Hi all, I am about to design a medical device. Which has a temperature sensor, whose value I need to send to an app via bluetooth. Initially I thought of using Stm32wb55ceu for it as it comes with bluetooth. But it has only 12 bit ADC. Msp430 has 24 bit Sigma delta ADC but it does not comes with bluetooth in it, I need to look for bluetooth IC. But with an op-amp and hx710b I can use stm32 I guess. So I am confused, which combination to choose considering price and performance. Please enlightenment me.

 

Well Ill play .., all about requirements and such. MCU is probable the easy part. Is it a one off.
How are you collecting the temp, what are its specks and accuracy of this part( or circuit).
What range and accuracy does your use here require. range/precision/speed ....

MCU will probable be the easy part, if you drive from requirements from detector circuit/ic to consumption layer a blue tooth app.
External Bluetooth or extern DAC or maybe a couple of op amps and precision voltage source ..
may be a footprint or $$$ drive decisions.

For DAC step ... You can always manipulate the signal into a range that mater ..
Again question is what range of temp do you care about and how much accuracy. 12 bit may be enough example a 10 degree range 10/4096 = 0.002 degrees ,can probable easily trust to 0.02 degrees. ( throw away 2-3 LSB , and oversample with a low pass ) have to understand

 

T

Well Ill play .., all about requirements and such. MCU is probable the easy part. Is it a one off.
How are you collecting the temp, what are its specks and accuracy of this part.
What range and accuracy does your use here require. range/precision/speed .

MCU will probable be the easy part, if you drive from requirements from detector circuit/ic.
External Bluetooth or extern DAC or maybe a couple of op amps and precision voltage source ..
may be a footprint or $$$ drive decision s.

For DAC step ... You can always manipulate the signal into a range that mater ..
Again question is what range of temp do you care about and how much accuracy. 12 bit may be enough example a 10 volt range 10/4096 = 0.002

The sensor which I will be using is TS318-11c55, which has range in mV. When the temperature is at 100 degree Celsius, it will show 8.4mV as output. This data is based on datasheet. The concern I have is, usually the temperature range will be in 35 to 43 degree Celsius, so to read those 12 bit might not be enough.

 

The sensor which I will be using is TS318-11c55, which has range in mV.

Hi E,
It is usual to have an amplifier when working with low level signals, then use an A2D to produce a digital value ,
You could then use a 10 Bit ADC.
E

 

Again just pondering , sounds like a fun project ..
But a buffer and amp/shift voltages will probable make every thing else simpler.
And 12 bit i would guess would be enough. Again what are requirements If 1/100 of a degree and a 10 degree range. Well that is 1000 .. 4096 should be enough even 2lsb or error.

When i queried the device at mouser ( Datasheet is their but lacking on a lot of details) . IT does indicate filters are available ,, so may be go fishing at vender site for info.

Interesting enough their is a "Development board" at Mosier and at least a narrative description reading examining and /or purchasing 16$ US , as it has done a lot or the RND for you.

https://www.mouser.com/datasheet/2/272/IR_Sense_5_Click-3476277.pdf
https://www.mikroe.com/ir-sense-5-click

They do use a I2C to talk off board no Bluetooth, all depends of what you up 2.

 

Again just pondering , sounds like a fun project ..
But a buffer and amp/shift voltages will probable make every thing else simpler.
And 12 bit i would guess would be enough. Again what are requirements If 1/100 of a degree and a 10 degree range. Well that is 1000 .. 4096 should be enough even 2lsb or error.

When i queried the device at mouser ( Datasheet is their but lacking on a lot of details) . IT does indicate filters are available ,, so may be go fishing at vender site for info.

Interesting enough their is a "Development board" at Mosier and at least a narrative description reading examining and /or purchasing 16$ US , as it has done a lot or the RND for you.

https://www.mouser.com/datasheet/2/272/IR_Sense_5_Click-3476277.pdf
https://www.mikroe.com/ir-sense-5-click

They do use a I2C to talk off board no Bluetooth, all depends of what you up 2.

So I guess, I should stick to 12 bit ADC and work on amplification first and it should be enough.

 

You can always trade speed for accuracy. Your temperature won't move quickly, so you can IIR filter which can increase the accuracy by four bits just by reducing the response time by a factor of 16, just by using noise as your dither signal.

 

So I guess, I should stick to 12 bit ADC and work on amplification first and it should be enough.

** More armature thoughts ***

IF you only 1/10 a degree absolutely 12 is enough and overkill .. 1/100 maybe all depends on how well you can precondition.

I would start with a buffer and amp stage. a simple voltage booster moving to 0-3.3v or what a-ref you are using.

I would pull out your favorite low offset , low drift, high input impedance , op amp and go from their . Will need to isolate range you care about or can tolerate in analog circuit. IT just math.

thinking tlv2364 as i just have a bunch on hand ( just a hobby junk rack ) .. They are 3$ or so if cost matters.

 

You can always trade speed for accuracy. Your temperature won't move quickly, so you can IIR filter which can increase the accuracy by four bits just by reducing the response time by a factor of 16, just by using noise as your dither signal.

Well I wanted to do the things as much as possible in hardware itself, but as you said I can definitely trade speed for accuracy.

 

** More armature thoughts ***

IF you only 1/10 a degree absolutely 12 is enough and overkill .. 1/100 maybe all depends on how well you can precondition.

I would start with a buffer and amp stage. a simple voltage booster moving to 0-3.3v or what a-ref you are using.

I would pull out your favorite low offset , low drift, high input impedance , op amp and go from their . Will need to isolate range you care about or can tolerate in analog circuit. IT just math.

thinking tlv2364 as i just have a bunch on hand ( just a hobby junk rack ) .. They are 3$ or so if cost matters.

Thanks for suggesting, but I need to do research in finding op amp, as the availability and lead time of certain IC 's need to match with requirement...

 

As i walk away thoughts ,, if it is a real medical device <---> that matters and or commercial venture .. Contact vender ( data sheet is terrible ) !!
That moves all thoughts into the what happens when it fails and how to fail soft , lets not kill anyone or get reading value wrong.