Sciencemadness Discussion Board - Accurate temp control with PIDs

Accurate temp control with PIDs

kaymatt - 7-7-2019 at 03:26

First-time caller, long time viewer here with a question about PIDs and heating mantles.

I have a made in China heating mantle without magnetic stirrer but it only has a hit-em-and-hope dimmer dial and power gauge throttling the voltage. It's fine for near enough is good enough applications, not so fine for distilling or refluxing at temperatures that fly perilously close to a solvent's degradation temperature. I have run a series of tests at each mark indicated around the dial pot so I know what temp neighbourhood I'm in, but have found manual adjustments aimed at specific temperatures to be frustrating and results mixed, to say the least.

So, rushing in where angels fear to tread, I picked up one of these PID controllers online. I understand they come preprogrammed with settings more suitable for a central air conditioning system than the inside of a round bottom flask refluxing at 180 degrees, but I am reasonably confident with mastery of the settings that I can get far more accurate control with one than without one. And at that price point, I'm not losing much in the process of finding out that they're garbage.

I do have a couple of questions for experienced heads though. Spoiler alert! I am an electrical novice of the first order so if I come across as irreversibly stupid you'll know why.

1. Assuming all work is performed accurately under controlled conditions and the thermocouple probe is situated in the optimal position, will wiring the PID and SSR between the mains outlet and the heating mantle's power inlet give the PID the control it needs to adjust voltage when it needs to? I really don't want to make any modifications to the mantle unless absolutely necessary and this is the no mess, no fuss option I'm aiming for.

2. I would like to add a mantle with a magnetic stirrer to my kit. Obviously, I would have to avoid interrupting the power to the magnets to make the PID useful under these circumstances. Without pouring over the electrical schematics, any suggestions on how to best proceed, this time with the modification gloves off?

3. Would it be unwise to always set the mantle's inaccurate voltage pot to its maximum setting to give the PID all the adjustment room it may require, or will setting it to an eyeball estimated level just above the temperature I'm trying to arrive at be sufficient? Will this affect what settings to have in the PID? I'm trying to avoid lengthy reprogramming of the PID settings as much as possible but accuracy and safety are higher up on my hierarchy of needs than convenience. Variation in temperature between processes will commonly be between 120 and 200 degrees C, though I may require temps down to just above room temp and as high as 350 degrees as needed.

4. How would you deploy the thermocouple probe to get as accurate a temperature reading as possible? I was hoping for a convenient flexible probe that lays flat between the flask and the heating element but I'm almost certain I'm going to need an available flask neck, a thermometer vacuum adapter, and a 6mm diameter probe long enough to break the surface of the liquid in order to get anywhere near ideal conditions. Not impossible, but if an easier way exists, I'm listening.

5. To ensure the PID and SSR's terminals are not exposed to potential accidents involving highly volatile solvents, does anyone have any housing suggestions? I was hoping to house them in a way that meant I could quickly, conveniently and safely plug or unplug the PID/SSR as needed. If anyone has any experience/advice/designs/warnings in this regard, I will be eternally grateful.

Many thanks in advance to my fellow artful lab dodgers.

Sulaiman - 7-7-2019 at 04:22

from experience ;

. the bi-metalic temperature regulators found in most 'temperature controllers' are ok for domestic cooking.

. for simple distillations they are ok, especially when using a sand or water bath.

. the hysteresis is large so temperatures fluctuate about the mean quite significantly,
making fine fractionation distillation almost impossible.

. a simple pwm 'dimmer' works very well for setting a constant heating power.

. to use your mantle with a pid controller I would turn the internal temperature controller to Max.
(actually I bypass the internal regulator as it is of no use to me)

. you should only try to directly control the temperature of heated liquids when the operating temperature is well below the boiling point.
When at the boiling point increasing power will only increase the boiling rate, not the temperature,
so attempting automatic temperature control with feedback is futile.

. retro-fitting magnetic stirring to an existing hotplate or heating mantle looks to be extremely difficult.
My diy 500ml heating mantle was supposed to incorporate magnetic stirring but it was easier to add overhead stirring.

. an overhead stirrer is useful for RBFs in mantles and beakers or flasks on a hotplate

. if you use a pid controller and monitor the temperature of the hotplate then you will get a constant temperature hotplate,
fluctuations in thermal conductivity between the hotplate and liquid will cause temperature fluctuations

monolithic - 7-7-2019 at 05:38

Quote: Originally posted by kaymatt

First-time caller, long time viewer here with a question about PIDs and heating mantles.

