Do turbulators actually work?

[engunear]

I've seen some beautiful condensers with turbulators on this website, and know a lot of people swear by them, but while trying to measure heat transfer coefficient I got side-tracked a little and realised my gear would allow me to measure their performance.

I was measuring the HTC by sticking a thermocouple up the outlet of a Leibig. What happens is there is a very clear transition between boiling point and cooler. From that you can work out the HTC (power / area / temp diff). My Leibig can be disassembled so I can do A/B comparisons. The hypothesis is that a turbulator should shorten the length of the region where the temperature is equal to the boiling point.

To cut a long story short, the turbulator didn't make a damn bit of difference. I get about 25cm without it, or with any of the designs pictured.

The pic shows the ugliest Leibig you ever saw (if a thing that makes whiskey can be ugly). At the very top we have a device for measuring coolant temperature at the outlet. Then we have half of the outer jacket, then the inner pipe with one of the turbulator designs with the second half of the outer jacket. Then a second turbulator design on a spare piece of pipe and a trusty tape measure. (There is no plastic/distillate contact.)

When measuring the active length and HTC, the flow gets cranked right up so that the inlet and outlet temperature are close so the linear average of these can be used.

By the way, I get about 2000 W/C/sq m for HTC. I have had several attempts at this and have got higher numbers before, up to 3000. But it is easy to lose heat from the setup so "results may vary". Mains voltage also changes. So I now work out the power for each experiment by measuring the flow, the coolant inlet temperature and the coolant outlet temperature. This tells me exactly what is happening in the condenser at that moment.

Digging around I noticed a good summary of Reynolds number posted by still_stirrin. It is interesting that adding a turbulator does not change the Reynolds number. So a turbulator may lead to laminar flow over an irregular surface rather than something that should be classed as "turbulence". This is not my area of expertise, I'm just picking at the definition.

A lot of people will take some convincing that these measurements are right. I'm one of them. The evidence needed is a second set of measurements that prove or disprove the results. I've done it several times and get results that vary enough that I wish they were better, but if turbulators are great the effect should be large and overwhelm errors.

[Truckinbutch]

My turbulator was built to be removable and fitted to the tube . Got no fancy meters and measuring gadgets . All I can say is that my rig runs better with the turbulator in place than it does without it . At the lowest denomination ; it ain't hurtin a damned thing by being there . More exposure to copper surface area .

[engunear]

Yep, that is a good point. My measurements did not show that they made things worse.

[Truckinbutch]

You got to factor in , too , that I hand hammered copper ground wire flat and then did a ribbon candy twist . Filed and hand fitted that into 50+" of vapor tube . It's not just a twisted piece of wire shoved down the hole . That could be a difference you missed .

[googe]

Good experiment engunear, ya hard doing defining experiments without having a professional testing center in your shed!, you did well. I've always said turbulators arnt necessary, sounds like it's.more of a.heat transfer issues and not turbulence.

[engunear]

Yep, hammered copper may be better. From what I read the primary benefit is increasing turbulence, which I aimed to do and its not clear to me that neater would be better for this. And, aside from it being a single experiment, I reckon that theory has a burden of proof that has not been met.

Increasing area in contact with water is a separate issue, but one that can still be reduced to a measurement.

I also tried increasing area by roughening the surface with a file, measuring and then repolishing and measuring again. No difference.

I'm treading delicately here, because there is a lot of beautiful work with turbulators. By suggesting they add no benefit I run risk of offending, but may save time for some others in future. Or if we get data they work, so much the better. But I looked and found none.

[skow69]

Thanks for your efforts, engunear. Personally I always try to quantify everything as much as possible. Sometimes I have a hard time believing in anything that doesn't have a number attached. BTW Engineering ToolBox shows the typical HTC for this type of condenser as 1000 - 4000 W/m2K, so your numbers are definitely in the ballpark. In fact they're smack in the middle of the ballpark.

Your experiment looks legitimate to me, and your conclusion valid, for as far as it goes. Do you have plans to test the turbulation of the vapor as well? I believe that is what Truckinbutch is referring to.

Anytime you write a post that runs contrary to the conventional wisdom you can count on drawing fire, which may or may not be presented in the spirit of advancing our collective knowledge for the betterment of the hobby, even though that is obviously the purpose of your thread. All you can do is put on your kevlar shorts and try to imagine how much shit they gave Copernicus.

