Thanks for the video! Interestingly the capacitance also varies with the temperature, a so-called temperature coefficient. It's possible to build an LC-oscillator and determine the temperature by measuring its frequency, which will change for about 100-500 Hz per 10-20 C or so. The fact that the frequency of an LC-oscillator varies with the temperature is a big problem when building an analog radio. There are special methods to solve this problem called temperature compensation. The idea is to connect two capacitors in parallel, one with a positive temperature coefficient and another with a negative temperature coefficient so that the net frequency shift would be about zero.
Great video again!:) Thank you! If you don't mind, I would extend what you have said in the case of ntc+adc. If you directly measure this resistor divider circiut with the adc and avoiding a buffer stage, you need to make sure that the sampling capacitor of the (SAR) adc charges up properly. Okay, it is small capacitor, but inproper setting of the adc can lead to false measurements in certain cases.
Very good conceptual review, thanks for sharing. Whenever you can, please post also the use of semiconductor-based sensors as, Si diodes and BJT (-2mV/C), LM35, and even Schoktty diodes (leakage current/C + 3xDarlington). A lot of applications with electronics could use their operating range. I have done some simple variable speed fan speed controllers using NTC (+transistors), or diode+OpAmp for power supply heatsinks. Based on your good way of explaining things, it would be a pleasure to see your ideas explained or demonstrated by you. Thanks for your channel and please keep these good postings!
Another great video! Thank you. In most of my applications I try to use PT1000 devices, since the wiring error is one order of magnitude less significant. It allows two wire applications when the necessary resolution is around 1/2 degree, only., while simplifying the input circuit and value conversion, due to the more linear behavior. PT1000 is a quite common device nowadays.
great video, the temperature meters and calibrators have used and repaired, all were PT100 based. They were 4 wires, often equiped with lemo connectors and the measurement current was pulsed. An important part of the accuracity is, like in a dmm, the voltage refenece used and the voltage measurement circuit. The PT100 used are insainly expensive and kept in nice velved lined wooden boxes .
More accurate measurement using the Thermocouple method is can be made buy using a reference junction thermocouple configuration this uses two or three Thermocouple connected to creates a electrical balance placing the reference Thermocouple(s) into a thermal reference such a melting water ice enables a calibrated measurement that now depends more on the accuracy of the volt meter. I have used this method in the lab and could reliably resolve 0.1 degree Celsius changes in the boiling point of solvents during distillation due to barometric pressure variation with the weather.
You are right. It's not just as common for temperature measurements as these 3; I guess at some point I might do a video on not so common temperature measurement methods. Do you know of any other more obscure ways?
One remark about things such as the "cold-junction compensation" - You go too fast through this for your viewers who don't already know about this, and the explanation is redundant for your viewers who do already know about this. Great tutorial, as always, nevertheless!
Your comments on the need for amplification for RTD and thermocouple signals are a bit out of date. Nowadays, very high resolution delta sigma ADCs are available at low cost, and they make it possible to directly connect a thermocouple to the input with no amplification. I am working on a product that involves measuring thermocouples this way, using a 24 bit ADC. It works very well.
I guess it has to do with the intended complexity and overall price of the system - an 8bit uC with a 10bit ADC coupled with a bit of amplification should be cheaper than a controller with a 24bit ADC. Anyway, a bit of amplification should help extend the measurement range and precision, even for the large adc.
@@FesZElectronics Admittedly in my product, the ADC is the single most expensive component. But that's because I went for one with lots of bells and whistles. Overall it is quite cost effective as the only other components on the inputs are some resistors/caps for an antialiasing filter. IIRC 18 bits is enough to get 0.1C resolution from type K thermocouples. So you could use something like the MCP3421 (18bit single channel differential input ADC with onboard precision reference) which is less than £2 in small quantities. Edit: it also has a PGA in it, but you wouldn't actually need to use it
Please explain to me too. I see no error. Note that the bridge on the left is mirrored compared to the bridge on the right. But the reference designators clarify everything.
Response from @Phillipe Paternotte: for me your bridge drawing as-is is only consistent for a compensated measurement if you adjust the bridge to have both branches currents equal and V=0. Then the RTD value is computed by resistors ratio; If you do not adjust the resistors and just measure V 0, the Rw is not compensated because the bridge is not equilibrated so you cannot compute the RTD value with just measuring V. I'm rather used to the common method that is used in industry, where a current generator is used and two voltages measured instead of a differential measure, so that the current in the third wire is null so its resistance doesn’t matter. The effective measure is something like RTD = (2*V2-V1) / Isource. An example may be found here: www.instrumentationtools.com/three-wire-rtd (but generally we measure both voltages relative to ground).
Yes. This circuit is actually used for compensation measurements. Thanks for your observation. It's also funny that you provided a link to instrumentationtools, where in rtd-questions-and-answers there is a section 'Three wire RTD Wiring' with almost the same picture and almost identical explanation as in the Fesz's video :)
