Searching \ for '[EE] increasing sensitivity of a thermistor' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=increasing+sensitivity
Search entire site for: 'increasing sensitivity of a thermistor'.

Exact match. Not showing close matches.
PICList Thread
'[EE] increasing sensitivity of a thermistor'
2012\02\07@153645 by alan smith

picon face
hi group..me again

I want to bounce something off the members.


So I have a board that I designed with a 10K
thermistor, but its feeding a 3.3V micro (and that is fixed...cant change this
part).  They have come back and asked for
a better resolution.   
Now a thermistor, it simply generates a voltage
as a divider with a source impedance of course.  As the temperature changes, the voltage that
is
generated across it changes, and the resolution
of what you can measure is going to be dependent on the source voltage as well
as the ADC resolution.
 
So the ONLY way to get better resolution is to
be able to increase the source voltage from 3.3V to 5V or use a higher
resolution ADC.  Now for either case, the
micro thats reading this is 3.3V, so using 5V is out of the question.
 
Sorta thinking....using a PIC running at 5V to
read the thermistor, and then either using a PWM output (with RC filter) to regenerate
a voltage of controlled steps over the temperature range they want to measure
(pretty sure its 0 to 100F) for the ADC in their part to measure.
 
Any  other possible clever ideas?

2012\02\07@155719 by Djula Djarmati

flavicon


On 07-Feb-12 21:36, alan smith wrote:
{Quote hidden}

What about using a constant current source?
What other higher voltages (or negative) are available? What is the resolution now and what is required? What are the resistances at 0 and 100 degrees F?

Djul

2012\02\07@160845 by Robert Rolf

picon face
Can you change the thermistor to one with a larger value so that %R change
is larger? Or change the lower divider resistor
to get a lower starting voltage in the divider.
This will make your reading less linear, but may be acceptable trade off if
all you are doing is looking for  a threshold to trigger at.

Can you can add a rail to rail op amp to increase the gain (-offset) on
your voltage divider so that you use the full input range of your
A/D? You could also use two op amps to split the input range (with higher
gain) across two A/D channels. One channel handles the lower range, the
other the upper range.

You could also make a boost converter to get 6V (or more) from your 3.3 V
using one of the Maxim RS232 parts (or 555 (or spare PIC pin)
and a cockcroft walton voltage multiplier (capacitor and diode tree). I
once used C-W booster to get 40V from 5V for a project).

On some PICS you can set the A/D reference voltage to a lower value (V
divider or Voltage ref IC) to get more resolution.
e.g. match the max voltage you expect from the thermistor cct.
The A/D resolution is Vref/ADcount.

Some of these options were discussed a couple years ago on PIC list, so
check the archives for 'increase resolution' thread.
circa 2009.

R

On Tue, Feb 7, 2012 at 1:36 PM, alan smith <spam_OUTmicro_eng2TakeThisOuTspamyahoo.com> wrote:

{Quote hidden}

>

2012\02\07@161254 by Denny Esterline

picon face
>
>
> Now a thermistor, it simply generates a voltage
> as a divider with a source impedance of course.  As the temperature
> changes, the voltage that
> is
> generated across it changes, and the resolution
> of what you can measure is going to be dependent on the source voltage as
> well
> as the ADC resolution.
>


What value is the fixed resistor in the divider circuit?

It's been a few years but I seem to recall a homework question about
picking the best fixed resistor to optimize the output delta of the
thermistor. As memory serves, the optimum choice was a less than intuitive
solution.

-Denn

2012\02\07@162221 by Marc Nicholas

picon face
On Tue, Feb 7, 2012 at 3:36 PM, alan smith <.....micro_eng2KILLspamspam@spam@yahoo.com> wrote:

{Quote hidden}

Are you sure the accuracy and repeatability of your existing circuit better
the resolution that you're being asked to provide?

Out of interest, is this an R25 thermistor? What's the beta value?

Thx.

-m


2012\02\07@173719 by Picbits Sales

flavicon
face
Totally off the top of my head but possibly worth a look ...... use the thermistor in part of an oscillator then measure the time period or number of pulses in a fixed time period.

You could also use the PIC output pins to switch in/out various bias resistors for the divider network depending on the measured temperature at the thermistor - you may not get the exact 3.3v at the pin but you should be able to work around it enough to get a reasonable reading.

Could be worth a look, could be a dead end.
Dom

{Original Message removed}

2012\02\07@180051 by Spehro Pefhany

picon face
At 04:12 PM 07/02/2012, you wrote:
> >
> >
> > Now a thermistor, it simply generates a voltage
> > as a divider with a source impedance of course.  As the temperature
> > changes, the voltage that
> > is
> > generated across it changes, and the resolution
> > of what you can measure is going to be dependent on the source voltage as
> > well
> > as the ADC resolution.
> >
>
>
>What value is the fixed resistor in the divider circuit?
>
>It's been a few years but I seem to recall a homework question about
>picking the best fixed resistor to optimize the output delta of the
>thermistor. As memory serves, the optimum choice was a less than intuitive
>solution.

