-
Categories
-
Platforms
-
Content
connection of a thermal sensor TC1047AVNBTR to the microcontroller MSP430F4618
- Thread starter Alex9
- Start date
KrisBlueNZ
Sadly passed away in 2015
All the information you need can be found in the TC1047 data sheet.
The TC1047Axxxxx requires a power supply between 2.5V and 5.5V DC and draws up to 60 uA from that power supply. It generates an output voltage that varies by nominally 10 mV per degree Celsius and is nominally 500 mV at 0 degrees Celsius.
It can measure temperature up to +125 degrees Celsius, and at this temperature its output voltage is nominally 1.75V. All of this information is clearly explained in the data sheet.
You can connect its output directly to any input of the ADC of the MSP430 device. The ADC's reference voltage must be at least 1.75V if you need to measure up to +125 degrees Celsius.
You can convert a raw reading from the ADC into an equivalent temperature using the following formula:
temperature_(degrees_Celsius) = (ADC12_raw_value / 4096 * ADC_VREF * 100) - 50.
If you want to calibrate the sensor, adjust the 100 to change the scaling and adjust the 50 to change the offset.
I recommend adding some simple digital filtering on the ADC value to reduce variation due to noise. A simple 1/16-new-plus-15/16ths-old averaging algorithm should work well.
If you have any other questions please ask them clearly.
The TC1047Axxxxx requires a power supply between 2.5V and 5.5V DC and draws up to 60 uA from that power supply. It generates an output voltage that varies by nominally 10 mV per degree Celsius and is nominally 500 mV at 0 degrees Celsius.
It can measure temperature up to +125 degrees Celsius, and at this temperature its output voltage is nominally 1.75V. All of this information is clearly explained in the data sheet.
You can connect its output directly to any input of the ADC of the MSP430 device. The ADC's reference voltage must be at least 1.75V if you need to measure up to +125 degrees Celsius.
You can convert a raw reading from the ADC into an equivalent temperature using the following formula:
temperature_(degrees_Celsius) = (ADC12_raw_value / 4096 * ADC_VREF * 100) - 50.
If you want to calibrate the sensor, adjust the 100 to change the scaling and adjust the 50 to change the offset.
I recommend adding some simple digital filtering on the ADC value to reduce variation due to noise. A simple 1/16-new-plus-15/16ths-old averaging algorithm should work well.
If you have any other questions please ask them clearly.
Thank you. 1/16-new-plus-15/16ths-old averaging algorithm - can you give me a link with article about it ? I dont know what is it. I have a possibility of connection of the sensor to a supply pinch (about 5 volt is programmed, and it is necessary for the sensor 3) microcontroller. Whether it is necessary in such to set a case the power conditioner and the filter? Whether the tightening resistor and where it is set on the diagram shown in the document of the sensor is required?
KrisBlueNZ
Sadly passed away in 2015
Actually no, I haven't been able to find any articles on that type of smoothing. Perhaps because it is very simple.
The input has a corresponding "smoothed value" variable, which has several more bits of resolution than the raw value; these bits are called "remainder bits".
New raw values coming from the ADC are incorporated into the smoothed value variable by calculating 15/16ths of the old smoothed value and adding 1/16th of the new value, although there are several tricks to simplify the calculation and avoid most of the bit-shifting.
In this application, with 12-bit raw values and a 16-bit MCU, you would use four remainder bits, so the smoothed value would have 16 bits allocated like this:
The top 12 bits are the smoothed equivalent of the raw value, and the bottom four bits are "remainder" bits.
The incorporation calculation would implement the 15/16ths and 1/16th calculations. You can of course use other numbers of remainder bits; three remainder bits gives 7/8ths old plus 1/8th new, which gives less smoothing than four remainder bits, and you can use more than four remainder bits for heavier smoothing.
Edit: If you perform any calibration by offsetting or changing the scaling of the raw value, you should do this before the smoothing is applied. The smoothing can give you more bits of resolution than the ADC actually has.
