Searching \ for '[EE] RTD - 3 wire and 4 wire types' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=rtd+wire+wire+types
Search entire site for: 'RTD - 3 wire and 4 wire types'.

Exact match. Not showing close matches.
PICList Thread
'[EE] RTD - 3 wire and 4 wire types'
2008\01\03@125627 by alan smith

picon face
I've used the 2 wire RTDs  before...essentially measure the voltage across the device for the temperature, but I've been asked to look at 3 wire or full bridge RTD's.  Now, they talked like they are one and the same but I think a full bridge is a 4 wire?  4 wire seems simple enough in that it has a constant current source to get better accuracy but what about 3 wire devices?  A current source that dumps into a common ground?  Just need some clarification on that.
     
---------------------------------
Looking for last minute shopping deals?  Find them fast with Yahoo! Search.

2008\01\03@132530 by Lamebert

flavicon
face
On Thu, 3 Jan 2008 09:55:38 -0800 (PST), you wrote:

>I've used the 2 wire RTDs  before...essentially measure the voltage across the device for the temperature, but I've been asked to look at 3 wire or full bridge RTD's.  Now, they talked like they are one and the same but I think a full bridge is a 4 wire?  4 wire seems simple enough in that it has a constant current source to get better accuracy but what about 3 wire devices?  A current source that dumps into a common ground?  Just need some clarification on that.

With 3 wires you can sense the resistance of one RTD wire and make the
assumption the other wire will be the same.

2008\01\03@134021 by Mike Harrison

flavicon
face
On Thu, 3 Jan 2008 09:55:38 -0800 (PST), you wrote:

>I've used the 2 wire RTDs  before...essentially measure the voltage across the device for the temperature, but I've been asked to look at 3 wire or full bridge RTD's.  Now, they talked like they are one and the same but I think a full bridge is a 4 wire?  4 wire seems simple enough in that it has a constant current source to get better accuracy but what about 3 wire devices?  A current source that dumps into a common ground?  Just need some clarification on that.
>  

With 4 wire, you apply current on one pair and measure voltage differentially on the other. Easiest
way is to apply a constant-ish current, and take the ADC reference as the voltage across a precision
series resistor in the excitation leg - this gives a direct reading of resistance independent (
within reason) of the excitation current, avoiding the need for expensive voltage references.

3-wire makes the reasonable assumption that the voltage drop on both excitation wires is the same.

Assume the wires to one end are A,B and the other is C

Apply current to A, with C grounded
Measure voltage across B & C - this is the RTD voltage plus the voltage drop on the C wire
Measure voltage between A and B - this is the voltage drop on the A wire. Subtract this value from
the BC reading to get the 'real' RTD voltage value.


2008\01\03@144237 by Spehro Pefhany

picon face
Quoting Mike Harrison <spam_OUTmikeTakeThisOuTspamwhitewing.co.uk>:

{Quote hidden}

Another approach is to use a second current sink that's twice the
value of the main one, and just apply that to the compensation
lead. That causes the leadwire resistance to cancel out without
any calculation, and with a single-ended input, provided, of course,
that the leadwires are of equal resistance. Otherwise there's an
error proportional to the difference in the leadwire resistances.
With 100R (or less) sensors, narrow sensing ranges, and long runs
of relatively small-gauge wire, the leadwire resistance can be
substantial compared to the sensing range, so differences from
different lots of wire can have a significant effect. Nonetheless,
4-wire sensing is rarely used outside of lab situations.

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


2008\01\11@170244 by alan smith

picon face
So it turns out its a 3 wire RTD.  So if I understand it correctly, one wire is the excitation voltage,  one is return and one is the voltage drop across the RTD ?
 
 
Spehro Pefhany <speffspamKILLspaminterlog.com> wrote:
 Quoting Mike Harrison :

{Quote hidden}

Another approach is to use a second current sink that's twice the
value of the main one, and just apply that to the compensation
lead. That causes the leadwire resistance to cancel out without
any calculation, and with a single-ended input, provided, of course,
that the leadwires are of equal resistance. Otherwise there's an
error proportional to the difference in the leadwire resistances.
With 100R (or less) sensors, narrow sensing ranges, and long runs
of relatively small-gauge wire, the leadwire resistance can be
substantial compared to the sensing range, so differences from
different lots of wire can have a significant effect. Nonetheless,
4-wire sensing is rarely used outside of lab situations.

