Hello Everyone
Does anybody have any experience with Thermistors?… I am planning to develop a circuit to regulate the temperature of a soldering iron and would like to use a Thermistor and a PIC to test what range the temperature is in:
<50 C
50 - 225
225 - 260
260 - 290
290 - 325
(resolution is not a major issue – precision could be as loose as plus or minus 10 degrees C).
Obviously I need a Thermistor with a temp range of at least 0 to +325 degrees C. and was recommended that to use with a PIC with no additional signal amp it should be around 1k at room temperature.
Does anybody know where I would find a Thermistor that can deal with these ranges? The most I can find only go up to 125 degrees C.
Also has anybody ever designed a temp testing circuit using Thermistors that could give me any pointers (this will be the first circuit I’ve designed as I have very little electronics experience).
Thanks everyone.
Adam.
Thermistors for Temperature Regulation
Hi there,
I have used thermistors before in circuits to compensate for
temperature changes that affects the operating point.
The thermistor is nothing more than a resistance that changes
with temperature, although they make tons of values and they
do publish the curves and provide formulas for predicting
the resistance change over temperature.
I've never used one that went higher than about 150 deg C though
so you will have to look around to find what you want there, or
set it up mechanically so that there is some heat loss before it
gets to the themistor.
One thing nice about temperature control is that you dont need
a hugely complex circuit to do it. A simple comparator will work
pretty well and provide very accurate temperature adjustment.
The thermistor is used with a fixed resistor in a resistive divider
circuit to provide a voltage to one of the comparator inputs.
The other comparator input is a pot which divides a fixed reference
voltage down to the same voltage as the thermistor causes when
it reaches the desired temperature. the comparator output then
swings negative (or positive) when the temperature goes under
(or over) the set point turning the heating element on (or off) as
needed. This can get you easily to 0.1 degree C accuracy even
without a PIC. The only thing to keep an eye on is the resolution of
the thermistor in the divider circuit. Since the thermistors resistance
does not change linearly as the temperature changes, this means
some set points will end up being more accurate than others.
You can calculate this by knowing the following:
1. The thermistors resistance at the desired set points (ohms)
2. The upper resistive divider value (ohms)
3. The reference voltage
4. The hysteresis of the comparator (also presettable, in volts)
They also make somewhat more accurate thermistors too that are
supposedly as good as 1% on the resistance.
If you are interested in using this kind of circuit i'll post more later.
Even if you do use a PIC however, you'll have to do some of these
kinds of calculations anyway to determine what values you want to
use in your circuit.
I have used thermistors before in circuits to compensate for
temperature changes that affects the operating point.
The thermistor is nothing more than a resistance that changes
with temperature, although they make tons of values and they
do publish the curves and provide formulas for predicting
the resistance change over temperature.
I've never used one that went higher than about 150 deg C though
so you will have to look around to find what you want there, or
set it up mechanically so that there is some heat loss before it
gets to the themistor.
One thing nice about temperature control is that you dont need
a hugely complex circuit to do it. A simple comparator will work
pretty well and provide very accurate temperature adjustment.
The thermistor is used with a fixed resistor in a resistive divider
circuit to provide a voltage to one of the comparator inputs.
The other comparator input is a pot which divides a fixed reference
voltage down to the same voltage as the thermistor causes when
it reaches the desired temperature. the comparator output then
swings negative (or positive) when the temperature goes under
(or over) the set point turning the heating element on (or off) as
needed. This can get you easily to 0.1 degree C accuracy even
without a PIC. The only thing to keep an eye on is the resolution of
the thermistor in the divider circuit. Since the thermistors resistance
does not change linearly as the temperature changes, this means
some set points will end up being more accurate than others.
You can calculate this by knowing the following:
1. The thermistors resistance at the desired set points (ohms)
2. The upper resistive divider value (ohms)
3. The reference voltage
4. The hysteresis of the comparator (also presettable, in volts)
They also make somewhat more accurate thermistors too that are
supposedly as good as 1% on the resistance.
If you are interested in using this kind of circuit i'll post more later.
Even if you do use a PIC however, you'll have to do some of these
kinds of calculations anyway to determine what values you want to
use in your circuit.
LEDs vs Bulbs, LEDs are winning.
Thanks Al, thats really helpful.
Will the set-up you described be suitable for multiple temperature ranges?
What I have in mind you see is a soldering iron that offers visual feedback via LED's of the temperature range the iron is in:
<50 C will tell the user the iron is cooled down enough to handle.