The implementation can be a bit messy so I won't get into details. There are many other approaches to smoothing noisy ADC conversion data if that is required. Have a look at these Wikipedia pages:
https://en.wikipedia.org/wiki/Smoothing
https://en.wikipedia.org/wiki/Digital_filter
I'm afraid I don't understand the rest of your post.
The input has a corresponding "smoothed value" variable, which has several more bits of resolution than the raw value; these bits are called "remainder bits".
New raw values coming from the ADC are incorporated into the smoothed value variable by calculating 15/16ths of the old smoothed value and adding 1/16th of the new value, although there are several tricks to simplify the calculation and avoid most of the bit-shifting.
In this application, with 12-bit raw values and a 16-bit MCU, you would use four remainder bits, so the smoothed value would have 16 bits allocated like this:
Code:
F E D C B A 9 8 7 6 5 4 3 2 1 0
s s s s s s s s s s s s r r r rThe incorporation calculation would implement the 15/16ths and 1/16th calculations. You can of course use other numbers of remainder bits; three remainder bits gives 7/8ths old plus 1/8th new, which gives less smoothing than four remainder bits, and you can use more than four remainder bits for heavier smoothing.
Edit: If you perform any calibration by offsetting or changing the scaling of the raw value, you should do this before the smoothing is applied. The smoothing can give you more bits of resolution than the ADC actually has.
The implementation can be a bit messy so I won't get into details. There are many other approaches to smoothing noisy ADC conversion data if that is required. Have a look at these Wikipedia pages:
https://en.wikipedia.org/wiki/Smoothing
https://en.wikipedia.org/wiki/Digital_filter
I'm afraid I don't understand the rest of your post.
Last edited:
My microcontroller is set on the experimental board http://www.ti.com/tool/msp-exp430fg4618. I would like to have the general supply of the sensor with the microcontroller, but for supply option from batteries didn't find information in board documentation about stabilizator and filter existence between batteries and the microcontroller that it was possible to be connected to a pure signal. Prompt according to the diagram on page 18 of the user guide of a board - whether there is there a stabilizator and the filter? Or it is possible to connect the sensor to a pinch of the 100th microcontroller, without using the stabilizator and the filter? Whether the additional reference supply source is required for microcontroller ADC12 ? Thank you.
KrisBlueNZ
Sadly passed away in 2015
I can't understand your questions.
The user guide has pages numbered 2~16; there is no page 18.
The TC1047Axxxxx requires a power supply in the range 2.5~5.5V DC and draws 60 uA. You can power it from VREF if the VREF voltage and current specifications are suitable, otherwise you should power it from AVCC.
AVCC is usually supplied from the same source as DVCC with a small resistor and/or inductor in series, and with one or more decoupling capacitors from AVCC to AVSS.
This arrangement is usually the same whether the power source is batteries or an external adapter or power supply.
If this doesn't answer your questions, please find someone who can explain your questions in clear English and I will try to answer them.
The user guide has pages numbered 2~16; there is no page 18.
The TC1047Axxxxx requires a power supply in the range 2.5~5.5V DC and draws 60 uA. You can power it from VREF if the VREF voltage and current specifications are suitable, otherwise you should power it from AVCC.
AVCC is usually supplied from the same source as DVCC with a small resistor and/or inductor in series, and with one or more decoupling capacitors from AVCC to AVSS.
This arrangement is usually the same whether the power source is batteries or an external adapter or power supply.
If this doesn't answer your questions, please find someone who can explain your questions in clear English and I will try to answer them.
I am limited in connection by RF1 and RF2 connectors of the experimental board. It isn't possible to find their characteristic to select to itself the necessary 2 outputs. Can you help me ? I looked for information in board and microcontroller documentation. I entered into search the words signed at pinches of these connectors on the diagram of a board.
KrisBlueNZ
Sadly passed away in 2015
Are you saying that can only use the RF1 and RF2 connectors? None of the pins on those connectors are connected to pins on the MSP430FG4618 that can be connected to ADC inputs, so you will not be able to connect your temperature sensor through those connectors.
Look at the MSP430FG4618 data sheet to find the pins that can be used as ADC inputs. Then look at the schematic of the circuit board and see where those pins are connected. You will need to connect to one of those pins, to get your analogue signal into the ADC.
