Temp readings with thermistor question

[rob]

Hey... I need to extend the length of the sensor wire of an el-cheepo
radio shack outdoor thermometer. How long before I run into a
problem??...also if I replaced sensor with another thermistor with
different properties will read out of temperature be off?? if so will
it be consistantly off so I could easily adust after I calibrated?

Thanks rob

 

[ChrisGibboGibson]

Hey... I need to extend the length of the sensor wire of an el-cheepo
radio shack outdoor thermometer. How long before I run into a
problem??

Been there, done it. About 60 feet total cable length.

...also if I replaced sensor with another thermistor with
different properties will read out of temperature be off??

Yes

if so will
it be consistantly off so I could easily adust after I calibrated?

No

Gibbo

 

[Baz]

I tried to extend the cable on a Jaycar indoor outdoor thermometer without
success.

From what I could work out the cable's capacitance is part of an RC
oscillator, as well as the thermistor. The frequency of the oscillator is
counted to obtain the temperature. Extending the cable (in my case an extra
10 meters or so using telephone cable) resulted in unusable readings. As
the same counter was used for both indoor and outdoor reading I couldn't see
any easy way around this problem so I made do without an extended cable.

This may not be the case with your unit.

Regards
Barry

 

[rob]

Thanks Guys....we are doing some solar air heating work and would like
to monitor 9 spots...on the cheap.....my idea is to make 8 sensors and
run wires to spot where temperature display unit is located....either
have wires run to switch or hook up manually one by one when readings
are taken....My understanding (limited) is that I can run thermistors
to a volt meter and temps can be calculated from resitance...is this
correct....we would like to eventually log them with a computer but
that is down the road a bit...thanks again.

Rob

 

[John Popelish]

Thanks Guys....we are doing some solar air heating work and would like
to monitor 9 spots...on the cheap.....my idea is to make 8 sensors and
run wires to spot where temperature display unit is located....either
have wires run to switch or hook up manually one by one when readings
are taken....My understanding (limited) is that I can run thermistors
to a volt meter and temps can be calculated from resitance...is this
correct....we would like to eventually log them with a computer but
that is down the road a bit...thanks again.

Rob

You have the right idea. The limitations using DC (immune to the
capacitance of the cable) are that the current that produces the
voltage drop across the thermistor must not heat the thermistor enough
to matter and must not drop enough voltage along the wire to matter.
Thermistors with a room temperature resistance of something like 10k
ohms make the wiring resistance pretty insignificant, and as long as
you pass a fraction of a milliamp through them do not have much self
heating. So if you can come up with a 100 microamp meter movement,
you can make a fair thermometer with only a voltage regulated supply
and some resistors and a multi pole selector switch. You will have to
make a custom meter face that indicates temperature on a nonlinear
scale. Later, a multiplexer and an A/D converter can connect the
thermistors to a computer.

 

[rob]

The limitations using DC (immune to the> capacitance of the cable) are
that the current that produces the

voltage drop across the thermistor must not heat the thermistor enough
to matter and must not drop enough voltage along the wire to matter.
Thermistors with a room temperature resistance of something like 10k
ohms make the wiring resistance pretty insignificant, and as long as
you pass a fraction of a milliamp through them do not have much self
heating. So if you can come up with a 100 microamp meter movement,
you can make a fair thermometer with only a voltage regulated supply
and some resistors and a multi pole selector switch. You will have to
make a custom meter face that indicates temperature on a nonlinear
scale. Later, a multiplexer and an A/D converter can connect the
thermistors to a computer.

THANK YOU John

Awesome advice
I happen to have a 10K @25°C thermistor in my hands and they are
readily available...stated temp ranges (-50 to 110°C) will be fine for
all but one location where we recorded 125°C last summer...melted the
plastic thermometer that was in there ...lol....Will going 110 ruin
the thermistor?....Should I use solid or stranded wires from
thermistor to display???....can I power it with an old cell phone
adapter??....5 VDC 260mA output or would a different one be
better?...

Thanks again
Rob

 

[John Popelish]

The limitations using DC (immune to the> capacitance of the cable) are
that the current that produces the

THANK YOU John

Awesome advice
I happen to have a 10K @25°C thermistor in my hands and they are
readily available...stated temp ranges (-50 to 110°C) will be fine for
all but one location where we recorded 125°C last summer

Are you talking about ambient temperature or some heated device? 125C
is well over the boiling point of water.

...melted the
plastic thermometer that was in there ...lol....Will going 110 ruin
the thermistor?

It might. It depends on the kind of thermistor and what it is
packaged in. The glass encased ones often handle this temperature.
But for that high a temperature, you will probably want to start with
something higher resistance than 10k. Does one thermistor have to
handle the whole range of temperatures? By the way, what range are
you looking at?

....Should I use solid or stranded wires from
thermistor to display???....