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


2008\01\11@220904 by Spehro Pefhany

picon face
At 05:02 PM 1/11/2008, you wrote:
>So it turns out its a 3 wire RTD.  So if I understand it correctly,
>one wire is the excitation voltage,  one is return and one is the
>voltage drop across the RTD ?

It depends how you drive it, as I said there are a couple of ways.
You're measuring resistance, so you'll usually be feeding it a
current.

With what you have .. one wire is attached to one side of the RTD,
two wires to the other.  Call them A, B and B'

If you ground B', drive -1mA into B and +1mA into A, you can just
read the voltage on A.

Otherwise you can ground B', drive +1mA into A, and read the voltage on A,
minus double the voltage on B (the subtraction can be done digitally
or by analog means. Whichever way you go, it's usually safe to assume
that the compensation voltage will change slowly so you can filter
the cr*p out of it.

And there's another way- ground A, feed 1mA into B and read the voltage
from B', subtracting double the difference between B and B'.

And yet another way.. ground B', drive +1mA into A and +1mA into B,
then read the differential voltage A - B.

(1mA above is just an example typical of low resistance RTDs, but
obviously other values are possible).

Best regards,

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



2008\01\11@225957 by Mohit Mahajan (Lists)

picon face
I can't help you with the exact details - Microchip has some pretty good
RTD application notes, which will help you with the theory and practice.


alan smith wrote:
{Quote hidden}

2008\01\14@021354 by Vasile Surducan

face picon face
The four wires are connected to 2 RTD terminals (two wires on one end
the other two on the other end). Ther 3 wires are also connected to 2
RDT terminals (two on one terminal and one to other terminal). RTD is
using the simple methode of measuring resistence using an ammeter and
a voltmeter, yes the one learned by everyone in the highschool, called
upstream and downstream methodes.
The thirs wire is used just to avoid dropout on long wires, based on
the principle that measuring voltage has a large input impedance.
On two wires feed the RDT with current. On the third wire  (or on the
other two) just measure the voltage across the RDT.

On 1/12/08, alan smith <@spam@micro_eng2KILLspamspamyahoo.com> wrote:
{Quote hidden}

> -

2008\01\14@114938 by Bob Blick

face picon face

--- Vasile Surducan <spamBeGonepiclist9spamBeGonespamgmail.com> wrote:

> The four wires are connected to 2 RTD terminals (two
> wires on one end
> the other two on the other end). Ther 3 wires are
> also connected to 2
> RDT terminals (two on one terminal and one to other
> terminal). RTD is
> using the simple methode of measuring resistence
> using an ammeter and
> a voltmeter, yes the one learned by everyone in the
> highschool, called
> upstream and downstream methodes.
> The thirs wire is used just to avoid dropout on long
> wires, based on
> the principle that measuring voltage has a large
> input impedance.
> On two wires feed the RDT with current. On the third
> wire  (or on the
> other two) just measure the voltage across the RDT.
>

What abnout the drop on the shared wire? I don't see
how this is very useful, why not just use four wires
and do it right?

Cheerful regards,

Bob

2008\01\14@121406 by Alan B. Pearce

face picon face
>What abnout the drop on the shared wire? I don't see
>how this is very useful, why not just use four wires
>and do it right?

The drop on the shared wire is assumed by calculation from sensing the wire
on the other end. I have yet to become aware of an installation where the
difference in wire length is enough to upset the calculation.

There is no reason why this is 'wrong'. 4 wires are really only needed when
measuring very low resistances or extreme accuracy is required.

2008\01\14@124625 by Bob Blick

face picon face
Hi Alan,

"Assumed by calculation" seems to me to have a certain
cost, just as an extra wire would have a certain cost.
I'm wondering if the availability of 3 wire sensors
has more to do with the engineer who came up with the
idea being less bright than he should have been. 3
conductor and 4 conductor wire, not much difference in
price.

Actually, using a 3 wire sensor could be much more
expensive than using a 4 wire sensor. With a 4 wire
sensor, you measure the voltage across two terminals.
With a 3 wire sensor, you also need to measure the
drop.

Definitely the 3 wire idea was from a
less-than-perfect engineer.

