Looking For Ideas - Temperature Control

idbruce · 2012-01-25 17:23

Hello Everyone

I am looking for some ideas pertaining to temperature control of a corrosive chemical that I am using in a set of experiments. My experiments consist of exposing and developing Premier PCBs from Philmore DATAK, and the corrosive chemical is a mixture of sodium hydroxide and water.

After many experiments, I have come to realize that the temperature of the developer (sodium hydroxide and water) is of vital importance. If the developer is too cold, the board will not develop, and if the temperature is too warm, it dissolves too much photo-resist off of the PCB.

I am currently using an old Mr. Coffee to heat the developer, but I need better temperature control. More specifically, I need a method of monitoring the temperature of the developer. I believe I can hack the coffee maker for on/off time periods, but what would be a good method for monitoring the temperature of this corrosive chemical and applying it to the on/off control of the coffee heater?

Thanks In Advance For Any Assistance You May Provide

Bruce

Unsoundcode · 2012-01-25 17:54

Tape an LM34 to the outside of the pot http://www.parallax.com/StoreSearchResults/tabid/768/txtSearch/temperature/List/0/SortField/4/ProductID/87/Default.aspx

Maybe even have an agitator.

Jeff T.

Circuitsoft · 2012-01-25 17:57

I think most thermistors are Lye-safe, and I suspect even (for example) an LM335 or MCP9700 would work. Being an epoxy case, it should be okay. Pot the wires in more epoxy and you'll probably be better yet.

Circuitsoft · 2012-01-25 17:59

Or, maybe non-contact? http://www.sparkfun.com/products/9570

frank freedman · 2012-01-25 17:59

idbruce wrote: »

Hello Everyone

I am looking for some ideas pertaining to temperature control of a corrosive chemical that I am using in a set of experiments. My experiments consist of exposing and developing Premier PCBs from Philmore DATAK, and the corrosive chemical is a mixture of sodium hydroxide and water.

After many experiments, I have come to realize that the temperature of the developer (sodium hydroxide and water) is of vital importance. If the developer is too cold, the board will not develop, and if the temperature is too warm, it dissolves too much photo-resist off of the PCB.

I am currently using an old Mr. Coffee to heat the developer, but I need better temperature control. More specifically, I need a method of monitoring the temperature of the developer. I believe I can hack the coffee maker for on/off time periods, but what would be a good method for monitoring the temperature of this corrosive chemical and applying it to the on/off control of the coffee heater?

Thanks In Advance For Any Assistance You May Provide

Bruce

Check out some of the DIY reflow ovens that some people have hacked out of toaster ovens. some pretty involved controls. Otherwise, you could use a couple of cores on the prop, one for PI controller for the temp reg. (no need for the diff part here system reaction time to slow to be of use), one for circulation pump to keep the mix uniform, and maybe one to control a valve and fish tank bubler to keep the chemistry at the surface of the board fresh.

Just random thoughts based on too much time surfing the low budget DIY sites.........

FF

idbruce · 2012-01-25 18:07

Thanks for the input guys.

I am now thinking that controlling the hot plate (coffee maker) through PWM might be a good idea as compared to turning the hot plate on and off based on liquid temperature.

ElectricAye · 2012-01-25 18:26

Just thinking out loud:

Not sure how strong your NaOH solution is, but perhaps you could use Pyrex (Borosilicate) glass tubing, seal one end of it with a propane torch (easy to do), thread a DS18B20 one-wire sensor into it, run the wires out the top of the tubing so the fluid doesn't get in, use one or more strategically positioned to give you a temperature profile. You can even gently bend the tubing if need be. Then, for heaters, think about using some non-inductive resistors, or resistors without too much inductance, place them inside similar types of Pyrex tubing, and heat them via PWM control. You should include a temperature sensor inside the same tube with the resistor heaters so they don't over heat. In other words, you've got sensors monitoring the fluid temperature and sensors inside the tubing monitoring the temperature of the resistor heaters.

If you need to stir the solution, consider the magnetic stir bars, the kind that are teflon coated. You can buy magnetic stir plates or DIY by putting a neodymium magnet on the end of a motor shaft and let it spin the stir bar in the solution.

http://www.onlinesciencemall.com/octagonal-stir-bar-64-x-10mm-w-pivot-ring-magnetic-stirbar.html

http://www.onlinesciencemall.com/magnetic-hot-plate-stirrer-hotplate-w-stir-bar.html

You might be able to get the entire package off of ebay - there are magnetic stir plates that heat, and the lab type are fairly accurate.

http://www.ebay.com/itm/Sybron-Thermolyne-Nuova-II-Magnetic-Stirrer-Stir-Plate-used-Work-Good-/190631234141?pt=LH_DefaultDomain_0&hash=item2c6281ca5d#ht_1378wt_1141

idbruce · 2012-01-25 18:42

I also have something similar to the burner shown below. Perhaps I could add a light dimmer for fine control.