Keep up the good work.

[Hound Dog]

I think Skow69 is right. Engunear and TB are on the opposite sides of the pipe. If I am correct Engunear is talking about turbulence in the water on the outside of the vapor tube and TB is talking about an insert inside the tube to cause turbulence in the vapor creating more contact time.

[shadylane]

An experiment I once did with a turbulator made the condenser less efficient.
The turbulator was made from a twisted copper strip.
My theory why it didn't work was the hot vapor was heating one end of it and the copper was conducting the heat to the other end.

[bearriver]

An experiment I once did with a turbulator made the condenser less efficient.
The turbulator was made from a twisted copper strip.
My theory why it didn't work was the hot vapor was heating one end of it and the copper was conducting the heat to the other end.

Seems to reason that if it got bad enough, then vapor collapse would ensue. Given that is what was happening.

[engunear]

Turbulators inside the tube ... a different story. I'm trying to think how easily to test this as this whole thing for me is a diversion from a study of coiled condensers ... which I think are more interesting because they actually can be turbulent.

It seems to me that people confuse the behaviour of that mix of air and water droplets which we all see when we have a shower and water vapour (without air). The former drifts about gently and can be directed by waving your hands and stuff, and the latter is quite different. True vapour, when it reaches a surface below its boiling point, condenses, creating a vacuum which sucks more vapour into the space. This continues until the surface heats to the material's boiling point. This is why thumpers thump (they suck air back up the steam delivery pipe) and why steam burns are so nasty. This is also why there is a clean transition between boiling point and lower in the condenser. So my gut feel (reinforced by the memory of burnt knuckles from opening a boiler pot too soon) is that a turbulator down the middle will not help. It may not harm but it won't help.

Its friggin hard getting repeatable measurements. The boiler pot is on a wooden table (previous experiments leaving it on a concrete floor showed 100W into the concrete), and is wrapped in an old bedspread. I made a bunch of measurements a week ago at around 15 cm for the active length, took the gear down, re-assembled it, got 20 cm for about 3 tests (with and without the turbulators), did some stuff and roughened the pipe and got 25cm. At this point I'm figuring "well thats interesting, rough pipe traps water and forms an added resistance". So I re-polished the pipe ... and got 25cm again. So my thinking is that heat loss into surroundings can be quite high and as the insulation heats up, then it falls so heat into the condenser rises. But I come back to that key piece of logic - if something has a worthwhile effect it is easy to measure. If the active length had, say, halved as a result of adding a turbulator then it would have overcome the measurement variation. My with and without turbulator tests were in succession, so I believe they are valid. Also, the Reynolds number is not changed by a turbulator, so the math is on the side of that argument.

[NZChris]

Interesting. I must try shoving a thermocouple up mine to see what is happening. It's a short one made out of scrap tube I'd been hanging onto and I wasn't happy with the condensate temperature at the spout, so I inserted a coil sized to fill the space, lengthen the path of the water and have it circling the tube, rather than creating turbulence. It works fine now.

[engunear]

There are two functions in a condenser - condensing gas and cooling liquid. They have different ideal geometries. A Leibig does both jobs in one. A spiral to lengthen the condensate path is fine idea. I tried it once and had no joy so I'm interested in your result. Not that mine was a carefully metered experiment. From my understanding of the theory it should not interfere with condensation at all.

BTW, I was reading "On Food and Cooking" last night and a certain Leibig was the guy who came up with the idea of searing meat to seal in the juices. He introduced the idea of using science in cooking. The meat searing theory has been debunked (see "The Curious Cook"). But Leibig played a part in the preparation of two essential food groups!

[shadylane]

I suspect making the turbulator out a material that doesn't conduct heat would increase it's efficiency.

[Ferment_It]

Cool experiment but I would argue that total water use per run is a better metric for efficiency than the HTC. You said that you measured flow rates but i didn't see these results. Do you have any water usage data?

[engunear]

I gotta get this problem out of my brain, day job and all.

But reading Wikipedia about Reynolds number, the equation has a "characteristic length" that, for a ship might be the width of the ship. Its describes the scale of what we are talking about. So it is not a turbulence vs no turbulence question, it is a question of what is the scale of the turbulence.

