conductivity check need help

[Fluffer88]

i need a way of heating calibartion fluid(1413uS/cm at 25degrees celsuis) in a conical flask +\_.1degree(plus or miunus 0.1 degree) for 20seconds. if any 1 can propose a simple list of sensors/idea of how i could achieve this it would be greatly appreciated. i need to put two sensors in fluid when 25degrees is reached i was thinking of a flask with major insulation and gel type gland to keep sensors in flask to minimise heat loss when set temp is hit. so i suppose i need a heater, accurate temp cut off, and i,ll use a stop watch to time, i also have an accurate fluke thermocouple which can go in gel gland as a second temp checker to insure 25 degrees is reached. i'm not gonna lie i'm kinda lost on this one??

i suppose having sensors in flask before heat is applied would be the right way

but i really just need a way of keeping temp constant for 20sec's

 

[jackorocko]

.1 degree accuracy? I can't solve your sensor problem, but just a thought when I was reading this. If you have something that had a high thermal mass to either contain your flask or sit underneath your flask it would seem a lot easier to maintain a temp for 20 seconds I would think.

I would certainly like to see what kind of answers this thread receives.

 

[OLIVE2222]

0.1°C seems achievable:
http://www.thermistor.com/directemp.php

maybe you can log your fluid warmed above 25° at room temperature to evaluate how he behave when cooling.
Then you can log your fluid from 20° and slowly warm it to evaluate how he behave when heating.
The idea is to find if warming up or cooling down give the more stable result.

Hi agree with Jackorocko that having high thermal mass would help.

maybe having the setup in a "cold box" can help, or having the room temperature close to 25°C.
This kind of regulations are generally done thanks to a PID but i don't know if it's applicable here.

Olivier

 

[Fluffer88]

what i was thinking but need your opinion on parts

i hope this picture makes sense and thank you for the link to usb thermometers. what i really need is a thermometer that will cut off at 25.1degrees and turn on when its below with good thermal mass this should do but i want to try make it automatic could you suggest parts??

 

[jackorocko]

If I understand you correctly then you will need to be able to measure the temp in hundredths of a degree to maintain a tenth of a degree accuracy. You need to know when to turn the heating element on before it cools 1/10 of a degree. Vice versa, when it is heating. If you don't need that good accuracy, then I still think your best bet is to use something with good thermal mass heat it up to the desired temp or a little more let it cool. If it is insulated good then it should cool very slowly and allow you time to get a measurement.

 

[Fluffer88]

Yes that is about right could you recommend what thermocouple would be best to use and how one would be best setting it up. what would be a good heating element to use seeing that it would have to be immersible. and thermocouple would have to turn off heating element at 25degrees and maybe i might need another thermocouple higher up connected in series in coductivty liquid. i like the look of the thermocouples with usb connection but i've never used anything like that before and how would that control my heating element. i was hoping to use a 12/24volt supply?! thanks again for your time!!

 

[OLIVE2222]

The picture make sense, basically the "cold box" way. They best bed is indeed that the cool down will be slow enough to ensure 20 sec of stability.
So a regulation "by is own".
Heating element can maybe be split to limit convection, by adjusting water volume and/or "venting" at the top optimization can be expected.
Logging's are still needed as first step, with a kind of default setup, to evaluate how close to the target we are.
Going for a real regulation will be a pain but it can maybe be used partially to speed up the process, just a guess.

However very few fluid can be qualified a day, don't know if it's for experiment or production purpose.

regarding the stuff to bought the thermometer will not control the heater, a thermostat can do that.
I don't drill down the thermometer manufacturer web site, maybe they can offer driver or software to control the heater via an i/o module. This site is just a first finding the should have others. A dedicated hardware can also be a solution.

Olivier


edit:

incredible ! +/- 0.017C precision at @ 0.01C
http://www.icllabs.com/Digital_thermometers.htm

also the fluke have a rs232 port, looks nice.
http://www.icllabs.com/pdfs/Product Spec Sheets/2011/1523-1524.pdf

 

[daddles]

You have a fairly challenging task and my first order of business is to state I've not worked on a problem like yours. But I would assume a properly-tuned PID controller would be able to do the heating control. The key is measuring the temperature correctly. Thermocouples just aren't gonna cut it (if you were a first rate instrumentation engineer, you might be able to get the requisite sensitivity with expensive equipment, but you're gonna do a lot of head banging). In fact, since you need to measure to 0.01 degrees to control to 0.1 degrees, I'd avoid thermocouples like the plague. Here's why: if you look at a type K TC table, you'll see that the voltage slope at 25 deg C is 41 uV per deg; that means you'll be trying to measure things to 1/100th of that or 400 nV. <snicker>

Instead, I'd suggest looking at the small thin film RTD devices that Omega sells. If your bath is something like mineral oil, then you could put the RTDs directly in the bath close to the device whose temperature you're trying to control. You're still going to have to measure small signals though and there's all kinds of noise and common mode signals wandering around just ready to mug your measurement system.

There are also thermometers that use e.g. quartz crystals to measure very small temperature changes (I believe HP made one that would measure to microdegrees about 30-35 years ago).

 

[OLIVE2222]

daddles,

The way I see the setup is not to measure 0.01 degrees to control the 0.1. I almost don't care about the exact water temperature. I do it indirectly. So the exact water tempature will be "the temperature needed to have 25° in the fluid".


Steps are:

* Heat up the water (measured with a thermoucouple, or anything in maybe the +/- 0.5°C precision range)
* Turn off the heater when, let say, 28°C is reached.
* Wait for the fluid 25° C trig measured with the +/- 0.1°C precision thermometer
* log the fluid conductivity and temperature for (minimum) 20 sec


This is based on the (reasonable) assumption that the cold box and the thermal mass will
provide a 20 sec of stabilty

However I agree that a PID can improve things if it can be implemented ,and it can according to your post.

Olivier

 

[daddles]

Hi, Olivier:

Well, this is one of those things that you have to figure out by experimentation. Your approach of heat and let the thermal mass stabilize things might work -- and it would be an easier approach than a closed-loop system. But the key to all of this is careful measurement of the temperature. For thermocouples, measuring a 0.1 degree C temperature change involves being able to measure a voltage change of 2.6 to 6.1 uV, depending on the thermocouple type. This can be challenging, especially in a noisy environment. Accuracy with thermocouples is a separate problem and requires a good ice point reference (I usually use ice in water rather than an electronic substitute).

I think I'd look at using a small precision thermistor as a sensor first. It will have more thermal mass than a small thermocouple junction, but with a sufficiently good thermal mass of the warming environment, this shouldn't matter. I'd recommend going to Measurement Specialties, as they produce thermistors that used to be made by YSI. You can get thermistors that are interchangeable to 0.2 deg C for $12, not a bad price.