I initially started roasting about 3 years ago with a hot-air popcorn machine, but quickly got tired of the small batch size.
I had been working on my car and lighting the area with a halogen floodlamp and getting really hot, which got me to thinking.
I did some quick research which indicated that more than 95% of the energy consumed by halogen lamps is emitted as infra-red radiation. Several experiments with lamps, reflectors, pyrex flasks, stainless bowls, etc etc later I had contructed a rudimentary 250 gram (half pound) roaster which showed the promise of this technique.

Construction:

And so I set off in search of a suitable drum for a larger-scale roaster. I found this at K Mart on special for $25, took it home, knocked off the spot-welded handles, drilled a hole in the bottom, and fabricated an adapter to mate the pot to the wiper motor (generic Japanese 12-volt continuous rotating reversible, which I had negotiated from the car wrecker for a mere $5).

I modified the pot lid by cutting a hole and fitting a scrap of stainless sheet as a loading chute, and screwed this to a little trapezoid assembly to assist in the alignment of the pot and lid (or drum and door, if you will).
One small high-temperature ball bearing was fitted to a bracket to support the weight of the front of the drum on the door.

I had acquired from the local scrap-metal merchant several aluminium extrusions which I used to fabricate the support beam for the drum and the base, lamp mounting bracket etc.
The case for the electrics (and main support housing) came from my "project stock" (or what my wife calls "All that old junk")
I re-used the reflectors from the floodlamps to focus the heat from the lamps onto the bean mass, and bent a small aluminium shield to prevent the beans from hitting the lamps or lodging on top of the reflectors and burning there.

The lamps were wired using glass-fibre insulated oven wiring - my background is as an electrical/electronics technician so none of this was much of a problem safety-wise.

I fitted a domestic 500 Watt lamp-dimmer with a larger heatsink to cope with the increased loading.
I used a 6 volt 2 amp switched-mode plug-pack DC power supply to power the motor, as I found that 12 volts rotated the drum too fast.

I experimented with the position of the lamps so that they were centred over the bean mass as the drum rotated, to maximise heat transference to the beans.

The roaster performed as hoped without any need for re-design or adjustment - all that I needed was to experiment with temperature settings to optimise the roast profile, but in fact the operation is very
straightforward, and has roasted without a hitch (apart from one replacement lamp at $3)

The tools that I used were everday home-workshop tools - hacksaw, files, electric drill, metal nibbler, soldering iron, bench vice, and hammer.

1) Capacity in pounds/ounces

0.5 kilo = 1.1 pounds

2) Type of heat

Electric radiant, using two 500 watt linear halogen lamps controlled by triac lamp-dimmer

3) Type of agitation

Rotating drum, fitted with one stainless steel vane and one stainless steel spiral wire guide for unloading

4) Roasting time at maximum designed load

18 minutes to 2nd crack (depending on ambient temperature and type of beans)

5) Roaster temperature at finish of max load

455 degrees F (235 C)

6) Roasting coffee

a) Plug in roaster and switch on
b) Switch drum to clockwise rotation
c) Set temperature control to max, preheat drum for 1 minute
d) Pour beans into loading chute
e) Go and do something else for 12 minutes or so
f) Back to watching roaster, wait for 1st and then (with smoke) 2nd crack.
The beans are easily visible through the loading chute (also easy to test temperature with IR thermometer), and there is (of course) plenty of light to see the beans to check the state of the roast.
g) When desired level of roast achieved, reduce temperature control to show about ambient light level, place the cooling tray under the front of the drum, crack the front of the drum to dump the beans into the tray, and switch the drum to counter-clockwise rotation to wind the beans out with the spiral wire.
h) When all beans have been dumped, close the front of the drum, switch to clockwise rotation, crank up the temperature to max, and load another batch.
i) Cool the roasted beans with updraft extractor fan

7) Cost

( $1 NZ = 78 cents US)

stainless cooking pot $25
windscreen wiper motor $5
halogen floodlamps (2) $30
DC power pack $10
lamp dimmer $15
stainless bolts, washers, strap $20
aluminium hardware (scrap) $5
switches, cables, knobs $12

I also used a few bits that I had lying around (e.g. the housing for the motor and electrics is the case from a power supply for a Franklin Ace - anybody know what that is - or was?)

Door cracked for dumping beans

Electrics...

Temperature profile (one pound of harrar)

Heater assembly..

Thanks so much for adding your roaster! It is a novel contribution. Good luck, Dan

Once again, Brainiac, very tasty! I read you have replaced a lamp already, so I figured you've had this roaster running for a while now.

Some pictures of your bean cooler would be nice, too.

And one more thing, while I'm writing, is the electric supply 220 volts over there or 110 volts?

CHAD

Hi Chad
thanks for the comments

We use 230 volt electricity here.

I have been using this roaster full-time since May 2006 (so more than a year now).

