Shell and tube heat exchanger

I guess its sorta like the hybrid heat exchanger that Palmer talks about in his book but more like conventional single pass counter flow shell and tube heat exchanger. 25 feet of 3/8 OD copper in a CFC yields 353 sq in, correct? Well a shell and tube heat exchanger with 1 foot 3/8" copper tubing (wort passage) would yield roughly 3 times the surface area if used with a 4" PVC pipe. About 80 tubes will fit inside the diameter of a 4" pipe. This is about 1100 square inches of cooling surface area. The hardest part of building something like this would be the manifold for the 80 or so individual tubes for tube side piping and securing it with in the PVC pipe. Stainless could also be used for shell side might be more costly. With the use of pumps pressure wouldn't be an issue and with the correct configuration the shell and tubes would all be filled to 100% capacity. I wonder why something like this isn't used more often. Simplicity of the other systems i guess but this would be way more efficient and cooler (bling factor). What is everyone take on a system like this. I have some experience in this area just wondering what everyone else opinions are.

Thanks and happy home brewing!

That thing would be a byotch to build.

Shell-and-tube is a common arrangement for boilers, large steam generators and reactors. Very strong. Stanley Steamer used them and racers routinely ran them at double the design pressure.

It is much simpler and cheaper to build a plate chiller for low-pressure applications, like wort chilling.

If you want to build one, the simplest way is to make end-plates where the holes are a few mills larger than the tubes outer diameter, then use a flaring tool to make a mechanical seal.

Beecha,
A couple of people have built shell & tube exchangers w/ pvc & copper tubing. Here's one I did
https://www.homebrewtalk.com/f51/high-volume-rapid-chiller-120861/
Also, you can do a search on High Volume Wort Chillers. Catt22 built one also. It's not nearly as elaborate as what you suggested, but it crash cools very quickly. Check them out and let us know what you're going to try. Luck - Dwain

beecha said:

I guess its sorta like the hybrid heat exchanger that Palmer talks about in his book but more like conventional single pass counter flow shell and tube heat exchanger. 25 feet of 3/8 OD copper in a CFC yields 353 sq in, correct? Well a shell and tube heat exchanger with 1 foot 3/8" copper tubing (wort passage) would yield roughly 3 times the surface area if used with a 4" PVC pipe. About 80 tubes will fit inside the diameter of a 4" pipe. This is about 1100 square inches of cooling surface area. The hardest part of building something like this would be the manifold for the 80 or so individual tubes for tube side piping and securing it with in the PVC pipe. Stainless could also be used for shell side might be more costly. With the use of pumps pressure wouldn't be an issue and with the correct configuration the shell and tubes would all be filled to 100% capacity. I wonder why something like this isn't used more often. Simplicity of the other systems i guess but this would be way more efficient and cooler (bling factor). What is everyone take on a system like this. I have some experience in this area just wondering what everyone else opinions are.

Thanks and happy home brewing!

Click to expand...

IMO the small diameter 3/8" OD tubing with only only about a 1/4" ID will have a huge resistance to flow with or without a pump. I built this prototype hybrid counterflow chiller with 1/2" rigid copper pipe for just that reason. It's 30" long and contains 8 pipes for a total length of approximately 20 ft. I just used some stuff I had on hand to build it and the length was completely arbitrary. It works amazingly well and far surpassed by expectations.

Here's some pics to give you the general idea. I think I could have fit two more pipes down the center for even better performance. It could also be made longer. A 36" long 4" PVC could contain approximately 30 feet of pipe. I don't know what the optimum configuration should be, but somewhere along the line there one would hit the inevitable point of diminishing returns. I wanted it to be an efficient quick chiller without being too bulky. The next one will probably be the 36" size with the 10 pipes inside, but to be honest, I don't know how much better it would function compared to the 30" design.

continued pics:









I don't know if this link will work to my Flickr pics but might be worth a try:

Flickr: Catt22's Photostream

There's more pics of the chiller build and some other brewing related pics.

IMO the key to a fast chill is a high wort flow rate coupled with a high volume water flow in the outer jacket. A pump is essential for max performance.

That's pretty slick, catt.

Are you using two pumps or running the cool side off the garden hose?

Thanks arturo. Dwain and samc are building similar chillers. It will be interesting to see what their results are.

I only use one pump and that's for the wort. I use the garden hose for the cooling water and if I'm making a lager I fill the elevated HLT with ice & water for the final stage using gravity flow. At that point, speed is not very important, so I slow down the wort flow to give the slower ice water flow time to do it's job.

