Matronics Email Lists :: View topic

Matronics Email Lists
Web Forum Interface to the Matronics Email Lists

Get Email Distribution Too! FAQ Search Memberlist Usergroups Register

Profile Log in to check your private messages Log in

Thermocouple leads
Goto page Previous 1, 2

Matronics Email Lists Forum Index -> AeroElectric-List

View previous topic :: View next topic

Author Message

ainut(at)knology.net
Guest

Posted: Thu Feb 02, 2012 8:04 pm Post subject: Thermocouple leads

I have the book, Bob.  I must say, you are Ohminous!

David M.

Robert L. Nuckolls, III wrote:

 	  Quote: 	

 <nuckolls.bob(at)aeroelectric.com>

 At 06:01 PM 2/2/2012, you wrote:

> 

>

> Also, any (new) junction will create an inaccuracy unless you 

> calibrate with the total installation.

>

> David M.

>

   A set of thermocouple leads can have quite a few

   extraneous junctions without affecting accuracy

   as long as (1) current in the leads is very low

   indeed approaching zero. Such is the case with most

   modern thermocouple instruments using specialized

   integrated circuits for thermocouple signal

   conditioning.  And (2) EACH junction needs to be

   repeated in number and kind on both sides of the

   thermocouple pair.

   This is described in more detail here:

 http://aeroelectric.com/articles/excerpt.pdf

   Bob . . .

-- 

Tell the truth.  Be honest.  Be responsible to and for yourself.

I liked America when it was free and it's people were responsible and had morals.

Every gram of cocaine you buy from elsewhere contributes to an innocent being murdered in Central and South America.  Grow your own or Stop taking it.

ainut(at)knology.net
Guest

Posted: Thu Feb 02, 2012 8:47 pm Post subject: Thermocouple leads

That runs contrary to my initial training (early 80's), but new things 

are discovered all the time!  Thanks, Paul.

David

Paul Millner wrote:

 	  Quote: 	

 Not exactly grue, David... if it's a chromel/chromel junction and a 

 alumel/alumel junction, there's no potential developed across it, so 

 nothing to calibrate out.   If, for instance, someone inserts some 

 copper wire, you'll have chromel/copper and copper chromel junctions, 

 and assuming they're separated, and at different temperatures, that 

 temperature difference CANNOT be assumed to be constant... so there's 

 no way to calibrate it out.

 So... either it doesn't matter, or it can't be known.  No middle ground.

 Paul

 On 2/2/2012 4:01 PM, David wrote:

> 

>

> Also, any (new) junction will create an inaccuracy unless you 

> calibrate with the total installation.

>

> David M.

-- 

Tell the truth.  Be honest.  Be responsible to and for yourself.

I liked America when it was free and it's people were responsible and had morals.

Every gram of cocaine you buy from elsewhere contributes to an innocent being murdered in Central and South America.  Grow your own or Stop taking it.

Dennis Jones

Joined: 17 Sep 2006
Posts: 45
Location: Poplar Grove, IL

Posted: Thu Feb 02, 2012 8:53 pm Post subject: Re: Thermocouple leads

It is an EPI-800 electronic system with a DPU.

millner(at)me.com
Guest

Posted: Fri Feb 03, 2012 9:56 am Post subject: Thermocouple leads

As Bob has shared, some really old systems (through, say, the 1950's) 

heavily loaded the junction, so the apparent voltage was much less than 

the theoretical voltage.  But, as long as those systems were calibrated, 

they were repeatable and accurate.  Splicing a wire, even with J or K 

wire, could increase resistance, and due to the relatively heavy current 

flow, could affect the voltage and hence calibration.  And at least in 

my industry, we were still using those dinosaurs into the 1980's, so ya 

gotta train to work with what you have, not what's being built today.



The predominance int he industry today, though, is high impedance 

devices that don't load the T/C junction... which has the side benefit 

of allowing multiple devices to share a junction.  A good thing.  A 

splice resistance is unimportant, as current flow is nearly nil.



