[EugeneOregon]

https://www.youtube.com/watch?v=2697QykYJg4

[SkyHighLer]

I assume there is a lab thermowell you could use in place of the thermometer, but couldn't you just stick a cheap K thermocouple stripped to the wires down into the flask through a stopper?

There are a couple of sellers of high temperature glass fiber insolation tubing at eBay, it's real glass fiber, not the fiberglass seen with most K thermocouples.


And I still think concisely controlling the wattage of the mantle rather than forcing a specific temperature limit is the way to distill based on watching the distillation temperature naturally oscillate while holding the wattage steady with a Variac and a power meter. It should be possible to understand at a deep scientific level how the molecules accumulate, reach a threshold and come over, cycling over and over, but is that really necessary? Maybe, but in the mean time my opinion is if you try and hold the temperature steady by pulsing energy spikes into it with a PID controller you're quite probably introducing counter swings.

[EugeneOregon]

Originally Posted by SkyHighLer

I assume there is a lab thermowell you could use in place of the thermometer, but couldn't you just stick a cheap K thermocouple stripped to the wires down into the flask through a stopper?

There are a couple of sellers of high temperature glass fiber insolation tubing at eBay, it's real glass fiber, not the fiberglass seen with most K thermocouples.


And I still think concisely controlling the wattage of the mantle rather than forcing a specific temperature limit is the way to distill based on watching the distillation temperature naturally oscillate while holding the wattage steady with a Variac and a power meter. It should be possible to understand at a deep scientific level how the molecules accumulate, reach a threshold and come over, cycling over and over, but is that really necessary? Maybe, but in the mean time my opinion is if you try and hold the temperature steady by pulsing energy spikes into it with a PID controller you're quite probably introducing counter swings.

I worked as a motion control control systems engineer for over a decade. Robotics like that uses a myriad of PID schemes for load positioning and accelleration control and so forth. I wrote my own custom versions of PID controllers via software to mate up with the electronics
I designed to control both motion control and heat processes in industrial environments.

I have an intimate knowledge of precise operation of PID control. PID is acronym for;
Proportional (amount of error)
Integral (duration of error)
Derivitive (Rate of error change)

These concepts apply to robotics, and a litany of other control systems including heating mantles. The PID control I use employs an off the shelf strategy that modulates the heat by regulating the pulse width and frequency of the pulses of voltage but not the pulse amplitude as you suggest.

Wattage is simply an expressiong of work and is computed as the voltage times the current and is expressed with an integral (time) factor generally as kilowatt hours. Your idea is unique and noteworthy but I suggest and am figuring out now from data learned that there is yet a different way to handle this than by controlling amplitude of voltage.

Robotics is faced with a similar problem when moving loads really. Measuring position for error at just the load position, or just the motor shaft position has limitations inherent to each and advantages to each. Load position gives you absolute accuracy and motor position gives you better velocity information among other considerations.

The solution is to incorporate dual encoders into the PID logic loop. The solution I think for precise control of my distillation rig also lies here, but with temperature we do not use encoders we use thermocouples. A dual thermocouple, one inside the boiling flask (load position) and one under the flask and above the heating element (motor position) would likely be the best way to tackle this problem and still maintain simplicity of pulse width modulation vs amplitude modulation as you suggest.

This will require a PID controller from off the shelf set up for dual feedback control. Most robotics anymore are all set up for that. This means that somebody has done the grunt work programming wise so with luck I can find something off the shelf when I choose to proceed.

The best course now though for me is to learn the "PID" control myself and essentially act as the second PID input by observing the flask thermometer and making set point adjustments. A dual input PID controller has some pretty technical math but if an off the shelf model exists when I next upgrade that part of the rig it is what I plan to do.

[stemznseedz]

I absolutely love what you are doing here! Keep making videos you have a fan in me for sure.I love seeing someone with the proper knowledge and skill set dive deep in like you're doing. I only understand a fraction of what your saying but it's inspiring me to learn!

[SkyHighLer]

^^ You got it, and I agree, a much more advanced control, completely in the digital realm is the way forward, and it may already exist, but I don't know where you'd find it outside of an engineering lab at this time.