topbanner.gif
Login
Username

Password




Not a member yet?
Click here to register.

Forgotten your password?
Request a new one here.
Shoutbox
You must login to post a message.

11/26/2020 6:37 AM
Everyone have a super Thanksgiving and may all your roasts turn out stellar woohoo

11/25/2020 4:20 PM
Howdy 1st CH and welcome to HR forum! Be sure to read in the forums to see if anything covers the topic and please post away in drum roaster forums Welcome

11/25/2020 3:35 PM
Hey everyone! I am new to the home roasters forum. Mainly here to read and learn! I’d also love to get advice about modifying my roaster (a Huky 500)

11/10/2020 5:29 PM
Welcome MJ, post away! Cheers

11/08/2020 6:05 PM
Hello All, New here, started with an air popper, now using fresh roast 800. I'm here to learn and connect with others Thank you

Users Online
Guests Online: 9

Members Online: 0

Total Members: 6,835
Newest Member: sangrai kopi
In Memory Of Ginny
Donations

Latest Donations
Anonymous - 5.00
Anonymous - 5.00
renatoa - 2.00
JitterzZ - 2.01
renatoa - 2.00

View Thread

Who is here? 1 guest(s)
 Print Thread
Cyclops - my 1kg drum build
allenb
The chart doesn't convey what I'm looking for. What is the temperature leaving the elements at a fan power level you expect to use to allow removal of chaff and smoke and at 90% or somewhere around there power level to the element. The chart is showing a maximum of 161C (321F) discharge temperature which will not get you where you want to go. Or, am I missing something?
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
pjvdl

Quote

allenb wrote:

The chart doesn't convey what I'm looking for. What is the temperature leaving the elements at a fan power level you expect to use to allow removal of chaff and smoke and at 90% or somewhere around there power level to the element. The chart is showing a maximum of 161C (321F) discharge temperature which will not get you where you want to go. Or, am I missing something?


No. You are not missing anything. I agree that it doesn't show what I would expect either.

The temperatures at the exit are definitely lower than what I would expect. It is unclear why this is the case. As an aside, this is why the table is incomplete, as I realised that I was not getting the results that I would expect (and I blew an element at 100%/100%).

I have a couple of theories as to why this may be. Either there is an issue with the responsiveness of the K-type thermocouple that I am using to take the measurements. Or, the airflow is too high at high fan %, causing significant variations in the temperature. I did notice if II moved location of the probe, the temerature reading did vary significantly, supporting this theory. I am not an expert in fluid dynamics, but I think the data also supports this, as you will notice the measured temperature at 80%/80% and 90%/90% is lower than 60%/60%, even though the colour indicates it should be hotter. I would be interested in other's thoughts on this.

Because I didn't trust the absolute temperature readings, I was guided more by the temperatures relative to each other, building up a picture of how the temperature curves develop as a factor of heat and fan.

I was also guided by the colour of the elements at their hottest point, which I did also record, although I neglected to include these in the table that I attached. I have now updated the table with the colour, so this might give a better indication (attached below). You will see that the colour at 100% fan / 100% heat is yellow. This is when I blew an element, ending my test :( On this basis, I am assuming that my "Orange/Yellow" is as high as I can go.

I did also run a couple of tests the next day with the tip of the thermocouple inserted into the end of the pipe as close as I could get to the live element without actually touching it - probably 5mm from the end of the live element coil. At this point I was measuring around 540C at 90% when the element was orange/yellow, which is more in line with what I would expect.

Chaff and smoke are removed when the fan is at least 50%, so I have some room to move in this regard.

As I mentioned, this settup is heating my roasting chamber to 250C in around 6 minutes, so clearly the temperature readings that I have recorded are incorrect for some reason.
pjvdl attached the following image:
screenshot_from_2020-07-30_09-07-06resized.png

Edited by pjvdl on 07/29/2020 6:58 PM
-- Paul

ECM Rocket espresso
Eureka Mignon grinder
500g BBQ rotisserie roaster
1kg drum roaster
pjvdl

Quote

renatoa wrote:

You should use 1 Hz PWM for at least 2% (1.67% on 60Hz) control resolution... 8Hz will degrade significantly your control precision.


I think I understand why you would get a 1.67% control resolution at 60Hz, but are you sure you would get a 2% resolution at 1Hz? To be fair, I struggle to get my head around the ATMega bit control theory of PWM, but from my testing, 1Hz PWM on a Arduino/TC4 setup switches the power on/off at a maximum frequency of 1Hz. To achieve a 2% control resolution you would need to include 50 cycles in your control cycle, which at 1Hz is 50 seconds for one cycle through. Or am I missing something here?