Dissimilar wire, especially with junctions at dissimilar temperatures, 

really plays havoc with even the open circuit voltage, and there's no 

way to calibrate that out, as the temperature of those splice junctions 

into the future typically can't be known.



Paul

nuckolls.bob(at)aeroelect
Guest

Posted: Fri Feb 03, 2012 1:20 pm Post subject: Thermocouple leads

At 09:39 PM 2/2/2012, you wrote:

 	  Quote: 	
		  



Hmmm, Bob... the physics doesn't support your assertion, unless you 

(mistakenly I'd think) assume that all those junctions are somehow 

held at the same temperature.  And how would you accomplish that?


	


    The PAIRS have to be adjacent. For example, ordinary

    gold over brass pins used to bring a t/c pair through a

    connector adds negligible effect to calibration because

    equal and opposite effects are inserted into each lead

    and in close thermal proximity to each other and they

    cancel each other out.



    A tech told me of a case where he found an overheated

    wire going through a connector because it was adjacent

    to a non-compensated thermocouple lead . . . the other

    lead in the thermocouple pair was not adjacent in the

    connector housing and, therefor, was running cooler.

    The warm-wire caused significant difference in temperature

    reading for the afflicted pair as compared to other

    t/c pairs in the same connector.



    So putting 'foreign materials' into the t/c system is

    not without risk . . . but easily managed with some

    planning and attention to details.

 	  Quote: 	
		  With the relatively common high impedance measuring devices, 

resistance in the leads isn't really the issue... but adding 

uncontrolled junctions at different temperatures can sure play havoc 

with accuracy.


	




    Absolutely. For example, a silver soldered joint adds

    two foreign junctions in immediate proximity. T/C-alloy

    to-solder offset by a solder-to-T/C-alloy right next

    to it. Two equal but opposite effects that cancel because

    while not 'controlled' they are at the same temperature.

    So my condition (2) was not completely explained. The

    foreign junctions must be in close proximity as in the

    soldered joint described above, or the t/c selector switch

    box described in Figure 14-7B.



    Splicing into t/c pairs to extend them will add loop

    resistance. UNPOWERED temperature readout instruments

    are still offered by manufacturers like ALCOR, Westach,

    and others in the automotive world (see Figure 14-10C in the

    'Connection).  These meters must be capable of reading

    thermocouple voltages directly . . . on the order

    of 10 to 20 millivolts FULL SCALE. Hence they are wound

    with few turns of heavier wire and draw considerable

    current (like 10-50 milliamps). In these cases, the

    termocouple wire tends to be heavy gauge just to get

    low resistance. If you check out the range of sizes

    available from Omega on type K wire:



http://www.omega.com/ppt/pptsc.asp?ref=XC_K_TC_WIRE&Nav=temh06



    . . . we see everything from 14AWG to 36AWG. To be sure,

    any gauge wire would work with a modern engine management

    system with internally conditioned inputs. But if you

    were driving an unpowered instrument over some distance

    in a refinery or smelting operation, the fatter wires would

    offer some options.

    When unpowered gauges were used on large piston aircraft,

    identical gauges and CHT/EGT thermocouples had to be individually

    compensated using an external resistor so that effects

    of lead length between instrument and engine could be

    calibrated out.



    This presupposes that the actual temperature value is of

    interest. In many airplanes, one uses such an instrument

    only to find "peak" on the mixture and then adjust in

    25F incremental departure from peak. So whether the peak

    occurs at 1/3rd scale or 3/4ths scale is not critical.

    Hence, no calibration needed.



    But if you've got a TIT limit to be observed, THAT system begs

    for calibration. When such instruments are supplied without

    provisions for exernal calibration, installation instructions

    will generally include an admonition not to change the length

    of the thermocouple lead.

   Bob . . .

steve(at)tomasara.com
Guest

Posted: Sat Feb 04, 2012 5:21 am Post subject: Thermocouple leads

Group,



Although it's made clear enough in the book, I don't think it's come 

through clear enough yet in this thread that when using dissimilar 

junctions you ALSO have to worry that you don't inadvertently move the 

reference (i.e. "cold") junction(s).



