Dear Mr. Wizard,
When racking to a secondary, I find that a lot of active yeast can be left behind, especially if it is done too soon. I don’t see much risk leaving beer on a small amount of sediment. It seems to me that racking to the secondary should take place when fermentation has reached a particular level, say 80% complete or when a hydrometer reads 1.01. What is your opinion on this?
Mr. Wizard replies:
Personally, I like to minimize the number of times wort and beer are transferred because with each transfer there is a risk of damaging the beer either by contamination or oxidation. I, like most brewers these days, use cylindroconical fermenters and the only time the beer is typically moved is after fermentation — either en route to the filter or directly to the serving tank for unfiltered beers. We do rack some of our beers to a secondary fermenter when we dry hop using whole hops or when we are making some beers that are aged in oak.
When racking into a secondary is deemed appropriate, I like to do the racking before fermentation is complete to help minimize oxidation since active yeast will quickly reduce the level of any oxygen introduced during transfer. If you wait until fermentation is complete and then rack, the likelihood of oxidation increases since yeast activity wanes after fermentation is complete. This can be especially problematic when dry hopping since whole hops have entrained air.
I agree with your rule of thumb of racking when the fermentation is about 80% complete and also agree that a small amount of yeast carry-over is not detrimental. In fact, when beer is transferred with very little yeast, I get concerned about oxidation and will use methods to remove oxygen from the vessel I am going into. At home when kegs are used, the easiest way to do this is to fill the keg with water and displace the water with carbon dioxide prior to filling.
I recently learned that many winemakers use pelletized dry ice to do the same thing. They place pellets of dry ice in a tank and allow the dry ice to sublime. This forms a nice blanket of carbon dioxide in the bottom of the tank and the wine is filled under the carbon dioxide blanket. This method is easy to use if you have access to small chunks of dry ice. This method requires attention to detail since dry ice in a closed container is a great way to make a little gas bomb. If you choose to try this method, do not place the dry ice in a closed vessel, rather leave the vessel vented to the atmosphere to ensure that pressure is not built up in the carboy, keg or whatever you are using.---------------------------------
Dear Mr. Wizard,
I love reading all the questions from fellow techno-geeks like myself. My homemade immersion chiller once worked wonders on extract brews. Since I’ve been mashing, its performance has dwindled. It usually 30–40 minutes to reach pitching temperatures. I’ve thought of building a counterflow chiller, but are they really that fast? Another thought I had, since I like the ease of immersion chillers, was to build another one using 1/2” copper tubing as opposed to 3/8”, so as to increase the surface contact area to the hot wort. Any thoughts?
Rochester, New Hampshire
Mr. Wizard replies:
So you want a techno-geek essay on heat exchangers? I’ll do my best here admitting up-front that mass and heat transfer are not subjects that I claim much expertise. But wort cooling isn’t rocket science and taking my advice, even if totally crazy, will not do anyone serious harm . . . so, yes, you can try this at home!
There are a couple handy equations that help communicate this subject. The first is Q=MCp?t, where Q expresses how much cooling we have to do, M is the mass of wort, Cp is the specific heat of wort (about 0.95) and ?t is the number of degrees the wort temperature will change. In English units, cooling 5 gallons (about 44 pounds) of wort from 210 ºF to 75 ºF requires about 5,600 BTUs. Another handy equation, Q=UA?t, relates this thermal load to the properties of the heat exchanger. Q is the cooling load (5,600 BTUs, for example), A is the surface area, U is the overall heat transfer coefficient (called “U-value” in heat exchange circles) and ?t is the temperature difference between the wort and the cooling medium.