Cheerful regards,

Bob



--- "Alan B. Pearce" <TakeThisOuTA.B.PearceEraseMEspamspam_OUTrl.ac.uk> wrote:

{Quote hidden}

> --

2008\01\15@064957 by Vasile Surducan

face picon face
On 1/14/08, Bob Blick <RemoveMEbbblickspamTakeThisOuTsbcglobal.net> wrote:
> Hi Alan,
>
> "Assumed by calculation" seems to me to have a certain
> cost, just as an extra wire would have a certain cost.
> I'm wondering if the availability of 3 wire sensors
> has more to do with the engineer who came up with the
> idea being less bright than he should have been. 3
> conductor and 4 conductor wire, not much difference in
> price.
>
> Actually, using a 3 wire sensor could be much more
> expensive than using a 4 wire sensor. With a 4 wire
> sensor, you measure the voltage across two terminals.
> With a 3 wire sensor, you also need to measure the
> drop.
>
> Definitely the 3 wire idea was from a
> less-than-perfect engineer.

No, :)

It doesn't need any "computation". All remote displaying devices using
3 wires to RTD have a "calibrating resistor" onboard.  Calibrating
algorithm is much easy than you imagine and does not need any
computation (OK, if you can't live without computations you may do
it...).

The temperature is measured on the field using the same three wire RTD
and the same displaying device but with very short wires (near zero
errors).
Or the RTD is replaced with a metrological two wire rezistive box .
The value read on the field (or the value of the set resistance from
the box)  is assumed as a "real temperature".

On the remote end,  the technician needs only to adjust the resistor
for one temperature (better near the end of scale) untill it has the
same readings as the field device (or the set resistance-temperature
dependence)

easy and clean,

Vasile


{Quote hidden}

2008\01\15@081133 by Spehro Pefhany

picon face
At 06:49 AM 1/15/2008, you wrote:
{Quote hidden}

Leadwires don't typically contribute more than a small percentage
of the sensor resistance, so we're talking about dealing with the last
few degrees of error (and, often more importantly, the drift of
that error since the leadwires are themselves RTDs and are
exposed to ambient temperature variations).

Often we can put a bit of electronics in the field and use only
two wires by converting temperature to a current loop, which is
better again. Then the RTD itself can be 2-wire or 3-wire, but the
leads are so short there is no concern. And in modern times, the
old current loops are giving way to various and nefarious digital field
buses.

Best regards,

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



2008\01\22@185044 by alan smith

picon face
So in searching around....LTC, TI, ADI and a various set of obsure vendors...
 
 Seems that the AS7793 is the chip of choice...built in current source, seems to be the perfect interface to a 3 wire RTD.  Anyone used these before?  Seems pretty straight forward to use.
 
 Now if the RTD cable has a ground shield, not terminated at the transmitter end, assume tying to the analog ground plane is what we want.  I am assuming that there is a ground shield...havent seen it yet.

     
---------------------------------
Looking for last minute shopping deals?  Find them fast with Yahoo! Search.

2008\01\23@003146 by Xiaofan Chen

face picon face
On 1/23/08, alan smith <RemoveMEmicro_eng2spam_OUTspamKILLspamyahoo.com> wrote:
> So in searching around....LTC, TI, ADI and a various set of
> obsure vendors...
>
>  Seems that the AS7793 is the chip of choice...built in current
> source, seems to be the perfect interface to a 3 wire RTD.
> Anyone used these before?  Seems pretty straight forward to use.

I think you mean AD7793. It seems to be a good part even
though the speed is a bit slow (but should be fine with RTD
application).

If you are looking at the performace of AD7793, the other
possibility is to look at Silabs C8051F060 which has
integrated best-in-class ADC for Microcontrollers
(2ch 16bit ADC up to 1Msps).

Xiaofan

2008\01\23@104958 by alan smith

picon face
The fastest I am reading is maybe once a minute.  The client wants...desires...24 bit accuracy so have to stick with an external ADC.
 
 Thanks for the feedback.

Xiaofan Chen <RemoveMExiaofancTakeThisOuTspamspamgmail.com> wrote:
 On 1/23/08, alan smith wrote:
> So in searching around....LTC, TI, ADI and a various set of
> obsure vendors...
>
> Seems that the AS7793 is the chip of choice...built in current
> source, seems to be the perfect interface to a 3 wire RTD.
> Anyone used these before? Seems pretty straight forward to use.

