All the Numbers You Need, and Then Some
by Tim Patterson
Lab work isn’t the most glamorous aspect of winemaking. Winery staffers rarely wake up in the morning energized about the prospect of spending a day running malos. Especially at harvest time, with untold batches of grapes showing up and needing numbers ASAP, testing can become more than trying. Sure, you can send everything out, but that costs a bundle, and you may have to wait a couple days for results. If only there was some machine that could do the job instead…
Sure enough, there are machines that can, apparently, do just that. Quite a number of sophisticated, high-capacity, multi-purpose, automated analyzers are on the market—many of them initially developed for other food, beverage, environmental and clinical purposes and now adapted with new software to meet the testing needs of the wine industry. You can bet that any Truly Big Winery has one of these, or more likely several, and they play workhorse roles at service laboratories.
But do they make any sense for small- and medium-sized wineries, producers of 10,000-40,000 cases? Are these wineries likely to have a spare $30,000-$50,000 to invest in a piece of equipment that doesn’t process grapes and wine, but scrutinizes analytes and generates numbers instead? If they took the plunge, would they be getting their money’s worth? And most of all, would they know what to do with all those numbers?
For such an investment to make sense for less-than-mega-wineries, it would have to meet at least three criteria:
1. Save staff time and trouble through higher capacity, better ease of use and reduction of human error;
2. Make sense economically, when all the costs of old and new testing systems are compared;
3. Produce results that are precise (that is, repeatable), accurate (that is, true) and actually useful to wine production.
Lab equipment that promises to fill this bill comes in several flavors based on differing underlying technologies and test methods. This column doesn’t try to be a feature review or comparative evaluation, but a quick look at the relevant technological categories is still useful.
Some of the available equipment concentrates on doing a few things well, over and over, reliably and in large quantities. This category includes the Sirius Vinotrate line of equipment and the MetrOhm Titrino line, automated titration systems primarily deployed to analyze pH and TA. Equipment in this category not only performs the titrations but dilutes samples automatically, de-gasses wine samples (removing CO2 that could throw off results) and runs tests on large numbers of samples at once. Flow analyzers, which inject the sample into a continuous flow of reagents, are frequently dedicated to measuring free and total SO2, something wineries do over and over again; an example here is the FIAStar technology from FOSS.
Others have the capacity to do multiple tests on one or more samples all at once. These automated analyzers—referred to as “segmented” or “discrete”—generally use well-established enzymatic techniques to produce chemical reactions, then read the final values using some form of spectrophotometry to assess color as the marker of the results. Entries here include the Chemwell system distributed by Unitech, the Astoria-Pacific “Discrete” system (basically the same machinery as the Chemwell, but with different software) and, at the highest end, the Konelab line of equipment from Thermoscientific. Chemwell, for example, offers a menu of 26 tests that can be run, including glucose, fructose, lactic acid, malic acid, tartaric acid, acetic acid, free and total sulfite, anthocyanins and total phenols.
Newer (at least for the wine industry) and quite distinct wrinkles are systems utilizing Fourier Transform InfraRed Spectroscopy (FTIR), deciphering the chemical components of a wine sample through the infrared portion of the light spectrum, a different region from standard spectrophotometers. This is an indirect, secondary method of measurement, which means the equipment needs careful calibration so that it “reads” the complexity of wine properly. The main player here is FOSS, with its high-end Winescan line and “entry-level” Oenofoss system priced around $30,000, comparable to the Chemwell. The Oenofoss can perform about a dozen tests on either juice or finished wine, all with one tiny wine sample and no need for reagents.
Jessup Wiley, a product manager at Gusmer, which distributes, installs and calibrates the Oenofoss systems, says they are making a concerted effort to appeal to wineries producing less than 50,000 cases. At the moment, he estimates that there are less than 100 units in operation, in wineries from very small to very large. Geoffrey Anderson, president of Unitech, which distributes the Chemwell, says their equipment is also in a range of wineries producing 15,000 cases and up.
Prices and features vary, but as a group, these high-powered machines have the potential to make life easier for wineries that do a high volume of testing. The amount of testing done annually is a better indicator of need than sheer case output: A 10,000-gallon tank needs one sample for testing, as might each one of hundreds of 60-gallon barrels. Wineries that make small batches of many different wines become good candidates. Crushpad, a micro-custom crush specialist now ensconced in facilities in Sonoma County, is an extreme example: Enologist and lab manager Justin Rose says that in 2009, they ran something around 22,000 tests on several hundred separate wines, most of them in single-barrel lots. Along the line, they invested in an Oenofoss.
The machinery holds out the promise of making testing simpler, faster and easier: Many samples can be run at once, or many aspects of a sample can be tested at the same time, or both. Most of these machines make sample preparation easier, doing some of the work staff normally does by hand. Knowing what to do with all those numbers, of course, and deciding how to translate them into action in the winery requires someone with a good deal of training, as does knowing for sure that the numbers are correct—a topic we’ll get to shortly.
