PT1000 sensors connected to thermistor inputs
-
Hi,
I just read this in the release notes of RFF1.2. Sounds too good to be true, but does that really mean I dont need to expensive converter board anymore?
Cheers
-
The support is for PT1000 sensors (not PT100) and the accuracy is not as good as the daughter board provides.
-
Might be a stupid question but why is it that platinum resistance sensors haven't been until now used directly connected. Its a case of resistance change related to temperature just like a thermistor?
-
It's question of ADC resolution and accuracy. With PT100 sensors, the voltages to be measured have to be very low, so the general purpose ADC in a microcontroller isn't suitable. It's better with PT1000 sensors, however the resistance change with temperature is small. The 10-bit ADCs on popular 8-bit microcontrollers don't have enough resolution. Even the 12-bit ADCs in 32-bit microcontrollers don't really have enough resolution, although oversampling can be used to try to get 1 or 2 bits of extra resolution. The gain and offset errors of the ADC cause inaccuracies in the readings too.
I implemented experimental PT1000 support at the request of an OEM customer, with a view to seeing whether we can make changes to the Duet design to make it work sufficiently well for regular use. I was quite surprised at how well it worked on a standard Duet, which is why I decided to include it as a standard feature in RRF. However, the calculations suggest that the worst case inaccuracies are quite large unless the ADC offset and gain are calibrated manually.
Most PT1000 sensors on the market are un-encapsulated thin film sensors, which are not suitable for use as-is in 3D printers. This one http://uk.farnell.com/webapp/wcs/stores/servlet/ProductDisplay?catalogId=10001&langId=44&urlRequestType=Base&partNumber=1653390&storeId=10151 is more suitable, but not cheap.
-
Thanks thats a great explanation thanks.
