Quality control in the brewery
by Michael J. Lewis
There are many definitions of quality; some are quite long. They fall into two general categories: (1) those definitions that refer to "excellence or fineness" and (2) those that simply refer to "defining characteristics."
Naturally you are drawn to the first sort of definition but promptly run into the problem of what is "excellence or fineness" and who gets to decide. You soon realize that "excellence or fineness" very much reflects a personal bias or one's life experiences and so has serious limitations as a definition for use in the working world of beer making. The second sort of definition avoids that pitfall and, although intrinsically duller, is more workable. Indeed the first definition of quality in Webster's dictionary is "an essential distinctive property, characteristic, or attribute." No nonsense about "fineness" there!
One good definition for a quality beer is therefore simply "a beer that consistently meets specification." These words are of course quite a mouthful. The idea of a specification immediately requires that someone, at sometime and someplace, has decided what the beer's defining character(s) should be and how it should be measured. The idea of consistency immediately requires a system of people, plant, and process who are able to repeat exactly what they do time and again. Therefore, these ideas about quality are complicated and expensive but most useful.
Many amateur beer aficionados and some craft brewers choose the "excellence and fineness" definition of quality. They can then decide that beer X has "quality" and beer Y does not based on well-entrenched biases. The "consistency meeting specifications" definition is based on scientific fact. You might argue that "terrible beer" can be "consistent and within specification" and therefore be a "quality" product. Quite right!
To nail down the point, whether or not you like a beer has nothing whatsoever to do with its quality and a definition of liking is quite different from a definition of quality. If you don't like a beer or don't approve of it for some reason, that is too bad, but that does not mean it has poor quality.
The "consistency" definition contains two points that need to be addressed: (1) What are and how do brewers establish specification and how do they find out if specifications have been met? (2) How do brewers set up the process to achieve consistency?
Setting up specifications is done all the time. Brewers decide on the basic properties of original gravity, color, and flavor and from this develop a formulation of raw materials and a process to extract what is wanted from them. Brewers should know how to read and use the specifications of the raw materials they buy. These specified items, however, become of singular importance when brewers must make the same beer many times and perhaps over many years, because specifications are nothing more or less than a list of beer properties that define the product.
Beer specifications and the analyses that go with them are of two general kinds: (1) those that can be perceived by the human senses and (2) those that require instrumental analysis.
Sensory methods: Sensory methods are not necessarily easy to apply (and often ill used) but are useful and quite cheap to do. They include an analysis of beer flavor (undoubtedly beer's most important attribute), beer clarity, color, and foam. Brewers who do not regularly and critically taste and visually examine their beers in a formal setting deny themselves much critical information.
Please note that tasting beers is different from drinking them and quite foreign to guzzling a few pints at going-home time. In a working brewery of any size, beers, and beers in production and all raw material (malts, hops, and water), should be tasted every day by members of a designated taste panel. The taste panel is the corporate memory of the flavor of the company's products. These people need to be trained to know what the beer should taste like and be sensitive to deviations from normal. If several tasters agree that a beer or a product in process deviates from normal, the brewer needs to make some decisions to correct the problem. A tasting card, with chosen descriptors on it, can be used routinely to describe beer flavors if preferred; in this way the direction of deviation and so clues to intelligent corrective action might be gained. Of course, it would be preferable to set aside a supply of beer as the perpetual flavor standard, but the ephemeral nature of beer flavor defeats this approach.
Beer color, on the other hand, can be measured in a comparator (just a light box set up for visually matching color - the human eye is much better at this than most instruments) or by quite cheap instruments, such as a tintometer. A standard beer set aside for color matching remains stable for quite a long time if kept cold and in the dark.
Haze is also easily detected by direct visual observation and can even be sufficiently quantified by shining a beam of light through the beer at right angles to the observer. Similarly, it requires very little genius to judge a beer's foam stability and other properties (such as cling) by a simple pouring test into a scrupulously clean vessel.
Observers can rate the beers on some sensory scale. Putting numerical values from instruments on flavor, haze, foam, color, and so forth is where the trouble starts, but that isn't really necessary for a simple quality-control program.
Some additional tests derive from these sensory observations. One of the most important qualities of a beer in the marketplace is that it should remain clear (bright) until consumed. Cycling a beer on some regular schedule (e.g. daily) between a warm place (60° C) and a cold one (40° C) will create haze; more stable beers withstand more cycles than less stable ones. It's not a difficult test, but brewers should know how many cycles their beer should normally withstand.
Similarly, storing a beer at 25° C in an archive (a fancy name for a warm cupboard) will encourage microbial growth and other sorts of beer breakdown. A sample of beer from each bottling run should remain in the archive during its expected lifetime in the marketplace. Most of these tests should yield the result "normal," but those that do not should raise the alarm and put corrective processes into effect.
Instrumental Analysis: The second kind of specification and analysis is not amenable to sensory testing. To decide whether these sorts of specifications have been met requires analysis by chemical tests or instruments. High on this list of "invisible" specifications has to be the original gravity (OG) and the degree of fermentability (hence alcohol content) of beers. These are most easily determined on wort but require an investment in some simple apparatus - a hydrometer and measuring cylinder.
The wort OG and fermentability are fundamental specifications for a beer, because beer is made from the fermentable portion of the wort. These values also allow a brewer to calculate extract yield from raw materials (brewhouse yield) and predict beer yield.
