Phidgets Forums

i'm using a 1135 phidget voltage sensor, I'm using the Raw value for maximum accuracy using this formula

voltage=(((value / 4.095 / 200.) - 2.5)/0.0681);

yet the result is not accurate at all, i think on a 12.36v battery (tested using a good voltmeter), the phidget reports 12.73v, which is completely off, with the error % reported in the product specs why am I getting such a difference or is this normal???

cizzi wrote:i'm using a 1135 phidget voltage sensor, I'm using the Raw value for maximum accuracy using this formula

voltage=(((value / 4.095 / 200.) - 2.5)/0.0681);

yet the result is not accurate at all, i think on a 12.36v battery (tested using a good voltmeter), the phidget reports 12.73v, which is completely off, with the error % reported in the product specs why am I getting such a difference or is this normal???

i get these bogus values with both the phidget interface 8/8/8 and the arduino duemillenove

Do you have the 8/8/8 in non-ratiometric mode?

-Patrick

patrick wrote:Do you have the 8/8/8 in non-ratiometric mode?

-Patrick

i don't know, how do I check the status of that, and how do I toggle it?

Well, depends on your language, but: The default is true, as most sensors are ratiometric, but 1135 is non-ratiometric.

-Patrick

patrick wrote:Well, depends on your language, but:

Code: Select all

ifkit.Ratiometric = false;

The default is true, as most sensors are ratiometric, but 1135 is non-ratiometric.

-Patrick

For my program I used

CPhidgetInterfaceKit_setRatiometric(ifKit, 0);

Which gave me the same result (I allready had it disabled in my code not knowing) and also tried setting it to 1.

I think the problem is just the accuracy of the device, I should have checked the specs in more detail before purchasing it for my application.

Are you running your conversion in the sensor change handler? Maybe you could try setting the sensitivity to 1.

The manual states a worst case error of 2% and typical 0.7%. The case that you mention has an error of 3%, assuming that your voltmeter is accurate, so you should be able to get better accuracy - or our specs are wrong, or the sensor is bad.

Also, does the sensor output 500 at 0v? There may be a consistent offset error that you could easily correct for.

-Patrick

patrick wrote:Are you running your conversion in the sensor change handler? Maybe you could try setting the sensitivity to 1.

The manual states a worst case error of 2% and typical 0.7%. The case that you mention has an error of 3%, assuming that your voltmeter is accurate, so you should be able to get better accuracy - or our specs are wrong, or the sensor is bad.

Also, does the sensor output 500 at 0v? There may be a consistent offset error that you could easily correct for.

-Patrick

I set the sensitivity to 1, I set the ratiometric to 0 and with 0V reference point I get values of 494 to 512.. still not precise, I just needed 2 decimal precision to monitor the state of charge of a deep cycle 12v battery

cizzi wrote:

patrick wrote:Are you running your conversion in the sensor change handler? Maybe you could try setting the sensitivity to 1.

The manual states a worst case error of 2% and typical 0.7%. The case that you mention has an error of 3%, assuming that your voltmeter is accurate, so you should be able to get better accuracy - or our specs are wrong, or the sensor is bad.

Also, does the sensor output 500 at 0v? There may be a consistent offset error that you could easily correct for.

-Patrick

I set the sensitivity to 1, I set the ratiometric to 0 and with 0V reference point I get values of 494 to 512.. still not precise, I just needed 2 decimal precision to monitor the state of charge of a deep cycle 12v battery

ok i managed to get it to -/+ 0.04V error margin which is acceptable.. by using the raw sensor value and adjusting it to the 0V reference point

Patrick wrote:Are you running your conversion in the sensor change handler? Maybe you could try setting the sensitivity to 1.

The manual states a worst case error of 2% and typical 0.7%. The case that you mention has an error of 3%, assuming that your voltmeter is accurate, so you should be able to get better accuracy - or our specs are wrong, or the sensor is bad.

Also, does the sensor output 500 at 0v? There may be a consistent offset error that you could easily correct for.

-Patrick

Hi!

I have the same problem as cizzi. I have 3 Precision Voltage Sensors (1135) and a SBC2 but every sensor shows a voltage of 6.755 V (result of the Phidget Control Panel) instead of 6.95 V (measured with different digital multimeters). This means an error of ~3%.

Ratiometric is set to 0 and sensitivity is set to 1. The output of every sensor at 0 V is ~500.

Any idea why the results are such inaccurate and how to improve them?