My new Liebig.

[HookLine]

Just finished off a new Liebig.

The old one is 3/4 inside 1", about 850 mm long, and is just too big and unwieldy for my purposes, and is not as efficient as I would like for a once through tap water system (it would be fine for a recirculating reservoir system).

The new one is 1/2 inside 3/4, and a bit shorter (and a lot lighter). This time I stripped a piece of 2.5mm^2 electrical copper cable, wound it around the 1/2 internal tube at about 35º, and brazed it on in several places. It is an easy but snug fit inside the 3/4 external tube.

Obviously the purpose is to make the coolant in the jacket spiral around the internal tube, not just travel straight up the side of it. This makes the coolant travel a lot further inside the Liebig, makes it more turbulent along the way, and ensures that all of the coolant will contact all of the internal tube on its path. This design feature is particularly useful when the Liebig is laying at some horizontal angle and/or is running hard, ie for pot stilling and stripping runs, because the heat tends to concentrate on the bottom side of the Liebig.

I had seen the basic idea mentioned before at HD, or maybe Yahoo, or both. Also when I used to fix medical and lab gear many centuries ago I saw some small specialised condensers that used some variation of spinning the coolant and/or vapour. Spinning can increase the length of the path travelled by the coolant/vapour by many times, without increasing the overall length of the condenser. Nothing new in the basic idea. It is not original from me.

The standard Liebig is not a very efficient heat exchanger, it's okay but not great, and this is one way of seriously improving the efficiency, get the coolant turning (or the vapour, or both, depending on the condenser design). The other is some turbulence in the vapour and coolant paths.

Well, that's my theory for today.

I report back on its performance when I run it.

Internal & External Tubes.jpg (78.25 KiB) Viewed 47729 times

Fitting together.jpg (38.19 KiB) Viewed 47729 times

Complete.jpg (40.13 KiB) Viewed 47719 times

[jaidormi]

Looks sweet. Nice torso shots too.

[rad14701]

Very nice... Great looking liebig... That spiral wire should help a lot...

[Tater]

Nice work.

[theholymackerel]

That should cause some turbulance and help with efficiency.

I'm lookin' forward to yer evaluation of it.

[HookLine]

Thanks, folks.

Looks sweet. Nice torso shots too.

That's my best feature, but only when I have a shirt on.

[junkyard dawg]

original idea or not, its brilliant.

I can't wait to hear how it works. That is a simple elegant solution.

[HookLine]

original idea or not, its brilliant.

I can't wait to hear how it works. That is a simple elegant solution.

I agree (if it works of course, ).

Mud said that this technique was suggested by Alex Bokakob, and I'm pretty sure that is where I got the idea from.

[SuperDavid]

I'm keen to hear about the results after a run with your new liebig.
I'm making my own now but may copy that style and make it shorter if it really is more efficient.

[snuffy]

A simple measure for efficiency is (flow rate * temperature rise) / power input. (*edit* I haven't had enough caffeine)

You can get flow rate with a bucket and a stopwatch. If you want to get fancy, a pressure gauge on the input to the condenser can be calibrated this way.

Temp rise is just thermometers on cooling input and output.

Power input is easy for electric immersion heating, more complex for other types of heating. In a pinch, boiler pressure and temperature will give an index. Pressure is going to be very low and those type of gauges are expensive. A manometer using mineral oil could be substituted.

We could have a race and start comparing condenser efficiencies. Extra points for least mass of copper per 1KW cooling capacity.

This could be the beginnings of a cult of turbulence - which could be a good thing.

* edit* Waitaminnit. temperature rise / (flow rate * power input). Sheesh.

Greatest rise scaled for least flow per power input. Corrections gladly accepted.

*stumbles off looking for more and stronger tea*

[HookLine]

Just did a test run on the new Liebig, and compared it to the old one under the same conditions.

Short version:

The older bigger straight flow Liebig seems to be slightly more efficient that the new smaller spiral flow version. But it is not certain, I don't thnk the results are consistent.

But the new one works well, should be able to cope with at least 5000 W of ethanol vapour, and has the big advantage of being smaller and much lighter.

The new one would work very well in a recirculating coolant system, but is no more efficient than the old one for using in a once through coolant system.

How much the spiral coolant flow affects performance either way, is not clear. We need another Liebig with the same basic dimensions as my new one, but without the spiral coolant flow, to be able to compare results more directly. (Or another Liebig with the same basic dimensions as my old one, but with the spiral coolant flow.)


Long version:


New Liebig dimensions: 1/2" inside 3/4", cooling jacket section 535 mm long, spiral water flow.

Old Liebig dimensions: 3/4" inside 1", cooling jacket section 830 mm long, straight water flow.