Lawson · 2012-01-25 19:21

If a couple of degrees is tight enough control I'd get a submersible fish tank heater in a glass vial and keep any plastic bits above the top of the caustic solution. If you need more precision I'd next look for the smallest thermistor that will survive submerged in the caustic solution and PWM control your current hot-plate. I would not recommend taping a temperature sensor to the outside of your container as this is going to have huge problems with ambient temperature changes.

Lawson

idbruce · 2012-01-25 19:32

@Lawson

Funny that you mention that...

I got one of those with my etching system and I just got finished cleaning it up. I believe that is what I am going to try next before making this a science project.

The board manufacturer recommends a developer temperature between 100 and 110 degrees fahrenheit. I am going to see if I can achieve this with the type of heater that you mention.

Bruce

kwinn · 2012-01-25 20:32

A simple proportional controller is usually good for something like this. As long as the temperature is more than x degrees below the set point 100% of the AC cycles are passed to the heater. As the temperature approaches the set point the percentage of AC cycles passed is reduced until none are passed at the set point. Lets say you can measure the temp to 0.1 degree C and the set point is 80C. Up until the temperature reaches 79C the heater receives every AC cycle. At 79.1C it receives 9 out of 10 cycles, at 79.2 it receives 8 out of 10 cycles, etc. until at 80C it receives no power.

Phil Pilgrim (PhiPi) · 2012-01-25 21:08

I can't imagine that it's necessary to cycle the power at a rate faster than once every second or two. PWM, cycle-by-cycle, and analog proportional control are gross overkill, IMO.

-Phil

idbruce · 2012-01-25 21:20

I would imagine that when I get real serious about producing PCBs, I will arrange a very sweet setup. However at this point in time, I am just trying to get through the testing stage and produce some very nice samples.

How about this...

Notice the pic of the hot plate above....

Couldn't I just hook up a step down transformer to this plate to significantly reduce the output wattage?

Phil Pilgrim (PhiPi) · 2012-01-25 21:23

Nah, just use a relay to turn it on and off with long delays between cycles (i.e. you don't want a buzzer). The physical time constant for a heater is extremely long.

-Phil

idbruce · 2012-01-25 21:27

Phil

That plate gets HOT very quickly on the lowest setting. I am unsure just how much of that heat would be needed to maintain the temperature of the container alone. But it wouldn't take long to heat it up.

Phil Pilgrim (PhiPi) · 2012-01-25 21:41

idbruce wrote:

That plate gets HOT very quickly on the lowest setting.

I'm sure it must. But the effective time constant is extended tremendously due to the thermal mass of the liquid it has to transfer heat to.

-Phil

Dr_Acula · 2012-01-25 21:51

PhiPi's logic makes sense and as such, one solution could be an old-skool 555 running a relay with variable PWM. If you tested it with water first you could measure your water temperature more easily than caustic, then work out the PWM settiing and if the ambient temperature doesn't change too much you would not ever need to change that setting.

Or a light dimmer (which might be easier and quicker to wire up).

I also wonder about LM34/LM35 type sensors. Neutral cure silicone over the solder joints, and the plastic coating on the wire should be caustic proof. Then you get to use a Propeller to control and display the temperature?

Oh, and thanks idbruce for the chuckle over on the GUI thread

idbruce · 2012-01-25 22:02

Dr_A

if the ambient temperature doesn't change too much you would not ever need to change that setting.

LOL I wish. That is one of the main problems in this type of request for help. In order to be an easy circuit, without being too complicated, the ambient temperature should be relatively constant, otherwise you have issues such as thermal transfer, in which case, the liquid really needs to be monitored. IMO

And you are welcome about the chuckle. I love beautiful women also

idbruce · 2012-01-25 22:08

Additionally, I was looking at the LM34 and the AD592

ElectricAye · 2012-01-25 22:24

idbruce wrote: »

Additionally, I was looking at the LM34 and the AD592

I'm just curious. Why do so many people on this forum use these devices instead of 1-wire devices, such as DS18B20's, which you can string up multitudes to a single Prop pin?

Bruce, whatever you do, work out the dynamics of your control system using only water at first. Don't rush into things and test your set-up with NaOH at the beginning. Also wear eye protection and rubber gloves. Boiling, splattering lye is kinda nasty stuff. It will turn your eyeballs into soap bubbles.

4x5n · 2012-01-25 22:35

idbruce wrote: »

Hello Everyone

I am looking for some ideas pertaining to temperature control of a corrosive chemical that I am using in a set of experiments. My experiments consist of exposing and developing Premier PCBs from Philmore DATAK, and the corrosive chemical is a mixture of sodium hydroxide and water.

After many experiments, I have come to realize that the temperature of the developer (sodium hydroxide and water) is of vital importance. If the developer is too cold, the board will not develop, and if the temperature is too warm, it dissolves too much photo-resist off of the PCB.

I am currently using an old Mr. Coffee to heat the developer, but I need better temperature control. More specifically, I need a method of monitoring the temperature of the developer. I believe I can hack the coffee maker for on/off time periods, but what would be a good method for monitoring the temperature of this corrosive chemical and applying it to the on/off control of the coffee heater?