From user.h file in the aArtisanQ_PID arduino shield (you will see that I have 8Hz uncommented):

////////////////////
// Time Base for slow PWM on OT1, OT2
// When NOT using PHASE_ANGLE_CONTROL option
// choose one of the following for the PWM time base for heater output on OT1 or OT2
//#define TIME_BASE pwmN4sec // recommended for Hottop D which has mechanical relay
//#define TIME_BASE pwmN2sec
//#define TIME_BASE pwmN1Hz // recommended for most electric heaters controlled by standard SSR
//#define TIME_BASE pwmN2Hz
// #define TIME_BASE pwmN4Hz
#define TIME_BASE pwmN8Hz
// The faster frequencies below are for advanced users only, and will require changes to the PWM16 Library
//#define TIME_BASE 15 // approx. 977 Hz
//#define TIME_BASE 7 // approx. 1.95kHz
//#define TIME_BASE 6 // approx. 2.2kHz
//#define TIME_BASE 3 // approx. 3.9kHz


Quote

renatoa wrote:
Are you using TC4 ? why not going with ICC ?


TBH, I started using PWM because I was initially learning how to work with an Arduino/TC4 and that is the default. I have been meaning to look at the alternative control mechanism, but haven't had the time. I think ICC is also dependent on the presence of a zero cross detector, which I have had a quick look for, but have not managed to source yet. This is definitely something that I want to try.
Edited by pjvdl on 07/29/2020 7:28 PM
-- Paul

ECM Rocket espresso
Eureka Mignon grinder
500g BBQ rotisserie roaster
1kg drum roaster
renatoa
When using PWM, the SSR command that comes from TC4 could be of an infinite degree of precision, but this not matter because is not sync'ed with the mains frequency, so you can chop (or miss) a random part of the sine.
As a result, when you dial for example 50% heater, and using 1 Hz cycle, the output could be anywhere in the 49-51 sine waves passed through the SSR.
From here, the 2% precision (for 50 Hz), and 1.67% = 1/60, for US mains.

I found this fact following this intriguing statement that you can read in any PID leaflet "...set the control period for SSR outputs to 2 seconds...".
After some digging, the explanation was been clear: this cycle time is required if you want 1% precision of control... elementary... if you see on screen 1...100%, then you want to feel on heater the difference down to 1% step.

For 8Hz PWM, for each PWM pulse you have practically only 50/8 = 6 sines available to control. So you have only 7 levels of power control:
0, 1/6 = 16.7%, 33%, 50%, 67%, 83% and 100%
This is dramatic decimation of control choices, that could led unusable my roaster for example... where 80% of roast time I play with figures between 72 and 77%, sometimes 1% in a minute change making a difference in final roast !
1% of power is 3 C degrees in hot air temperature, how much this matters depend on your specific machine inertia, quite critical for FB, acceptable for heavy drums.

With ICC the power modulation is done by passing or blocking the entire half-sine, (thus the need for a zero cross detection, as you well guessed)
So if you you have 100 half pulses for 50 Hz mains, then you can have 1% resolution/precision, and this is true instantly, in any moment.

Another advantages of using ICC:
- don't exhibits transients, because the switching is always near zero of the sine, so less or no perturbations in the sensible house equipment.
- the heater is less stressed with temperature swing, as for on-off
- for halogen heaters, less annoying pulsing light
danst
I see no problem with precision degree of 1 Hz with SSR and it is well suited for resistance heaters and their longevity. For heating control of electric drum roaster is 5% or even 10% scale of power enough.

The nichrome heater is usable (for classic drum roaster) only when it holds all the power without airflow. It must be so designed, not as these airflow-heatings from heatguns etc. (this means several power units in series and parallel or one with thick wire, but long for reducing power- amperage and load, all this results in bigger space requirments)
In a classic roaster, it must be possible to go on max even without forced airflow, above all in drying phase.
With common kind of nichrome (what you are using) is the airflow too high for good power usage.
The idea with inductive heating is for me a bit risky. Not easy for DIY.
Proved solution are tubular elements. It must be space for it, just like for gas heating.
I believe, it is a good idea to copy professional roasters.
1000g Ugly roaster, QM Alexia EVO, Bezzera BZ09, Niche Zero, Fiorenzato Doge 63, Mazzer Jolly(Mestre), Handground, T. Moccamaster, Aeropress.
allenb
Something to consider with resistive wire forced air convection heaters as you are using. If engineered correctly and established heater watt density is not exceeded, you will be able to increase cfm to a point where the element will show little to no visible glow in a normally lit room at design full power. If you are not exceeding the design voltage then you're air flow is insufficient at full power. I can take any 1500 watt nichrome forced air convection heater and flow slightly more than optimum designed cfm and end up with no visible color.