One other thing. The baffle/spacers shown aren't really necessary IMO. There's sufficient turbulance in the water jacket without them. The spacers are handy to hold the pipes in place while sweating the fittings, but something like a couple of beer cans or some 2" pvc could be used for that purpose. As I said, this was an off the cuff prototype and I was improvising everything as I built it. Next one will be more well thought out. I also plan to use 3/4" fittings for the cooling water inlet and outlet of the PVC jacket to improve the water flow even more. I used 1/2" for this first go around. The whole thing is relatively easy to put together and fairly inexpensive too.

I'm thinking a lot of these will be built once the word gets around from some other builders.

I am really surprised that this is something others have not experimented more with. A pump would pretty much be mandatory since the higher flow rate of cooling water you can pass through the shell side the more wort you can cool in the tube side (ideal for large batches). Another key design factor is space, these things can be made rather small and compact and fit nicely on the back of a brew rig, pump the wort through it and straight into the fermenter. The 1/4" ID copper would work but like catt said the resistance would be rather high. A pump with a high head capacity would work nicely for something like that and in a six inch pipe one could cram 100+ tube channels through it. I am glad a similar design has been thought of and tested and I am curious as to what other designs others have come up with.

Thanks for the replys guys.

Actually, this design concept has been around for a long time. I first came across something similar in a book called Brew Ware by Mark Stevens and Karl Lutzen written way back in 1996. So, the basic idea is nothing new. What is different about my version is the use of the rigid copper pipe and a pump.

Regarding the 100+ tubes you are proposing will have much more resistance than you might guess. Once you jump up to 6" diameter PVC, the end caps become a lot more expensive and short sections of the pipe are much harder to come by. The six inch pipe is also a lot heavier and bulkier. Lutzen's design used 6" dia pvc and a tight coil of 3/8" soft copper. The pvc jacket was shorter; only 24" long IIRC and he was using only gravity flow. Pumps were not commonly used back in the olden days of 1996.

Here's a link to the book: Brew Ware : Karl F. Lutzen & Mark Stevens & Randy (ILT) Mosher

Back in the old days, the advantages of a rapid chill were not fully realized by home brewers. At least not commonly.

thanks for the link Ill be sure to check it out!!

To me the hot wort is circulated down and back 6 times between water that would be 70 deg(your cold water input) and 120+ deg (your hotwater exit) on top of that your final path is out the same side the hot water is , so a pool of hot water at one end, and cool at the other, so in the end it would be rewarmed slightly on its exit.

Not sure if this makes sense or if its this double ipa.

MadWeezel said:

To me the hot wort is circulated down and back 6 times between water that would be 70 deg(your cold water input) and 120+ deg (your hotwater exit) on top of that your final path is out the same side the hot water is , so a pool of hot water at one end, and cool at the other, so in the end it would be rewarmed slightly on its exit.

Not sure if this makes sense or if its this double ipa.

Click to expand...

Yes, it makes perfect sense and you are correct in the sense of the ideal configuration. The reality in a practical sense is that it probably won't make a whole lot of difference. IOW, it works really well as is. Could it improved? Sure and I've already mentioned a couple of changes I will make on the next go around. I suspect though, that the performance improvement won't be very impressive if even noticeable at all. This chiller has a high volume of chill water flowing through it in a very turbulent pattern. There isn't much time or stability for temerature gradients to form. Certainly there is some, but I don't think it's very large except perhaps in the very early short duration initial stages. The wort temperature drops precipitously in one pass which only takes about four or five seconds. The wort temp is constantly dropping all the way through the circuit. It just cools more slowly when it passes through the warmer zone. It doesn't increase in temperature at all. The rate is slowed but not reversed as the wort moves through the circuit. I think this is the case until the wort is near the cooling water temperature and at that point is doesn't matter much. The wort is traveling through the circuit quite fast at about 3.5 gpm. Considerably faster than I think you realize. I'm not sure of the exact flow rate. I measured it crudely once, but need to do it again more precisely. Obviously, decent tap water pressure is essential.

Laminar Flow vs. Turbulent Flow. For those who are curious.

Along with surface area, and flow rate, flow characteristics are a key factor in a liquid cooling system. (disclaimer) I'm not a scientist by any means. Just trying to provide a simple explanation.


Laminar Flow

For our purposes, the hotter wort in the center of our copper pipe stays in the center, the cooler wort says against wall of the pipe. The wort in the hot center is not allowed to transfer it's heat away.

[ame=http://www.youtube.com/watch?v=KqqtOb30jWs&NR=1]YouTube - Laminar flow in pipe[/ame]


Turbulent Flow
There are no hot or cold zones, everything is mixed. All of our wort is allowed to transfer it's heat.