The analytical horsepower built into automation testing can eliminate many forms of human error in the lab setting. The technology ensures that samples get similar preparation and treatment, temperature gets controlled and no harried under-assistant-w inemaker has to eyeball one red color shift after another in a test tube. When these machines are on their game, winemakers sleep better.
Smaller wineries, naturally, have smaller staffs, often with everyone wearing multiple hats. Perhaps the biggest payoff from simplifying and speeding up routine testing could be that skilled, experienced, highly trained winemakers can get out of the lab and onto the winery floor, sniffing and tasting and poking the grapes, fine-tuning the treatments for different lots, finding ways to implement particular stylistic goals on a torrent of fruit.
With all those potential advantages in mind, working out the comparative costs is a complicated exercise. But the combination of saving money on reagents, being able to do more testing, cutting out some potentially expensive errors and freeing up skilled staff can add up to a powerful argument for certain smaller wineries.
Quantity and quality
So far, two of our three criteria seem feasible: Automation cuts down on muss and fuss, and for wineries that do a high volume of testing, the math could work out. That leaves just one question: Can you trust all those results, and what do you do with them?
Fancy technology always carries with it an aura of invincibility. The more your lab starts looking like the CSI lab, the greater the presumption of correctness. With all the research and development that goes into these machines, you would think they could figure out whether a wine sample has any residual sugar. But how do you know for sure?
Geoffrey Anderson of Unitech is quite enthusiastic about the potential in the Chemwell, but he also sounds a cautionary note. “We only recommend automation for labs that already have a significant background in enzymatic testing; otherwise, they can get buried in numbers.”
After toiling more than 30 years in the vineyard of wine testing, Gordon Burns of ETS Laboratories captures the issue this way: “The key point, which is hard to communicate, is that analytical results for wine are not a commodity. You can buy and sell bushels of corn with the presumption that they all have the same value. But that number on paper—generated in-house, or by an independent lab, or with an advanced analytical instrument—can have significantly different value, depending on the accuracy and precision. And there is a big downside if your results are not of the same value.”
Repeated surveys of winery lab work have found significant variability in test results, even among labs using the same techniques on identical samples. The range of results occurs not because of faulty testing technology, or a lack of knowledge of the correct procedures, but because too few wineries invest the time and money to make sure they are getting high-quality results.
The irritating fact is that every testing program needs a testing program to make sure it is capable of delivering results that meet a winery’s production needs. Every testing program needs a quality-control system running in tandem, specifying and implementing procedures for careful and consistent sampling, guaranteeing that any water used in analysis is itself free of analytes, cleaning equipment early and often, using reference wines to check test values on a frequent—probably daily—basis, calibration of equipment, keeping extensive records, and so on. Having strong confidence in a single test result often means running a half-dozen other tests. Otherwise, having an upscale automated analyzer perform dozens of tests in a single run could produce dozens of results that are all off, possibly by a distance that matters.
Does it require more time?
It could be, in other words, that powerful, automated testing requires more time and energy spent by more highly trained staff, rather than allowing the whole business to be turned over to the harvest intern.
Michelle Bowen of Vinquiry, another major wine industry testing lab, emphasizes that you “need to know your answers are correct. You need reference chemistry samples and check wines, but that check wine is only good for a month. Whites differ in testing from reds, sweeter wines from dry, and even wines from area to area. There are massive databases out there for calibration. It’s a big project for a small winery.”
“Years ago,” says Patricia Howe, a longtime observer of laboratory practices now working for ETS in its Roseburg, Ore., office, “the work of wine analysis was done by analytical chemists. Can a fancy toy give you the results an analytical chemist using reference methods would give you? To make sure the results are good, you need a more trained person to check on them. If you do have trained staff, these technologies can be good stuff—but smaller wineries using them as a black box, that’s not so good.”
The warnings coming from the testing services can sound a little self-serving: “Unless your winery lab can match our deluxe, ISO-certified standards, you have no idea what your numbers mean.” Just as the high-powered, automated machines can be a little intimidating, so can the testing elite, always asking, “But still, how do you know your results are correct?” Am I really going to pick a fight with Gordon Burns about my wine’s pH?
The skeptics, however, seem to have two points right. First, a winery has to have a clear idea of what values need to be tested, how often and with what degree of accuracy. Is it good enough to have pH readings within a tenth of a point? Does an error of a tenth of a gram per liter of malic acid matter? Knowing the quality of results needed is essential for making the economic calculations about investment in automated testing equipment. And so is having the appropriate level of quality control on the in-house testing, regardless of the program.
Moving to fancier equipment may well require an upgrade in staff training and laboratory procedures. The high-end automated testing technologies can be valuable workhorses for wineries with a heavy testing schedule—but only if some flesh-and-blood humans are riding herd on them.
In short, forget about the fantasy of having a machine do all that damn testing.
Tim Patterson is the author of “Home Winemaking for Dummies.” He writes about wine and makes his own in Berkeley, Calif. Years of experience as a journalist, combined with a contrarian streak, make hi m interested in getting to the bottom of wine stories, casting a critical eye on conventional wisdom in the process.
SEE OTHER EDITIONS OF THIS COLUMN �
CURRENT COLUMN ARTICLES »