Danst brings up a good point about being able to operate during the drying phase with little to no non-recirculating air flow which allows a humid environment during the drying phase. Studies have been done, but don't ask me by whom, that show a more even and beneficial heat transfer during the first few minutes when steam is allowed to surround the beans while being driven from them. This obviously only applies to mechanically agitated roasters and recirculating fluidbeds.
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
pjvdl
After a bit of a hiatus due to other commitments, I have had some time to revisit the heater in my 1kg drum build. My preference based on the input of members of this group was a tubular element, but the manufactures that I contacted in Australia were not able to build one to the specs I require in the space available (approx 70mm high x 200mm wide x 250mm deep).

Without this as an option, I decided to design an open-coil resistive wire heater. My calculations showed I needed a gauge of around 15 AWG. I have not been able to source Nichrome 80 in any gauge approaching this, but I could get Kanthal A1, which isn't ideal, but I figured I'd give it a try.

Design parameters:

Material: Kanthal A1
Gauge: 15 AWG
Max Voltage: 156V (65% of 240V)
Resistance: 15ohms
Power density: 2.09 W/cm2 (13.5W/inch2)

Winding my own open coil elements, I have mounted two of these in parallel on a S/S and mica frame. The fabrication is a little rough, but it does for a test while I wait for ceramics to fabricate a permenant solution.

My tests showthat I can operate this element at full power with no air flow. Using it, I have roasted 1kg of beans to first crack in around 7 minutes as a first test, so plenty of power.

This needs some more testing, but looks like it should be a good solution.
pjvdl attached the following images:
20201022_154632-small.jpg 20200912_115052.jpg 20200912_115041.jpg

Edited by pjvdl on 10/25/2020 7:41 AM
-- Paul

ECM Rocket espresso
Eureka Mignon grinder
500g BBQ rotisserie roaster
1kg drum roaster
allenb
Wow, I can see you spent more than a couple of hours designing mica standoff plates and path of all the connections! That had to be some major work but looks good.

A couple of questions. Will there be mica sheet beneath the Kanthal in case of melt down and element dropping? would be a good safety measure. What is the air flow path now? Is it able to access most of the coils before entering the rear of the drum?

Great choice going with the Kanthal as it is reported to be very robust.
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
pjvdl

Quote

allenb wrote:

Wow, I can see you spent more than a couple of hours designing mica standoff plates and path of all the connections! That had to be some major work but looks good.

A couple of questions. Will there be mica sheet beneath the Kanthal in case of melt down and element dropping? would be a good safety measure. What is the air flow path now? Is it able to access most of the coils before entering the rear of the drum?

Great choice going with the Kanthal as it is reported to be very robust.


Thanks Allen.

I agree that mica beneath the element would be a good idea. I have not done this at this stage, because I have limited mica supply and want to reserve it for my final element build (using ceramic insulators), and I can only get mica sheets from aliexpress - which takes around 2 months to get to Australia! My testing has also shown that with the 80mm spaced supports, my 15 AWG element doesn't droop more than a few mm and I have a 10mm gap beneath it.

I am still working on airflow. I previously had a baffle between the top of the element and the underside of the drum to force air across the element. With the current element design I have very little space though, so have removed it for the moment. This doesn't seem to be causing a problem though, as I am running my fan (max 160 cfm - overpowered I know!) at around 10 - 20% for the majority of the roast.

For those that are interested, I have included a link to the spreadsheet that I used from my heater element calculations below:

Resistive heater element design spreadsheet for Nichrome 80 and Kanthal A1 wire

https://docs.google.com/spreadsheets/...sp=sharing
Edited by pjvdl on 10/25/2020 6:31 PM
-- Paul

ECM Rocket espresso
Eureka Mignon grinder
500g BBQ rotisserie roaster
1kg drum roaster
allenb
Thanks for sharing the link for calcs for the resistive element. One nice outcome of your current design is the heavy gauge of the wire which would be quite a feat to melt it and cause an issue where it could drop against the roaster body. With proper grounding, it would pop the circuit breaker immediately.

So, does the air path go from an opening at the rear and blow across the coils towards the front of the roaster and then work it's obvious path around and back to the drum openings? This would be optimum.
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
pjvdl

Quote

allenb wrote:
So, does the air path go from an opening at the rear and blow across the coils towards the front of the roaster and then work it's obvious path around and back to the drum openings? This would be optimum.


