Thermocouple How-To?

Started by Scott Lingerfelt November 12, 2005
Hi Everyone,

I am looking for a single channel ADC that allow the least amount of
external components and allows between 12-14bit resolution. I need low
power for battery operation. The temperature range will be the
industrial standard -40C - 85C. Do any of the ADC IC's have the amp
built-in? I want the ability to read type K, E, T, J, and N
thermocouples. Of course the PIC will be programmed for each!

Can I use the PIC's analog pins and add an amp to get the same? I am
not a hardware guy!

Any help would be greatly appreciated!

Thanks in advance, Scott



Here is a 12 bit K type in one package.
http://pdfserv.maxim-ic.com/en/ds/MAX6675.pdf

Converting the input from any thermocouple sounds like a complete
design to me, but you may dig and find something similar to the MAX
part that is programmable in the analog portion. --- Scott Lingerfelt <slingerfelt@slin...> wrote:

> Hi Everyone,
>
> I am looking for a single channel ADC that allow the least amount of
> external components and allows between 12-14bit resolution. I need
> low
> power for battery operation. The temperature range will be the
> industrial standard -40C - 85C. Do any of the ADC IC's have the amp
> built-in? I want the ability to read type K, E, T, J, and N
> thermocouples. Of course the PIC will be programmed for each!
>
> Can I use the PIC's analog pins and add an amp to get the same? I am
>
> not a hardware guy!
>
> Any help would be greatly appreciated!
>
> Thanks in advance, Scott
>


My software has no bugs, only undocumented features.



--- In piclist@picl..., Scott Lingerfelt <slingerfelt@c...> wrote:
>
> Hi Everyone,
>
> I am looking for a single channel ADC that allow the least amount of
> external components and allows between 12-14bit resolution. I need low
> power for battery operation. The temperature range will be the
> industrial standard -40C - 85C.

There is no such industry standard. There is a desired range, but
thermoncouples do not have such a small range.

Also, most circuits and chips you find will be 0 to some positive
temperature. there are fewer examples of a negative swing.

Read up on thermocouples. They have huge ranges with tiny outputs.

RTD's would be better in the range you listed. The have larger
outputs for smaller inputs.

A E type would have the most sensitive span, But, type T is more
common, being able to accuratly measure from -200 C to 350 C or there
abouts. but if you want accuracy, it is good to about 1 deg C in
resolution.

Do some math. figure the voltage you can handle, it may be 0 to 5 on
your external ADC, and then divide that by the bits to get how many
counts you can expect. a 12 bit will offer 4096 counts. since you
want a range of 125 degrees, you might think you can get a sensitivity
of 0.03 degrees, but thermocouple are just not that accurate. > Do any of the ADC IC's have the amp
> built-in?

Not to that level, but you can buy a temperature transmitter. It is
the deivce that would connect to the thermocouple, have all the
internal citcuitry and output a mA or voltage signal. The Maxim chip
goes from zero C on up. It also delivers the range of the junction
and therefore will not have the smaller range you are looking for. > I want the ability to read type K, E, T, J, and N
> thermocouples. Of course the PIC will be
> programmed for each!

Each thermocouple has a different range, so each needs a different
calculation.

It might be possible to use a jumper or switch and select the range
(thermocouple) you want and have one circuit for any variety of
thermocuples. > Can I use the PIC's analog pins and add an
> amp to get the same? I am
> not a hardware guy!


the PIC input is 10 bit, or can count to 1024 counts. so, after you
convert the raw input to a 0 to 5 range, the PIC will be able to see
the range and break it into 1024 slices. with your 125 degree range,
you could get a pretty accurate signal. >
> Any help would be greatly appreciated!


A question, What do you want to do with the signal ?

If all you are looking for is an output of voltage or mA, then buying
a temperature transmitter would be the simplest way.

If you want to output or data log or control, then bringing the
reading into the PIC would make sense. Even for the beginner, there should be enough on Google to get you
started or more confused. A starter page might be Tracey Allen's page.
Although he does BasicStamp stuff, his writing is univeral.

http://www.emesystems.com/OL2therm.htm

Hope this helps.

Dave



Hi Dave,

I apologize to all I should have been a little more detailed.