As an illustration, my Long-EZ, when received, had a short pigtail on 

the under-sparkplug CHT T-couple.  It was transitioned to copper pretty 

much adjacent to the neighboring cylinder.  The copper ran all the way 

to a powered (i.e. high impedance) Westach gauge.  The Westach is 

designed (best I can tell) so that the cold junction(s) (i.e. the final 

terminations to non-T-couple wire) are actually made by the junctions 

that result from the crimps between what is supposed to be T-couple wire 

and fast-on connectors right at the back of the gauge.  This allows a 

temp sensor in the gauge to compensate for the voltages at the cold 

junctions.



By running copper from the T-couple wire pig-tails all the way to the 

gauge, there would usually be a significant temperature difference 

between the actual cold junctions (next to a cylinder) and the intended 

cold junctions (behind the panel) which would show up directly in error 

on the reading of the gauge.  That's all fixed now...



Steve Stearns

Boulder/Longmont, Colorado

CSA,EAA,IAC,AOPA,PE,ARRL,BARC (but ignorant none-the-less)

Restoring (since 1/07) and flying again (8/11!): N45FC O235 Longeze 

Cothern/Friling CF1 (~1000 Hrs)

Flying (since 9/86): N43732 A65 Taylorcraft BC12D

nuckolls.bob(at)aeroelect
Guest

Posted: Sat Feb 04, 2012 7:43 am Post subject: Thermocouple leads

At 07:15 AM 2/4/2012, you wrote:

 --> AeroElectric-List message posted by: Steve Stearns <steve(at)tomasara.com>



 Group,



 Although it's made clear enough in the book, I don't think it's come through clear enough yet in this thread that when using dissimilar junctions you ALSO have to worry that you don't inadvertently move the reference (i.e. "cold") junction(s).



 As an illustration, my Long-EZ, when received, had a short pigtail on the under-sparkplug CHT T-couple.  It was transitioned to copper pretty much adjacent to the neighboring cylinder.  The copper ran all the way to a powered (i.e. high impedance) Westach gauge.  The Westach is designed (best I can tell) so that the cold junction(s) (i.e. the final terminations to non-T-couple wire) are actually made by the junctions that result from the crimps between what is supposed to be T-couple wire and fast-on connectors right at the back of the gauge.  This allows a temp sensor in the gauge to compensate for the voltages at the cold junctions.



 By running copper from the T-couple wire pig-tails all the way to the gauge, there would usually be a significant temperature difference between the actual cold junctions (next to a cylinder) and the intended cold junctions (behind the panel) which would show up directly in error on the reading of the gauge.  That's all fixed now...



   Exactly. I've often thought that we should perhaps

   forsake the term "cold junction" for "reference junction".

   Classically, the reference junction was any point in the

   system where the transition from thermocouple alloys

   and instrumentation was known with accuracy. It could

   be hotter or colder than the temperature of interest.

   I recall the first thermocouple array I ever saw installed on

   a flight test airplane at Cessna. The reference junction

   for perhaps a dozen switched thermocouples was submerged

   in a Thermos bottle filled with crushed ice made with

   distilled water. Pretty durn close to 0 degrees C.



   But the self powered instruments couldn't be fitted with

   ice baths so an alternative reference junction compensation

   scheme had to be crafted. I've been meaning to dig through

   the old patents someday and see what the popular techniques

   were 70 years ago. In that test setup, a flight test technician

   had to measure each thermocouple one at a time using a precision

   potentiometer bridge in his lap. The millivolt readings were

   written to a test data sheet and converted to temperatures

   later when he wrote his test report.



   Tedious . . .



 [img]cid:.0[/img]



   A modern incarnation for implementing a reference junction

   is illustrated here:



 [img]cid:.0[/img]



   I think this is out of an analog devices AppNote illustrating how

   an AD592 temperature sensor is situated in thermal proximity to a T/C-to-Copper

   reference junction. Thus, while the "reference" can be wandering around

   with ambient temperature, it's effect on output is known and

   compensated for. Even this technique is outdated since Analog Devices

   and others have created thermocouple conditioner chips that feature

   built in reference junction sensing features and offer user friendly

   10 mv/degree output signals. 