Look at the MSP430FG4618 data sheet to find the pins that can be used as ADC inputs. Then look at the schematic of the circuit board and see where those pins are connected. You will need to connect to one of those pins, to get your analogue signal into the ADC.
Hello,
I have two pins - 3V (from batteries) and ground.
In documentation of ADC12 the generator of reference voltage is described.
It can be switched off, and also can give 1,5 volts or 2,5 volts.
But I didn't find, than to be guided in case of a choice of an operation mode of this generator for my diagram.
Explain than I can be guided still or on what pages about it in addition to read?
(I need to select SREFx bits of ADC12)
I had suspicions that reference voltage isn't necessary for me because of minimum voltage of an output of the sensor is 100 mV.
temperature_(degrees_Celsius) = (ADC12_raw_value / 4096 * ADC_VREF * 100) - 50 - from where such formula turns out ?
How to express the minimum voltage variation which shall fix ADC12 ?
Thank you.
I have two pins - 3V (from batteries) and ground.
In documentation of ADC12 the generator of reference voltage is described.
It can be switched off, and also can give 1,5 volts or 2,5 volts.
But I didn't find, than to be guided in case of a choice of an operation mode of this generator for my diagram.
Explain than I can be guided still or on what pages about it in addition to read?
(I need to select SREFx bits of ADC12)
I had suspicions that reference voltage isn't necessary for me because of minimum voltage of an output of the sensor is 100 mV.
temperature_(degrees_Celsius) = (ADC12_raw_value / 4096 * ADC_VREF * 100) - 50 - from where such formula turns out ?
How to express the minimum voltage variation which shall fix ADC12 ?
Thank you.
Last edited:
KrisBlueNZ
Sadly passed away in 2015
I don't understand your question.
The temperature sensor does not have to be connected to AVREF.
You want to connect the output from your temperature sensor to an input of the ADC in the MSP430FG4618, right? So you need to connect the temperature output to a MSP430FG4618 pin that can be internally connected to the ADC.
The temperature sensor does not have to be connected to AVREF.
You want to connect the output from your temperature sensor to an input of the ADC in the MSP430FG4618, right? So you need to connect the temperature output to a MSP430FG4618 pin that can be internally connected to the ADC.
KrisBlueNZ
Sadly passed away in 2015
Well, I tried Googling TC1047A PSPICE model but didn't find anything. I found a Microchip document that says they have PSPICE models for their op-amps and comparators, but it doesn't mention models for their transducer devices, and the document does cover them, so I take that to mean that they don't have models for them.
Of course you could just model the TC1047A as a voltage source.
As for modelling the MSP430, I don't think it's meaningful to simulate a microprocessor in PSPICE - you would need to simulate the whole thing including the firmware. I've looked on the TI web site and couldn't find anything, though of course they do have SPICE models for their linear devices.
So it seems there's no answer for your question. I don't really think it's a meaningful question.
Of course you could just model the TC1047A as a voltage source.
As for modelling the MSP430, I don't think it's meaningful to simulate a microprocessor in PSPICE - you would need to simulate the whole thing including the firmware. I've looked on the TI web site and couldn't find anything, though of course they do have SPICE models for their linear devices.
So it seems there's no answer for your question. I don't really think it's a meaningful question.
KrisBlueNZ
Sadly passed away in 2015
I don't understand your questions.
The ADC does have input capacitance and resistance. It will also generate pulses of current at its input during the successive approximation process. It's common to connect external capacitance (e.g. 0.1~1.0 µF) from the ADC input to 0V to absorb these current pulses so they don't disturb the sensor output and cause measurement errors.
The sensor produces an output voltage that's proportional to temperature. If you simulate it as a voltage source, the voltage must be controlled if you want to simulate more than one temperature. If you just want to simulate the sensor at a fixed temperature, the voltage source can be a fixed voltage.