The circuitry that I am using other than the sensor has an operating
range of -40C to 85C not the sensor. The total circuit must withstand
the same environment as the sensor thus if I go outside the range I must
utilize an external sensor complicating the device.

My device must have the ability to allow the customer to attach his type
of sensor required for the application. Type K will be the most
commonly used.

The total design must be run from a single 3.6V battery and have the
ability to last 5 years no less. Less components will aid in reducing
the overall battery usage.

Due to the size requirements the device must be as small as possible
which I must eliminate as many components as possible without
compromising performance. Of course COST is always an issue so
eliminating components will help!

Based on which sensor the customer selects the PIC will use the correct
calculation which is triggered by the plug.

Thanks for your help,

Scott Dave Mucha wrote:

>--- In piclist@picl..., Scott Lingerfelt <slingerfelt@c...> wrote: >>Hi Everyone,
>>
>>I am looking for a single channel ADC that allow the least amount of
>>external components and allows between 12-14bit resolution. I need low
>>power for battery operation. The temperature range will be the
>>industrial standard -40C - 85C.
>>
>>
>
>There is no such industry standard. There is a desired range, but
>thermoncouples do not have such a small range.
>
>Also, most circuits and chips you find will be 0 to some positive
>temperature. there are fewer examples of a negative swing.
>
>Read up on thermocouples. They have huge ranges with tiny outputs.
>
>RTD's would be better in the range you listed. The have larger
>outputs for smaller inputs.
>
>A E type would have the most sensitive span, But, type T is more
>common, being able to accuratly measure from -200 C to 350 C or there
>abouts. but if you want accuracy, it is good to about 1 deg C in
>resolution.
>
>Do some math. figure the voltage you can handle, it may be 0 to 5 on
>your external ADC, and then divide that by the bits to get how many
>counts you can expect. a 12 bit will offer 4096 counts. since you
>want a range of 125 degrees, you might think you can get a sensitivity
>of 0.03 degrees, but thermocouple are just not that accurate. >
>
>>Do any of the ADC IC's have the amp
>>built-in?
>>
>>
>
>Not to that level, but you can buy a temperature transmitter. It is
>the deivce that would connect to the thermocouple, have all the
>internal citcuitry and output a mA or voltage signal. The Maxim chip
>goes from zero C on up. It also delivers the range of the junction
>and therefore will not have the smaller range you are looking for. >
>
>>I want the ability to read type K, E, T, J, and N
>>thermocouples. Of course the PIC will be
>>programmed for each!
>>
>>
>
>Each thermocouple has a different range, so each needs a different
>calculation.
>
>It might be possible to use a jumper or switch and select the range
>(thermocouple) you want and have one circuit for any variety of
>thermocuples. >
>
>>Can I use the PIC's analog pins and add an
>>amp to get the same? I am
>>not a hardware guy!
>>
> >the PIC input is 10 bit, or can count to 1024 counts. so, after you
>convert the raw input to a 0 to 5 range, the PIC will be able to see
>the range and break it into 1024 slices. with your 125 degree range,
>you could get a pretty accurate signal. >
>
>>Any help would be greatly appreciated!
>>
> >A question, What do you want to do with the signal ?
>
>If all you are looking for is an output of voltage or mA, then buying
>a temperature transmitter would be the simplest way.
>
>If you want to output or data log or control, then bringing the
>reading into the PIC would make sense. >Even for the beginner, there should be enough on Google to get you
>started or more confused. >A starter page might be Tracey Allen's page.
>Although he does BasicStamp stuff, his writing is univeral.
>
>http://www.emesystems.com/OL2therm.htm
>
>Hope this helps.
>
>Dave >
>to unsubscribe, go to http://www.yahoogroups.com and follow the instructions
>Yahoo! Groups Links >
>



Hmmm...
I think you need to think thoroughly trough this, and do the math.

A:
different TK series have different connectors, so there goes one of your requirements.

B:
there is probably no way of electronically detect what sensor is connected, they all are
virtually zero ohms and output about the same very low voltage (uV range)...so there goes another one

C:
5 years @ 3.6V ?  Doable...it all depends on the size of the battery ! If the circuitry is to
run not 'on demand' but continuously then expect an 'impractical' battery size ( I assume the
gizmo is to be portable?)
 

