Who is here? 1 guest(s)
 Print Thread
Skywalker roaster mods
billsey
I'm going to have to test, but I assume something like this: https://www.amazo...B085BJRZN2 would also work with Skywalker Spy? And potentially allow the display to be used with normal Skywalker control? I'll check the wiring when they arrive, but I believe all four wires are duplicated on both sides.
Edited by renatoa on 06/02/2024 2:21 AM
 
renatoa
Yes, will work.
And indeed offer an easier way to tap the signals to the board.
Instead surgery in the middle of the cable, you just cut one of the female plugs and solder to Arduino.
Good find ! ThumbsUp

Btw, if interest I can post a new redesigned version of Spy sketch, using interrupts instead actual pulse waiting/blocking code, allowing a more reliable following of the protocol.
We are talking about one packet lost during 10 seconds (2-300 packets) instead actual packet skipping method, that let pass un-captured about 2/3 of packets (12-14 of 20 skipped, per second).
Edited by renatoa on 06/02/2024 8:20 AM
 
billsey
That seems like it would be better... I've got the SkyCommand sketch flashed right now, but haven't yet run any beans through. The splitters arrived today, and once I ohm them out I think I'll flash the spy sketch instead and run my first roast under the stock controls, but monitor using Artisan. Once I'm comfortable with that I'll give full control a try with the splitter in place, so I can see what the stock controller does when Artisan is actually in control. My thought is that unless you try to use both as control at the same time it ought to work...
 
Mike_Mathis

Quote

renatoa wrote:

Btw, if interest I can post a new redesigned version of Spy sketch, using interrupts instead actual pulse waiting/blocking code, allowing a more reliable following of the protocol.
We are talking about one packet lost during 10 seconds (2-300 packets) instead actual packet skipping method, that let pass un-captured about 2/3 of packets (12-14 of 20 skipped, per second).


I am a fan of Skywalker Spy. I find it my preferred way to roast coffee with this roaster. If you decide to post a redesigned version, I would be interested.
Edited by Mike_Mathis on 06/07/2024 8:50 PM
*Kaffelogic Nano 7
*Skywalker V1
*2023 BC-2
 
renatoa

Quote

billsey wrote:

...Once I'm comfortable with that I'll give full control a try with the splitter in place, so I can see what the stock controller does when Artisan is actually in control. My thought is that unless you try to use both as control at the same time it ought to work...


You can't do this with any (known) version of josh code.
Can be done, but requires massive redesign.

Anyway, if you raised the ball, let's brainstorm a bit, everyone is invited for an opinion, how could be operated such contraption, i.e. roaster taking commands from both panel and/or Artisan.
I don't mean the code, but could be the user experience.
For better understanding, will list below what can be done and what not...
- change of control arbitration - can be done. I mean if at a moment in time I have 65% power on control panel display and I move the heater slider in Artisan to 70%, the roaster will get 70%. Later, If I dial in control panel 60%, the roaster will get 60%
- display feedback of actual values - only partial, not both sides. In the example above, when an Artisan command will change the heater to 70%, the control panel display will continue to show 65%, it can't be notiffied that a change happened in the outside world, and other side is in control. Conversely, when the control panel would lower the power to 60%, this can be seen in Artisan, because the actual power value can be captured from panel control packet. This value will show and plot correctly in Artisan.
But this value can't force the Artisan slider to move itself to 60% from 70%, where it was previously moved manually.

Confusing enough ? Grin
Does such a mode of operation, allowing control from both sources, but having actual values ​​only visible in Artisan, make sense?
What could be the advantages of such setup though ?
At first glance I see the fact that you can do preheat and cooling from control panel, easier and safer than from Artisan, mmm... ?
Other thoughts ?

End of rant.
 
MaKoMo
Artisans event quantifiers, introduced in 2014, can be used to have a device channels data move a slider in Artisan. This mechanism has been added, and is in use, by machines like all supported Probat, Giesen and many other machines to synchronize the Artisan sliders with power inputs over their control panels.
See https://artisan-r...fiers.html
Edited by renatoa on 06/08/2024 6:10 AM
 
billsey
That would be the best of both worlds, I think. Artisan monitoring what the panel sends as well as overriding as needed. Could be a bit complex to implement though. For now I'd like to control either by Artisan or the panel, with both the panel and Artisan displaying the status. That allows me to use the panel yet get graphs from Artisan or have Artisan in control and the panel shows status. My guess is that should be good to go without changes...
 