The optimum choice is to have the series resistor equal to the
thermistor at the temperature of interest. Resolution declines
on either side of that. If you need high resolution over a wide
range of temperature, and the ADC resolution isn't sufficient,
then you need to do something else.

Consumer thermometers use a RC timing arrangement with a precision
(well, 1%-ish isn't really precision these days) reference resistor.
You could probably implement that using the SR FF that's in some PICs.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
speffspamKILLspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

2012\02\07@222633 by Bob Ammerman

flavicon
face
{Quote hidden}

Perhaps...

Go ahead and use +5V at the top of your voltage divider. If you pick the source resistor so that the voltage out of the divider never exceeds the 3.3V of the PIC over the range of possible values of the thermistor you'll be Ok. If you are concerned that the voltage will jump up if the thermistor is disconnected or reaches a higher resistance than you thought, then you can include a schotky diode from the middle of the voltage divider to the +3.3V rail.

-- Bob Ammerman
RAm Systems

2012\02\07@234713 by Dave

picon face
Wheatstone bridge
Bob Ammerman <.....picramKILLspamspam.....roadrunner.com> wrote:

{Quote hidden}

>

2012\02\08@045244 by alan.b.pearce

face picon face
> So I have a board that I designed with a 10K thermistor, but its feeding a 3.3V
> micro (and that is fixed...cant change this part).  They have come back and asked
> for a better resolution.
>
> Now a thermistor, it simply generates a voltage as a divider with a source impedance
> of course.  As the temperature changes, the voltage that is generated across it
> changes, and the resolution of what you can measure is going to be dependent on the
> source voltage as well as the ADC resolution.
>
> So the ONLY way to get better resolution is to be able to increase the source
> voltage from 3.3V to 5V or use a higher resolution ADC.  Now for either case, the
> micro thats reading this is 3.3V, so using 5V is out of the question.
>
> Sorta thinking....using a PIC running at 5V to read the thermistor, and then either
> using a PWM output (with RC filter) to regenerate a voltage of controlled steps over
> the temperature range they want to measure (pretty sure its 0 to 100F) for the ADC
> in their part to measure.
>
> Any  other possible clever ideas?

Microchip have an app note where they obtain quite high resolution measuring a thermocouple, by having the ADC 'window' on a range of the total voltage swing. They adjust the window point using the PWM output to set the base end of the window. Maybe you could use a similar method to achieve what you want. Look for AN1306 on the Microchip web site.




-- Scanned by iCritical.

2012\02\08@054602 by Isaac Marino Bavaresco

flavicon
face
Em 7/2/2012 18:36, alan smith escreveu:
{Quote hidden}

Can't you simply set the AD's Vref- and Vref+ to the lower and higher
limits of your input signal? This way, you won't lose any codes. That
is, when the input signal is at its minimum value (=Vref-) the AD will
read 0 and when the input signal is at its maximum (=Vref+) the AD will
read 1023.

Beware, if Vref- and VRef+ voltages are too close to each other the
circuit will be more susceptible to noise and will be less precise.

Another solution would be using op-amps to scale and shift your signal
so it fills the range 0V to 3.3V.

Isaac

2012\02\08@061217 by Michael Rigby-Jones

flavicon
face


{Quote hidden}

If you have (or can introduce) at least 0.5LSB's worth of noise into the thermistor circuit you can simply oversample to gain increased resolution e.g.. sample 4 times to gain one extra bit, 16 times to gain 2 extra bits etc.

You could use several different fixed resistors as part of the potential divider, connected to port pins (if available!).  This way you can have switched ranges by either tri-stating pins or setting them as outputs.  This may complicate linearisation however.

There is a Microchip app note somewhere for measuring resistance by charging a cap through the resistor and timing the voltage rise.  A fixed resistor can be switched in as a reference to remove the capacitance change over temperature (i.e. a dual slope ADC).   With a 16 bit timer you may be able to achieve better resolution than a 10 bit ADC.

16 bit Delta-Sigma ADC's are fairly cheap and can be had in small packages with an I2C or SPI bus e.g. Microchip MCP3425 ($1.57).  These tend to have differential inputs so use a Wheatstone bridge with the thermistor in one arm.