BTW, here is a picture of my MK 1 prototype

I cool my beans in a mesh-bottomed commercial circular sieve, which sits on a wire rack. I use an externally-vented 6-inch extractor fan (noisy and powerful) to alternately vent the smoke & chaff during roasting, then to draw cool air up through the beans - which also clears the remainder of the chaff.

the picture isn't that clear, but shows the general configuration...

some more pictures and descriptions (from about a year ago) are here

http://www.dbr.co...diant.html

Thanks for your comments Ed - praise from you guys means an awful lot to me ( s:1 I'm not worthy! I'm not worthy! s:1)

BTW, I use a non-contact IR thermometer for my temperature measurement - just point through the chute at the beans and press the button. This gives surprisingly good results (see my chart on an earlier post).

Brian

This is one very clever machine. There is so much to love about it.
I have already been scheming about building something similar.

At the risk of short-changing you on the rave reviews that you richly deserve, I'll just go to my two questions.

I see that you have a way to evacuate the smoke from the roaster, but what of the smoke's effect on the halogen bulbs? Does it coat them with goo as I am imagining? or does the intense heat just burn it off?

You probably can't answer this one, but I am also wondering how this will do on our 110 volt systems.

For me the only drawback to your entry was the limited "How-to-do-it" information in your article. I would like to have seen more step-by-step "info. The link to your web page helped, but unfortunately this couldn't be counted as part of the contest entry. Our Forum software's limitation of one picture per post has been a handicapping factor for a number of folks, and it may have been a frustrating factor for you also.

Bottom Line: A very, very nice entry, Brian. s:1 ;)

This is a well-conceived and well'executed design. It is clear you through out traditional designs and started from scratch. I like how parts are neatly integrated and work as a unit, not just a bunch of loose parts. It's also a design that would respond well to automation. Your article was well written, and although this is a complicated build, I think a skilled person could duplicate it with what is here. Some of the photos are poorly lighted. Additional photos of the parts would be useful to builders.

Dan

David, Dan: Thanks for the feedback, guys.

I have had no build-up of residues on the halogen lamps at all - the surface temperature of the quartz glass is so high (up to 900 F according to some sources).

As for 110Volt operation, 500 Watt per lamp is 500 Watt regardless of the voltage (you just use twice the current vs. 230V according to Ohm's Law) so no problems there.

I acknowledge a shortage of how-to-do-it info in my posts. I built the roaster a whiles back, and paid no attention to documenting the build (took photos subsequently but only a few). I will disassemble the roaster after my next session, and take some well-lighted (thanks Dan;)) photos which should fill in the gaps.

Brian

Brian:

I wanted to take this little thing in my arms right away and speed off before you could see me let alone catch me!!

Sweet, Sweet and more Sweet!

You got way high marks for your beautiful piece of art. Talk about form and function.

Your construction list/article was incomplete in terms of "how to" for me. I really wish you had taken a few minutes to put all of your fantastic photo's into the entry sans the link to another place. I need to look at the photo set up here. If you found posting photo's an issue please let me know.

How you put the electrical setup together was missing; though it looked well executed, clean, neat and basic even if photo's were missing.

You have created a real winner of form and function.

I want one.

Just to hold for a little while before it goes into the Smithsonian.

Love to you and family,

ginny

thanks Brian for the entry and you kindness to homeroasters.org

Brian,

The fit and finish of your roaster is outstanding and thoroughly impressive; it looks like a production model that one could buy on the market. That fact that you accomplished this with the halogen lamps is equally impressive to me considering the 1.1 pound capacity. s:2

Although I lack the knowledge and understanding to be able to construct a roaster of this caliber, I do hope your will be able to get some more pictures, as mentioned previously and include the nitty-gritty detail so that others could build this roaster. Your kind, courteous and unassuming style of prose would lend very well to an excellent treatise. s:1s:1s:1

Your participation in this contest is very much appreciated and I sincerely thank you.

As promised, a little more detail on construction of the halogen roaster.

First step: get your hands on a stainless steel stockpot - the one I used was 8 inches in diameter, 8 inches tall. It has straight sides and a lid that sits down into the body - this is important as it gives about 7 mm of latitude for movement before beans start falling out!

Spot-welded handles are usual on the cheaper pots - it is easy to knock off the handles without leaving holes.
(shown is an example with riveted handles)

Take stockpot (it was clean when I bought it!!). Remove all handles. Drill hole in centre of bottom (normally the concentric turned finish will make it easy to find dead centre).

Mounting hardware for the drum. The dark-coloured square washer is in fact stainless steel, and mounts inside the drum. The galvanised washer mounts outside the drum, and between them the two washers support the thin stainless-steel drum well.

Drum mounting hardware on the outside of the drum

Inside of drum. The main agitating vane is curved to speed up dumping of beans. I don?t know if more vanes would make any difference to the evenness of the roast? at present the motion of the bean mass under the radiant heat source is slide-and-tumble rather than continuous tumble, and I believe that provides the optimum heat transfer.

The hole for the loading chute was cut in the door using a sheet metal nibbler. the chute itself is a circle of scrap stainless steel and is held in place with bent tabs.

The door mounting allows accurate alignment of the door to the drum (the original plan was to use this mechanism to open the door for dumping the beans, but the top-mounted hinge and lock knob proved easier to use)

Most of the weight of the drum (and beans) rides on a small ball bearing, shown here, mounted on the top of the door

Bearing viewed from inside....