What’s important is that the capacity of a heat exchanger to cool can be changed by affecting the U-value, A or ?t. The most influential variable is the U-value and is composed of several important components, including material thermal conductivity, material thickness, exchange rate on the medium side, exchange rate on the wort side and a fouling factor (dirty units don’t work as well as clean units). Thermal conductivity is a property of the material of construction and the best material used for heat exchangers with respect to this value is copper. Stainless steel, by comparison has a much lower thermal conductivity. Material thickness also affects U and as the thickness of the material (your copper tube wall thickness) increases, the U value decreases. This is why cooling fins on radiators are so very thin.
The U-value also depends on liquid flow across the surface of the exchanger. High turbulence on either side of the exchanger increases U. Turbulence can be affected by decreasing fluid viscosity or by adding shapes that induce turbulence. If you have ever looked at a heat exchanger at the local brewpub you probably have noticed that the plates look like lasagna noodles. The shape of the surface is important because of its effect on the U-value. In some shell and tube heat exchangers (garden hose around a copper tube, for example) the inner tube or tubes carrying the product have a spiral or ridged pattern that increases the U-value.
Finally, there is the difference between the wort and the cooling medium (usually water or glycol) and the area. Heat transfer increases as the difference between the product and coolant temperatures increase. Since the temperature difference between product and coolant changes as heat transfer happens, a sort of average temperature difference called the log mean temperature difference or LMTD is used in calculations. Adding surface area increases how much of the product and coolant is in contact with the heat transfer surface.
So how does this geek Greek help your cooling challenge? For starters, it empirically helps answer the question about the tube diameter. While it may seem at first glance that a fatter cooling tube will benefit cooling because of the larger diameter, the increase in area comes with a potential reduction in the U-value unless the cooling flow rate is also increased to maintain good turbulence. So increasing the tube diameter and coolant flow (more BTUs per minute of coolant) is certainly one way to increase cooling.
When you measure the temperature of the water coming out of your heat exchanger and compare it to the wort temperature, you will ideally discover that the two temperatures are not more than a few degrees apart, provided you have good flow. Note that if the flow is very low the coolant may be nearly the same temperature as the wort, but very little heat transfer is occurring. The idea is to remove heat quickly and efficiently. You want the wort temperature to be at your target and the water temperature exiting only a few degrees cooler.
If you have a good coolant flow rate and have a big difference between the coolant outlet temperature and wort you know that the flow of heat into the coolant is poor. This can either be caused by low U-value or insufficient area. Generating turbulence on either side of the cooling coil will improve the U-value. If you use an immersion chiller, stirring the wort is one way to improve U. If you focus on the coolant side you can increase turbulence and contact between the coolant and cooling surface by decreasing tube diameter, but this limits the amount of flow you can push through the exchanger.
A chiller with supply and return headers connected by multiple, small diameter cooling coils increases the U-value and area while maintaining the flow rate of coolant required to rapidly chill the wort. This would be a pretty slick little unit and could be easily built with standard parts from a hardware store. In principle, this is how plate heat exchangers used in most sectors of the food and beverage world operate. It’s also how cooling jackets on big beer fermenters are designed.
No matter what type of immersion chiller you use, it becomes clear that the coolant flowing through the unit is more effectively heated than the wort that is sitting on the opposite side is cooled. This is because the U-value is so much better on the coolant side. For this very reason, it is the norm to put the product being either heated or cooled on the tube side of a shell and tube heat exchanger and to put the heating or cooling medium on the shell side.
This is why a counterflow heat exchanger, be it a shell and tube or plate design, is so much more efficient than batch cooling in a kettle, even without having really cold water. The key to using these units is control. Ideally you should be able to throttle the wort flow and the coolant flow. If you have a ball valve on your kettle you can use that to control wort flow and the spigot on the hose can be used to control coolant flow. Of course measuring the wort temperature is required and with a few simple tweaks you are guaranteed to have the wort cool enough for pitching provided that the cooling water is at least 5 ºF (3º C) cooler than the target wort temperature. If you find that it takes too long to chill the batch and increasing coolant flow does not help, then increasing the chiller capacity by adding more tube length is required. Stay cool!