PT1000 tempsensor design

Go To Last Post
11 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Guyzz i'm building a tempsensor for a "Mini Oven" SMD soldering.

I am going to use the below "kit" & sensor.

PT1000 measureboard (German , but schematic is "universal)
Comps pg3 , Schematic pg6
http://www.pollin.de/shop/downlo...

PT-1000 Sensor (400 C)
http://www.pollin.de/shop/dt/NDY...

I have Zero experience in RTD's , have only played with NTC's or DS1820's.

And was wondering if one of the more experienced with PT100/1000 would have a look at the schematic to see if the circuit is OK (Sometimes Pollin designers make some bad shortcuts).

As it's for the reflow curves i guess "ultra precision isn't needed" , but i'l like to know what i can expect.

I'll use an AVR , but do i need one with differential if i use the "Kit" , if not i have some M328's else M32

I have some (alot of) MCP1541 SOT-23
MCP1541 is a low power, high precision voltage reference. It provides a precise output voltage of 4.096V
http://ww1.microchip.com/downloa...

Would they be of any significant use on the aref pin ?

Any hints would be appreciated , and also ... was there something about cold junction ? or ???

/Bingo

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The circuit from Pollin does not look good. Not like a bad shortcut, but more like a lot of effort to get a bad solution. A simple bridge circuit with a single OP is better. It does not need a stable ref. voltage - so no need for one. The circuit could look like this:
http://www.rn-wissen.de/index.ph...
(again in German, but the circuit is quite simple). One may want to use a slightly better OP, maybe a Rail/Rail type.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

@Kleinstein

Thanx .. I'll have a look

Btw: I think i found the "Ref Document for the Pollin circuit"
http://www.mikrocontroller.net/a...

And the German thread describing a PT-1000 I/F
http://www.mikrocontroller.net/t...

They say that a LM317 is not a good 1mA constant current chip , it drifts with temp and have an Imin of 3.5mA

:-(

Damm ... Sometimes i hate Pollin (I have bought) ...

Edit:
Another I/F thread
http://www.mikrocontroller.net/t...

Edit2:
This is from C't ... A magazine i normally take serious
They usually doesn't design useless things , and the schematic looks like the Pollin one. Could the Pollin design be ok (useable) anyway ?

http://www.heise.de/ct/artikel/S...

/Bingo

Last Edited: Sun. Jul 3, 2011 - 05:56 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I have some (alot of) MCP1541 SOT-23
MCP1541 is a low power, high precision voltage reference. It provides a precise output voltage of 4.096V
http://ww1.microchip.com/downloa... ... 21653b.pdf

Would they be of any significant use on the aref pin ?

A precision voltage reference for your ADC is always a good thing. The AVRs internal one is stable, does NOT have a trimmed voltage (the voltage will be somewhere around the gauranteed value). See datasheet for details.

Quote:
I'll use an AVR , but do i need one with differential if i use the "Kit" , if not i have some M328's else M32
I don't understand the things in the datasheet, but at a quick glance the point of the kit is to convert the 1k ohm resistance into voltage of a reasonable level - 0 to 5V. Why differential in that case?

You'd need (or more precisely - take advantage of ) a differential measurement if you'd use a fourwire connection of a PT1000 directly. Which I would recommend in a combination with a precision current source.

The current source part though is weird. Part of the current will go through R9 and part through the RTD. But the amount of current through the RTD will vary with temperature. I'm kind of boggled by what it should be doing.

Anyway, the whole setup will rely on being calibrated via the trimmers. One of the great things about the PT1000 is that you don't need to calibrate it, you can rely on it giving you the resistance at a temperature.

There are pointy haired bald people.
Time flies when you have a bad prescaler selected.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ADS1248

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

@daqq
Thanx for the explanation (i think ...) i'm an analog noob

@Zauberer
Thanx looks nice , but i'd like to use what i allready have bought if possible.

@All

I just saw this Maxim app-note pg-8 (it uses a 4.096 vref .. i have the above Microch... ones that can deliver 1..2 ma) , and was wondering if i could replace the LM317 in the Pollin kit with a 4.096 vref) and some resistor magic ?

Or build the schematic as the one on pg-8 , but can i use the TLC2264 opamp (came with the kit) as "difference amplifier" ?

I guess i dont need more than well 2 or ?? degrees accuracy for the reflow profile.
So i prob don't even need to do compensation or ?

Thanx for helping out

There is 2m wire on the Sensor , that should be enough.
Else i'll extend with CAT5 cable ... Can i use that ?.

/Bingo

http://pdfserv.maxim-ic.com/en/a...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

For the measurement for the reflow process it does not have to be so accurate. One may need to do a few test runs anyway because the temperature of the sensor and that of the board is not the same.

One should not use a constant current source to measure resistance or a resistive sensor. The natural reference for measuring a resistor is a resistor. So use a simple voltage divider or a bridge circuit for higher resolution.
If you use the differential input of the AD in the Mega32 or similar you don't even need am extra OP-amp to use a bridge. Its just 3 resistors and one PT1000. No need for a stable ref. either - just Vcc is Ok, even if it is directly from a battery pack (like 3,5-4,5 V).

The TLC2264 is Ok as an OP-amp, if you need one for the bridge circuit. So you can do 4 channels.

With a PT1000 the cable is not so critical compared to PT100. So there is no real need for a 3 or 4 wire connection and the 2 m of network cable are Ok.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kleinstein wrote:
For the measurement for the reflow process it does not have to be so accurate. One may need to do a few test runs anyway because the temperature of the sensor and that of the board is not the same.

But the difference can be reduced with a simple technique.

Kleinstein wrote:
... and the 2 m of network cable are Ok.
... providing that the cable and solder connecting the sensor exposed inside the oven can tolerate the reflow temperatures ... a burning cable or melted solder joint might be embarrassing. :lol:

Cheers,

Ross

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kleinstein wrote:
So use a simple voltage divider or a bridge circuit for higher resolution.
If you use the differential input of the AD in the Mega32 or similar you don't even need am extra OP-amp to use a bridge. Its just 3 resistors and one PT1000. No need for a stable ref. either - just Vcc is Ok, even if it is directly from a battery pack (like 3,5-4,5 V)..

Thanx all

@Kleinstein

A schematic or a few wiring hints for a M32 would be 100% welcome

Please ...

For reflow i guess 0..300 deg-C would be the span/scope

/Bingo

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Circuit using the internal diff. amplifier inside the M32 is simple. For 0..300°C one get a resistance in about a 1000 to 2000 Ohms range. The internal amplifier has a 10 fold gain, which means no more than a 0,5 V range. So the logic choice to complete the bridge circuit are two resistors of 22 K (easier to get than 20 K) and one 1 K resistor. For a 5 V supply this gives about 0,45 V to 0.9 V for the divider with the PT1000 and about 0,45 V for the fixed divider (1 K and 22 K). These to voltage go to the inputs ADC1 and ADC0 or ADC3 and ADC2. The ADC is configured as differential input with a gain of 10 and VCC as ref. voltage.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Bingo600 wrote:
LM317 is not a good 1mA constant current chip

It is out of the specification.
Bingo600 wrote:
dont need more than well 2 or ?? degrees accuracy

I do not think you can get as close as +-5*C accuracy within a 200-300*C range using AVRs 7-bit differential ADC and not using sophisticated filtering methods this way. Use the fact you need a very limited bandwidth (filtration and oversampling) with acceptable delay (I think 5s is not a problem in your process). If you are planning to use a stabilizing controller, you need to check the stability of the loop.

No RSTDISBL, no fun!