Created Date: May 11,2017

Created by: schda12

The P2-08NTC is a thermistor input module for the Productivity 2000. Hardware configuration "temperature input range " has options "2252, 10K-AN Type 3, 10K-CP Type 2, 5K, 3K and 1.8K ohm ". These imply common temperature/resistance curves. One reference I found is at the RTD-Thermistor PDF link below. Reliably matching thermistors to these setting seems challenging, in part due to limited, missing or hidden AutomationDirect documentation.

Is the chart I found accurate? Do labels like "10K Type 3 " align with similar settings ( "10K-AN Type 3 ")?

I'm considering sensors from Amphenol/Thermometrics sold by Mouser. These have many options documented in data sheets such as http://www.mouser.com/ds/2/18/AAS-92...web-850743.pdf . These imply resistance curves via a "material system " code, which is documented at https://www.mouser.com/pdfDocs/AAS-9...091614-web.pdf . This does not reference common curve names directly. Instead, it provides parameters for the formula used to convert resistance to temperature and it provides a chart mapping temperature to the percentage of the devices documented resistance at 25 degrees C (e.g. 10,000 ohms).

Is there a more direct way of reliably matching a sensor to a setting on the P2-08NTC? For example, is there documentation listing the constants used in the conversion formula for each of the supported "temperature input range " settings?

Thanks,

David

Created Date: May 11,2017

Created by: g.mccormick

AN is type 3

CP is type 2

The 10K thermistor is the only supported type that has different (AN or CP) t ypes. They are very close, but specifing the wrong type could result in several degrees of error depending how far from 77F you are. My main complaint is that you cannot have different types for different channels.

{ "data-align ": "none ", "data-size ": "medium ", "data-attachmentid ":106185}

Created Date: May 13,2017

Created by: schda12

It's interesting that the interpretation of the "standard " 10K curves varies so much. The errors vary significantly between Amphenol's Thermometrics specs for material F (corresponding to 10K type 2 in their 10K thermistors) compared to the Veris and the Mamac charts. There is somewhat less variation for 10K type 3 and Amphenol material type Y. For example, see Amphenol Type 65 . Other Amphenol material types appear to be much farther from the 10K charts.

If the P2-08NTC honors the Mamac charts, the biggest conversion error reaches 0.24% at 0 degrees C for type 3 and 0.40% at 80 degrees C for type 2.

If the P2-08NTC honors the Veris charts, the biggest conversion error reaches only 0.10% at 70 degrees C for type 3 and 0.24% at 80 degrees C for type 2.

This is based on a spreadsheet I prepared but am unable to share. I can't rule out the possibility of spreadsheet errors.

If I'm interpreting the error correctly, it means an error of a couple degrees C for a sensor that's supposed to be reliable to 0.1 degrees C. If I use exactly the same sensor throughout my application, this might not be that big a deal. However, it seems troubling.

There will be additional error from the resistance of the sensor leads, which will vary in length, and especially from the way the sensor contacts the item being measured, which will be far from ideal.

Created Date: September 01,2017

Created by: DanLuttrell

There is a real problem achieving accurate temperature readings from this module, which is a shame since thermistors are very sensitive and stable over time. Somewhere in the module the thermistor resistance is converted to temperature using the Steinhart-Hart equation. This require 3 coefficients that are determined from the thermistor's resistance/temperature characteristic within the temperature range of interest. If we had access to these coefficients, we could tailor them for each thermistor type and for the temperature range we are working in. There are many 10K types out there and the options we have (2) do not adequately cover the range of sensors.

What would be most helpful is to make the actual resistance measurements available to the controller so that we could implement the Steinhart-Hart equation in the program - a very easy thing to do with the P2k. This is available on one of the Click input modules. It should be straightforward to do this with the P2-08NTC.

Once we solve this problem, perhaps we could address the filtering issue (that is, what's going on and how to turn it off!).

Anyone else having issues with this module? Quantization of the measured temperature below about 0.4 degC is also a problem.

Created Date: September 01,2017

Created by: ControlsGuy

What would be most helpful is to make the actual resistance measurements available to the controller so that we could implement the Steinhart-Hart equation in the program - a very easy thing to do with the P2k. This is available on one of the Click input modules. It should be straightforward to do this with the P2-08NTC.

This is what I was thinking. Get the resistance (not necessarily even with a thermistor module, you could use a voltage or ma module if you want), and do the math yourself. It's especially easy in a case like this where you already have the equation, but even if you didn't, you can easily derive the underlying equation even from tabular data.