This is my first post.. I've been researching the hobby for the past few months, and it might be time to post
I remember reading the main disadvantage of LM was 'just when you need reflux the most, you get a decrease in reflux ratio'
The advantage of VM in contrast, 'constant reflux ratio througout the run'.
If I understand correctly, this is because of the needle valve: When the hearts phase is finished, there is a decrease in liquid volume . There is less liquid in the head, but you take out the same amount of liquid out for product. That way, same ml product and less ml total condensate means less reflux, so less reflux ratio.
What I was thinking, why not use this reduction in reflux ratio, but in reverse?
If we use the needle valve to control what goes back to the column, and take the rest as product (just the reverse of what we do in a conventional LM head), wouldn't we get an increase in reflux ratio when the hearts are finished?
here's a diagram to help visualize, it's similar to a parrot's beak (I took the design for it from a plan on the yahoo group), and the overflow goes out as product.
I would like to hear what you think about it.. do you reckon it would work?
reverse LM.gif (12.31 KiB) Viewed 13758 times
I hope the post wasn't too long
Cheers
Yonch (Jonathan)
Looks a bit messy but the idea would be correct.
If you have a set quantity being returned as reflux, the rate at which the product was being collected would gradually slow down over the course of a run until it completely stopped.
Cheers.
2"x38" Bok mini and
Pot still with Leibig on 45 litre boiler
Welcome to the forum, Yonch and good to have another enquiring mind on here. i need all the help i can get to copy from you smart guys
rad14701 wrote:My take on your drawing is that you're trying to reinvent the Nixon Stone offset head which can be used to control both reflux and take-off...
The N/S was exactly what popped into my head as i was reading the description.
Hi again
thanks everybody for your replies! it sure is fun to have people to consult with!
the earlier diagram probably made it hard to understand the difference, and yes I was thinking of the Nixon-Stone design when thinking about it..
punkin, rad: were you referring to the original Nixon-Stone or is there a variant that does "reverse" LM? If so do you know where I can find a design, I'm curious to see how to implement it elegantly and not in a mess
so this diagram might show the "difference" between N/S and reverse LM. I took the "inline" N-S version from one of the posts in the forum..
reverse LM2.gif (13.8 KiB) Viewed 13693 times
hope this might make it clearer?
or was it clear enough the first time?
rad14701 wrote:I guess I'm just not getting what it is that you're trying to achieve, other than making the column more complex than it needs to be...
I was trying to improve on the advantage of VM and get not just a constant reflux ratio (as in VM) but even an increase in reflux ratio towards the end of the run.
It may complicate an LM design, but it might be simpler than a VM (no need for two condensers)
Still trying to wrap my head around the theory behind what you're suggesting... Is it really any different than what I'm using in my Small Scale still...??? I'm not seeing true VM instead of LM, which my head uses... Perhaps part of the problem is your understanding of how improved Nixon Stone heads work... They are LM and can regulate both output and reflux independently - to an extent... The two reducer design, and its variants, use a single needle valve fore liquid control, which reduces complexity and cost... But, then again, it controls "liquid" and not "vapor"...
yonch wrote:Hi again
thanks everybody for your replies! it sure is fun to have people to consult with!
the earlier diagram probably made it hard to understand the difference, and yes I was thinking of the Nixon-Stone design when thinking about it..
punkin, rad: were you referring to the original Nixon-Stone or is there a variant that does "reverse" LM? If so do you know where I can find a design, I'm curious to see how to implement it elegantly and not in a mess
so this diagram might show the "difference" between N/S and reverse LM. I took the "inline" N-S version from one of the posts in the forum..
reverse LM2.gif
hope this might make it clearer?
or was it clear enough the first time?
Yonch
I can see that it'd work but you'd have to have a siphon outlet or you'd get vapor out there when its not full of liquid. Most of us that have run LM columns still use a small liebig 'cause the distillate is too hot and subject to evaporation without cooling. I think you're right about the RR increasing automatically at the end of the run. You're gonna build this contraption soon I hope??
Last edited by minime on Mon Nov 10, 2008 4:21 pm, edited 1 time in total.
I'm with minime... Hit the workshop and butcher up some copper... It's the only real way to prove or disprove theory... Speculation and physics only gets you so far...
It works allright. You just add a vaporlock to the line that allows product to flow out above the needle valve. The issue is that you are still left with a sensitive needle valve to deal with. They tend to get hot and change adjustments which still causes the need to do fine adjustments during the run. Been there done that. Now how about inventing a better control method?
This is quite known design in my country.
E.g.
After setting by valve reflux remains constant. Product stream decreases during the process (less alcohol in keg and vapour).
It could be supplemented by electronic control of heating giving real ARC. The control algorithm is simple and output power is only a function of the temperature in keg. Because reflux is constant, power decreasing decreases product (it means reflux ratio RR increases).
E.g.
keg temp. [ºC] 78 93 96 97 98 99 100 102
power [%]_____75 70 65 63 59 62 65 65
Where full reflux is for 50% of power, 65% gives 4/1 RR (the same 65% gives 8/1 RR for low percentage of alcohol in keg). Additionally, at the end of process, head temperature is taken into account. If it increases above 0.05ºC power is reduced giving 100% reflux. After head temperature drop to its initial value 8/1 RR is forced by increased power. This last loop is repeated until in full reflux temperature can not drop.
I only said what I could find on our forum. I have only translated the picture and regulator idea but it is reported as a proven concept.
cemik1 wrote:This is quite known design in my country.