[ame=http://www.youtube.com/watch?v=-WNKn81eOgU]YouTube - Turbulence Field[/ame]

The same principle applies to the cooling water circuit.


Looking at catt's design, turbulent flow is created for the wort every time it rounds one of the corners. The cooling water is likely turbulent throughout the tube. The design that started the thread has a buttload of turbulence.

Compare this to an IC. Although the copper is coiled, the wort flow is more laminar than turbulent. By stirring the wort while running an IC, you are creating more turbulence in the wort, thus improving the heat transfer.

With a CFC, the wort flow through the copper is mostly laminar, similar to the water in an IC. This system works well because of the high flow rate of cooling water.

The plate style heat exchangers are the most efficient because they bring together surface area, flow, and turbulence.

Again, I'm no scientist. Please correct me if I missed a point.

To summarize, turbulent flow good. Laminar flow bad.

sounds good to me!

arturo7 said:

Laminar Flow vs. Turbulent Flow. For those who are curious.

Along with surface area, and flow rate, flow characteristics are a key factor in a liquid cooling system. (disclaimer) I'm not a scientist by any means. Just trying to provide a simple explanation.


Laminar Flow

For our purposes, the hotter wort in the center of our copper pipe stays in the center, the cooler wort says against wall of the pipe. The wort in the hot center is not allowed to transfer it's heat away.

YouTube - Laminar flow in pipe


Turbulent Flow
There are no hot or cold zones, everything is mixed. All of our wort is allowed to transfer it's heat.

YouTube - Turbulence Field

The same principle applies to the cooling water circuit.


Looking at catt's design, turbulent flow is created for the wort every time it rounds one of the corners. The cooling water is likely turbulent throughout the tube. The design that started the thread has a buttload of turbulence.

Compare this to an IC. Although the copper is coiled, the wort flow is more laminar than turbulent. By stirring the wort while running an IC, you are creating more turbulence in the wort, thus improving the heat transfer.

With a CFC, the wort flow through the copper is mostly laminar, similar to the water in an IC. This system works well because of the high flow rate of cooling water.

The plate style heat exchangers are the most efficient because they bring together surface area, flow, and turbulence.

Again, I'm no scientist. Please correct me if I missed a point.

To summarize, turbulent flow good. Laminar flow bad.

Click to expand...

Not a bad explanation for a non-Chemical engineer. If anyone wants to ensure turbulence during the design phase use the Reynolds number and shoot for a value greater than around 4,200. Reynolds number - Wikipedia, the free encyclopedia

The equation isn't really that intimidating it just needs the internal diameter of the pipe, the velocity of the fluid, the viscosity of the internal fluid (this changes with temperature!), and the density of the fluid (also changes with temperature!). Make sure that at boiling and 55F that the Reynolds number is above 4,200 and you'll maximize heat transfer.

Regarding the earlier posts which theorized that a shell and tube heat exchanger would have back pressure from flow; not a chance. The velocity through any one tube would be quite low and the length of the tubes is very short. We could work it out easily enough but the pressure drop across the HEX would be much less than 1 psi.

ChemE said:

Regarding the earlier posts which theorized that a shell and tube heat exchanger would have back pressure from flow; not a chance. The velocity through any one tube would be quite low and the length of the tubes is very short. We could work it out easily enough but the pressure drop across the HEX would be much less than 1 psi.

Click to expand...

Yes, I agree. I was overlooking the fact that all of the tubes were connected to manifolds at each end instead of being continuous and looped back or coiled. A bigger problem with the design having some 80 tubes crammed into a 4" ID PVC pipe would be having enough room around those tubes for sufficient cooling water flow. That, and it would be a major challenge for a DIY'er to actually fabricate that thing. I also think it would work as well or possibly even better with fewer, but longer tubes.

I wanted to come up with a simple high performance design that most anyone could build easily and economically with common, off-the-shelf pipe and fittings and I think I have done that. Keeping it relatively compact was also a major design consideration.

I have a tube-in-hose store bought CFC that worked OK, but not even close to fast enough for my needs. I have avoided buying a plate chiller mostly due to the clogging issues and to some degree the cleaning issues too, not to mention they are rather pricey.

I think that a high performance tube-in-hose CFC could be built using 25 ft or so of 1/2" ID soft copper tubing and a large diameter outer hose with an ID of maybe 1-1/4", but that bad boy would be both bulky and heavy too. I don't know what the cost would be for that kind of a design, but I'm thinking it won't be cheap. It might also be a hassle to find and fabricate the necessary fittings.

Edit - Posted to wrong thread