225-260 C will be the idling temperature.
When the user picks the iron up it will then switch to a using temperature range and regulate between 290-325 C.
The only Thermistor I can find that gets anywhere near 325C is a small bead ceramic thermistor that goes to 300C, it's the only one I can find anywhere: http://www.ysitemperature.com/techdocs/ysi-beads.pdf
I had originally thought of using a thermocouple but somebody mentioned just like you that a thermistor would require a less complicated circuit.
Thanks
Adam
Will the set-up you described be suitable for multiple temperature ranges?
What I have in mind you see is a soldering iron that offers visual feedback via LED's of the temperature range the iron is in:
<50 C will tell the user the iron is cooled down enough to handle.
225-260 C will be the idling temperature.
When the user picks the iron up it will then switch to a using temperature range and regulate between 290-325 C.
The only Thermistor I can find that gets anywhere near 325C is a small bead ceramic thermistor that goes to 300C, it's the only one I can find anywhere: http://www.ysitemperature.com/techdocs/ysi-beads.pdf
I had originally thought of using a thermocouple but somebody mentioned just like you that a thermistor would require a less complicated circuit.
Thanks
Adam
Hi again,
If you want to light multiple LEDs depending on temperature you
would have to use one comparator per LED. The LM339 comes
with four comparators in one 14 pin package.
You would use another comparator for the temperature control,
where you have a pot adjusted to the correct temperature
with a resistor connected to the top terminal. When the user
picks up the iron a sensor would have to trigger something to
short out this resistor, so the temperature setting goes higher.
Or you could put the resistor at the bottom of the pot, and
have it shorted out and the pot adjusted to idle temperature.
When the sensor detects the iron has been picked up, it open
circuits this resistor which raises the voltage at the pot arm
which raises the temperature.
The output of that one comparator is used to drive the heat
element, probably through a triac.
The triac needs a circuit of it's own however, so maybe a PIC
would be a good idea anyway. In this way you could control
either the firing angle (sync the PIC up to the line 60Hz) or
you could simply count cycles and blank out enough cycles to
regulate the temperature. I would prefer controlling the
firing angle, and the frequency is low enough so this shouldnt
be a problem. Doing it with the PIC of course means you
have lots of variety in the way in which you control the heat
element. For example:
1. Measure the thermistor resistance using a AD input and
a reference voltage with resistor to divide (or look up the
application note Microchip has for accurately measuring
thermistor inputs).
2. Use the on chip comparator for the temperature control,
using one output to PWM a RC circuit to set the temperature,
based on thermistor measurement.
Of course when using the PIC you can drive the LEDs with
digital outputs, no problem there. Three LEDs means three
i/o pins, and one more for the thermistor, and one for the
control output, and one more for the line sync input, and
one more for the temperature setting unless you hard code
it into the PIC. That's 6 with hard coding, 7 with pot adjustable
setting (another AD port reads the pot to set the temperature).
One thing nice about having the pot to set the temperature is
that you can adjust it to whatever you want when the element
is to be driven at it's high end. Of course another pot to set
the idle temperature if you care to do that.
This project is starting to sound very interesting. I might just have
to try one myself. I've built several analog triac control circuits
and they work really great, so i can imagine how easy it would
be to do with a pic.
BTW, the 60Hz line sync input can come from a full wave
rectifier output without capacitance. This of course means
that you cant use a dc wall wart, but have to use an ac type
and rectify it yourself. All the analog circuits i've designed this
way used this technique, and to get around not being able to
use capacitance on the output a simple 1N4004 diode can be
used to isolate the full wave output from the filter cap, meaning
you can get both the pulsing full wave output AND the filtered
dc to run the circuit, all with a common ground.
Just to note again, the thermistor is very non-linear, so at the high
temperature end it may be hard to accurately measure temperature
unless you choose the series resistor (used to get the divided voltage
that represents the temperature) carefully. Since in your app it
sounds like the temperatures where you need accuracy not only
dont have to be that accurate anyway, they are closely spaced
where some accuracy is needed (idle and normal settings) so i
would expect the simple thermistor and series resistor circuit
to suffice.
If you want to light multiple LEDs depending on temperature you
would have to use one comparator per LED. The LM339 comes
with four comparators in one 14 pin package.
You would use another comparator for the temperature control,
where you have a pot adjusted to the correct temperature
with a resistor connected to the top terminal. When the user
picks up the iron a sensor would have to trigger something to
short out this resistor, so the temperature setting goes higher.