   The situation in your airplane as-received could be illustrated

   like this:



 [img]cid:.1[/img]



    The indiscriminate insertion of a "rogue" junction added new but

    unknown effects that the instrument would not be able to resolve

    resulting in wildly inaccurate readings.



 

   Bob . . .

52cdff.jpg

Description:

Filesize:

97.75 KB

Viewed:

7568 Time(s)

52ce7c.jpg

Description:

Filesize:

32.01 KB

Viewed:

7568 Time(s)

nuckolls.bob(at)aeroelect
Guest

Posted: Sat Feb 04, 2012 8:11 am Post subject: Thermocouple leads

At 03:53 AM 2/2/2012, you wrote:

Another "lightbulb" reply, thank you. I know I've 

read this in the Connection, but still, a reminder doesn't go amiss.



Is this effect mitigated by having a common 

ground on the firewall, with a braided strap to 

the engine? Â If the instruments are grounded to 

the forest of tabs, and that's only a short 

distance along a fat piece of copper to the engine...



    Yes. If you're using a forest of tabs firewall

    ground and have the fat braid to the crankcase,

    I'd be surprised if you have a ground loop issue

    with the thermocouples.

At 01:20 PM 2/1/2012, you wrote:




I'll second that. On the Dynon an improper ground 

on 1 device/sensor can/will impact the reading on 

another. I used to have an old Cadillac which 

when you moved the electric seat the high beams 

flashed. They called it a feature.



    Hmmmm . . . how do you mean "improper ground" on a

    sensor? Are we talking thermocouples here? In the

    case of a thermocouple welded down to it's thermowell,

    the installer has no control over the existence or

    quality of that ground.



    When considering devices like temperature or pressure

    sensors, if they're fitted with ground wires, then

    that wire should be brought back to the point where

    the instrument grounds . . . either the forest of

    tabs on the firewall or the instrument panel ground.



    If the sensor gets grounded by virtue of it's being

    attached to the engine (common with automotive devices)

    then again, the installer has little control over

    any voltage differentials that the the sensor ground

    verus the instrument ground . . . except to bring

    the two a close together as possible (forest of tabs)

    and use a fat bonding jumper.



    If improper delivery of signal to one channel of

    the instrument can upset readings of other channels,

    then the instrument at least suspected of poor

    design. I'd find that hard to believe of Dynon.

   Bob . . .

nuckolls.bob(at)aeroelect
Guest

Posted: Sat Feb 04, 2012 8:13 am Post subject: Thermocouple leads

The second image got hosed during a cut-n-paste. Here's

 how the message SHOULD have looked.  -BN-

 -



 At 07:15 AM 2/4/2012, you wrote:

 --> AeroElectric-List message posted by: Steve Stearns <steve(at)tomasara.com>



 Group,



 Although it's made clear enough in the book, I don't think it's come through clear enough yet in this thread that when using dissimilar junctions you ALSO have to worry that you don't inadvertently move the reference (i.e. "cold") junction(s).



 As an illustration, my Long-EZ, when received, had a short pigtail on the under-sparkplug CHT T-couple.  It was transitioned to copper pretty much adjacent to the neighboring cylinder.  The copper ran all the way to a powered (i.e. high impedance) Westach gauge.  The Westach is designed (best I can tell) so that the cold junction(s) (i.e. the final terminations to non-T-couple wire) are actually made by the junctions that result from the crimps between what is supposed to be T-couple wire and fast-on connectors right at the back of the gauge.  This allows a temp sensor in the gauge to compensate for the voltages at the cold junctions.



 By running copper from the T-couple wire pig-tails all the way to the gauge, there would usually be a significant temperature difference between the actual cold junctions (next to a cylinder) and the intended cold junctions (behind the panel) which would show up directly in error on the reading of the gauge.  That's all fixed now...



   Exactly. I've often thought that we should perhaps

   forsake the term "cold junction" for "reference junction".

