somewhere in the logger/program?

Don't know. I would check for plausible logger input voltages and corresponding log recordings.

The sensor seems O.K., I checked this before.

You problem might be that you're reading the current pulses from the Solar controllers PWM output in a random manner.

May be. It would be the data loggers job to average the input according to the recording speed. If it's behaviour is unclear, a RC lowpass may be required. But even pulses won't explain out-of-range recordings.

 

[sz-aiden]

I am using the CR5210-35 transducer.

 

[Instrument_Rookie]

Thanks for the possible solutions guys. The problem is the sensors are an hour away but I am going there tomorrow to try and troubleshoot it. I will take lots of measurements and run raw voltage inputs to see how it behaves. I am also going to bring a different logger down that I am more familiar with. I will let you know how I make out.
The charge controller uses Max Power Point Tracking to charge the batteries if that makes a difference in the Pulse Width Modulation idea.

 

[nsaspook]

I have a PIC18 hall current sensor solar energy project to monitor off-grid power generation. The waveforms I get with a hall current sensor make it pretty hard to get stable full range readings without long sampling periods and digital filtering. The actual peaks are high compared to the averaged signal. A RC filter could smooth it out but you might lose some of the original signal and accuracy. Good luck.

The top waveform is the current ADC measurement gate (software generated sync signal), the bottom waveform is the raw signal from the hall sensor at low current draws from the panels into the Charge controller.

Hall sensor PWM waveforms. | Flickr - Photo Sharing!
CLIP0036 | Flickr - Photo Sharing!

DSCN2112 | Flickr - Photo Sharing!
Project Software Page: solar-monitor - Microchip PIC18F8722 Solar battery/charge monitor MBMC - Google Project Hosting