|
|
Post by julius on Mar 16, 2017 13:02:42 GMT -5
Hi,
I would like to know exactly how the Lakeshore 335 (including the model 3060 thermocouple option) firmware performs the room temperature compensation.
Since temperature curves are often not linear, the naive approach of adding the measured temperature, based on the measured voltage, and the reference temperature leads to an error.
Best regards
Julius
|
|
|
|
Post by Lake Shore Ryan on Mar 17, 2017 17:41:01 GMT -5
Hi Julius,
Yes, you're right, if we just add a temperature offset as you say, the accuracy would only improve close to the calibration point.
The older version of our Curve Handler software (look for the link at the bottom of the page that says it's for the Model 330 and older) is capable of applying a permanent adjustment to thermocouple temperature response curves based on measured values in your setup. These adjusted temperature curves can then be loaded onto the Model 335 and applied to an input.
The image below shows the effect of adjusting based on a calibration point above 0°C.
You can see that the points below 0°C are not adjusted, but the points above are gradually shifted to smoothly intersect the new calibration point.
I would hope that this is also how the 335 handles the calibration point, but I'm not certain. I'll check with our engineering team on Monday and let you know. If you have more information for me about which temperature ranges you're interested in, that would be helpful too. Thanks!
|
|
|
|
Post by Lake Shore Ryan on Mar 20, 2017 17:41:30 GMT -5
Hi Julius,
It looks like the instrument uses a simplified method for adjusting the temperature reading. The instrument's temperature compensation applies a simple voltage offset to have the instrument report the anticipated voltage at the given temperature. This offset is maintained for all readings from that input. So using the room temperature compensation on the 335 will result in the error you're talking about.
So if you require a more accurate thermocouple temperature conversion, you can use Curve Handler version 8.3 or higher and use the Make Adjustment To Curve File option to create a more gradual adjustment to the curve. Click on the image below to see where this button is located in the program.
Once you have a new adjusted temperature curve, you can load it into the 335 and apply it to your input. This should result in a slightly more accurate temperature conversion over a wider temperature range than what is possible using the onboard room temperature compensation.
Let me know if this answers your question. Thanks!
|
|
|
|
Post by julius on Mar 24, 2017 9:30:42 GMT -5
Hi Ryan,
Thank you for this detailed answer!
So basically the 335 does this: Total U = U(room_temperature)+ U(device_temperature)
and converts it to a temperature with the given curve?
Can you give an estimation of the error that can be expected by using this method?
We are interested in a temperature range from 300K to 1000K.
|
|
|
|
Post by Lake Shore Ryan on Mar 27, 2017 23:39:23 GMT -5
Hi Julius,
Sorry for the delayed response, I wanted to get more familiar with how this scenario would play out on our instruments and software.
Firstly, the equation is more like: Temperature = f(Ureading) where Ureading = Umeasured - Uoffset and Uoffset = Umeasured @ calibration temperature - Uof curve value @ calibration temperature
Uoffset will be fixed for as long as you use that calibration. So it is most valuable when you are most interested in accuracy over a narrow temperature range.
If you are interested in a large temperature range like you are, the generated voltage from the thermocouple from 300 to 1000 K could vary greatly from 1 to 30 mV respectively for a K type thermocouple. So calibrating an offset at a single point would give a poor result.
Instead use Curve Handler version 8.3 or higher and use the Make Adjustment To Curve File option to create a more gradual adjustment to the curve like I showed in the previous post. Since you have such a wide temperature range, I'd suggest adjusting the curve with multiple points (if you have access to a more accurate thermometer than a thermocouple) by clicking the Next Point button to add additional points.
I'm not sure if you're considering it, but I couldn't recommend our thermocouple wire for your application which has plastic insulation due to our focus on cryogenic applications. Instead, I'd suggest searching for ceramic beaded thermocouple wire (if you haven't already).
If you can find an appropriate type K thermocouple, you could use the inbuilt temperature curve in the 335 for this thermocouple type and apply adjustments as I mentioned above.
When working with thermocouple though, you should know that thermocouples are notorious for not being very repeatable over multiple thermal cycles. So you may need to recalibrate your thermocouple regularly if you are concerned about temperature accuracy. The 335 can be capable of handling whatever level of accuracy you require. However, the thermocouple itself will probably be the greatest source of error in your measurement. They are difficult sensors to use accurately, which is the reason they aren't the preferred sensor to use unless they're the only sensor capable of surviving the measurement temperature.
I hope this helps.
|
|