Your plan is brilliant IMO and I really like the idea of the condenser using horizontal space. Anything to save on height and get more packing in there.
rkr wrote:It works allright. You just add a vaporlock to the line that allows product to flow out above the needle valve. The issue is that you are still left with a sensitive needle valve to deal with. They tend to get hot and change adjustments which still causes the need to do fine adjustments during the run. Been there done that. Now how about inventing a better control method?
Cheers, Riku
Snuffy has located a true multi-turn needle valve at a very reasonable price. Would work much better than the hardware store models. https://www.surpluscenter.com/item.asp? ... A&catname=" onclick="window.open(this.href);return false;" rel="nofollow
Thought cemik1's design looked familiar. Found these pics a couple of years ago on the net somewhere, sorry don't remember where. Those clever Poles. It is where I got the idea for the angled designs I drew up a while back (like the 3 drawings at the bottom). Click on the images for larger, clearer version.
These are nice designs - very similar in principle; controlling the return of condensate to the column. There is some advantage to keeping the reflux as hot as possible to conserve energy.
I've built and tested (with water) an LM design based on Bokabob's lesser known tee-overflow design. The takeoff needle valve is at the bottom of the product condenser so it doesn't have the heat-creep issue Riku mentioned. The output from the valve goes to a vented overflow. This makes measuring the takeoff rate much easier, since it surges every 10 - 15 seconds. Much easier than counting drops per second.
Concept picture showing the detail of the head (prior to adding the overflow tube.)
the overflow tube picks liquid from the bottom of the reservoir.
In operation.
I made the reservoir large to experiment with heads compression. Still haven't gotten around to that one....
These all allow diversion of product to control the reflux rate. they differ in geometry.
I like the angled head. That's nice and I hadn't seen one before.
The design I presented allows to extend an effective column length. It is also cheap, because only one cut and soldering is required. In a classic (HookLin's drawing above) version:
Cooling inlet, outlet and ventilation. The last one is also used for cleaning (water splash).
If it is interesting you could also see a boiler connection
Now I would open it but there is my first stuff on the pictures. Soldering also fixed coil in the center of condenser. I have also changed flanges from 2mm copper to 5mm brass.
From what I understand, reverse LM, is MUCH more like ARC, than a standard LM.
In LM, you have a pretty much constant take off volume (unless you 'fiddle' with the valve and reduce it). However, the actual amount of condensed distillate is reduced as the run continues. Thus the volume being returned to the column is reduced, and thus, the reflux rate drops. This is a pretty well known side effect, and the reason why a still head like a boka inline requires adjustments to be made to the takeoff valve, and at times, that valve might have to be closed to re-equalize.
In the RLM, you have a constant return volume (the reflux). Thus, early on, your take off is more, but later in the run, when the amount of condense is reducing, your take off will start to fall also, but the return rate will stay the same. At some point, your take off will simply stop, as the amount being condensed is the same as the amount being returned.
It is a VERY interesting idea. I personally prefer LM over VM, however, it does need more monitoring than a VM to get a quality run. In this model, I think you are getting the best of 'both' types.
H.
Hillbilly Rebel: Unless you are one of the people on this site who are legalling distilling, keep a low profile, don't tell, don't sell.
that last design you posted is still not RLM. It is normal LM. It has a 'constant' takeoff, and the overflow goes back to the column.
The difference in the RLM is you have a 'constant' RETURN. Then any extra is taken off. Thus, when the total vapor rate drops off in later part of the run, this exta will be less. But the 'constant' return is still the same amount.
Thus the RR will actually go up as the run progresses in a RLM, while in a 'normal' LM, that RR will reduce.
Lets throw some made up numbers at it.
Here is LM:
At the start, condensing 60 'parts' per minute, but taking off 20 parts. Thus the RR is 3:1 Later, you are condensing 50 parts, but still taking off 20 parts (you have not 'touched' the valve). Thus, you are getting a RR of 2.5:1, but at this time, probably needing more RR, since the ethanol is less in the boiler. Later still in the run, you are only condensing 40 parts, but still taking off 20 parts. Thus, you are only getting 2:1, and are likely pulling tails.
Now, in the RLM you start out condensing 60 parts. You put 40 parts back into the boiler, and are taking off 20 parts. You are getting RR of 3:1. Later, when condensing 50 parts per minute, you are still returning 40 part, and now taking off only 10. Thus, you are running at 4:1, and running at high ABV, but possibly a little slower than you 'might' be able to run. By the time you get to where you are condensing 40 parts, the takeoff would stop.
Now, to get RLM setup and 'tuned' properly, it will take some work to do, vs something like ARC, which does this 'on the fly'.
RLM has some upside to simple LM. I do not see it as a 'perfect' fit, but I do think this allows more of a controlled management of the RR, keeping the LM working producing a higher ABV, with fewer tails, without needing to be babysat as much as the 'original' LM.
Personally, I have done pretty well, babysitting the original LM (I actually have a nixon offset), using temp readings about 14" down in the packing along with upper head temps. When the temp lower in the packing starts to rise, I lower the takeoff rate. This allows me to keep the tails out of the 'compressed' upper layer I am removing.
The RLM method, appears to do some of that without having to have user interaction. Not sure which one will do better in the end. I personally like to control things in my runs, but would like to see how this method operates in the real world.
H.
Hillbilly Rebel: Unless you are one of the people on this site who are legalling distilling, keep a low profile, don't tell, don't sell.