I think you mean AD7793. It seems to be a good part even
though the speed is a bit slow (but should be fine with RTD
application).

If you are looking at the performace of AD7793, the other
possibility is to look at Silabs C8051F060 which has
integrated best-in-class ADC for Microcontrollers
(2ch 16bit ADC up to 1Msps).

Xiaofan

2008\01\23@183517 by Xiaofan Chen

face picon face
On Jan 23, 2008 11:49 PM, alan smith <EraseMEmicro_eng2spamspamspamBeGoneyahoo.com> wrote:
> The fastest I am reading is maybe once a minute.  The client wants...
> desires...24 bit accuracy so have to stick with an external ADC.

I see. But you won't get 24bit accuracy anyway. Not even 24bit resolution
according to the datasheet (maximum effective resolution is 23bits in
the slowest update rate).

Xiaofan

2008\01\23@183617 by Xiaofan Chen

face picon face
On Jan 24, 2008 7:35 AM, Xiaofan Chen <RemoveMExiaofancKILLspamspamgmail.com> wrote:
> On Jan 23, 2008 11:49 PM, alan smith <micro_eng2STOPspamspamspam_OUTyahoo.com> wrote:
> > The fastest I am reading is maybe once a minute.  The client wants...
> > desires...24 bit accuracy so have to stick with an external ADC.
>
> I see. But you won't get 24bit accuracy anyway. Not even 24bit resolution
> according to the datasheet (maximum effective resolution is 23bits in
> the slowest update rate).
>

And if you use the internal gains, the effective resolution is even lower.

Xiaofan

2008\01\23@221413 by Spehro Pefhany

picon face
At 12:31 AM 1/23/2008, you wrote:
>On 1/23/08, alan smith <spamBeGonemicro_eng2STOPspamspamEraseMEyahoo.com> wrote:
> > So in searching around....LTC, TI, ADI and a various set of
> > obsure vendors...
> >
> >  Seems that the AS7793 is the chip of choice...built in current
> > source, seems to be the perfect interface to a 3 wire RTD.
> > Anyone used these before?  Seems pretty straight forward to use.
>
>I think you mean AD7793. It seems to be a good part even
>though the speed is a bit slow (but should be fine with RTD
>application).
>
>If you are looking at the performace of AD7793, the other
>possibility is to look at Silabs C8051F060 which has
>integrated best-in-class ADC for Microcontrollers
>(2ch 16bit ADC up to 1Msps).

If we get signal the temperature is changing fast enough to need a
1Msps ADC I think somebody set up us the (nuclear) bomb.

>Best regards,

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



2008\01\23@223805 by Apptech

face
flavicon
face

>>If you are looking at the performace of AD7793, the other
>>possibility is to look at Silabs C8051F060 which has
>>integrated best-in-class ADC for Microcontrollers
>>(2ch 16bit ADC up to 1Msps).

> If we get signal the temperature is changing fast enough
> to need a
> 1Msps ADC I think somebody set up us the (nuclear) bomb.

But, no ...
Unless perhaps it's an incoming one, as Mach 6.

>From a recent list posting to another list
The 1 Msps will get you down to 1/16" resolution at Mach 6
:-)

       R

{Quote hidden}

& Russell said:

An update every quarter of an inch or so at Mach 6 :-)


2008\01\28@112011 by alan smith

picon face
Thats all true....he can get by with 16bit, but wanted to use something better than 16bit ADC to make sure he can obtain it.  So even at 20bit, it should be fine.

Xiaofan Chen <EraseMExiaofancspamEraseMEgmail.com> wrote:  On Jan 24, 2008 7:35 AM, Xiaofan Chen wrote:
> On Jan 23, 2008 11:49 PM, alan smith wrote:
> > The fastest I am reading is maybe once a minute. The client wants...
> > desires...24 bit accuracy so have to stick with an external ADC.
>
> I see. But you won't get 24bit accuracy anyway. Not even 24bit resolution
> according to the datasheet (maximum effective resolution is 23bits in
> the slowest update rate).
>

And if you use the internal gains, the effective resolution is even lower.

Xiaofan

More... (looser matching)
- Last day of these posts
- In 2008 , 2009 only
- Today
- New search...