The degree of fermentability can be determined by a rapid fermentation test in which a high population of yeast cells, with frequent agitation, rapidly ferments out the wort. The difference between the starting and ending gravity in this test tells the degree of attenuation to be expected in the brewery fermentation and is a good predictor of the alcohol content of the beer. Consistent values for these specifications bespeak consistent mashing and so consistent products and give much information for little investment of time and money.
At the same time, wort flavor and clarity can be noted. A sample of wort, taken under aseptic conditions and set aside in the archive, will reveal its microbiological status in a few days and tell a good deal about the sanitary status of the brewhouse. During fermentation, daily hydrometer readings on all fermentations in process are not advisable. But at the end of a fermentation you can expect the present gravity to reach that determined by the rapid-fermentation test and should probe for this information.
On this same sample of beer-in-process, you can test for diacetyl (if important to the product) by heating the beer to develop and volatalize the diacetyl aroma. When the beer reaches end gravity, it's time to chill the beer to cellar temperature. These strategies give lots of information practically free of charge.
Beer analyses (other than those sensory methods for foam, color, flavor, and haze already mentioned above) are less justifiable than wort analyses. Beer pH is easily measured with an instrument of modest cost (as these things go). Beer bitterness would be a nice measure to have but requires a good spectrophotometer and involves solvent extraction. Besides, you can expect your trained taste panel to give reproducible and sufficiently accurate bitterness data.
Package beer, on the other hand, must be analyzed for CO2 content (carbonation) and bottle "air" for flavor stability. These are both measured on a device called a Zahm-Nagel, named for the company that makes it. Adaptations of the device allow it to be used on beer in tanks. Similarly, for a serious microbrewer an oxygen meter would be a good investment. Use it to check on air pick-up, particularly downstream from the fermenter, such as during beer transfers and filtration and during final processing into packages.
Analysis of beers for alcohol content, real extract, and so on won't add much to the wort analysis already mentioned, and if they are to be done properly would require a good (read expensive) analytical balance.
The microbiological status of a packaged beer, especially one destined for a distant market, is of prime concern for beer flavor and for the safety of the consuming public (potential for exploding bottles). Clarity is no guarantee of "sterility" (a much over-used word in the microbrewing trade, especially as in "sterile filter") because bright beer might contain many hundreds of thousands of yeasts and bacteria. The only satisfactory microbiological test is to pass at least 100 ml of beer through a 0.45 micrometer membrane, then plate the membrane on media (such as MRS) under conditions (for instance anaerobic at about 25° C) capable of detecting the target organisms in low numbers. A quick squint at a beer sample under a microscope doesn't cut it. Release of beer to the trade should await a report on its microbiological status.
The other aspect of the original definition of quality is "consistency" as in consistency of process, people, and product. Brewers generally behave in consistent ways when making beer. For example they repeat quite exactly the weights of malts and hops used, and the times and temperatures at which they are brewed (although temperature can be hit or miss, especially in some infusion systems, brewers should learn how to calculate - predict - mash temperature and manipulate the wort data mentioned above).
Generally, brewers are well aware of the need for specification and consistency (quality) in the brewhouse operations. That might be said with less confidence in cellar management -- in wort aeration, fermentation temperature, and so on -- and there is one glaring shortcoming in many breweries that is worth mentioning: control of the yeast pitching rate.
The rate and extent of yeast growth intimately affects beer flavor, and yeast growth (all else being equal) arises from pitching rate. Ideally, the amount of yeast pitched should be a consistent number of living yeast cells (commonly 106 cells per milliliter per degree Plato) that can be determined directly by haemocytometer count under the microscope. This requires a decent laboratory and appropriate skills to go with it.
There are simpler techniques of consistent yeast addition, of which the packed cell volume is most practical. In this case a known volume of the yeast slurry is centrifuged and the percent solids read off. If, based on experience, we know that 25 percent solids requires 1.5 liters of yeast slurry per hectoliter, then 15 percent solids needs 25 ÷15 x 1.5 = 2.5 liters of slurry. Without calibrating the packed cell volume to yeast counts, the method permits consistency without exact numbers. In a recent experiment the students enrolled in the Master Brewers Program at the University of California, Davis found that one yeast slurry of 17 percent yeast solids contained 8 x 108 (800 million) cells per milliliter and so required 1.5 liters of yeast slurry per hectoliter of wort.
Keeping records is the foundation of a quality control program. There is very little point in making quality-control measures on worts and beers unless the values obtained are actually used in some sensible way to control quality.
This involves recording the results in a continuous fashion such as on a graph, so that up or down trends in extract yield or fermentability, for example, can be detected readily. You can establish upper and lower limits of acceptability as guidelines for corrective action. There are many statistical methods for setting up such plots, but they don't have to be complicated to be effective and useful.
A separate issue from that dealt with here is quality assurance. Quality assurance (QA) has the same relationship to quality control (QC) as good financial records have to preparing one's tax return. Thus QA has to do with the general status of the plant, process, and people as well as the product. QC deals with specific issues of specification and analysis focusing on the product only.
For example an excellent program of plant sanitation, or the decision to hire educated brewers, is a QA program with the objective of assuring low counts of foreign organisms in the beer. The QC analysis determines that low counts have been achieved. QA programs are essential to success in a brewery enterprise. The three pillars of QA are (1) specified raw materials, (2) consistent processing, and (3) rigorous sanitation. Michael J. Lewis, Ph.D. F.I.Brew, is a Professor Emeritus, University of California, Davis. He is also a principal in the consulting firm Lewis Twice and Shellhamer.
Published in the June 1997 issue of BrewPub Magazine.