Or you could put the resistor at the bottom of the pot, and
have it shorted out and the pot adjusted to idle temperature.
When the sensor detects the iron has been picked up, it open
circuits this resistor which raises the voltage at the pot arm
which raises the temperature.
The output of that one comparator is used to drive the heat
element, probably through a triac.
The triac needs a circuit of it's own however, so maybe a PIC
would be a good idea anyway. In this way you could control
either the firing angle (sync the PIC up to the line 60Hz) or
you could simply count cycles and blank out enough cycles to
regulate the temperature. I would prefer controlling the
firing angle, and the frequency is low enough so this shouldnt
be a problem. Doing it with the PIC of course means you
have lots of variety in the way in which you control the heat
element. For example:
1. Measure the thermistor resistance using a AD input and
a reference voltage with resistor to divide (or look up the
application note Microchip has for accurately measuring
thermistor inputs).
2. Use the on chip comparator for the temperature control,
using one output to PWM a RC circuit to set the temperature,
based on thermistor measurement.
Of course when using the PIC you can drive the LEDs with
digital outputs, no problem there. Three LEDs means three
i/o pins, and one more for the thermistor, and one for the
control output, and one more for the line sync input, and
one more for the temperature setting unless you hard code
it into the PIC. That's 6 with hard coding, 7 with pot adjustable
setting (another AD port reads the pot to set the temperature).
One thing nice about having the pot to set the temperature is
that you can adjust it to whatever you want when the element
is to be driven at it's high end. Of course another pot to set
the idle temperature if you care to do that.
This project is starting to sound very interesting. I might just have
to try one myself. I've built several analog triac control circuits
and they work really great, so i can imagine how easy it would
be to do with a pic.
BTW, the 60Hz line sync input can come from a full wave
rectifier output without capacitance. This of course means
that you cant use a dc wall wart, but have to use an ac type
and rectify it yourself. All the analog circuits i've designed this
way used this technique, and to get around not being able to
use capacitance on the output a simple 1N4004 diode can be
used to isolate the full wave output from the filter cap, meaning
you can get both the pulsing full wave output AND the filtered
dc to run the circuit, all with a common ground.
Just to note again, the thermistor is very non-linear, so at the high
temperature end it may be hard to accurately measure temperature
unless you choose the series resistor (used to get the divided voltage
that represents the temperature) carefully. Since in your app it
sounds like the temperatures where you need accuracy not only
dont have to be that accurate anyway, they are closely spaced
where some accuracy is needed (idle and normal settings) so i
would expect the simple thermistor and series resistor circuit
to suffice.
LEDs vs Bulbs, LEDs are winning.
Again, thank you very much.
I have to admit though I understand maybe 1 in 10 of your words! I am studying Product Design and have had very little specific experience in pure electronics, this soldering iron is my final year project so I think I'll take your fantastic description to my electronics professor and maybe ask him to translate for me!
It's a lot more complicated than I expected. I also have limited physical space inside the product. Perhaps I could get some more advice once I've made a start breadboarding the circuit and understand it all a bit better.
Adam
I have to admit though I understand maybe 1 in 10 of your words! I am studying Product Design and have had very little specific experience in pure electronics, this soldering iron is my final year project so I think I'll take your fantastic description to my electronics professor and maybe ask him to translate for me!
It's a lot more complicated than I expected. I also have limited physical space inside the product. Perhaps I could get some more advice once I've made a start breadboarding the circuit and understand it all a bit better.
Adam
Hi again Adam,
[See also added real-life tests below]
Oh ok sure.
Just to add one more small note, and that is that the thermistors
also have a limit as to how much current you can run through
them, so that means the value of the series resistor is limited
to high enough values to avoid this problem.
I too have not been able to find any themistors that go above
300 deg C. I thought about using copper wire as the resistance
would double from room temperature to 300 deg C, but unfortunately
it changes very little from 270 deg C to 300 deg C. This would mean
some sort of amplification would be needed and i have doubts
as to how well this would work anyway.
Another idea is to ensure some heat loss before the thermistor.
If the thermistor was mounted inside the handle it could perhaps
sense the 'cool' end of the metal part of the irons heat element.
This end would not be as hot so perhaps a standard themistor
could be used. The drawback would be that there would be some
delay introduced in the feedback and some inaccuracy. For example,
say the user moves the iron back and forth, the thermistor will cool
down a little more than it was even though the tip might not change
as much. There might also have to be a layer of insulation completely
around the thermistor except where it touches the metal so that it
doesnt matter if the thermistor is on top, on the side, or on the bottom.