If you calculate the water thickness that is equivalent to the heat transfer coefficient value of 2000W/sqm/C we are talking 0.25mm or something. So we have to have turbulence with a scale less than 0.25mm to make a difference. My turbulator had a scale of about 3-4mm, so it probably created turbulence at that scale, but not at the scale of 0.25mm, thus it made no difference.

[engunear]

... commenting on Ferment_It's comments:

Just as Reynolds number has a characteristic length, each house has a characteristic water use that varies with the number of teenage girls living there. But jokes aside, there is a tradeoff between making a condenser efficient and making it light, and somewhere in there is an optimum that depends on your circumstances. To find that optimum it helps to know what the HTC is. And adding a turbulator takes time so if it makes no difference then it adds weight and is time you could spend in other endeavours.

These measurements were made at high flow rates (~2 liters per minute) to make the math come out simpler. My understanding aligns totally with the parent website calculator, but not the HTC recommended there. With the right HTC, you can make the tradeoff to suit yourself. I started with a 800mm Liebig that was heavy and a pain to use. Its now shorter, maybe too short. But it still works.

[Truckinbutch]

Why are we arguing this point ? Number crunching vs instinct vs CW and it all comes out the same : Every flippin one of us is going to continue to do what works for them .

[engunear]

Sorry, didn't mean to come across like that.

[Truckinbutch]

Sorry, didn't mean to come across like that.

Not dissin you , friend . Just a commentary on quantifying stilling versus individuals following their own beliefs . There is such a thing as over complicating a relatively uncomplicated hobby .

[engunear]

No probs.

There is such a thing as over complicating a relatively uncomplicated hobby .

We are just different people here. I'm addicted to the question "why". I love how quickly the hobby switches from filing metal to 19th century physics, and I'm only on this website to find out where my understanding is wrong, and to find a better way.

[engunear]

To give a quantitative answer to questions about what flow rate I used, and flow as a metric, see attached graph. It shows the length of the region where condensing is taking place vs flow. This graph lets you choose whatever flow you like to run at and compare how a condenser with an HTC of 850 with one with 2000. No matter how it is made flow cannot go below 358 ml/min for this power (2kW) - at this power even an very long condenser will spew vapour. This is a critical limit.

For each flow, there length for which condensation takes place, which you can measure. If your condenser is longer than this length, then the extra length is cooling condensate and adding weight. If shorter it spews vapour.

Accept for flows close to the critical limit, whatever the flow you choose, having an HTC of 2000 lets the condenser be less than half the length of one for 850, as you can see from the graph.

The purpose of a turbulator is to make the HTC higher so the condenser can be shorter and thus lighter. For the same length, flow, and power, by making the active region shorter the coolant comes out cooler.

The red dots are some measurements I made a week ago that I can't seem to reproduce. The blue dot matches closely the measurements made last weekend at a very high flow rate. The red line has points from the website calculator. The rest is from my equations. Enjoy!

[shadylane]

I'd recommend a simpler experiment.
To eliminate any variations in the experiment, Run an internally heated boiler at 1375w with a water charge.
The flow and temperature of the condenser water needs to be constant.
When a liebig is adjusted properly, you can run your hand up and down it and feel a point were the temperature changes rapidly.
Adjust the flow to were the temp change is in the middle of the liebig and leave it there for the remainder of the experiment.
Shove a turbulator up the liebig's ass and see if the point were the temperature changes rapidly, moves up or down.

[engunear]

I'm about to stop thinking about this problem. But I've been trying to figure out why we measure an HTC of 2000, rather than any other number. What follows is engunear's maths – using equations but without complete understanding of the concepts. I'd love to find an adult to talk to this about. If the understanding is right it makes the following predictions:

1. A turbulator can improve performance if it increases the surface area by a big enough margin. For example, if you soldered square section copper as per the diagram (distance between strips equal to their thickness) then you increase the surface area by 3X, so you should only need 1/3 the length of condenser. This is not via turbulence, but simple conduction.
2. HTC depends on turbulence and turbulence depends on coolant velocity. So the tighter the fit between the outer and inner pipe, the better the performance.
3. HTC depends on coolant velocity and the dimensions of the system. So a Nixon (/Dimroth) condenser should have a much higher HTC than a Leibig (I have seen this in preliminary measurements.)
4. A turbulator might work better if its dimensions were sub-cm, so the scale of the turbulence is small.