   Classically, the reference junction was any point in the

   system where the transition from thermocouple alloys

   and instrumentation was known with accuracy. It could

   be hotter or colder than the temperature of interest.

   I recall the first thermocouple array I ever saw installed on

   a flight test airplane at Cessna. The reference junction

   for perhaps a dozen switched thermocouples was submerged

   in a Thermos bottle filled with crushed ice made with

   distilled water. Pretty durn close to 0 degrees C.



   But the self powered instruments couldn't be fitted with

   ice baths so an alternative reference junction compensation

   scheme had to be crafted. I've been meaning to dig through

   the old patents someday and see what the popular techniques

   were 70 years ago. In that test setup, a flight test technician

   had to measure each thermocouple one at a time using a precision

   potentiometer bridge in his lap. The millivolt readings were

   written to a test data sheet and converted to temperatures

   later when he wrote his test report.



   Tedious . . .



 [img]cid:7.1.0.9.0.20120204100504.01e3b368(at)aeroelectric.com.2[/img] 



   A modern incarnation for implementing a reference junction

   is illustrated here:



 [img]cid:7.1.0.9.0.20120204100504.01e3b368(at)aeroelectric.com.3[/img] 



   I think this is out of an analog devices AppNote illustrating how

   an AD592 temperature sensor is situated in thermal proximity to a T/C-to-Copper

   reference junction. Thus, while the "reference" can be wandering around

   with ambient temperature, it's effect on output is known and

   compensated for. Even this technique is outdated since Analog Devices

   and others have created thermocouple conditioner chips that feature

   built in reference junction sensing features and offer user friendly

   10 mv/degree output signals. 



   The situation in your airplane as-received could be illustrated

   like this:



 [img]cid:7.1.0.9.0.20120204100504.01e3b368(at)aeroelectric.com.4[/img] 



    The indiscriminate insertion of a "rogue" junction added new but

    unknown effects that the instrument would not be able to resolve

    resulting in wildly inaccurate readings.



 

   Bob . . .

6ef574.jpg

Description:

Filesize:

97.75 KB

Viewed:

7567 Time(s)

6ef584.jpg

Description:

Filesize:

112.22 KB

Viewed:

7567 Time(s)

6ef620.jpg

Description:

Filesize:

32.01 KB

Viewed:

7567 Time(s)

ainut(at)knology.net
Guest

Posted: Sat Feb 04, 2012 12:21 pm Post subject: Thermocouple leads

Bob, I'm using the AD594/595 and it is my understanding that even with 

all the magic inside the chip, one has to compensate for the extra cold 

junctions aka reference junctions.



David



Robert L. Nuckolls, III wrote:

 	  Quote: 	
		   At 07:15 AM 2/4/2012, you wrote:

 

 <steve(at)tomasara.com>



 Group,



 Although it's made clear enough in the book, I don't think it's come 

 through clear enough yet in this thread that when using dissimilar 

 junctions you ALSO have to worry that you don't inadvertently move the 

 reference (i.e. "cold") junction(s).



 As an illustration, my Long-EZ, when received, had a short pigtail on 

 the under-sparkplug CHT T-couple.  It was transitioned to copper 

 pretty much adjacent to the neighboring cylinder.  The copper ran all 

 the way to a powered (i.e. high impedance) Westach gauge.  The Westach 

 is designed (best I can tell) so that the cold junction(s) (i.e. the 

 final terminations to non-T-couple wire) are actually made by the 

 junctions that result from the crimps between what is supposed to be 

 T-couple wire and fast-on connectors right at the back of the gauge.  

 This allows a temp sensor in the gauge to compensate for the voltages 

 at the cold junctions.



 By running copper from the T-couple wire pig-tails all the way to the 

 gauge, there would usually be a significant temperature difference 

 between the actual cold junctions (next to a cylinder) and the 

 intended cold junctions (behind the panel) which would show up 

 directly in error on the reading of the gauge.  That's all fixed now...



   Exactly. I've often thought that we should perhaps

   forsake the term "cold junction" for "reference junction".