The thermistor could be fixed with some thermal epoxy (Arctic Alumina
comes to mind). Some of the metal tubes are hollow so perhaps the
thermistor could be mounted inside the tube after the handle has
been removed.
If i get a chance a little later i'll tack a standard 100 deg C thermistor
to the cool end of my soldering iron and see how hot that end gets.
Of course a better way would be to have the sensor right inside the
tip, where it's temperature change could be measured quickly.
I guess a thermocouple would have to be used however.
There are lots of things that change their dynamic resistance as the
temperature changes, such as diodes, but they are not rated up
to temperatures that high either. I also looked up NiChrome wire,
and although that has a higher initial resistance (than copper wire)
its sensitivity to temperature is only something like one-tenth of
that of copper wire, so i dont see that working either.
Another idea would be to investigate how commercial irons with temperature
regulation work...what kind of sensor they use, or perhaps someone
else on this forum even knows already and can share?
Perhaps i'll ask...
ADDED LATER:
Ok, i was able to set up a quick experiment to see what would happen
if a thermistor (110 deg C max type) was mounted at the 'cool' end
of the soldering iron. The cool end is the end where the heat element
tube attaches to the handle of the iron.
What i noticed right away was the first problem:
Since the thermistor is mounted at the cool end, that puts the
heat element BETWEEN the tip and the thermistor, probably the
worst place it could possibly be mounted. This in itself is probably
good enough cause to scrap the whole 110 deg C thermistor idea
in itself, because the heat travels in one path to get to the tip and
another path to get to the thermistor. This means that the thermistor
does not really sense tip temperature. It might have 'something' to
do with the tip temperature, but another problem quickly showed its
ugly face...
The response time of the thermistor is extremely slow. I mean really
slow. The settling time for a change in element voltage of 10 volts
is about 5 full minutes, and the time it takes to change even a little
(5 ohms out of 2k) to a change of 110v to the heat element (big big
change) is a full minute. This doesnt mean this control method wont
work at all, but would be a rather shabby approach.
There is also the question of what the thermistor senses when the
tip is used to heat a large object...it's temperature will drop fast,
within a second, and what has to happen next in order for the
thermistor to sense anything at all is the element temperature has to
drop, because that is considered a 'node' sitting directly between
the tip and the thermistor. Technically, the element temperature should
actually drop, and then that means less heat getting to the thermistor,
but that's going to take much time, and even then we still have to take
into consideration the time constant of the thermistor, which could be
as much as 20 seconds.
So it's back to square one: try to find a better sensor that will sense up
to the irons highest tip temperature so the tip temperature itself can be
measured directly. Im starting to see this as the only right way to do
this. If we are going to bother to do it at all, sense tip temperature.
I should add that the thermistor at the cool end technique could be used
for limited temperature measurement of the iron, say to indicate when
it was safe to handle the iron. It might also be used to indicate the
general 'mode' of the iron (full, idle, off), because the slow time constant
would not matter for this application. It's only when we try to regulate
the tip temperature that this idea will not work very well.
[See also added real-life tests below]
Oh ok sure.
Just to add one more small note, and that is that the thermistors
also have a limit as to how much current you can run through
them, so that means the value of the series resistor is limited
to high enough values to avoid this problem.
I too have not been able to find any themistors that go above
300 deg C. I thought about using copper wire as the resistance
would double from room temperature to 300 deg C, but unfortunately
it changes very little from 270 deg C to 300 deg C. This would mean
some sort of amplification would be needed and i have doubts
as to how well this would work anyway.
Another idea is to ensure some heat loss before the thermistor.
If the thermistor was mounted inside the handle it could perhaps
sense the 'cool' end of the metal part of the irons heat element.
This end would not be as hot so perhaps a standard themistor
could be used. The drawback would be that there would be some
delay introduced in the feedback and some inaccuracy. For example,
say the user moves the iron back and forth, the thermistor will cool
down a little more than it was even though the tip might not change
as much. There might also have to be a layer of insulation completely
around the thermistor except where it touches the metal so that it
doesnt matter if the thermistor is on top, on the side, or on the bottom.
The thermistor could be fixed with some thermal epoxy (Arctic Alumina
comes to mind). Some of the metal tubes are hollow so perhaps the
thermistor could be mounted inside the tube after the handle has
been removed.
If i get a chance a little later i'll tack a standard 100 deg C thermistor
to the cool end of my soldering iron and see how hot that end gets.