The research ...

The thermal resistance equation can be turned around to find the thickness of a stationary water layer that is equivalent to any HTC we measure. This is 0.3mm for HTC=2000. So what is 0.3mm thick? Reading about boat hulls, they talk about a boundary layer of 25 microns. Thats too low.

The standard way to interpret Reynolds number (R) is as a threshold where turbulence sets in. But turbulence is a fractal concept; a fish 1cm long can have laminar flow around it while in a 5m eddy from a ship. Asking if the flow is laminar or turbulent depends on your point of view, or in the jargon, your “characteristic length”. So if we set R=2300 (threshold for turbulent flow) and turn the equation around, since we know the flow, and kinematic viscosity (Google is your friend) we can find out the characteristic size of the eddies. I get around 10mm for the kind of flow rates I used and the pipe dimensions. This is too big to explain the HTC.

Then if you read about boundary layers and thermal flow, you find that even Fourier beat his head against this problem. So I am not be ashamed if I don't get it. I can't find an equation for what we are doing. But there is an equation for the thickness of the thermal boundary layer … and assumes you have laminar flow over an infinite flat plate. But it also has a length term; if we put in the length of the condenser, we get nothing like 0.3mm.

But if we put in the characteristic size of the eddy, then the number comes out about right. I get 0.5mm. So the model is we have eddies impinging on a flat plate. The size of the eddy is about 1cm and beneath that we have a boundary layer. I don't know if its right, but the maths is about right and its plausible.

[Hound Dog]

You put a lot of effort into thinking this through. It can be a bit frustrating when the vapor just doesn't want to listen to your equasion isn't it? There are a lot of variables too. Are you taking vapor speed, alcohol content, humidity, ambient air temperature and barometric pressure all into consideration? Really man, it's a tube inside a tube with water running between them to condense vapor. If you're having fun though, keep it up because that's really what this is supposed to be about.

[Haus]

Is there any reason besides it would hard to build, could a smaller secondary line running up thru the center of the vapor tube and plumbed into the outside water jacket? Sort of a an outside and inside cooling type thing? Kind of off topic but not that far.

[ranger_ric]

A lot of variables have been discussed here and I admire the OP'er for his efforts. Out here in the desert the biggest takeaway that I have found in this discussion is the following quote from the OP

"if a thing that makes whiskey can be ugly"

There you go boiled down to a nutshell . . . . .

Carry On
RR

[jedneck]

[quote="Haus6565"]Is there any reason besides it would hard to build, could a smaller secondary line running up thru the center of the vapor tube and plumbed into the outside water jacket? Sort of a an outside and inside cooling type thing? Kind of off topic but not that far.[/quote http://homedistiller.org/forum/viewtopi ... 87&t=48945 I built one. Short and effeciant.

[engunear]

Someone gave me a glass one like Haus6565 suggests. Its never been used but it is beautiful.

There are a lot of variables too. Are you taking vapor speed, alcohol content, humidity, ambient air temperature and barometric pressure all into consideration?

Yeah, checked those. Also hemisphere (the direction of the distillate spiral down the condenser tube varies you know). Its amazing how friggin complicated two lousy pipes are.

If all we learnt is to plug 2000 into the website calculator, and add a safety margin to that, then I think will get good size condensers and the effort was worth it. Otherwise you can have safety margins on safety margins and its all a mess to get to stay in place against gravity (been there).

The next project is a condenser that is as light and efficient as I can possibly make it. It will be an offset Nixon and the plan is to engineer it rather than just guessing. Now guessing also works and makes damn fine whiskey, not being critical of guessing. And yeah, its fun.

[NZChris]

Without using a calculator, I estimate that the water path in my condenser, with a coil that completely fills the water jacket, is roughly three times the length of the jacket, but with no change to the vapor tube length.

How important that is, probably has a lot to do with how close you are to a limit and bugger all to do with turbulence. Any effect might be related to the angle of the condenser. I would think it would have little effect if vertical.