The ADC does have input capacitance and resistance. It will also generate pulses of current at its input during the successive approximation process. It's common to connect external capacitance (e.g. 0.1~1.0 µF) from the ADC input to 0V to absorb these current pulses so they don't disturb the sensor output and cause measurement errors.
The sensor produces an output voltage that's proportional to temperature. If you simulate it as a voltage source, the voltage must be controlled if you want to simulate more than one temperature. If you just want to simulate the sensor at a fixed temperature, the voltage source can be a fixed voltage.
Whether there is a difference in what will control a voltage source - current or tension? And what it is better to select and why?
This sensor has has output resistance and output capacity. Where it is possible to find their numerical values? In documentation on the sensor I didn't find such values.
This sensor has has output resistance and output capacity. Where it is possible to find their numerical values? In documentation on the sensor I didn't find such values.
KrisBlueNZ
Sadly passed away in 2015
I don't really understand your questions.
The drive capability of the sensor's output is defined as a maximum allowable source and sink current of at least 100 µA. I suggest you connect a capacitor between the sensor's output and the 0V rail, i.e. across the sensor's output voltage. This will provide a relatively low AC impedance source to the ADC to minimise measurement error due to currents injected back into the ADC input pin during the successive approximation process.
Your equivalent circuit makes no sense to me. You have drawn capacitors in series; they are actually in parallel. Also you have drawn Vout as a current source not a voltage source. That diagram means nothing to me.
If you start from the beginning and explain in detail what you are trying to do, I may be able to help you. But I doubt it.
The drive capability of the sensor's output is defined as a maximum allowable source and sink current of at least 100 µA. I suggest you connect a capacitor between the sensor's output and the 0V rail, i.e. across the sensor's output voltage. This will provide a relatively low AC impedance source to the ADC to minimise measurement error due to currents injected back into the ADC input pin during the successive approximation process.
Your equivalent circuit makes no sense to me. You have drawn capacitors in series; they are actually in parallel. Also you have drawn Vout as a current source not a voltage source. That diagram means nothing to me.
If you start from the beginning and explain in detail what you are trying to do, I may be able to help you. But I doubt it.
Thanks for the help. I want to figure the equivalent circuit of connection of the sensor with the microcontroller. The vendor told me that completely to simulate it it won't turn out because of high complexity. Therefore the simplest diagram which I could draw in the program Orcad Capture (and test it in PSpice) is necessary to me. One output of the sensor is connected to an ADC input.Also there is the common wire - the ground. The input circuit of ADC has resistance of 2000 Ohms and capacity of 40 picofarads. The sensor will transform temperature to tension. To me the formula of its conversion is known. That is the voltage source on the diagram which is controlled something is necessary. It will play temperature role.Also the sensor has own output resistance and output capacity. I drew them offhand. Here and all my diagram. I need to execute the following analysis types in this program:Bias Point Delay, DC Sweep,AC Sweep,Transient,Fourier Analysis,Parametric,Monte-Carlo,Sensitivity,Transfer Function,Worst-Case.It is necessary to clarify how my diagram works. If it doesn't take away from you a lot of time, draw the correct simple diagram. It is difficult to understand according to what text description specifically it shall be. For me there is unclear a determination of resistance and capacity of an output of the sensor. Thank you.
KrisBlueNZ
Sadly passed away in 2015
WHY? Can you explain exactly WHY you want to do that? And how would a simulation be useful to you?
Yes, I would tend to agree with that. Depending on what you really want to do, you may need to simulate the entire MSP430 microcontroller, including its firmware!
I will describe the circuit that you should BUILD. This is NOT the simulation circuit. I don't think there's much point trying to simulate it.
Connect the sensor across the supply rails. Place a decoupling capacitor (0.1 µF X7R ceramic) across the power supply pins of the sensor, as close as possible to the device.
Connect the sensor's output to a pin on the MSP430 that is ADC-capable.
Connect another 0.1 µF capacitor from that pin to the AGND rail (analogue ground). That capacitor should be placed near the MSP430 but it doesn't need to be extremely close.
I don't think I will be able to help you simulate it.
Please answer the first question: WHY do you want to simulate it, and what do you hope to learn by simulating it?
Similar threads
