If you are ballin on a budget though - what would be the better setup? Is it that common to have a vac leak when using a thermocouple adapter? Am I better off using my current glass thermowell with no stir bar (but this time add oil to the well)? Would my frustration level be higher due to vac leaks vs. temps?
I mostly try and watch the reactions of the solution vs stare at an inaccurate number anyway, but I was gonna buy a laboy 2 neck 1000ml boiling flask and thermo adapter for like $50 vs my current 1 neck chemglass bf with thermowell -
when using a thermometer adapter is it recommended to get a glass thermowell vs the rubber top ones? the glass adapters i have seen seem TOO short in depth for a boiling flask.
PID overshoot or undershoot is a problem across industries. Robotics is all about PID control.
In the case of an oil we are trying to heat and heat just enough to not overshoot but not so little it takes hours to get there, then the best analogy for this case that I like is a bowling ball and rubber band analogy. A bowling ball can be pulled along a bowling lane and slide. If you connect a rubber band to the bowling ball you could pull hard enough to get it to slide a precise speed.
The bowling ball speed is the heat of the oil and the rubber band stretching is the heating mantle. In order to know when to stop pulling you need a sensor. If you place the sensor on the ball you will know the exact speed of the ball only. If you place the sensor on the rubber band you will know how hard you are pulling the ball only.
Now imagine pulling on the rubber band. At first the ball doesn’t even move because of friction. So you pull the rubber band tighter and tighter until the ball begins to move. The rubber band is stretched now and the ball is accellerating down the lane but it is still too slow. If your sensor is on the ball your controller would see only that it was taking a long time to get going so would stretch the rubber band tighter. At some point the ball will reach the target speed with the rubber band stretched so tight it just keeps going fast owing to a sling shot effect. The controller would see the overshoot but can do nothing about it except just stop pulling.
If the sensor was on the rubber band instead measuring how tight the band was stretched then it could easily be tuned (all PID controls need to be tuned for the load) to equate a certain force of pull with speed of the ball so would be adjusting the force of pull based only on how hard you were already pulling and not how fast the actual ball was already rolling. In essence you only set the force at which you pull and not the speed of the compound (ball) directly.
This analogy really fits heating up this oil because it takes a long time for the heat to transfer into the compound. If the sensor is at the mantle a PID controller can be programmed to never stretch the rubber band overly tight so to speak but if the sensor is in the muck and represents the ball in the scenario then a PID controller simply has no current data to act from and is always working from a highly delayed effect while the “rubber band” gets stretched. Most likely SOPs that come with kits employing a single in compound sensor inform the operator to adjust setpoints at various times during disitllation and such in a way that tries to mitigate the problem but still, it is what it is.
Complex robotics almost always employ both sensors in the scenario above. One at the force and one at the load and then a complex PID control can really get tightly adjusted but that is not something a heating mantle app would ever need.
So, I’ve started placing the thermo probe between the flask and mantle with different results due to what was in the flask, but it seems to have calmed down my overheating issues. The funniest thing happened this last run though. It was second pass and i preheated the dist in the roto and got the bf to the same temps and transferred from roto to spd. It was heating up normally when things got weird. My vapor temps outran my mantle and hit 185 and main body with the mantle temp at 155. This then stayed fairly steady up to 192 vapor and 160ish mantle. And the flow was insane. I was very surprised. Never seen anything like that before. I don’t know how this happened, anyone else seen this before?
Edit: I fortunately had the multi cow on and was able to separate main and tails. This lead to the lightest colour I’ve achieved so far. I did do degumming in etho without brine wash, just winterized again and filtered. But this weird run was by far my most successful.
I got a new wire probe. Tested it yesterday. Gonna do a run with it later today im hoping. Whatever happened the control was on point and i was moving through the fractions based on vapor temp and what i saw happen in the head.
Even with the cold spot the mantle still reacted as it should, never overshot or fucked up. Went down when i lowered temp and up when i raised it.
seconded. just because it has a digital read out doesn’t make it correct. verifying the accuracy of every measuring device you use is good practice.
Once upon a time, the boss was using a 30/30 (30" vac to 30PSI) gauge on our CLS. Gauge read vac, so he opened the system, turns out it was at 100PSI and the gauge had wrapped around. nobody got hurt. but it could have gone very differently.
I’ve had a clamp brush my ear on the way by, due to a gauge that was lying.
Jeeze! I have signs hanging that say “always assume every vessel is under pressure, open vent before removing clamp”. Never had an issue with that “knock on stainless steel”. We have had issues where people didn’t tighten both sides of the clamp, I saw trickle of liquid and hauled ass over there real quick with a torque wrench. That made my day.
For sure, but mid run i couldn’t lift the bf, so i didn’t trust the temp reading at the bf. Instead i relied on the vapor temp and flow. I’m gonna do a test today with the new wire and old probe to see how they correlate.
Edit: i have two extra probes between flask and mantle for total of 3.
Readings are 109.7 for original probe, 109.1 for test probe 1 and 106.8 for test probe 2. I can confirm that moving the original probe i found cold spots, so that’s most likely what happened. Nevertheless, amazing results for me.
Why must minions be so careless! Imagine what could have happened if you weren’t on hand to tighten that leak… what do you torque your CLS clamps too? I am assuming that you run your CLS nice and cold too?
I do 38 in/lb on a 4" envelope gasket. Human error is 99.9% of the equation here. That’s why good sops with checklists are important. Same goes for putting your minions in a two man crew. The buddy system has come in handy a few times.
