Fluid-Bed Roasting: The Path Less Taken

By Arno Holschuh


In the 1960s, Mike Sivetz was a chemical engineer working at a polyurethane plant. One of the process steps he designed involved drying magnesium pellets by floating them in a column of heated air. Sivetz saw those pellets, and in an impressive leap of the imagination, he realized that the process could be adapted to the drying and heating of another important commodity: Coffee.

Thus fluid-bed roasting was invented. Long known as “Sivetz roasting,” this roasting architecture saw rapid adoption by a minority of the industry in the 1970s. It offered specific advantages: It was a faster way to roast, with profiles being shortened by as much as 50%, which made your equipment that much more productive on a pounds per hour basis. The roasting equipment was much simpler in design, consisting primarily of a burner or electric heater, a blower, and a roasting chamber. A cottage industry of tiny manufacturers was started; my first roasting experience, in the ‘90s, was on a home-brewed fluid-bed roaster. It had been constructed out of a heater coil from a GE electric stove, a blower from a Chevy, a candy thermometer, and some thin sheet metal. (Legend had it that the design came out of a Mother Earth catalog).

Specialty coffee grew through its infancy in the 1980s, a syrup-drenched adolescence in the 1990s, and grew to some form of maturity throughout the last two decades. Through this maturation, fluid-bed roasting has fallen out of favor. It’s clear advantages in terms of design simplicity, cost, and productive capacity did not speak to our singular focus on coffee quality. The most talented roasters and the most trusted manufacturers tended to ignore the technology and devote all their time to improving their skills on open-flame drum roasters, most notably the venerable Probat UG series. It became an agreed-upon “truth” that fluid-bed roasters couldn’t deliver quality, because they didn’t use any conductive heat transfer.

This “truth” is now worthy of re-examination – and challenge. A well-designed (or thoughtfully-modified) fluid-bed roaster can produce coffees with amazing clarity and sweetness. Your fruit and floral cupping notes will be present in the cup to an extent that simply isn’t possible with traditional drum roasters. The key to achieving these results: A great technician who is capable of helping the roasting staff overcome, or work around, the inherent design challenges.

Convection Is Not The Enemy

When you think about how fluid-bed and drum roasters work, the most obvious difference between the two is the method of heat application. Drum roasters famously use conductive heat (from the beans touching the hot steel of the drum), radiative heat (from the superheated steel and iron of the drum and end plates), and convective heat (from the beans being bathed in hot exhaust gases from the burners). Fluid-bed roasters normally only use convective heat – the only heat into the roasting chamber comes from the flow of hot air that both heats the coffee and mixes it around.

This definitely makes a difference. The higher the amount of conductive heat in a roasting system, the more body your coffee will have. At the same time, it is important to note that even very traditional roasters will be using primarily convective heat – something around 75-85% of the heat transfer is convective. So in a sense, all good roasters are air roasters.


What has happened in our perception is that we have conflated convection with poor roasting results. A lot of fluid-bed roasters really do provide sub-par roast quality, but it has nothing to do with their core technology. It is, rather, a failure to get good temperature and airflow measurements and apply adequate control systems. The most obvious proof of this point is the exceptional roast quality that’s provided by Loring Smart Roasters. Fluid-bed roasters may not be everyone’s favorite, but it is hard to argue with the delicate, sweet coffees they can produce. And they are almost purely convective in their heat transfer – the flame is located well away from the roasting chamber, with all heat being carried in by air.

What’s The Problem?

The common problems with Sivetz roasters can be separated into two categories: Inherent challenges and poor execution. Let’s tackle the baked-in problems first.

All fluid-bed roasters use a column of air to both float and heat beans. That air flow always has to be fast enough to move the coffee around. If it does not, then the beans sitting right on the airflow inlet at the bottom of the roast chamber will be constantly exposed to super-hot air, which makes them scorch and then eventually ignite. Having done it a couple times, I can tell you that this event counts as a Very Bad Day. So, you need really robust airflow for proper mixing.

On the other hand, the rate of heat transfer into the beans is also dependent on airflow. The faster the air washes over the beans, the faster they get heated. So there are times when you’d really like to slow down your heat transfer to the beans, but you cannot do it by slowing down the airflow.

The other inherent challenge for fluid-bed roasters is that they require a high-pressure flow of air. In drum roasters, the air has a lot of space to flow as it goes through the beans. In fluid-bed systems, they have to physically lift the coffee to get through. This means that fluid-bed roasters require blowers that can stand up much greater back-pressures – like as much as 30 or 40 times greater. We know from working with steam that high-pressure systems are prone to leaks.