Inconsequential, unless there will be lots of movement of the wires.
Phone wire or cat 5 cable (6 conductor network cable) would probably
work fine.

can I power it with an old cell phone
adapter??....5 VDC 260mA output or would a different one be
better?...

That is probably plenty of voltage, but it should be regulated or add
a regulator to it, so that changes in line voltage will not shift your
calibration.

 

[Nicholas O. Lindan]

not drop [too much] voltage along the wire
[from the thermistor] to the meter.

Real world - industrial control and laboratory precision - thermistors
are three or four wire. Four wire thermistors have current on one set
of wires and voltage sensing on the other. A 3-wire set up measures the drop in
one of the wires, doubles it and uses this to correct for total wire drop.

Most measurements are made ratiometrically, either manually using a bridge
or automatically with a voltage divider. The thermistor is in series with
a high-precision high-stability resistor and the ratio of the voltages across
the thermistor and the sense resistor follows the ratio of the resistances.
These methods make the measurement relatively immune to changes in excitation
current or voltage. In harsh environment instrumentation the sense resistor
may have a thermistor attached to it to allow removal of it's thermal drift.

Most industrial thermistors are 100 ohm, though there is really no limit on
value. 1mA is the standard excitation. The itty-bitty fly-shit sized
thermistors used in precision instrumentation are often 10K as are some
other laboratory types.

Self heating is mitigated with tight coupling of the thermistor to the
process and having the thermistor thermal mass << process thermal mass.
Hence the fly-shit size of some.

You will have to make a custom meter face that indicates temperature
on a nonlinear scale.

You can also use the tables produced by the National Bureau of Standards.
They publish a telephone directory sized book with all the thermistor and
thermocouple tables.

Later, a multiplexer and an A/D converter can connect the thermistors
to a computer.

Well, what else would you do? If it isn't virtual and digital it just
doesn't exist these days.

There are formulas for converting resistance to temperature. The NBofS
gives the coefficients everyone has agreed on using and you just use
the values in your software.

You can use any odball material you want for a thermistor, thin iron wire
works well. In this case a polynomial curve fit can be made from a few
calibrated values. Often one is only interested at holding a temperature
steady and a 1-point calibration works fine.

 

[Spehro Pefhany]

Most industrial thermistors are 100 ohm, though there is really no limit on
value. 1mA is the standard excitation. The itty-bitty fly-shit sized
thermistors used in precision instrumentation are often 10K as are some
other laboratory types.

You're thinking of Pt RTDs, which are primarily 100-ohm, DIN curve.
The term "industrial thermistor" is practically an oxymoron. There are
a few company-specific "standard" RTDs used in HVAC applications that
are made from base metals. There's a few almost-standard thermistors
used in HVAC applications.

You can use any odball material you want for a thermistor, thin iron wire
works well. In this case a polynomial curve fit can be made from a few
calibrated values. Often one is only interested at holding a temperature
steady and a 1-point calibration works fine.

Thermistors are made from bulk semiconductor material. RTDs are metal
film or wirewound bulk metal or alloy devices. They both change
resistance with temperature, but RTDs are PTC only and have a change
per degree C about an order of magnitude less than thermistors (and an
order of magnitude more than base-metal thermocouples).


Best regards,
Spehro Pefhany

 

[rob]

John

This projoct is a 120 sq ft (6X20') solar air heater. Prior to
entering heater, air gets preheated by flowing between new black steel
roof and old tar-gravel. After going through actual heater, heat is
blown down 3 floors to basement to a room that is 6x7x8 and FILLED
with 45 gal. drums and large plastic pop bottles full of water. This
thermal mass of water allows the heat to be stored and released at
night.

Are you talking about ambient temperature or some heated device? 125C
is well over the boiling point of water.

Yea...we have burst steel drums, melted plastic, baked plywood to a
golden brown color similar to Thankgiving Turkeys, and, broken LOTS of
non-tempered plate glass...lol ....Our new mantra now is "high temp
materials"

It might. It depends on the kind of thermistor and what it is
packaged in. The glass encased ones often handle this temperature.

But for that high a temperature, you will probably want to start with
something higher resistance than 10k. Does one thermistor have to
handle the whole range of temperatures? By the way, what range are
you looking at?

My Thermistor is a radio shack unit....no idea what it is packed
in...I can use different ones as I am still in design phase..lol..Here
in Atlantic Canada night time winter lows can get to -25°. The heater
was just built when we got the 110 but did not have fans hooked up
soooo...now that I think about it a 10K should be fine....

Inconsequential, unless there will be lots of movement of the wires.

Phone wire or cat 5 cable (6 conductor network cable) would probably
work fine.
Done


That is probably plenty of voltage, but it should be regulated or add
a regulator to it, so that changes in line voltage will not shift your
calibration.

Ahhhh...startung to get a mind grip on this...