Of course a better way would be to have the sensor right inside the
tip, where it's temperature change could be measured quickly.
I guess a thermocouple would have to be used however.
There are lots of things that change their dynamic resistance as the
temperature changes, such as diodes, but they are not rated up
to temperatures that high either. I also looked up NiChrome wire,
and although that has a higher initial resistance (than copper wire)
its sensitivity to temperature is only something like one-tenth of
that of copper wire, so i dont see that working either.
Another idea would be to investigate how commercial irons with temperature
regulation work...what kind of sensor they use, or perhaps someone
else on this forum even knows already and can share?
Perhaps i'll ask...
ADDED LATER:
Ok, i was able to set up a quick experiment to see what would happen
if a thermistor (110 deg C max type) was mounted at the 'cool' end
of the soldering iron. The cool end is the end where the heat element
tube attaches to the handle of the iron.
What i noticed right away was the first problem:
Since the thermistor is mounted at the cool end, that puts the
heat element BETWEEN the tip and the thermistor, probably the
worst place it could possibly be mounted. This in itself is probably
good enough cause to scrap the whole 110 deg C thermistor idea
in itself, because the heat travels in one path to get to the tip and
another path to get to the thermistor. This means that the thermistor
does not really sense tip temperature. It might have 'something' to
do with the tip temperature, but another problem quickly showed its
ugly face...
The response time of the thermistor is extremely slow. I mean really
slow. The settling time for a change in element voltage of 10 volts
is about 5 full minutes, and the time it takes to change even a little
(5 ohms out of 2k) to a change of 110v to the heat element (big big
change) is a full minute. This doesnt mean this control method wont
work at all, but would be a rather shabby approach.
There is also the question of what the thermistor senses when the
tip is used to heat a large object...it's temperature will drop fast,
within a second, and what has to happen next in order for the
thermistor to sense anything at all is the element temperature has to
drop, because that is considered a 'node' sitting directly between
the tip and the thermistor. Technically, the element temperature should
actually drop, and then that means less heat getting to the thermistor,
but that's going to take much time, and even then we still have to take
into consideration the time constant of the thermistor, which could be
as much as 20 seconds.
So it's back to square one: try to find a better sensor that will sense up
to the irons highest tip temperature so the tip temperature itself can be
measured directly. Im starting to see this as the only right way to do
this. If we are going to bother to do it at all, sense tip temperature.
I should add that the thermistor at the cool end technique could be used
for limited temperature measurement of the iron, say to indicate when
it was safe to handle the iron. It might also be used to indicate the
general 'mode' of the iron (full, idle, off), because the slow time constant
would not matter for this application. It's only when we try to regulate
the tip temperature that this idea will not work very well.
LEDs vs Bulbs, LEDs are winning.
Wow, thanks for your help Al.
Yeah, I spoke to one of my lecturers today and I think a thermocouple is the only feasable route, my original plan was to hold it at the tip and I think i'll be sticking to this becuase I have some extra design features at the tip that would easily accomodate this, possibly without the need for any thermal epoxy (to allow for tip changes etc). Also by the time the iron cools down after switch off, all areas of the iron's shaft reach 50 Degrees at about the same time due to the thermal equilibrium being reached so once the tip is safe to touch it's a safe enough bet to say the rest of it is too.
I should have a thermoouple - to - digital - converter chip on the way from a company called Maxim-IC, I can then link this up to a PIC input to control my LED and heating element relay outputs (that is once I've taught myself how to program PICs!!!).
Thanks for all your help, I'll keep you posted on how the project is getting on if you like.
Adam
Yeah, I spoke to one of my lecturers today and I think a thermocouple is the only feasable route, my original plan was to hold it at the tip and I think i'll be sticking to this becuase I have some extra design features at the tip that would easily accomodate this, possibly without the need for any thermal epoxy (to allow for tip changes etc). Also by the time the iron cools down after switch off, all areas of the iron's shaft reach 50 Degrees at about the same time due to the thermal equilibrium being reached so once the tip is safe to touch it's a safe enough bet to say the rest of it is too.
I should have a thermoouple - to - digital - converter chip on the way from a company called Maxim-IC, I can then link this up to a PIC input to control my LED and heating element relay outputs (that is once I've taught myself how to program PICs!!!).
Thanks for all your help, I'll keep you posted on how the project is getting on if you like.
Adam
Who is online
Users browsing this forum: No registered users and 27 guests