I have a made in China heating mantle without magnetic stirrer but it only has a hit-em-and-hope dimmer dial and power gauge throttling the voltage. It's fine for near enough is good enough applications, not so fine for distilling or refluxing at temperatures that fly perilously close to a solvent's degradation temperature. I have run a series of tests at each mark indicated around the dial pot so I know what temp neighbourhood I'm in, but have found manual adjustments aimed at specific temperatures to be frustrating and results mixed, to say the least.

So, rushing in where angels fear to tread, I picked up one of these PID controllers online. I understand they come preprogrammed with settings more suitable for a central air conditioning system than the inside of a round bottom flask refluxing at 180 degrees, but I am reasonably confident with mastery of the settings that I can get far more accurate control with one than without one. And at that price point, I'm not losing much in the process of finding out that they're garbage.

I do have a couple of questions for experienced heads though. Spoiler alert! I am an electrical novice of the first order so if I come across as irreversibly stupid you'll know why.

1. Assuming all work is performed accurately under controlled conditions and the thermocouple probe is situated in the optimal position, will wiring the PID and SSR between the mains outlet and the heating mantle's power inlet give the PID the control it needs to adjust voltage when it needs to? I really don't want to make any modifications to the mantle unless absolutely necessary and this is the no mess, no fuss option I'm aiming for.

2. I would like to add a mantle with a magnetic stirrer to my kit. Obviously, I would have to avoid interrupting the power to the magnets to make the PID useful under these circumstances. Without pouring over the electrical schematics, any suggestions on how to best proceed, this time with the modification gloves off?

3. Would it be unwise to always set the mantle's inaccurate voltage pot to its maximum setting to give the PID all the adjustment room it may require, or will setting it to an eyeball estimated level just above the temperature I'm trying to arrive at be sufficient? Will this affect what settings to have in the PID? I'm trying to avoid lengthy reprogramming of the PID settings as much as possible but accuracy and safety are higher up on my hierarchy of needs than convenience. Variation in temperature between processes will commonly be between 120 and 200 degrees C, though I may require temps down to just above room temp and as high as 350 degrees as needed.

4. How would you deploy the thermocouple probe to get as accurate a temperature reading as possible? I was hoping for a convenient flexible probe that lays flat between the flask and the heating element but I'm almost certain I'm going to need an available flask neck, a thermometer vacuum adapter, and a 6mm diameter probe long enough to break the surface of the liquid in order to get anywhere near ideal conditions. Not impossible, but if an easier way exists, I'm listening.

5. To ensure the PID and SSR's terminals are not exposed to potential accidents involving highly volatile solvents, does anyone have any housing suggestions? I was hoping to house them in a way that meant I could quickly, conveniently and safely plug or unplug the PID/SSR as needed. If anyone has any experience/advice/designs/warnings in this regard, I will be eternally grateful.

Many thanks in advance to my fellow artful lab dodgers.

1. Yes, but a PID/SSR does not control voltage. It's just a fancy electrical switch, either on or off.

2. You need to split the stirring and heating circuits. The easiest way would be to look at the heating element and cut the leads, then reconnect those leads to your PID output. Otherwise, as you noted, when your PID turns on/off it will also turn on/off your stirring.

3. I always set the voltage pot to maximum on my hot plate. I'm not certain but I think most PID units also have some sort of learning algorithm, where it attempts to understand the thermal load (time the circuit is closed versus the measured temperature change) to better control the process temperature. I doubt it would make a huge difference, but adjusting the voltage pot from reaction to reaction might make it a bit harder for the PID to predict the correct on/off timing to reach a given process temperature?

4. For distillations, a thermocouple well works decently. You can get them from Chinese lab suppliers on eBay for pretty cheap. You can use any insulated (with bare end) thermocouple wire with such a setup since there's no need to worry about corrosion. If you want to directly measure the temperature of a liquid then a stainless steel thermocouple probe may work. If not, some companies make glass-coated probes. You could probably make one yourself, if you're good with glassworking. Just make sure you leave a small vent hole in the glass if you do make one so the hot/expanding gas doesn't explode the glass coating.

5. PVC junction boxes are good for quick and ugly enclosures. https://www.homedepot.com/b/Electrical-Electrical-Boxes-Cond... A more elegant solution would 3D printing a custom enclosure. I doubt PVC stands up well to extreme solvent abuse, but it's better than nothing. Usually these junction boxes are sealed with an o-ring, and cable glands/grommets are pretty easy to find if you want pigtails for plugging your PID/SSR into the wall outlet and your mantle. Romex is a pain in the ass to work with, so just buy a cheap 10-20 amp extension cord and hack it up.

[Edited on 7-7-2019 by monolithic]

kaymatt - 7-7-2019 at 08:29

Quote: Originally posted by Sulaiman

from experience ;

. the bi-metalic temperature regulators found in most 'temperature controllers' are ok for domestic cooking.