Common test conditions and results were:

Plain water steam.

Boiler well insulated. Also insulated the pot column well to minimise heat loss.

Air temp: 29-30 ºC (at midnight, after it had been raining heavily for 2 hours! That's the monsoon tropics for ya. Sheesh.)

Power in: 2400 w

Head temp: 99.5 ºC

Condensate output rate: 3000 ml per hour

Temp differential between coolant in and coolant out (new Liebig was 50.5-30.2 = 20.3 ºC, and old Liebig was 50.3-29.9 = 20.4 ºC)

All measurements were made twice, or were constantly monitored during the test.

Two different thermometers were used to measure head temp. One gave 99.5 ºC, and the other 100.4 ºC. For consistency all measurements were taken with one thermometer, the 99.5 ºC one.

I used 50 ºC coolant out temp as my reference point, set the coolant flow rate to give me that temp (as close as I could get), and then took all the other measurements. (Coolant out temp was measured directly at the output of the Liebig, not at the end of the coolant drain hose.)


The differences between the two designs were:

New Liebig:

Coolant flow rate: 1550 ml per minute

Condensate output temp: approx. 40 ºC

Condensate output stream constant


Old Liebig:

Coolant flow rate: 1450 ml per minute

Condensate output temp: approx. 38 ºC

Condensate output stream spurts quite a bit in a more-or-less cyclical fashion


On the face of it, the new spiral Liebig is slightly less efficient. It used about 6.5% higher coolant flow rate, and had a slightly higher condensate output temp (though this temp was difficult to measure accurately).

But this does not actually make sense to me given that the change in coolant temps (ie the amount of heat shifted) for both condensers was practically identical. Not really sure what happened there. Maybe snuffy has an idea or two.


On the plus side the new spiral Liebig uses a lot less copper mass, and has a smaller overall size. It easily handled 2400 w of water steam. Given a higher coolant flow rate (or lower coolant in temp) I don't doubt it could comfortably handle twice that amount of water steam, which means it will certainly handle at least 5000 w of ethanol vapour, which is plenty for our purposes.


It is possible that the minor efficiency difference may not be constant for other conditions (ie different heat input, coolant flow rate, coolant in temp, etc). I might have just picked a set of test conditions where the spiral version was not at its best, and I suspect that it may do better when running at high vapour speeds (as a % of its maximum). But the 50 ºC coolant out temp reference is close to what I use when doing normal alcohol runs (I set coolant out temp to mid 40s), so this was a real world test and comparison.


Interestingly the coolant out temp in the spiral version seemed a touch more stable than on the old one, which moved about ± 0.3 ºC, indicating that the coolant out on the old Liebig was not as well mixed as spiral version. If real, this effect is subtle, and will need more sensitive measurement tools than I have to hand. But it also does not fit in with the lower efficiency results.


One thing that was quite different was that in the new Liebig the temperature gradient as you moved along the condenser was much more evenly distributed around the jacket. In the old one the temp was unevenly distributed, with maximum heat on the top side of the jacket, (at the vapour input end). This difference is not surprising, indeed it was expected and it would have been surprising if it had not been evenly distributed on the new one.


Keep in mind this test is with water steam which has a far higher latent heat capacity than ethanol, so the coolant flow rate to condense ethanol steam would be much lower. I used plain fresh water for the vapour source because it is a stable reference point that anybody anywhere can use, so we can compare test results more easily. If we try to compare results using vapour with ethanol in it we have to start allowing for different concentrations of ethanol in the vapour. Water is water and is the same everywhere. The only factor that has to be allowed for with water is the altitude.

It is not yet clear how much difference the spiral water flow really makes. Somebody needs to build a Liebig the same as my new one, just without the spiral inside, to really test it properly.

It was an interesting experiment, and apart from anything else it is a lot smaller and lighter. But I gotta admit the slightly lower efficiency in its use of coolant was both disappointing and puzzling. It works well, but no better than the old big one. However, if you are using a recirculating coolant system, a small difference in efficiency is not important, because the coolant flow rate is going to be a lot faster than the minimum required.

One important measurement I do not have yet is the weight of each condenser. But there is no doubt about which one wins hands down there, for up to (at least) 5000 w of ethanol vapour, the spiral one is a much more efficient use of copper (and money). The bigger old one is overkill in basic size for most of our set-ups.

(Just realised, I did not use anything inside the vapour tubes to help with turbulence, they were just clear all the way though. So that may or may not effect the results.)

[rad14701]

Good info, Hookline.. Well thought out comparison...