Exactly. In the photo, you can just see the opening at the back of the element (partially obscured from view by one of the mica supports)

Out of interest, I am using a fan for push air through, rather than sucking from the cyclone. Seems to be quite effective and I don't need to worry about the fan having to deal with hot air.
-- Paul

ECM Rocket espresso
Eureka Mignon grinder
500g BBQ rotisserie roaster
1kg drum roaster
allenb
The reason most drum roasters are designed with pull through the drum versus push is due to the difficulty of sealing all possible alternate path openings that the pushing source could send the air through which, depending on how many gaps were left during construction could allow a good % of the total fan output to end up bypassing the intended path rather than be strictly routed past heat source and straight to the rear end of the drum and through the drum and out. By pulling at the funnel, there is much less chance of large variations in total cfm across the coffee in the drum. As long as the air from the fan isn't able to make a quick turn up and into the drum's rear perf plate, which would be allowing cool ambient air to be flowing across the beans and subsequently cooling the beans while the drum surface is heating them, then pushing is fine and can be just as effective. OTOH, even a pull design can end up with this problem if the air entry point doesn't allow directing the air across the heat source prior to entering the drum.
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
pjvdl

Quote

allenb wrote:

The reason most drum roasters are designed with pull through the drum versus push is due to the difficulty of sealing all possible alternate path openings that the pushing source could send the air through which, depending on how many gaps were left during construction could allow a good % of the total fan output to end up bypassing the intended path rather than be strictly routed past heat source and straight to the rear end of the drum and through the drum and out. By pulling at the funnel, there is much less chance of large variations in total cfm across the coffee in the drum. As long as the air from the fan isn't able to make a quick turn up and into the drum's rear perf plate, which would be allowing cool ambient air to be flowing across the beans and subsequently cooling the beans while the drum surface is heating them, then pushing is fine and can be just as effective. OTOH, even a pull design can end up with this problem if the air entry point doesn't allow directing the air across the heat source prior to entering the drum.


This is an interesting point, although I think another advantage of using a push method with a DIY build is that it is more obvious where air is leaking, because you can actually feel it being pushed through the gaps. My experience is that sucked air is more difficult to perceive by hand. During my build, I could easily feel where air was following an undesirable path, because I could feel it by hand - particularly if I turned the heater on to heat the air prior. I found a number of gaps using this method, which allowed me to modify my design along the way.

I have been left with a couple of tiny gaps between SS sheets that are bolted together, but expand (and therefore move slightly) when heated. I haven't worried about these yet, but if they become a problem, I plan to create some gaskets with high temperature gasket silicone.

The other reason I decided on a push method was more of an asthetic one. If I went with a pull method, I would have needed to mount the fan on the top of the cyclone, which I personally felt wouldn't look as good. Certainly secondary, but was minor consideration :)
-- Paul

ECM Rocket espresso
Eureka Mignon grinder
500g BBQ rotisserie roaster
1kg drum roaster
renatoa
Don't forget the air temperature in both cases... pull = hot air, more expensive and quicker wear fan.
allenb

Quote

The reason most drum roasters are designed with pull through the drum versus push is due to the difficulty of sealing all possible alternate path openings that the pushing source could send the air through...


I failed to qualify this statement. This obviously only relates to electrically heated roasters or indirectly heated gas fired roasters utilizing a heat exchanger.
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
pjvdl
Ok. So i have spent the last couple of days running the machine with its new heater through its paces. I have now completed a number of roasts:

500g x 2
700g x 1
1000g x 2

and am starting to get a feel for the machine with my new improved heater. Observations:

- Seems to have plenty of power for 1kg
- copes well with 500g as well, although need to turn the power down, otherwise I am getting to FC in under 5 minutes
- I am only using 10 - 20% fan through most of the roast, so should be a lower percentage of the heat through convection, with most through conduction/indirect IR

I can't say I've noticed much difference in terms of flavour yet, but I am certainly no cupping expert, and my tests to date are by no means controlled! Time will tell as I better understand the nuances of the machine I think.

I have also started playing with PID control (Artisan + Arduino + TC4+), with only limited success tuning the control loop so far. I need to get my head around this through more trial and error, but the slow response time of the machine makes it difficult to create a control loop that is responsive, but doesn't oscillate. I started with the default values in Artisan, but these were terrible for my machine, which has a response time of around 30 seconds (measured from time when the ET peaks to the time when BT peaks). Anyway, more time is needed on this ...
-- Paul

ECM Rocket espresso
Eureka Mignon grinder
500g BBQ rotisserie roaster
1kg drum roaster
Jump to Forum:

Similar Threads

Thread Forum Replies Last Post
Roaster build forced air and agitator Building a Coffee Roaster 3 11/25/2020 10:53 PM
My 3kg propane fluid bed roaster build Fluidbed Roaster 9 11/20/2020 11:52 AM
Dereks 1-2Kg drum roaster build Drum Roasters 115 11/17/2020 4:37 AM
KKTO build questions***FOR SALE*** Turbo Oven Roasters 118 11/02/2020 12:28 PM
Fluid Bed Roaster Build - bye bye Heat Gun and Flour Sifter Fluidbed Roaster 161 10/25/2020 2:36 PM
Homeroasters Association Logo, and all Content, Images, and Icons © 2005-2016 Homeroasters Association - Logos are the property of their respective owners.
Powered by PHP-Fusion copyright © 2002 - 2020 by Nick Jones.
Released as free software without warranties under GNU Affero GPL v3.
Designed with by NetriX