Organikum - 29-6-2004 at 11:37
I am in need for a layout of a working temperature control circuit for to use with a type "K" thermocouple.
It should have the following features:
- variable hysteresis
- include a relais or an output capable to drive a relais
- include an output which can be used on a alphanumeric display.
- temperature control with hysteresis - aka the two points where on/off is triggered.
- simple final shutoff at a certain temperature.
I had a look on the net already and found HEAPS of circuits. It is impossible for me to sort the scrap out and to decide whats what I need.
A circuit for a alphanumeric display which can show the temperatures of several thermocouples - output derived from a circuit named above would be
nice. It has to show only one output at a time of course, a simple switch to change to another one suffices.
If I have expressed something not clear enough please feel free to ask. I guess this will be unavoidable, my english skills on electronics are lousy.
regards
ORG
Specifications
HRH_Prince_Charles - 29-6-2004 at 13:52
Org.
If you are going for this level of complexity, do you want 'bang-bang' control (on/off power via relay)? If you are going this far, you
could go for a pulse-width-modulation circuit and include closed-loop proportional control - that way you get more accurate and smoother temperature
control.
When you say 'shutoff at certain temperature', do you mean some kind of thermal over-temperature switch - if the system overheats, it will
cut power?
I am looking at something similar for hotplate/oilbath control.
Charles.
IvX - 29-6-2004 at 23:06
Can you use/spare a computer for this?If so you'd get a lot of things done.Aaroncake has some very good idea's for controling things with
computers.
The best aproach IMHO would be to make something like a solar engine then possibly code your behavious with a computer.Even without the computer there
are probably a good few solar engines that can do/aproach what you need.
solved
Organikum - 30-6-2004 at 12:12
I found what I was after.
Analog Devices 1
Analog Devices 2
These two pdf files tell it all.
On my search I found something what might be interest for the pyromaniacs:
Measure Projectile Velocity Optically With An Ohmmeter
Nice, isnt it?
ORG
The_Davster - 30-6-2004 at 12:50
Yes, quite nice.
Can't load the 2nd link
HRH_Prince_Charles - 30-6-2004 at 14:55
I can't load the 2nd link. Is this what you were referring too:
http://www.analog.com/UploadedFiles/Data_Sheets/664361174AD5...
trilobite - 30-6-2004 at 15:40
I've been thinking about a similar device, maybe with PID control (proportional integral derivative, has to be calibrated of course) or maybe
only a simple thermostat. In my opinion this is a job for a microcontroller such as the AVR series from Atmel. They are cheap, quite powerful
considering they are 8-bit, the programmer is a piece of cake compared to many other similar devices and open source development tools (avr-gcc) and
libraries are available. With such a device doing calculations on the data would be easy, such as linearization of the thermocouple output, given you
have some experience in coding.
Would come much cheaper than designing and building equivalent circuitry from components but I'd still include a watchdog built from discrete
components to make sure there would be some security measures against the software failing.
But it's going to take some time, so much to do anyways.
Edit: The AD595 looks like a very nice chip.
[Edited on 1-7-2004 by trilobite]
ad596
HRH_Prince_Charles - 1-7-2004 at 04:38
Page 7 of the datasheet for the AD596 has a circuit for a temperature controller with digital readout.
The only AD596/7 I can find in the UK is a calibrated version for £400. AD595's are common and about £10 each.
IvX - 1-7-2004 at 04:46
Yes the links are very nice
Say trilobite you say coding with a computer would be very hard?All you have to do is connect it to the printer port and you have something that
exactly the same(well not exactly) of using the microcontroler.Hell you could even use BASIC whihc I presume you'pd have to use anyway
Aaron Cake's main website
The electronics page
[Edited on 1-7-2004 by IvX]
Organikum - 1-7-2004 at 06:40
Thanks Prince Charles, that was the link.
A interface to a PC is not the problem. The circuit shall provide standalone functionality. The cheaper AD595 will do nicely, of course it will need
calibration, but thats unavoidable anyways as I intend to build or at least assemble the circuit by my own.
ORG
trilobite - 1-7-2004 at 13:46
Coding probably is hard if you haven't done it before. Why would I have to use basic when I can code in C? Basic sucks *shudder*. It is the
language I used when I was 12 years old. An RS232 interface would be a piece of cake with an AVR model that includes an UART.