Thanks In Advance For Any Assistance You May Provide

Bruce

I already posed this question on another thread on controlling temp with a prop. I was pointed toward the process control tutorial sold by Parallax. It's based on pbasic and the basic stamp but it does go through the process of writing PID temp control. Not to hard to translate into spin.

jmg · 2012-01-25 22:39

idbruce wrote: »

I am currently using an old Mr. Coffee to heat the developer, but I need better temperature control. More specifically, I need a method of monitoring the temperature of the developer. I believe I can hack the coffee maker for on/off time periods, but what would be a good method for monitoring the temperature

If you want precise temperature, you want better than bang-bang control.
A good example of what they called Proportional control was the old Philips TDA1023.
This applies a slow, small amplitude ramp to the error signal, so within a temperature range, the average power
is proportional to the temperature error. Call it slow PWM if you like.
(outside that range, it is 0% or 100% just like a bang-bang controller)

It is also a good idea to not have too much thermal overdrive, aka too many watts, or the energy in the system
still flows to push UP the liquid temperature, even after you have removed the power.
So you get more deviation in your control point, than you would otherwise have managed.
ISTR systems with DUAL heaters for this reason. The big one gets you up in temperature faster, and then the
smaller wattage one effectively holds it there.

idbruce · 2012-01-25 22:58

@ElectricAye

Nice pic... I already have a couple stinging burns near the fingernails.

I'm just curious. Why do so many people on this forum use these devices instead of 1-wire devices, such as DS18B20's

I can't speak for everyone else, but as for me, I just have not taken the time to learn about 1 wire devices. Additionally, a lot of the older documentation is not really geared for 1 wire devices.

@4x5n

Yea, I was researching Applied Sensors and Process Control several hours ago. After doing some research, I always like to come here and get some input.

@jmg

Thanks for the input.

@Phil, Dc_A, and all others that have replied. I truly appreciate all the input.

idbruce · 2012-01-25 23:19

@4x5n

I went back to process control to review something, and I definitely noticed that I missed a lot of good reading material that applies to heaters. Thanks for the tip, otherwise I may have overlooked it again.

Bruce

JonnyMac · 2012-01-26 20:39

I recently did a commercial design that required monitoring four thermocouples; the MAX31855 made reading the TCs very easy. Something to consider if you're wanting some precision.

ElectricAye · 2012-01-26 21:33

JonnyMac wrote: »

I recently did a commercial design that required monitoring four thermocouples; the MAX31855 made reading the TCs very easy....

Very interesting! I haven't had to use a thermocouple in many years, but this looks awesome, especially for the higher temperatures. Does anybody have a Propeller object for this that they wouldn't mind sharing with lesser minds?

idbruce · 2012-01-27 00:50

@JonnyMac

Thanks for the input Jon. I must admit that I have never messed with thermocouples, but it sure sounds good, and from a glance at the datasheet, it looks like a very simple interface for getting temps. There was a time when Maxim would send me sample chips, but now they have become stingy.

I believe this is probably the setup I will use.

@Everyone

For the circuit boards that I am working with, 106 degrees fahrenheit seems to be the ideal temperature the developer.

T Chap · 2012-01-27 06:59

Some years ago I used a thermocouple inside a convection oven for feedback on generating the temp profile. I used an SSR to turn on and off the oven elements, driven by the Prop. The elements needed to be turned on fully for a few seconds at least per cycle. It was like a very slow PWM. The profile worked very well and it was not hard to maintain a fairly narrow range of temp with some tuning. The thermocouples I found online had about 2 feet of wire, and I just ran the wire up inside through a little hole, and positioned the tip near the PCB surface.

ElectricAye · 2012-01-27 07:00

Bruce,

many people here have given their opinions about what sort of control approach to use (proportional, PWM, bang-bang, etc), but so much depends on the physical characteristics of what you're doing. For example, the so-called "thermal mass" of your system often affects such a decision. If you are working with small volumes of fluid that must hover right around 106 F in a small tray, then you probably want small, relatively low-wattage heaters, stirring of the fluid, several temperature sensors, and PWM control of the heaters. On the other hand, if you have a large vat of fluid and don't care about convective currents causing thermal disparities over your process, then you could probably get away with bang-bang control and a simple hysteresis concept. Generally speaking, you can often get away with simple proportional control for heating applications and not need to worry about the integral-derivative aspects of PID. Though PWM might be overkill, it probably wouldn't hurt no matter what you decide to do, and it's something the Propeller is well-equipped to handle. Maybe if we knew more about the size of your system, or could even see a photo, brainz on the forum could offer more targeted advice.

Bean · 2012-01-27 07:53

You will probably be just fine with a "bang-bang" controller.
If the temperature is below a certain temperature turn the heater on. If the temperature is above then turn the heater off.
Sure, the temperature will cycle a couple of degrees, but with a range of 100 to 110 you should be fine.

Bean

JonnyMac · 2012-01-27 08:17

Does anybody have a Propeller object for this that they wouldn't mind sharing with lesser minds?

Here's mine -- freshly updated.

Looking For Ideas - Temperature Control

Comments