Again...Thank you
Rob

 

[John Popelish]

John

This projoct is a 120 sq ft (6X20') solar air heater. Prior to
entering heater, air gets preheated by flowing between new black steel
roof and old tar-gravel. After going through actual heater, heat is
blown down 3 floors to basement to a room that is 6x7x8 and FILLED
with 45 gal. drums and large plastic pop bottles full of water. This
thermal mass of water allows the heat to be stored and released at
night.



Yea...we have burst steel drums, melted plastic, baked plywood to a
golden brown color similar to Thankgiving Turkeys, and, broken LOTS of
non-tempered plate glass...lol ....Our new mantra now is "high temp
materials"

My Thermistor is a radio shack unit....no idea what it is packed
in...I can use different ones as I am still in design phase..lol..Here
in Atlantic Canada night time winter lows can get to -25°. The heater
was just built when we got the 110 but did not have fans hooked up
soooo...now that I think about it a 10K should be fine....

It is probably dipped in epoxy. It will last a while at 100C. But for
the final design, glassed encased ones would probably last longer.
And the lead wires into the hot area should probably be teflon
insulated. You can change to cheaper wire at a terminal box that is
at more reasonable temperature outside the heater.

Ahhhh...startung to get a mind grip on this...

Again...Thank you
Rob

A hint about the selection of the resistors that set the range of the
reading. The network will have the most expanded scale (change of
current versus temperature) if the resistance of the network equals
the resistance of the thermistor at that temperature. So you can
expand the temperature range of interest and compress (but still get a
reading) for extremes far from that range.

Biasing the back side of the meter with a second divider allows the
meter to read zero (no current) at an arbitrarily low temperature,
even though the thermistor will still have some high resistance and be
passing some small current. This allows you to cut off the scale at a
low temperature you don't care about.

So the circuit might be something like this: (view with fixed width
font, like courier)


+5V ---+----------------------------+
| |
R1 Rth
| - + |
+-Meter--+-------------------+
| |
R2 R3
| |
0V-----+--------+

The divider R1 and R2 set the temperature where the meter reads zero
current. They can be a decade lower resistance than the thermistor,
to keep from reducing the meter sensitivity (adding series resistance
to the meter current path). If you want to reduce the sensitivity
(expand the upper end of the range) raise the values of these
resistors.

R3 (in parallel with the meter) sets the load resistance in series
with the thermistor, to determine what temperature is most expanded on
the meter scale. Rth is the remote thermistor.

To keep from pegging the meter negative each time you switch
thermistors, you could use a separate R3 for each thermistor and put
the selector switch between the meter + terminal and each R3. Break
before make contacts.

Once you have tentatively selected a meter thermistor temperature
range combination, you have a bit of math to work through to determine
what resistor values give you what you want, and even if all your
objectives are reachable. But it is just an application of ohm's law
and the table of resistance versus temperature for the thermistor.

 

[Nicholas O. Lindan]

we are doing some solar air heating work and would like
to monitor 9 spots...on the cheap.....my idea is to make 8 sensors and
run wires to spot where temperature display unit is located....either
have wires run to switch or hook up manually one by one when readings
are taken....My understanding (limited) is that I can run thermistors
to a volt meter and temps can be calculated from resitance...is this
correct....we would like to eventually log them with a computer but
that is down the road a bit...thanks again.

Most, if not all answer to temperature measurement can be found at:

http://www.omega.com/temperature/Z/zsection.asp

Omega is very good source for temperature measurement equipment and
will sell to anyone with a credit card.

ebay is not a bad source for industrial instrumentation and control,
with the stuff often going for ten cents on the dollar.

If you can severely over-heat a building with the present
setup (~100C) I would advise not doing anything 'on the cheap'.

I had posted before assuming you were using an RTD
when you used the term 'thermistor'. If you are indeed using
a true thermistor then please ignore most of the previous postings.

OTOH, I would advise using an RTD or thermocouple to measure
temperature with acceptable accuracy and stability. Feed the
temperature sensors into strip-chart data logger - ebay. You should
also have an emergency shut-down and alarming system.

 

[John Popelish]

:
(snip)

... You should also have an emergency shut-down and alarming system.

....based on an independent measurement of temperature, so that a
single device failure will not destroy property.

 

[Rich Grise]

John

This projoct is a 120 sq ft (6X20') solar air heater. Prior to
entering heater, air gets preheated by flowing between new black steel
roof and old tar-gravel. After going through actual heater, heat is
blown down 3 floors to basement to a room that is 6x7x8 and FILLED
with 45 gal. drums and large plastic pop bottles full of water. This
thermal mass of water allows the heat to be stored and released at
night.

Hope this isn't dumb or insulting, but is the basement room insulated?

Yea...we have burst steel drums, melted plastic, baked plywood to a
golden brown color similar to Thankgiving Turkeys, and, broken LOTS of
non-tempered plate glass...lol ....Our new mantra now is "high temp
materials"

Eek! That's a lot o' solar energy!

....

Ahhhh...startung to get a mind grip on this...

Cool! (so to speak ;-) )

Cheers!
Rich