Gonna toss this out there even if it sounds a bit retarded... Perhaps a side by side comparison of the two designs would be in order... What I'm referring to is a column or pot still head with a T at the top with both liebigs attached, one on each side of the T... Can't get much more equal and fair than that... No variances do deal with, just overall performance... One design will be a clear winner...

[HookLine]

Thanks rad.

This test was in effect a side by side comparison. Problem is because they were different sizes it did not test the effectiveness of the spiral water flow. Need one the same size, just without the spiral in it, to properly test it.

[SuperDavid]

thanks for doing that write up HookLine!

[rad14701]

Thanks rad.

This test was in effect a side by side comparison. Problem is because they were different sizes it did not test the effectiveness of the spiral water flow. Need one the same size, just without the spiral in it, to properly test it.

That's how I read it... I was referring to running two liebigs off the exact same vapor stream at the same time... Can't get a better comparison than that... Although it would take some time and effort to achieve such a comparison it would sure help in moving liebig development forward... Don't go out of your way on my account, however, as it was just a hypothetical suggestion...

<rant>
Argh...!!! How I hate apartment life and the inability to do "mad lab" experiments on a whim...!!! I did crap like this, day in and day out, years ago, on a daily basis... The work shed was where I spent my time from early morning till the wee hours of the morning the following day... Sometimes it went on for days before a few hours sleep came due to utter exhaustion... Getting rich and retiring early just doesn't appear to be in the cards... Grrr...!!!
</rant>

[HookLine]

Well, the first test results have been bugging me, coz it was not a proper comparison, I was not controlling for Liebig dimensions. As I said, to test it properly another Liebig identical to the spiral version is needed, except just without the spiral in the coolant jacket, ie a straight coolant flow. So I scrounged around my bits box and found I had most of the parts for another one, a quick trip to the hardware store, an hour or two on the tools, and I had that Liebig.

I tested them (full test results below), and the only significant difference is in the condensate output temp, with the spiral version 4.1 C lower than the non-spiral version. This is not surprising and was one of the main differences I was expecting to see.

Nonetheless, at a practical level I would have to say that at this point there seems to be no major difference between the two versions.

At least, not under these test conditions.

I suspect it is quite probable that any major difference in efficiency may not show up until much higher power input is used. Or more accurately, much higher heat input to coolant flow rate ratio.

To put it another way, the spiral version may have a higher absolute capacity than the non-spiral version. Unfortunately I do not have any more heat I can put into the boiler.

However, one possible test I could do is to repeat the experiment with one change: start with 1200 w, and set the coolant out temp to the same value (or maybe slightly lower, say 45 C, to give me some headroom). Take all the other measurements. Then increase the heat to the full 2400 w, without changing anything else, and see if that gives a significant difference between the two versions.

I asked snuffy about all this, and he agreed I was heading in the right direction and suggested to just repeat the test but turn the coolant flow rate right down until the coolant out temp was getting near the vapour temp, say about 95 C, and see what difference it made. In other words, drive the Liebigs to near their limit.

Sounds good to me. So, one more round of testing. But will leave it to next weekend. Gotta spirit run of rum to do. Want some real action, not just water steam substitute, getting withdrawal symptoms.



Common test conditions and results were:

Liebig dimensions: 1/2" inside 3/4", cooling jacket section 535 mm long.

Plain water steam.

Boiler well insulated. Also insulated the pot column well to minimise heat loss.

Air temp: 28 C (± 0.2)

Power in: 2400 w

Head temp: 99.7 C

Condensate output rate: 3000 ml per hour

I used 50 C coolant out temp as my reference point, set the coolant flow rate to give me that temp (as close as I could get), and then took all the other measurements. (Coolant out temp was measured directly at the output of the Liebig, not at the end of the coolant drain hose.)


The individual results were:


Non-spiral Liebig

Coolant in temp: 29.9 C

Coolant out temp: 50.6 C

Coolant flow rate: 1555 ml per minute

Condensate output temp: 48.8 C

Temp differential between coolant in and coolant out was 50.6-29.9 = 20.7 C



Spiral Liebig

Coolant in temp: 30.1 C

Coolant out temp: 50.6 C

Coolant flow rate: 1550 ml per minute

Condensate output temp: 44.7 C

Temp differential between coolant in and coolant out was 50.6-30.1 = 20.5 C

[Ugly]

Wow.... talk about deja vu. I see what you're saying now about someone else liking the spirals Hook. The Liebig on my fuel still is 1" over 3/4" (using the same spirals and turbulator I posted for the VM build) and is 30" long. It can knock down five Imperial gallons an hour of methanol from an oak charcoal fire under a 55 Imperial gallon boiler with ease reducing the distillate temp to no more than 30c at output using plain ol swamp water. I can't measure the BTU/hr (KW) of the fire, but I know from experience I'm putting out close to 250000 Btu/hr from the draft induced charcoal (75KW equivalent energy) . Not all that heat goes to the boiler but a lot of it does, it boils that stinking mass of methanol and congealed cellulose liquids in under an hour. Heat control is handled by moving the firebed (steel with ceramic casting) up and down away from the boiler and/or by sliding a ceramic faced plate across the top of the fire.