You'll also probably need lookup tables for each sensor type as they are slightly non-linear,
and this is different for each type.
The temp range for the cold junction may be a challenge, and this one will need a different lookup
table on its own...one for each of the sensor types...

How's that for encouragement? :o)

I wouldn't take on a job like this...
 
 

Scott Lingerfelt wrote:

 Hi Dave,

I apologize to all I should have been a little more detailed.

The circuitry that I am using other than the sensor has an operating
range of -40C to 85C not the sensor.  The total circuit must withstand
the same environment as the sensor thus if I go outside the range I must
utilize an external sensor complicating the device.

My device must have the ability to allow the customer to attach his type
of sensor required for the application.  Type K will be the most
commonly used.

The total design must be run from a single 3.6V battery and have the
ability to last 5 years no less.   Less components will aid in reducing
the overall battery usage.

Due to the size requirements the device must be as small as possible
which I must eliminate as many components as possible without
compromising performance.  Of course COST is always an issue so
eliminating components will help!

Based on which sensor the customer selects the PIC will use the correct
calculation which is triggered by the plug.

Thanks for your help,

Scott

--
*******************************************
VISIT MY HOME PAGE:
<http://home.online.no/~eikarlse/index.htm>
LAST UPDATED: 23/08/2003
*******************************************
Regards
Eirik Karlsen
 

Hi Eirik,

Yes, I to am questioning this project but it can be done but for the price!

Eirik Karlsen wrote:

> Hmmm...
> I think you need to think thoroughly trough this, and do the math.
>
> A:
> different TK series have different connectors, so there goes one of
> your requirements.
>
The manufacturer has specified the different connectors they will be
using in their facilities. This will be five total! Each sensor will
have their own port for connection.

> B:
> there is probably no way of electronically detect what sensor is
> connected, they all are
> virtually zero ohms and output about the same very low voltage (uV
> range)...so there goes another one
>
I use the same matching connector for each sensor. Each are slightly
modified that each has a spring loaded switch that is monitored by the
PIC. If any have been triggered I will know which sensor has been attached.

> C:
> 5 years @ 3.6V ? Doable...it all depends on the size of the battery !
> If the circuitry is to
> run not 'on demand' but continuously then expect an 'impractical'
> battery size ( I assume the
> gizmo is to be portable?)
>
I always target a power source were possible but the customer request
that all units to be wireless and battery operated. I am also exploring
the possibility of using rechargable batteries.

I have many sensor that I have developed over the years lasting up to 10
years. I implement a cyclic sleep and each are programmable from 10s up
to 5min. The battery is sized on the 10s sample rate and other
factors. The device can be commissioned a de-commissioned in place or
they can also be mobile. Depending on the product line the manufacturer
is running will determine the location.

>
>
> You'll also probably need lookup tables for each sensor type as they
> are slightly non-linear,
> and this is different for each type.
> The temp range for the cold junction may be a challenge, and this one
> will need a different lookup
> table on its own...one for each of the sensor types...
>
Yes, this will be a tough one but possible! I have been testing a few
equations that I have generated from a curve fitting program. I am very
impressed with the accuracy of around 99.9%. I feed in the curve for
the sensor and it generates linear and non-linear equations. This
eliminates the need for tables.

> How's that for encouragement? :o)
>
> I wouldn't take on a job like this...
>
Yes, thanks for the encouragement! The challenge is what attracted me
to this project.

Scott

> Scott Lingerfelt wrote:
>
>> Hi Dave,
>>
>> I apologize to all I should have been a little more detailed.
>>
>> The circuitry that I am using other than the sensor has an operating
>> range of -40C to 85C not the sensor. The total circuit must withstand
>> the same environment as the sensor thus if I go outside the range I must
>> utilize an external sensor complicating the device.
>>
>> My device must have the ability to allow the customer to attach his type
>> of sensor required for the application. Type K will be the most
>> commonly used.
>>
>> The total design must be run from a single 3.6V battery and have the
>> ability to last 5 years no less. Less components will aid in reducing
>> the overall battery usage.
>>
>> Due to the size requirements the device must be as small as possible
>> which I must eliminate as many components as possible without
>> compromising performance. Of course COST is always an issue so
>> eliminating components will help!
>>
>> Based on which sensor the customer selects the PIC will use the correct
>> calculation which is triggered by the plug.
>>
>> Thanks for your help,
>>
>> Scott
>>
> --
> *******************************************
> VISIT MY HOME PAGE:
> <http://home.online.no/~eikarlse/index.htm
> <http://home.online.no/%7Eeikarlse/index.htm>>
> LAST UPDATED: 23/08/2003
> *******************************************
> Regards
> Eirik Karlsen > to unsubscribe, go to http://www.yahoogroups.com and follow the
> instructions >
> >. >