Cheers

Mike

=======================================================================
This e-mail is intended for the person it is addressed to only. The
information contained in it may be confidential and/or protected by
law. If you are not the intended recipient of this message, you must
not make any use of this information, or copy or show it to any
person. Please contact us immediately to tell us that you have
received this e-mail, and return the original to us. Any use,
forwarding, printing or copying of this message is strictly prohibited.
No part of this message can be considered a request for goods or
services.
=======================================================================

2012\02\08@082953 by Spehro Pefhany

picon face
At 11:47 PM 2/7/2012, you wrote:
>Wheatstone bridge

What he's doing is effectively a bridge a la Sir Charles, with half
the bridge composed of series resistor and thermistor and the other half the
switched capacitor array in the SAR ADC. The ADC measures the unbalance
(ratiometric to the supply voltage, conveniently, if the references are the
supply rails).

BTW, the more current you put through the thermistor, the more voltage
per degree you get, but also the more self-heating at a given temperature-
this should be considered to make sure it is not causing unacceptable errors.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
@spam@speffKILLspamspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

2012\02\08@121048 by MarcoG

picon face

Hi Mike,  I have an old code snippet related to your first suggestion
(oversampling). It was from Basic Micro forum but now I don't see it anymore
on the web. I don't know if this code is valid or not, however  I attach it
if anyone interested.

Also the guy who has written this says that a pre-requirement is at
least 1 LSB of noise on ADC reading. In this example the resolution is
declared as increased from 10 to 16bit, but real resolution may be lower.

a good Appnote for ADC oversampling: http://www.atmel.com/dyn/resources/prod-documents/doc8003.pdf


regards,
Marco


-------------------------------------------------------------
.......
........
ADCON1 = 10001110 ;right justified

ADmain                    ;main loop for A/D
clrf OutBuff               ;make 0
clrf OutBuff1
clrwdt
movlw 0x40              ;load cnt with 64(dec)
movwf cnt
st
decf cnt                     ;cnt = cnt - 1
bsf ADCON0,GO    ;start A/D
l3
btfsc ADCON0,GO ;A/D done ?
goto l3
bsf STATUS,RP0    ;bank 1
movfw ADRESL      ;get A/D low byte
bcf STATUS,RP0    ;bank 0
addwf OutBuff1       ;add to low byte
btfsc STATUS,C     ;if carry
incf OutBuff             ;inc high byte
movfw ADRESH     ;get high A/D
addwf OutBuff         ;add to high byte
movf cnt
btfss STATUS,Z      ;is count = 0
goto st                      ;no loop again


--------------------------------------------------------





----- Original Message ----- From: "Michael Rigby-Jones" <KILLspamMichael.Rigby-JonesKILLspamspamoclaro.com>
To: "Microcontroller discussion list - Public." <RemoveMEpiclistTakeThisOuTspammit.edu>
Sent: Wednesday, February 08, 2012 12:12 PM
Subject: RE: [EE] increasing sensitivity of a thermistor




> {Original Message removed}

2012\02\08@163948 by YES NOPE9

flavicon
face
>
> On Feb 8, 2012, at 10:10 AM, MarcoG wrote:
>
>
> Hi Mike,  I have an old code snippet related to your first suggestion
> (oversampling). It was from Basic Micro forum but now I don't see it anymore
> on the web. I don't know if this code is valid or not, however  I attach it
> if anyone interested.
>
> Also the guy who has written this says that a pre-requirement is at
> least 1 LSB of noise on ADC reading. In this example the resolution is
> declared as increased from 10 to 16bit, but real resolution may be lower.
>
> a good Appnote for ADC oversampling:
> www.atmel.com/dyn/resources/prod-documents/doc8003.pdf
The requested page cannot be found.
The page you have requested might have moved or has been archived. Please use our Product Finder or site search to locate your document or page.
>
> regards,
> Marco

2012\02\08@164953 by MarcoG

picon face

arghhh.... there was a bad character!  :-(

 http://www.atmel.com/dyn/resources/prod_documents/doc8003.pdf





{Original Message removed}

2012\02\08@175154 by IVP

face picon face
>  http://www.atmel.com/dyn/resources/prod_documents/doc8003.pdf

Microchip have AN1152. Search for 'oversampling

2012\02\09@061800 by Electron

flavicon
face
At 12.12 2012.02.08, you wrote:
{Quote hidden}

Delta-Sigma cleverness. :)

2012\02\22@094305 by alan smith

picon face
Just a followup.  Think I found the issue...and of course if there is a way for someone to hook up something wrong..they will.  So the value being read is supposed to be across the thermistor with the upper resistor tied to +V...they put the thermistor first and then the resistor to ground.  So waiting for parts to come in and verify, because the spread sheet said it should work.

Thanks for all the great ideas tho!

More... (looser matching)
- Last day of these posts
- In 2012 , 2013 only
- Today
- New search...