View Thread

directtraderoasting wrote:

Maybe I missed it, but what is the drum thickness and what rpm?

Nice build. I have a custom 2kg, but looking to make a smaller one in the future as well.

JackH wrote:

Welcome to Homeroasters!
Nice job on the roaster! Thanks for posting your build.
Hang on a bit, someone here should be able to help you with the element problem.

pjvdl wrote:



Thanks.

No you didn't miss it

The fire extinguisher that I used has some type of carbon steel body. Wall thickness is approx 1.7mm thick, so hopefully about right. I have welded a perforated stainless backplate on (only because I had plenty of stainless left over from the build).

I am using a re-puposed 12V DC, 2-speed windscreen wiper motor from and old Ford for the drum (picked up from a wreckers for $25). Published speeds for this are 65 rpm and 45 rpm, although by reconfiguring the wiring I can also get 130 rpm. I am currently using 65 rpm, which I have tested with 500g and 1kg bean loads and it seems to be giving the beans plenty of air, so hopefully should be about right. I also has plenty of torque, even when beans get a little caught up against the face of the roaster. If the speed isn't quite right, might I might need to take advantage of the Arduino TC4+'s phase angle control to tune the speed, but I am not going to go there unless I need to.

allenb wrote:
You mentioned possibly mounting the element in free air (not within a tube). This will not provide the needed amount of heat transfer from element to air unfortunately and will cause the nichrome to exceed the maximum rated wire temp.

allenb wrote:
It appears from what you stated that you are not using the same blower that was originally attached to the element within the heatgun? Shoot us some photos showing the arrangement of the current blower and how it connects to the element tube. This will typically only happen when there is insufficient air flow.

allenb wrote:

One way to provide you with needed feedback from the heater is to monitor the air temp leaving the end of the tube. As long as the voltage feeding the element is no higher than what it received on highest setting of the heatgun and you never exceed the highest discharge temperature claimed by the heatgun manufacturer then all should be good.
Typical causes for element meltdown are

insufficient airflow

uneven flow

lack of sufficient and uniform gap between element coils and tube (or mica insulator material). If an element is not centered within the tube, it can reduce the gap to where very little air is passing across a portion of the element and while most of the element is ok, part of it is going beyond the maximum rated temperature.

Exceeding watt density by using higher than design voltage

Short to ground

danst wrote:

Hi, I believe open nichrome elements are cheap, but more problematic then heating rods.

danst wrote:

The needed air flow for wires can be too high for a drum roaster. I can say, 2x 1800W for 70% can be too low power for 1 kg roaster. Better 4 x 1800W at max 50-60% or 3x for 70% .
I am very curious if this thing goes good and for longer working time. Good luck with it, I hope I am wrong with my conclusions.

danst wrote:

Yes, some chinese roasters have 2.5 kW for 1 kg coffee. I say no profiling roasters. I mean, with so low power it is difficult to really "manage" the roast and to keep profile what you want. The more powered gas roasters are more flexible.
In this type of home made roaster, with a lot of metal, it is a problem not to waste the heat. All the body take it. And the ambient too. And it has a huge influence on the roast.
Probably the solution is less body mass, more insulation, good hot air distribution - and this is not a classic roaster. I think this goes towards the fluid bed concept, like Gene Cafe and others with more convective heating.
With less mass is more power needed - it looses the accumulation of the heat for first stages of the roast. And so we have the problem again. Less power - longer roasts, significantly lower chance on controling RoR.
I like it classic and there is a need of roughly 4-5 kW / 1kg greens for a good responsive roaster. No matter gas or electric.
And this has nothing to do with quality of the result. A roast from underpowered roaster can be very good. The operator must to fit the roaster properties, however.
On the other hand, It can be a Rolls-Royce or Mustang. (DIY)

renatoa wrote:

For the drum approach I would not rely on responsiveness to control something related to profile... I mean based on real time feedback from machine temperatures.
All the pros I know roasting on drum machines are using recipes based massively on machine natural curve and their high inertia, with only 3-4 power change points and nothing else.
Something like: preheat, charge keeping the same power level as in preheat, at TP add 20%, one minute after add another 20%, keep that level until DE and 2 minutes after, lower 10%, then another 10% after one more minute, and that's all. Just an example
What we call profiling is to find the exact moments when to apply these power changes, and how much. Not a computer controlled power change based on BT evolution feedback, because the machine inertia simply don't allow this.
Think to at least 30 seconds lag time for a step response, and you will have a better reason why.
Overpowering will not change this behaviour.
You don't have have tyres and brakes enough to stop an overshot of an overheated machine

Related to cheap machines, here, at the gates of the Orient, predominant are the Turkish brands. same price range as Chinese and several centuries more advanced in coffee roasting experience make their choice obvious.

renatoa wrote:

Beware how to evaluate inertia, the empty machine could make you believe it's enough responsive, but the beans mass adds a lot to heat inertia.
In my TO machine - 6 seconds lag time - after an empty preheat, the (air) temperature drop is 120 C per minute, from 185 C to about 60 C (skin test)
Loaded with beans it's a different story, needs 5 minutes cooling down to same level. After the first minute the temperature is still in the 180-190 C range.
This was just a test I did in the beginnings, to find if cooling inside the roast chamber, as for FB, is an option.

directtraderoasting wrote:

My 10" diameter x 12" drum is around 3mm thick (a little over) and it's pretty responsive, but you still have to be ahead of what the bean is going to do when profiling.

You have to do a few roasts to know how the bean responds then tweak your adjustments based on that bean and how is going to act in certain stages of the roast.

I run 2 temp probes. One in the bean mass and one near the opposite side up by the bean loading chute (hot air in the drum).

I also have an analog thermometer in my bean loading chute and chaff collector (mainly to keep an eye on things/backup).

Edit: if I ever build another 2kg/2.5kg, I may go a little thicker on the drum wall. The next standard option I can easily order is a .250" wall pipe, which is 6mm. I think that may be better overall. I do like that the 3mm drum is quick to heat up and cool down though and it's responsive.

I think the slightly thicker wall might be easier though, just make the adjustments earlier

renatoa wrote:

For the drum approach I would not rely on responsiveness to control something related to profile... I mean based on real time feedback from machine temperatures.
All the pros I know roasting on drum machines are using recipes based massively on machine natural curve and their high inertia, with only 3-4 power change points and nothing else....

pjvdl wrote:


I am interested in your suggestion of heating rods. What do you mean by these?

Thanks for the clarification. Yes, I understand. I don't want to hijack Cyrus's thread, so might be better to respond in the other thread I started, but I had hoped to use nichrome to provide a blend of convective and radiant on the drum. Unfortunately my testing with nichrome is showing this is probably not practical with the forced air requirements of the nichrome. Might need to look at tubular elements or gas heating if I wnat more radiant on the drum. I am also considering infrared halogen elements to heat the drum (not directly on the beans).

Do you have a tubular element that you would recommend? I am also wondering how best to control a tubular element using a TC4/Adruino combination. . Do you know if slow PWM still works, or do I need to look at ICC or phase angle control?