. for simple distillations they are ok, especially when using a sand or water bath.

. the hysteresis is large so temperatures fluctuate about the mean quite significantly,
making fine fractionation distillation almost impossible.

This would go some way to explaining the dissatisfactory results of a fractional distillation I attempted to carry out using the mantle's manual controls. Either it was proceeding far too quickly or not at all. I systematised a process where I would make tiny increases on the dial every 15 minutes and note any changes observed but couldn't get it to settle on or about the target drip rate of 2-3 drips per ten seconds and gave up after three attempts.

Hence the desire to plug in a PID and let rip on a saved program setting after I've determined the optimal temp setting to achieve the desired flow rate.

From the rest of your reply it appears that I only need to measure the temperature at the point of contact between vessel and heat source, rather than within the liquid itself. Cheers for your insights and wish me luck

Ubya - 7-7-2019 at 08:43

Quote:

1. Yes, but a PID/SSR does not control voltage. It's just a fancy electrical switch, either on or off.

But it switches it on and off fast enough to give a PWN signal, the result is the same as a regulated voltage

Sulaiman - 7-7-2019 at 10:03

Quote: Originally posted by kaymatt

From the rest of your reply it appears that I only need to measure the temperature at the point of contact between vessel and heat source, rather than within the liquid itself. Cheers for your insights and wish me luck

Sorry I did not make myself clearer,
to maintain a constant liquid temperature (below boiling point)
the sensor should be at the liquid temperature.
Sensing/controlling a hotplate surface is generally useful,
but measuring the liquid temperature will be more accurate.
(glass is a poor conductor of heat so there are usually tens of degrees difference between hotplate and liquid)

S.C. Wack - 7-7-2019 at 17:05

Any enclosure needs to account for the SSR...a fan or external heat sink.

Will using a relay eventually burn out the mantle control somewhere?

Some people and textbooks will tell you that fractionating without a variac is impossible. For low BP a liquid sensor seems sensible but for high boiling heat sensitive material without stirring I'm not sure that one wants to heat full blast for extended periods. I have said fractionating can be done with a thermocouple under the RBF, and sometimes the built-in thermocouple wire on old Glas-Cols can be looped around the PID and controlled thusly, having done these things. In response at another board, it has been posted that I'm totally, completely full of shit, my advice burned the contents, and I can go fuck myself, IIRC.

[Edited on 8-7-2019 by S.C. Wack]

kaymatt - 7-7-2019 at 17:17

Quote: Originally posted by Sulaiman

So, if I am to understand you succinctly, a PID measuring the temperature of a liquid undergoing distillation can approach a boiling point, but not exceed it to produce the desired flow rate. If it measures the hotplate temperature it is only limited by the hotplate it's controlling, but I would have to factor in the difference in temperature on either side of the glass boundary to exercise any degree of control over the flow rate.

So, I need a formula. Or is this where the art resides?

Sulaiman - 7-7-2019 at 20:06

I am still confusing you, sorry.

Using a pid controller for maintaining a temperature below b.p. (i.e. NOT for distillations or refluxing)
I would measure the liquid temperature.

For distillations, the temperature sensor for your pid controller
should NOT sense liquid or vapour temperature,
you need a constant heat flow to the liquid.
This requires a constant power input to your mantle (e.g. a 'dimmer')
or, slightly better,
a constant temperature heat source,
e.g. your pid controller with its thermocouple embedded in a hotplate.

I have not tried using a mantle with a pid controller's thermocouple embeded in the mantle heating element,
it should work to give near consant temperature on the outside of an RBF, hence constant heating power,
but I use a 'dimmer' to give constant heating power.

AFAIK there is no simple way to accurately determine the temperature of the liquid from the external (heating element) temperature,
you need to measure liquid and/or vapour temperature directly.

XeonTheMGPony - 8-7-2019 at 05:55

step 1: https://www.youtube.com/watch?v=wkfEZmsQqiA

step 2: https://www.youtube.com/watch?v=NVLXCwc8HzM

Learning more about the PID and sensors, even if you are familiar, refreshers help!

Different style videos:

https://www.youtube.com/watch?v=UR0hOmjaHp0

https://www.youtube.com/watch?v=6OH-wOsVVjg

Tuning:

https://www.youtube.com/watch?v=3viD5ij60EI

Check the switch speed and type of the SSR, and is your dimmer a 0 cross over trigger? ferro resonant?

I am building mine as PID with K type sensor and I elected to make the input switchable, default is embedded in mantle for constant heat as I do tons of distillations, but can switch to in flask sensor for reaction temp holding

I configured mine to drive as pwm out put via SSR, if possible avoid mechanical relays they are a hassle, use over aggressive heat sinks, will extend operating life/conditions