> The circuitry that I am using other than the sensor has an operating
> range of -40C to 85C not the sensor. The total circuit must withstand
> the same environment as the sensor thus if I go outside the range I
must
> utilize an external sensor complicating the device.

OK, that makes a lot more sense.

That range is still pretty horrific to find and make parts for.

espically in these low sensor ranges. >
> My device must have the ability to allow
> the customer to attach his type
> of sensor required for the application.
> Type K will be the most
> commonly used.

possible with a specific units. ie: A 'T' type unit and an "K' type
unit, not a universal unit. > The total design must be run from a
> single 3.6V battery and have the
> ability to last 5 years no less.
> Less components will aid in reducing
> the overall battery usage.

Wow, nothing like raising the bar is there ?

Obviously, you are going to connect to something else no ?

can you use power from that remote thing ? the whole 4-20mA concept
gives you that 4mA to power your micro-amp circuit. life goes out
decades... battery will die in the cold well before it dies of demand.

Duracell rates it's Akalines as having 85% life after 4 years storage.
not useage.
http://www.duracell.com/oem/primary/alkaline/alkaline_manganese_over.asp

But, supposing your 5 year drain is able to be less than the battery
power, I am not sure if you can run the thing constantly. a solar
cell would help immensly.

I assume you could have the thing sleep 90% of the time ?

is the output going to be serial or analogue ?
> Due to the size requirements the device must be as small as possible
> which I must eliminate as many components as possible without
> compromising performance. Of course COST is always an issue so
> eliminating components will help!
>
> Based on which sensor the customer selects the PIC will use the correct
> calculation which is triggered by the plug.

The whole thermocoulple, dissimlar metals thing is based on the cold
temperature jucntion. if you have copper-constatine wire and connect
it to a copper-iron plug, you have to also have a cold junction
circuit to montior the conditons at the connections. I can't help but wonder if you can power the thing from remote power ?
I assume it will be connected to some other devices. and those would
have the ability to supply power ?

If it is for a data logger, you can wake and sleep and write to memory
for years on a button cell.

I had a single channle HOBO run for a couple years on a button.
http://www.onsetcomp.com/Products/indoor_guide.html

Dave



A little hint about batteries..

I've developed something similar, an RF tracking system, battery operated
with a minimum 2 years life.

As battery, I've found out that the best choice was using a Lithium-Thyonil
Chlorure battery. The energy stored in these battery is huge, compared to
alkaline and other chemicals, you can get 2700 mAh @ 3.6V from an AA size
battery. The discharge curve is quite good, the voltage is near 3.6V for the
whole lifetime and it falls down only when the battery charge very low. The
drawbacks of these battery are the low peak current (about 20 mA, but this
is your case luckily) and the high cost, I got my SAFT 2300 mA batteries at
3€ each from a local distributor here in Italy. You can get a lower price
(2.2€ maybe) for bigger quantities.
Good manufacturer of Lithium Chlorure Thyonil batteries are SAFT and Maxell.
There are others but I've found out that these two are the cheapest and
looking at the datasheets they also provided the best batteries.

Of course, the key for a long life, is the CPU sleeping time. In my
application it was 97%.

Hope this helps.