I wouldn't think of using a microcontroller unless these particular chips weren't so cool.
I also take back what I said about the price, it only applies if you want to have a stand-alone controller with linearization. I'd use the
'595 or similar even with a microcontroller. By the way, that chip is even recommended by Horowitz and Hill in The Art of Electronics.
[Edited on 1-7-2004 by trilobite]
Hotplate/Oilbath
HRH_Prince_Charles - 2-7-2004 at 14:40
I think it is quite clear where O. is going. Hotplate/stirrers are rip-off pieces of equipment, assembled in the Far East by slave labour for pennies
and flogged in the West for many hundreds of Euros.
It is possible to put together something that beats the commercial spec. for far less money. It would be great if everyone gets together to develop an
'open source' spec. for a hotplate stirrer. In the UK, a basic hotplate/stirrer will cost £300 new. For less than half that, you can put
together a better spec. machine, and it is all yours. You expend more labour, but it is better, it is more fun and Asian children aren't being
screwed building it.
Org. has discovered some excellent circuit drivers. Missing are suitable elements, stirrer designs and implentations.
I've been looking at PC fan drivers to get a good constant speed magnetic stirrer.
Anyone got any other ideas?
Charles
Clarence House,
London
Organikum - 3-7-2004 at 12:13
Type "K" thermocouples are available at every electronics supply and not very expensive.
I would love to find a source of the ADXXX IC already put on a PCB designed after the suggestions of AD in the papers. With or without thermocouple
and with potentiometers for the settings (and jumpers). This MUST exist and shouldnt be much more expensive than to buy the IC alone for the
manufacturer gets the much better prices. If somebody gets hold of this - please tell!
Thermocouples are preferred "blank" without any insulation, they can be put in glasstubes and those filled with MgO. This makes chemical
resistant devices for cheap.
On the stirrer: It depends on the use. I made the experience that magnetic stirring is not suited for my purposes as most of my reactions are
heterongenous or/and produce viscous gunk together with the product. Magnetic stirrers are just not powerful enough to handle this.
I use an electrical drill with flexible shaft by now what works but not perfect. I am still after a good solution for the seal to the flask and of
course there is the problem that the neck of the flask is to small for a suitable stirrer. I will have to buy me some with moveable TEFLON blade I
guess. 
I already localized a manufacturer of TEFLON double-lip seals - those would be perfect. Have to find a seller but have not searched intensly by now.
Heating mantles are always better than hotplates but of course not so easy to find for free like old hotplates are.
Damn. Where is HERMES with his project of selling labware?
IvX - 8-7-2004 at 02:28
Before you buy anything you should look at old models of HP Jornada and such as the 53(not sure)is about 50 bucks.Much more bang for the buck than
buying even industrial
Mantles
HRH_Prince_Charles - 15-8-2004 at 15:19
It depends what temperature you need, but rswww.com (uk & international) sell some flexible silicone rubber heating mats. Good to 200 Celsius.
Considering the cost of a good thermometer, a plain K-type thermocouple set in a glass tube is a bargain.
I've been playing with a magnetic stirrer based on a PC fan using pulse-width-modulation control. With PWM, you can control the fan down to 10%
normal speed. I've been cutting the casing and blades off and fixing a pair of rare-earth magnets on to the hub. For a stirbar, I've been
using a small alnico bar magnet. Will certainly stir 1 L of water so the vortex hits the bottom. Would like to increase the gap between magnets to
allow for a heated oil bath. Retrieved a pair of magnets from an old disk-drive. These are phenomenally powerful but too big for your average PC fan.
Working on a pully driven axle for them. I might just resort to buying some egg stirbars to save the hassle of fabrication.
The best ghetto oil for the heated bath I can find is avocado - it has a smoke point of 270 Celsius. Bit more expensive than peanut oil but less
smoke! Anyone have any better suggestions?
Temperature Control
HRH_Prince_Charles - 19-8-2004 at 15:16
I was getting ready to build the circuit from the spec. that Organikum found when I stumbled on temperature controllers: they are everywhere and
cheap.