   Classically, the reference junction was any point in the

   system where the transition from thermocouple alloys

   and instrumentation was known with accuracy. It could

   be hotter or colder than the temperature of interest.

   I recall the first thermocouple array I ever saw installed on

   a flight test airplane at Cessna. The reference junction

   for perhaps a dozen switched thermocouples was submerged

   in a Thermos bottle filled with crushed ice made with

   distilled water. Pretty durn close to 0 degrees C.



   But the self powered instruments couldn't be fitted with

   ice baths so an alternative reference junction compensation

   scheme had to be crafted. I've been meaning to dig through

   the old patents someday and see what the popular techniques

   were 70 years ago. In that test setup, a flight test technician

   had to measure each thermocouple one at a time using a precision

   potentiometer bridge in his lap. The millivolt readings were

   written to a test data sheet and converted to temperatures

   later when he wrote his test report.



   Tedious . . .



 Emacs!



   A modern incarnation for implementing a reference junction

   is illustrated here:



 Emacs!



   I think this is out of an analog devices AppNote illustrating how

   an AD592 temperature sensor is situated in thermal proximity to a 

 T/C-to-Copper

   reference junction. Thus, while the "reference" can be wandering around

   with ambient temperature, it's effect on output is known and

   compensated for. Even this technique is outdated since Analog Devices

   and others have created thermocouple conditioner chips that feature

   built in reference junction sensing features and offer user friendly

   10 mv/degree output signals.



   The situation in your airplane as-received could be illustrated

   like this:



 Emacs!



    The indiscriminate insertion of a "rogue" junction added new but

    unknown effects that the instrument would not be able to resolve

    resulting in wildly inaccurate readings.



   Bob . . .




	


-- 

Tell the truth.  Be honest.  Be responsible to and for yourself.



We want our freedoms back.  Strike the illegal legislations called: 1) obamacare, 2) "Patriot (HA!) Act", and 3) 'presidential orders' that affect anyone besides gubmnt worker bees.

Hate crime laws?  Really?  Thought police?  Orwell would be proud.



Every gram of cocaine you buy from elsewhere contributes to an innocent being murdered in Central and South America.  Grow your own or Stop taking it.

nuckolls.bob(at)aeroelect
Guest

Posted: Sat Feb 04, 2012 4:54 pm Post subject: Thermocouple leads

At 01:48 PM 2/4/2012, you wrote:

  	  Quote: 	
		  --> AeroElectric-List message posted by: David <ainut(at)knology.net>



 Bob, I'm using the AD594/595 and it is my understanding that even with all the magic inside the chip, one has to compensate for the extra cold junctions aka reference junctions. 	


    No. See



  http://tinyurl.com/6xp4xc 



    In the features list right on the front cover we see

    the statement "Built-In Ice Point Compensation".



    In Figure 13 we see an interesting use of the 594/595

    as a Celsius thermometer. Just short the t/c pins together

    and the chip reports it's own internal temperature (very

    close to ambient).



    In the data acquisition signal conditioner I illustrated

    at:



  http://aeroelectric.com/Pictures/Data_Acquisition/Weeder_Module_DAS_4.jpg 



    . . . one of those thermocouple conditioner chips was used

    for just that purpose. This gizmo was installed in the nose

    of a Beechjet for some pitot temperature studies at high

    altitude. Some of the integrated circuits were qualified

    down only to 0 degrees C. If you look at the picture

    you can see a power resistor peeking out from under the

    perf-board. That resistor and three others were used as

    heaters controlled by one of the T/C chips to maintain

    10C inside the box. 



    So, I'm happy to report that these critters are very complete

    and require no ice-point reference junctions. 



 

   Bob . . .    [quote][b]

natsuto
Guest

Posted: Tue Feb 14, 2012 5:33 pm Post subject: Re: Thermocouple leads

I have seen improper ground, cause all to read. Engine to ensure that the equipment grounding, rather than the ground bus.

Display posts from previous:

	Matronics Email Lists Forum Index -> AeroElectric-List	All times are GMT - 8 Hours Goto page Previous 1, 2
Page 2 of 2

Jump to:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum