|
|
|
This page created with Open Office Writer. Best viewed with Mozilla Firefox. Rotisserie Toaster Oven Retirement: 11 years old forever.............................................
|
|
|
Comments and questions? |
Downloads for this project Schematic: RTO-sch.pdf Parts List: see schematic |
||
|
Rotisserie Toaster Oven |
|||
|
Summary: This coffee roaster project modifies a thrift-store rotisserie toaster oven to produce a 16oz batch of roast beans in 15 minutes. The overall cost came to about $120.00. A schematic and parts list with sources are provided. The text includes a little science behind roaster problems and solutions. The unusual features presented here are: 1) an easy method for adding insulation; 2) an adjustable vent fan to facilitate indoor roasting and profile control; 3) a new "High-Lifter" drum for faster roasting; and 4) an easy electrical modification to allow simple profile control between 1st and 2nd crack.
 Welbilt TR660AS Rotisserie Toaster Oven (modified)
All right, so it ain't pretty. It's a... prototype!
My
first roaster was (and still is) a respectable Zach & Dani's
(now called the Nesco Pro).
It's a very friendly and automatic little roaster t Of course, none of the off-the-shelf available roasters meet all the requirements. Though now the Behmor 1600 comes close. But when I started this hobby, the Behmor was just beginning production and would not start shipping for another year. I thought it strange that the home-roaster market contained such a gaping hole, no low cost machines for 1# batches. I kept looking, thinking surely I must be missing something. Eventually, I realized I would have to build my own. Happily, there was a large and active community of home roasters on the internet. Thank you, everybody, for the friendly advice and for the encouragement of your successes! 
I did what any self-respecting home roaster would do... I headed to the nearest thrift store. Since I wanted some measure of control over roast profile, an RTO (rotisserie toaster oven) seemed like the best option. Coffee roasting needs higher temperatures than are normal for small ovens, somewhere around 550°F. I selected an RTO with the highest power rating I could find, 1600 watts. The Welbilt machine offered a small clue that its design and materials might stand up to higher temps. The temperature dial didn't stop at 400° or even 450°, it had an additional setting after 450° called "MAX". That doesn't necessarily mean anything, but it's kind of hopeful. 1(a). Roasting a pound of beans in 15 minutes on 120VAC/15A Or: Keeping da heat in da oven! Time? 15 minutes is an arbitrary time limit for roasting coffee. But we need some standard to compare roasters. Some large continuous industrial roasters and many air-spout roasters can do a 5 minute roast. That seems incredibly short but they do achieve an even roast level through to the center of the bean. For mid-sized (5-70K) industrial roasters, 12 minutes seems like a more typical roast time. Yet I know many home roasters who follow profiles out to 20 minutes or more with good results in the cup. It's a matter of personal taste. It's also one of those many choices we make that can vary with the characteristics of specific beans and specific roasters. My own choice after trying a fair number of long and short profiles on an assortment of beans is to shoot for no more than a 15 minute roast. This is partly for flavor and partly so I can maybe do three roast cycles per hour. Power? One standard household electrical socket in the United States can give 15 amps at 120VAC. The real limit is the breaker in your fuse box. So we can expect a maximum of 1800 watts. Blowing a breaker or fuse in the middle of a roast can ruin a lot of good beans. It's possible to build a roaster with separate heating elements using two power cords on two separate power breakers for a total of 3600 watts. But that shouldn't be necessary according to the numbers presented in Coffee Technology by Michael Sivetz. If I calculate correctly from the charts, his chemical and physical analysis says that raising 16oz (weight) of green coffee from 80°F to 380°F (actual bean body temp, city roast) requires only 284 watts in a 15 minute profile. Using that figure I should be able to roast 6 pounds of coffee at home without blowing a fuse! Unfortunately, that doesn't include all the heat losses that occur in real-life ovens. So if batch size matters, it really comes down to a question of... ROASTER EFFICIENCY. By itself, the Welbilt TR660AS oven could not roast more than a few beans effectively. According to my kill-o-watt meter, the actual power consumed by the Welbilt is only 1380 watts (hot). Less than the rated 1600 but, judging by the extremely high temperatures I found on outside surfaces, most of the power was wasted by directly heating the room without ever spending any time inside the oven. Not very efficient. I guess this is typical for most counter-top ovens? So it's likely that a little extra insulation will dramatically raise the oven temp. I disassembled the oven looking for ways to add a layer of high temp insulation. The rear wall was nothing but a single sheet of metal, the top nothing but a removable black metal pan. Clearly the two biggest heat losses. The sides and bottom were a little better, having two layers of metal providing air-gap insulation. The glass front door provides an air barrier but it leaks a little around the edges and it offers little resistance to infrared (IR) energy. All the inside metal surfaces were black-coated and were probably better IR absorbers than reflectors. In most ovens there are two kinds of heating going on at the same time. First and probably most important is conducted heat through hot air and water vapor. Second is the radiated IR heat from the heating elements. The amount of each varies with the design and materials in different ovens, but they are always both important. An efficient oven will keep in as much of the hot air as possible -and- reflect as much of the IR as possible until it hits the food. IR is just like visible light, only lower frequency. Bright shiny metal can be a great mirror for IR. Double-walls with free-air gaps are definitely better insulation than a single wall, but they still allow the trapped air to convect, naturally rising and falling around hot and cold areas even in small spaces. So heat is carried between the inside-outside walls, and IR can radiate directly between them. Physical insulation is much better than air since it prevents most heat transfer. McMaster and Carr sells a good assortment of high-temp oven insulation. Their 1/8-inch material was thin and easy to cut with scissors. Common oven insulation is a fiberglass mat. It should be handled with sturdy rubber or plastic gloves, not the thin disposable kind. Tiny pieces of the fiberglass threads are nearly invisible but they can cause some mighty annoying pokes and itches. Handle the insulation away from yourself to avoid dropping tiny bits on your clothes, and don't wipe them into your eyes or mouth with a careless finger. Fiberglass mats are a standard industrial oven material and once the fiberglass is in place behind a wall it's not a problem.  Roast Chamber: New inside insulation behind bright metal walls. I could have installed the insulation between the existing walls (where there were double walls). But the roast chamber walls were black, and I wanted to cover them with a bright, more reflective metal facing the beans. So I just put the new insulation on the inside of the oven, then covered the insulation with shiny metal. Very painless. Measuring the inside walls, I cut two panels of insulation that could loosely fit around the heating elements on the left and right sides, and also cut a panel to fit against the back wall. Insulation is good at stopping heat flow out of an oven but, unless you can find foil-backed insulation, it doesn't reflect any IR. And since it's a fiber mat, it can soak up a lot of smoke, tar and oils during roasting. Something reflective and washable is needed. Most home improvement stores sell a very thin aluminum sheet for roof flashing. It's extremely cheap and has a very low thermal mass which allows the oven to heat up or cool down faster. So I tried using it for a somewhat shiny new inside wall that would cover the insulation and reflect IR back on the beans. The aluminum flashing is thin (9.2mil) and it was easy to cut with scissors. Each side wall was held in place by a single stainless steel (SS) screw. The rear wall was held in place by being cut wide to fit behind the side walls, no screw needed. An open slot (on the left side only) through both the aluminum and the insulation was necessary to insert and remove the rotisserie rod. The side walls, like the insulation, had to slip over the existing heating elements. So I drilled round holes to fit over the heating rod insulators, but I cut them as slots. [But after a year of use the aluminum became dark brown with corrosion that could not be cleaned off. No good for IR reflection. I replaced the three aluminum walls with 16mil stainless steel sheet. Much better.]  Top-view: new ceiling insulation over new SS ceiling panel. The ceiling is the hottest thing in an oven, next to the heating elements themselves. I cut a ceiling panel from a 25mil SS sheet and attached it across the top edges of the original oven walls with SS screws. Above that I laid a piece of insulation. The SS sheet gets extremely hot but most of the heat stays in the oven thanks to the insulation. And the very shiny SS is a great IR reflector. Then I used the original black metal roof pan to provide a final air collector on top. I mounted it upside-down on the outer metal cover with an array of SS screws/nuts. The 12 holes through the ceiling are the smoke vents. The air flow pulled through these holes is limited to just enough to keep the smoke under control and remove as little heat as possible. 1(b). High-Lifter Drum Helps Efficiency Inside an oven, temperatures vary considerably. The top of all ovens is hotter than the bottom, the back hotter than the front, usually. The rotisserie drum helps by carrying the beans around more of the area inside the oven to soak up more available heat. However, most coffee drums use simple straight vanes to lift and stir the beans. Straight vanes allow the beans to drop early, long before the vane rotates to the top of the drum. So with a standard drum, the area occupied by the moving beans can be less than half the volume of the drum. This wastes the hottest part of the oven and a lot of bean heating energy goes unused. Some homeroasters have built drums with vanes on the outside to scoop air in towards the beans. Certainly a very clever and effective adaptation. Also, increasing the rotating speed to 60 RPM can help a little by carrying the beans a little farther, but the effect is highly dependent on the diameter of the drum. Making a typical drum carry the beans all the way around would require rotating fast enough for the beans to feel zero-g at the top of their arc. For a 7" diameter drum, that would require 100 RPM, a 6" drum would need 108 RPM. But the beans would be plastered against the walls, little mixing would occur and small differences in the distribution of the beans would likely vibrate the drum. 
Just to be different, I thought I'd try making the drum diameter as large as possible and also carry the beans all the way around, let them experience all the available heat first hand. The risk is either that temperatures at the top of the oven or the IR nearer the heating elements will be too much, and the beans will be carbonized on the outside, undone on the inside. So far though, none of the roasts have shown uneven internal heating. But I'm sure I must be close to charing beans. The "High-lifter" drum is 7-3/16" in diameter and 9" long. If it were any larger I couldn't get it out between the upper and lower heating elements. I had to cut the notch in the left oven wall lower by another half inch so the rod and the drum could be dropped down a little when maneuvering in and out of the oven. The drum assembly is just a pop-riveted aluminum frame covered by SS screen. The screen is a fine 16/inch mesh of 9 mil wire that is flexible, very strong and provides 73% permeability for increased air flow and IR exposure. Perforated metals would only provide 63% at most and they have a higher thermal mass. Even so, I did use punched metal for the end plates. 
The skeleton of the drum is formed from 1/8" x 1/2" aluminum stock, two rings and six ribs. For rigidity, four of the ribs are actually two U-shaped frames that fit together from opposite ends, offset by 60°, so that the end segments allow for the rotisserie rod hole (square). The rod holes are reinforced by wide fender washers. The only thing I'll claim bout this drum is that it's CBE (crude but effective). Perhaps more important, instead of straight vanes the high-lifter has 3 scoops. The scoops are designed to hold at least 1# of green beans total to insure good mixing. The total volume of the drum (about 350 in3) was intended to handle 2# of green, which it does very well. (I really had high hopes!) But the oven can't roast them fast enough. A 2# batch just creeps into 2nd crack after about23+ minutes. 
The scoops only begin to drop beans at the top of their arc. Most beans fall and mix at the back of the previous scoop. Originally I had designed it for six scoops but I dropped back to three being afraid there wouldn't be enough bean mixing in the very short distance between scoops. The scoops themselves are made of the same SS screen as the outside of the drum. Can't have too much air flow around the beans. After trying the drum with a load of beans it looked like the scoops were carrying the beans too far, and their might not be enough mixing. So I bent down the front-center edge of each scoop slightly into a squat "V" shape to allow some beans to start dropping early. Overall, the result is a very even roast. (see photos at left) Some handy SS hardware finished the drum. The latch with the spring-loaded ring from McMaster-Carr is easy to use even with my big oven mitts on. I didn't provide any way to clamp the drum on the rotisserie rod, and fortunately it has never shown any tendency to walk into the side walls. Turns out it's handy to slide it around while inserting and removing the fat drum from the oven. 
One more modification is always necessary for proper drum operation: a faster motor. Stock rotisserie motors only rotate at 3 RPM which I don't think is fast enough to insure good mixing and an even roast. Most of the comments I've read in the forums agree. Most folks use a motor in the 40-60 RPM range but for some real good reason I can't recall now, I used a 30 RPM motor. Maybe availability. Actually, it's a 24 RPM motor at 24 volts, but my simple power supply puts out about 30V so the motor runs a little faster. Torque is also important. I assumed if I did have 2# of beans and they were all clustered on one side of the drum, the heaviest motor load would be: 3.5" x 32oz = 112 in-oz. This doesn't allow for inertial jolts at starts, stops and stalls, but it's a good ballpark figure. My motor of choice provides 105 in-oz and it has been working fine. Physically, the motor shaft was a little shorter than the original. So I had to shorten the legs on the motor mount bracket. The easiest way was to just bend a knee in both legs. Attaching the original square rotisserie rod socket to the new motor required carefully drilling a 0.125" hole through the motor shaft for the drift pin. 
The new motor is 24VDC. DC motors provide much better starting torque and a definite direction of rotation. Plus, the direction is reversible. I thought running the drum backward might be useful for small loads where I wouldn't want one scoop taking up most of the beans. Running backwards, the scoops just function like standard blades. To feed the DC motor I had to include a transformer/rectifier/capacitor and a new directional switch. The new motor switch mounted easily in place of the old now-unused oven function switch. There was plenty of extra room in the right side for the new DC power supply. The transformer, rectifier, and capacitor mounted easily to the rear panel. I used some junk-box parts which were way bigger than necessary, providing about twenty times more current than the motor needs. Typical operating power is only about 100mA at 30V, 3 watts. Circuit details are in the download files available at the top of this page. I was a little concerned that the high oven temperatures would make the right side area too hot for the new DC power supply components and wiring, even with the new roast chamber wall insulation. So just in case, I added another layer of insulation (white) against the original oven wall, cut to fit around the motor and the ends of the four heating elements protruding through the wall. Can't have too much insulation. 2. Smoke Control for Indoor Use Unfortunately, the temperatures necessary to turn sour green beans into dark brown coffee also produce smoke, water vapor, a long list of not-so-pleasant volatiles and combustion gases, mostly CO2. Ideally, none of these should stay in the oven chamber around the beans. And the mix is not something we should breath if we had the choice, in some ways similar to cigarette smoke. But venting all those unwanted roasting byproducts will carry heat away too. Not good for efficiency. The only thing to do in a simple (non-recirculating) oven is minimize the amount of vented air to only what's required to remove most of the troublesome smoke. Fortunately, it doesn't take much air flow. And the beans don't start smoking until after 1st crack, so air flow isn't needed very much until about last quarter of a typical roast cycle. Heavy smoke generation doesn't occur until after the start of 2nd crack. 
I control airflow with a high-temperature vent fan and an adjustable air valve (air-can). The fan sits on the top of the roaster, draws oven air upward and blows it out a window. Hot air and smoke from inside of the oven are drawn through twelve small ceiling holes (already shown), and then through a hole in the black top pan (visible in photo at right), and then up through the air-can. The intake on the fan has five holes. I drilled a matching pattern in the bottom of a tin can (green beans) chosen for a good fit on the fan casing. The can sits upside-down so all five holes can align to allow maximum airflow. But if the can is rotated, its holes no longer align with the fan's holes and air flow is reduced. When the can is rotated off by 45° the outer four holes are closed off allowing air to flow only through the center hole. The ratio of these hole areas (all five open to just center open) is 4:1, giving me an air flow adjustment of about 4:1. The maximum flow is somewhat constricted by the size of the 12 holes in the ceiling as well as similar holes in the bottom of the oven, but I do get some control of airflow.  Air flow set to max. Side marks are minimum. The vent fan came complete with it's own cooling fan decoratively perched on top to cool it's own motor. The main fan is all cast aluminum to easily withstand oven air, but by the time the air leaves the fan it has cooled considerably thanks to the black metal pan surface, the air-can and the body of the fan itself. A 4" vent adapter screwed to the square face of the fan allows a standard flexible clothes dryer hose to direct the smoke outside through a standard dryer vent. Some day all these parts will need cleaning, but so far they've vented over 100# of roast coffee with no significant accumulation of crud. Strangely, the crud looks exactly like instant coffee. I wonder... 3. Consistent Time and Temperature Profiles The original wiring allowed the thermostat (top knob) to turn the upper heaters on or off. The timer (bottom knob) controlled everything on/off. For the first trial roast I allowed the oven to run on max power from start to finish, which drove the beans quickly between 1st crack (1C) and 2nd crack (2C) with barely a pause. The second roast trial showed that that shutting down the top heaters at 1C caused too drastic a drop in temperature. The top heating elements are significantly higher power than the lower elements (947W vs. 610W). Any slowly changing on/off scheme like a thermostat or by hand is too slow to avoid big drops. That's probably fine for pizza and french fries, but big drops are not fine for coffee beans. The only possibilities were: 1) a faster electronic control that can switch faster than the heating element thermal time constants; or 2) don't drop the power so drastically. I decided to try the second way first, although I had no assurance that it would produce a useful roasting profile. As shown on the schematic, two rectifier diodes connect across the thermostat switch. One diode would have handled the current but I like conservative designs. When the thermostat switch is closed, the top elements get full power as always. But when the thermostat switch opens, current can still flow through the diodes. The diodes allow just a tiny bit more than exactly half-power to the top heating elements, so the drop in temp is not so drastic, 1000 watts total instead of 1380. Just by chance, this works very well. 
The oven allows two ways to vary temperature, the air flow tin can and the thermostat switch. The two controls are easy to change and the roasts are identical from batch to batch in respect to time and bean color for any given bean lot. In practice I do a simple roast profile in three periods: First period is warm-up. I rotate the air-can to minimum and set the thermostat control for "300°F". I load the drum empty and turn on the rotisserie motor so things heat up evenly. Then I set the timer control to about 20 minutes (arbitrary) to connect power to the heaters. I use the timer just as a free safety device that will turn the oven off if I drop dead. With these settings the oven is running at full power, 1380 watts, and I watch the neon indicator. In approximately 12 minutes the oven temp will reach above 450°F where I consider it warmed-up. Since the roast chamber now has extra insulation, the thermostat sees only a fraction of the actual internal heat, and it finally tries to switch off even though it was set for only 300°F. And the neon light flickers differently telling me to move on to second period. Second period needs some beans. I load about 20oz (weight) of good green beans in the drum and start it turning again. The air flow stays at minimum, but the thermostat control gets set to "MAX". The safety timer gets turned back to 20 minutes again to keep things running. From here on, the thermostat will be just a switch for either full or half power on the top heaters. The oven is now running at high power just as it was during warm-up, but the thermostat set at max will never cut back by itself to half-power. In this mode the oven brings the beans up to 1st crack as fast as it can, usually in about 11 minutes. Third period begins somewhere in the middle of 1st crack. I rotate the air flow can to maximum and turn the thermostat down to "zero" forcing the top heaters run at half-power. This overall reduction in heat is just enough to allow the beans to coast through a respectable 2-3 minute break between 1st and 2nd crack. I wish I could say this was by design, but in truth it's just luck. Had this power level not produced a manageable break between the cracks I would have had to start some more serious oven modifications.
4. Cost Approximately $120 x. Lessons Learned and Future Improvements Insulation: The glass front door could benefit from some aluminum foil or flashing sheet to reflect IR. It shouldn't be too hard to attach it or hang it so a window is still clear to view the beans. However, judging by the new internal oven temperatures, even the cheap aluminum flashing and 1/8 inch insulation has raised oven temp by about 150°F! Now, setting the original thermostat for "250°F" warms the oven to 400°F. I use that setting as my warm-up. Post-note: I tried 6" band of household aluminum foil, folded in half and hanging over the glass door. This created a new 3" inner reflective surface on the inside of the oven. Surprisingly, it created a dramatic increase in bean heating. In fact it became difficult to prevent scorching the beans! As mentioned above, the top elements are higher power. So the foil reflecting more of the IR from the front-top element must have pushed the heating over the edge in this oven. I removed the foil. But it's certainly an effect to remember. Half-power Indicator light: The new neon bulb I added above the thermostat knob doesn't give much of an indication when the top heaters go in/out of half-power mode. I tried a wide variety of resistor values to set the current down near the limit of neon conduction, thinking It would give a bigger difference when the diodes cut the voltage in half. Not so good. It works, but I need to think of a better indicator circuit. Chassis: I don't like the roof chamber formed by the black top pan. It's part of the vent duct but It's difficult to open up for cleaning. I'm afraid that over time chaff dust and tar will collect there and possibly start a fire. So far I haven't seen anything building up, but I'll need to check it once in a while. A better design would have all the exit venting easily accessible for cleaning. Materials: The flashing sheets used for interior walls buckled up a little after going through some initial heat cycles. Not a serious problem, they just look a little funny. I did allow extra lateral room so they could expand without pushing on anything. So I'm sure the wavy effect is due to the cheapness of the flashing in that the material thickness (and content) probably varies more by % than normal dimensioned aluminum sheets. Next time I'll use thin SS which would reflect better and clean better. Post-note: I have replaced the aluminum flashing wall panels with thin stainless. Very nice IR reflection an no build-up of gunk yet. After two months of weekly roasting sessions the walls still need only vacuum cleaning. Air Flow: It's clear from the little wisps of smoke that escape around the top edge of the glass door that I'm not pulling any real negative pressure on the oven, even with the can valve fully open. Enlarge the holes in the SS ceiling panel to increase the limited exit aperture. In measuring the total hole area at the inlet to the fan I realized that I should have made the roof holes at least 0.23" instead of 0.2". To make sure the flow turbulence through the 12 holes does not present a greater restriction than the fan inlet itself, I'll drill the holes out to 0.25". Post-note: I drilled the inside roof holes out to 0.27". It helped a little but not much. I think I'm now limited by the vent fan. It's rated at 6 CFM max. Drum: The high-lifter drum is a functional success but my construction has a lot of little problems. There are always 5-10 beans stuck in various locations around the drum. Clean-out takes several minutes between roast cycles, so re-warming the oven between batches requires a full 10 minutes. I didn't do a good job creating tight joints inside the drum so beans would have nowhere to get caught. Also the large diameter of the drum makes it a little more awkward to insert and remove than normal. I guess that's unavoidable. Assessment: Coffee is great. Construction was fun. Since this one works well I don't need any other oven, but... Ye-hah! Build more roasters!!! Have fun, Jim
|
|||