These are a very efficient design and easy to build. Don't forget the turbulator .

[HookLine]

Third Test:

Well, I spent a frustrating 2 hours last night trying to get reliable data using high coolant out temps (90-95 C), but was unable to as it was impossible to get the temp to stabilise at very low coolant flow rates (ie higher coolant out temps). I only have a 1/2" brass gate valve for my coolant flow control valve, and it may just not be accurate and stable enough at very low flow rates.

I have run out of time for this experiment, for now at least. I will hang onto both of the smaller Liebigs, and hopefully will be able to get back to this later in the year.

The only real difference we have so far between the two Liebigs is in the condensate output temp in the second test (4.1 C lower in the spiral version) which is a nice hint, but not a clear result.

[rad14701]

It may just be that the spiral liebig allows the condensate to run a little slower than its straight slanted liebig counterpart... Probably not much additional voodoo aside from that... Always nice to see comparison data, however... Thanks for your efforts...

[HookLine]

The spiral is in the coolant jacket, not the vapour/condensate path.

I have not given up yet, just do not have spare time for a while to pin this one down.

Cheers
Hook

[Hawke]

Hook,
My new liebig is pretty close to the same dimentions as yours. Just finished a stripping run on some rum.
Had seen where some thought a kinked wire in the vapor path may improve efficiency. I took another aproach.
I cut a copper strip 3/4x4 inches and gave it a 180 degree twist and knife edged one end. Slid this into the input end of my liebig with the sharp end toward the boiler.
I was able to reduce my water flow by a good 30% over my previous runs.

[HookLine]

Hook,
My new liebig is pretty close to the same dimentions as yours. Just finished a stripping run on some rum.
Had seen where some thought a kinked wire in the vapor path may improve efficiency. I took another aproach.
I cut a copper strip 3/4x4 inches and gave it a 180 degree twist and knife edged one end. Slid this into the input end of my liebig with the sharp end toward the boiler.
I was able to reduce my water flow by a good 30% over my previous runs.

I did actually have a gently coiled strand of copper electrical cable sitting loosely inside the vapour tube, running the length of the coolant jacket. Forgot to mention that. I was gonna test the difference between having this coil, and not, as well as the difference between the two Liebigs. But didn't get there. For now.

I'd also like to test a turbolator, like your twisted plate, but running the full length of the condenser, or most of it.

[Hawke]

I was a little suprised that it worked as well as it did. Have another charge of rum to do, will see if it's repeatable.

[Eugene]

Hi all,

Hookline, we hade a Liebig a litle while ago, used 5/16 copper, made a tight spiral around a 1/2 mandrel and put it inside some 1 inch copper with caps etc to finish, never ran it because we were worried about the back pressure, but it got me thinking.

How about using 1/2 inner, with a coil on the outside of 1/4 copper, like your condensor coils, but not that tightly wound, slide this over and solder as required to the inner 1/2", then run a 1" outer as ussual, this would make the coolant water follow the spiral, not just introduce turbulance, but it would ensure it went around and around etc.

I dont think you would wnat them too tight or flow restriction may be an issue, but the small copper tube soldered to the inner would transfer heat into the water as well as the wall of the inner tube transfering heat too.

Just an idea.

[Hack]

It's the vapor you want to spiral not the coolant.

[HookLine]

Hi Eugene

All I can say is that even if the spiral in the coolant jacket does give some back pressure and restrict coolant flow, then it is only by a small amount and does not adversely affect performance. It slightly out performed the non spiral flow version (though, of course, I did not manage to conduct full testing).

The spiral winding is both to make the coolant turn around the centre tube, and give more turbulence in the coolant.

It's the vapor you want to spiral not the coolant.

I would suggest you want both to spiral.

[Hack]

I would suggest you want both to spiral.

After a little thought on my part, that makes sense.

[Hawke]

I'd also like to test a turbolator, like your twisted plate, but running the full length of the condenser, or most of it.

My concern was whether that would cause pooling, that's why I went with the short bit. Of course, you could cut some small notches or drill holes to drain it.

[manu de hanoi]

same same but stainless. On this one (my design mistake) the end fittings were a bit too large.

[eddie]

Could the larger condenser be performing better due to the greater surface area of the inner tube and the increased amount of cooling water in contact with it due to the increased volume provided by the larger jacket?