Leaks cause two problems: smoke and loss of energy. If the leak in your hot-air path is after, in, or near the roast chamber, what leaks out will likely be smoke. If you can smell roasting coffee during the roast cycle, you probably have a leak. Over time, this can lead to some pretty poor air quality for the roast staff (and customers, should your setup be in a café). Energy loss is a concern for the roaster, as every cubic inch of hot air they lose is a bit of energy they do not have available for roasting. If your roaster appears to be lagging, this is the first place to troubleshoot.

Clearly we need to be prepared to find and repair any air leaks. Leak-finding techniques include listening for whistling, to running your hands along the air ducts to feel for any currents, to using a tool called a “smoke pen” that generates a thin column of smoke that is visibly disturbed when in the presence of an air current.

The most common leak in fluid-bed roasters hot-air path will be the roasted-bean exit valve. Once the coffee’s been roasted, you have to get it out, and whatever valve, gate, or door lets roasted coffee out can also let hot air and smoke out when the roaster is in operation. Make sure the sealing surfaces of this valve are clean and haven’t been pitted or scratched; make sure whatever holds it closed, be it a spring, weight, or solenoid, is working properly.

Trust… But Verify

The biggest single factor that leads to poor results in fluid-bed roasters is poor control. Coffee gets put into the roaster, heat is applied as quickly as possible given the machine’s limitations, and the coffee comes out when tan, brown, or black. On some roasters, this is all the stock equipment will allow you. This is not true profile roasting, which by definition involves being able to apply different amounts of heat at different times in the roast cycle. If you want to profile roast, even by doing such basic things as stretching an espresso after first crack, you’ll need a way to dial the heat application up and down, and you’ll need to know what your temps are.

Really quick thermodynamics lesson: Heat is not temperature. Temperature is how hot something is; heat is how much energy is being added. Heat is watts (or joules, or calories, etc.), temperature is degrees Fahrenheit or Celsius.

The amount of heat that will be imparted to a batch of coffee depends on the speed and temperature of the air flow. The speed, as noted above, is likely to be static, or something you have to set to move the coffee adequately (but not blow it out of the chamber). The temperature of that air column is what we can try to control.

First: Measure it. What you want to know is what the air temp is going into the bean mass, and in the bean mass itself. This will allow you to determine how hot the air column has to be to yield a certain amount of heat application. As an example, you might measure these at the beginning of the roast and find that the air temp going in is 350°F, and the bean temp 125°F. The difference between those two is 225°F. Now measure those temperatures again in 30 seconds: Air temp in is 400°F, bean temp 175°F.

Here’s the deal. If you know what temperature difference between bean and air yields what rate of rise, you can work backwards with a skilled roaster to figure out what inlet temperature you want at any point on the roast curve.

Bear with me on this story problem, this is important – you just learned that if the air is 225°F hotter than the beans, your rate of rise is 100°F/minute. (If this doesn’t make any sense to you, it probably will to your roasting friends.) Now if you were to aim for the air temp to be 125°F hotter than the bean temp, the rate of rise will be less (but still existent). At some point, the difference between the air and bean temps is so small that the rate of rise will be zero; that’s your “stall point.”

Anyway, all that head-scratching above is mostly for roasters. For techs, the question is how to measure it. The answer is simple and familiar to anyone who’s worked on modern brewing equipment: Thermocouples. If your roaster does not provide you with the data above, and you can, I suggest drilling holes to get a measurement. You can get fittings for most thermocouples from any reasonable industrial supply house, and the thermocouples themselves are also common these days. You’ll need a hand-held reader… or you can now start looking at using Cropster, or another roast-data capture product.

One last word about thermocouples: Keep them clean! Temperature readings can vary really widely if the steel housing of the probe gets gunked up. Scrape them off, use steel wool, wrap them in a rag soaked in espresso detergent… all will work with time.

Limitations We Cannot Overcome

So what if you can measure the temperature, you and your team know what you want your inlet temperature to be, but you have no way of achieving it? Too many fluid-bed roasters lack variable heat application. A really good tech can start modifying equipment to replace, say, a simple relay with a PID controller. But that’s outside the skill set of many techs. It’s definitely outside the scope of this article. Just know that with variable heat application, skill, and accurate measurement, your fluid-bed system can stand tall against any drum roaster out there.