Regards

Vito ----- Original Message -----
From: "Dave Mucha" <dave_mucha@dave...>
To: <piclist@picl...>
Sent: Monday, November 14, 2005 4:42 AM
Subject: [piclist] Re: Thermocouple How-To? >
> > The circuitry that I am using other than the sensor has an operating
> > range of -40C to 85C not the sensor. The total circuit must withstand
> > the same environment as the sensor thus if I go outside the range I
> must
> > utilize an external sensor complicating the device.
>
> OK, that makes a lot more sense.
>
> That range is still pretty horrific to find and make parts for.
>
> espically in these low sensor ranges. > >
> > My device must have the ability to allow
> > the customer to attach his type
> > of sensor required for the application.
> > Type K will be the most
> > commonly used.
>
> possible with a specific units. ie: A 'T' type unit and an "K' type
> unit, not a universal unit. > > The total design must be run from a
> > single 3.6V battery and have the
> > ability to last 5 years no less.
> > Less components will aid in reducing
> > the overall battery usage.
>
> Wow, nothing like raising the bar is there ?
>
> Obviously, you are going to connect to something else no ?
>
> can you use power from that remote thing ? the whole 4-20mA concept
> gives you that 4mA to power your micro-amp circuit. life goes out
> decades... battery will die in the cold well before it dies of demand.
>
> Duracell rates it's Akalines as having 85% life after 4 years storage.
> not useage.
> http://www.duracell.com/oem/primary/alkaline/alkaline_manganese_over.asp
>
> But, supposing your 5 year drain is able to be less than the battery
> power, I am not sure if you can run the thing constantly. a solar
> cell would help immensly.
>
> I assume you could have the thing sleep 90% of the time ?
>
> is the output going to be serial or analogue ? >
> > Due to the size requirements the device must be as small as possible
> > which I must eliminate as many components as possible without
> > compromising performance. Of course COST is always an issue so
> > eliminating components will help!
> >
> > Based on which sensor the customer selects the PIC will use the correct
> > calculation which is triggered by the plug.
>
> The whole thermocoulple, dissimlar metals thing is based on the cold
> temperature jucntion. if you have copper-constatine wire and connect
> it to a copper-iron plug, you have to also have a cold junction
> circuit to montior the conditons at the connections. > I can't help but wonder if you can power the thing from remote power ?
> I assume it will be connected to some other devices. and those would
> have the ability to supply power ?
>
> If it is for a data logger, you can wake and sleep and write to memory
> for years on a button cell.
>
> I had a single channle HOBO run for a couple years on a button.
> http://www.onsetcomp.com/Products/indoor_guide.html
>
> Dave >
> to unsubscribe, go to http://www.yahoogroups.com and follow the
instructions
> Yahoo! Groups Links



Hi Vito,

Thanks for the reply! I have been using the Tadiran and Saft for a
couple of years and found them to be very reliable in harsh environments
and they have a lot of power. I have been using the Tadiran AA and it
also has around 2400 - 2700 mAh @ 3.6V. I am also creating an RF
solution that will be sleeping majority of the time until polled. I
have a circuit that can be activated to wake up the chip similiar to an
RFID tag would.

I have a couple of A/D chips coming within the next couple of days to
start testing and prototyping. I will let everyone know the outcome.

Thanks,

Scott

Vito Russo wrote:

>A little hint about batteries..
>
>I've developed something similar, an RF tracking system, battery operated
>with a minimum 2 years life.
>
>As battery, I've found out that the best choice was using a Lithium-Thyonil
>Chlorure battery. The energy stored in these battery is huge, compared to
>alkaline and other chemicals, you can get 2700 mAh @ 3.6V from an AA size
>battery. The discharge curve is quite good, the voltage is near 3.6V for the
>whole lifetime and it falls down only when the battery charge very low. The
>drawbacks of these battery are the low peak current (about 20 mA, but this
>is your case luckily) and the high cost, I got my SAFT 2300 mA batteries at
>3€ each from a local distributor here in Italy. You can get a lower price
>(2.2€ maybe) for bigger quantities.
>Good manufacturer of Lithium Chlorure Thyonil batteries are SAFT and Maxell.
>There are others but I've found out that these two are the cheapest and
>looking at the datasheets they also provided the best batteries.
>
>Of course, the key for a long life, is the CPU sleeping time. In my
>application it was 97%.
>
>Hope this helps.
>
>Regards
>
>Vito >----- Original Message -----
>From: "Dave Mucha" <dave_mucha@dave...>
>To: <piclist@picl...>
>Sent: Monday, November 14, 2005 4:42 AM
>Subject: [piclist] Re: Thermocouple How-To? >
>
>>>The circuitry that I am using other than the sensor has an operating
>>>range of -40C to 85C not the sensor. The total circuit must withstand
>>>the same environment as the sensor thus if I go outside the range I
>>>
>>>
>>must
>>
>>
>>>utilize an external sensor complicating the device.
>>>
>>>
>>OK, that makes a lot more sense.
>>
>>That range is still pretty horrific to find and make parts for.
>>
>>espically in these low sensor ranges.
>>
>>
>>
>>
>>>My device must have the ability to allow
>>>the customer to attach his type
>>>of sensor required for the application.
>>>Type K will be the most
>>>commonly used.
>>>
>>>
>>possible with a specific units. ie: A 'T' type unit and an "K' type
>>unit, not a universal unit.
>>
>>
>>
>>
>>>The total design must be run from a
>>>single 3.6V battery and have the
>>>ability to last 5 years no less.
>>>Less components will aid in reducing
>>>the overall battery usage.
>>>
>>>
>>Wow, nothing like raising the bar is there ?
>>
>>Obviously, you are going to connect to something else no ?
>>
>>can you use power from that remote thing ? the whole 4-20mA concept
>>gives you that 4mA to power your micro-amp circuit. life goes out
>>decades... battery will die in the cold well before it dies of demand.
>>
>>Duracell rates it's Akalines as having 85% life after 4 years storage.
>> not useage.
>>http://www.duracell.com/oem/primary/alkaline/alkaline_manganese_over.asp
>>
>>But, supposing your 5 year drain is able to be less than the battery
>>power, I am not sure if you can run the thing constantly. a solar
>>cell would help immensly.
>>
>>I assume you could have the thing sleep 90% of the time ?
>>
>>is the output going to be serial or analogue ?
>>
>>
>>
>>
>>
>>>Due to the size requirements the device must be as small as possible
>>>which I must eliminate as many components as possible without
>>>compromising performance. Of course COST is always an issue so
>>>eliminating components will help!
>>>
>>>Based on which sensor the customer selects the PIC will use the correct
>>>calculation which is triggered by the plug.
>>>
>>>
>>The whole thermocoulple, dissimlar metals thing is based on the cold
>>temperature jucntion. if you have copper-constatine wire and connect
>>it to a copper-iron plug, you have to also have a cold junction
>>circuit to montior the conditons at the connections.
>>
>>
>>I can't help but wonder if you can power the thing from remote power ?
>> I assume it will be connected to some other devices. and those would
>>have the ability to supply power ?
>>
>>If it is for a data logger, you can wake and sleep and write to memory
>>for years on a button cell.
>>
>>I had a single channle HOBO run for a couple years on a button.
>>http://www.onsetcomp.com/Products/indoor_guide.html
>>
>>Dave
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>to unsubscribe, go to http://www.yahoogroups.com and follow the
>>
>>
>instructions >>Yahoo! Groups Links
>>
>>
>>
>>
>>
>>
> >
>
>to unsubscribe, go to http://www.yahoogroups.com and follow the instructions
>Yahoo! Groups Links >
>




>--- In piclist@picl..., Scott Lingerfelt <slingerfelt@c...> wrote:
> >
> > Hi Everyone,
> >
> > I am looking for a single channel ADC that allow the least amount of
> > external components and allows between 12-14bit resolution. I need low
> > power for battery operation. The temperature range will be the
> > industrial standard -40C - 85C.

Maxim APPLICATION NOTE 991 - The 1-Wire Thermocouple
<http://www.maxim-ic.com/appnotes.cfm/appnote_number/991>

Pretty low component count but I assume the chip is expensive (Maxim).

dwayne

--
Dwayne Reid <dwayner@dway...>
Trinity Electronics Systems Ltd Edmonton, AB, CANADA
(780) 489-3199 voice (780) 487-6397 fax

Celebrating 21 years of Engineering Innovation (1984 - 2005)
.-. .-. .-. .-. .-. .-. .-. .-. .-. .-
`-' `-' `-' `-' `-' `-' `-' `-' `-'
Do NOT send unsolicited commercial email to this email address.
This message neither grants consent to receive unsolicited
commercial email nor is intended to solicit commercial email.