Even the basic ones take a thermocouple input, have PID control (good!) and can supply relay or proportional control circuitry. Just dial in the
temperature and they control the element appropriately. Far better control than simple on/off.
http://www.qis-uk.co.uk/Select_Page.html
You can find these very cheaply if you look around.
agitator shaft seals
Magpie - 19-8-2004 at 18:32
Slightly off topic - but mentioned tangentially by Organikum:
I don't know how good a seal you (Organikum) need but a spring-loaded mechanical seal would give a better seal than a lip seal, especially if
there is a significant pressure differential with the lab atmosphere. These can be purchased in small diameter for not too much I would guess. Most
I've seen are ceramic vs carbon. But many other material pairs are available. Of course affixing this to glassware may be a challenge.
Shaft Seals
HRH_Prince_Charles - 20-8-2004 at 00:32
You can get magnetic seals using ferromagnetic fluid. Expensive to buy new but can sometimes find cheap on auction sites etc.
thefips - 31-8-2004 at 10:19
Instead of using a temperature control with a thermocouple,you can use a cirquit with a triac and a variable resistor.I have built one and now I can
set every temperature by turning the variable resistor.You can use ist for every kind of heating device,for example hotplates,heating mantles,... If
you use a thermometer,you can draw the resulting temperature on the variable resistor.
Here is a cirquit for this: http://home.t-online.de/home/510018878710/schaltungspages/l%...
It is for 220V~.
And here is a temperatur switch: http://www.produktinfo.conrad.com/datenblaetter/175000-19999...
The language is German,but there was no english datasheet.But the cirquit is the same.The temperature can be set up to 100°C,but if you take other
variable resistors,you can use your thermocouple.But you have to test it and compare it with a normal thermometer.
Quince - 10-3-2005 at 01:28
A PID can be built with a single chip (quad opamp). See here (figure 1):
http://www.sensorsmag.com/articles/0504/52/main.shtml
I have a thermocouple but I don't know what type it is. If I remember, the guy in the store said it was to 350*C, so it's probably a T.
I'm thinking of using a triac through an optocoupler (I have these in my parts box), with an NE555 configured as a PWM. I want to set the
temperature within a degree or two. I have a part, LM335, that is on its own a sensor, however in the datasheet it shows circuits where it is used
for thermocouple compensation:
http://www.national.com/ds/LM/LM135.pdf
Looking at pages 8-10. However, I'm not sure how this actually compensates the thermocouple, and what they mean by "cold junction
compensation". Does that mean when both the LM335 and thermocouple are at ambient temperature, i.e. it's to calibrate the setup before use?
The top circuit on page 10 is interesting to me, because the output is calibrated to a C scale. However, it says this:
"Terminate thermocouple reference junction in close proximity to LM335."
I'm not sure what that means, and why it's not on the other circuits. And what type of thermocouple is that, marked with CHROMEL/ALUMEL?
I figure I can have an opamp compare the output to the voltage given by a potentiometer wired as a voltage divider (which I set), driving the PWM.
Edit: if I limit the max temperature to 200*C, I assume I can just encase the thing in Teflon, right? But I'm not sure how likely I'm to
need a higher temperature. I don't know how to do glass... 
One more thing: It is OK to put a glass beaker on a camping propane cooking burner, if I put a wire gauze under it? I tried it with a really low
flame, but I was a fraid to use a high flame as I don't know if the beaker needs to be raised above the gauze or can rest on it.
Sorry for the long message, but I didn't want to have a series of posts. I hope that if I get partial answers, the other questions won't be
ignored by not being at the bottom of the thread (or I'll have to repost them...)
[Edited on 10-3-2005 by Quince]
Quince - 14-3-2005 at 17:30
Anyone?
Twospoons - 14-3-2005 at 18:29
CHROMEL/ALUMEL is a type K thermocouple, probably the most common in existance (must be cheap or something).
A thermocouple operates on temperature differences, not absolute temperature. There are two junctions in a couple, one hot, one cold, where the two
materials join. One join is obvious - at the end of the wire, where you want to measure the temperature. The other join is actually at the connector
where the thermocouple plugs into whatever measuring gear you're using. Cold junction compensation is essentially measures the connector
temperature (absolute) and adds this to the temperature difference seen by the thermocouple between it's two ends. Hence the desire to
"Terminate thermocouple reference junction in close proximity to LM335." so that this compensation is as accurate as possible.
Can't help you with the glass over propane question though