billsey
It turns out the splitter has something internal other than just a connection between the ends, which is problematic when used on the Skywalker, the temperature probe throws an error with in in the circuit. I'm going to do some testing to see if where the issue is, but will likely just have to build one up myself.
 
renatoa
Thank you for the report. Sorry to hear this.
Checked with an ohmmeter if all 4 wires are connected as expected ?
Tried on both sides of the Y ?
 
billsey
I got some USB breakout boards in so could test conductivity. One side of the female connectors only has power and ground hooked up, and the wire itself only has two conductors. I've got more breakout boards on the way so I can wire up a Y cable that has all four lines. It was just a 50:50 chance I'd choose the wrong side when I first fired it up, since when I connect the other side of the Y the Skywalker can see the temperature normally. Now that I look at the detail on the Y cable, the fine print near the bottom says "One USB 2.0 Female is JUST for Power Charging". :(
 
renatoa
Voila... the fine print...
 
billsey
OK, I've created a Y adapter myself, it's nice and compact and seems to work as expected:
Dang it, I tried to attach images and it says the .jpg files aren't allowed, even though they say they are in the attachment dialog. :( Drag and drop also doesn't work.
 
renatoa
This post from FAQ could help: https://homeroast...?cat_id=16
 
billsey
Yeah, that's pretty much what I tried. I get to the Attachment; Browse button, select a file (such as PXL_20240626_201552585.jpg, 354KB), the attachment is registered as C:\fakepath\PXL_20240626_201552585.jpg and when I select the Post Reply button it returns a attachment filetype not allowed, even though .jpg files are in the list of supported file.
 
billsey

Quote

billsey wrote:

Yeah, that's pretty much what I tried. I get to the Attachment; Browse button, select a file (such as PXL_20240626_201552585.jpg, 354KB), the attachment is registered as C:\fakepath\PXL_20240626_201552585.jpg and when I select the Post Reply button it returns a attachment filetype not allowed, even though .jpg files are in the list of supported file.


OK, let me try again, I reduced the file size to closer to 100KB each. Same error, but one of the files actually worked that time. Weird...
billsey attached the following image:
pxl_20240626_201552585.jpg
 
renatoa
Nice approach... but my personal disaster scenarist whispers to me that the weight of the cables, pulling down the male plug, could led to something undesirable... in time... Shock
 
billsey
My thought is that if others like the compact design, someone will do a 3D printed case to hold it, and perhaps the Leonardo, against the roaster to alleviate the potential for the weight causing torque on the connector. By itself it is really low weight, the cable they are using to go to the display is also both light weight and flexible, if you choose a cable to go to the Arduino that is also flexible and light there should be minimal issues even without a case. If I had mad CAD skillz I'd do a case for myself... smile
 
Dan N
There's no reason to use a big Leonardo unless you already have one, the smaller Nano works just fine. I am intrigued by the Pro Micro which is even smaller than the Nano but has the exact same chip as the Leonardo. That's the one I'd make a case for.

EDIT: I do NOT recommend using the Pro Micro. Its power connections have subtle differences from the other Arduinos that adds complexity and isn't beginner friendly.
Edited by Dan N on 07/06/2024 8:48 AM
 
renatoa
Nano has two chips, the main controller and the usb chip.
The Micro has both in the single ATmega32U4.
Because of that the Micro has a larger boot sector and needs larger libraries for the Serial communication. Thus there is less memory for the sketch.
But not enough less to impact loading an even larger Skywalker code.
 
billsey
I think I've got a nano around here. With any of the small ones like those you wouldn't need the second USB-A female port, just wire directly in the adapter and mount them all together. You would have a little box that holds the Arduino and both male and female USB-A ports. You plug that into the back of the Skywalker and plug the display into that, then plug your laptop into the port on the Arduino. Clean...
It looks like a Pro Micro clone with USB-C runs under $10. How about I make you a deal? I'll build a USB-A pass through with four header pins available and include a Pro Micro card, send it to you, and you can design a box for it. smile
 
Tomoaki Nakamura

Quote

sloppyjosh wrote:

The readings from the roaster are inversely proportional to the temps but I haven't put a meter directly on the probe.

I used a 4th degree polynomial and got it working pretty well. I ran a preheat cycle and it the values output by my Arduino seemed in line with the controller. The controller applies some smoothing so it's kind of difficult to get a direct 1:1 agreement but it looked really good.

I just need write the code to indicate the status of the fan and heater, put it in the correct format for Artisan and give it a go.


If you write temp vs time curve, temp is over smooth even when R and V changes. This is because Skywalker already applied rolling average on the raw temp. It is not a raw temp and quite low compared to raw temp.

I installed RTD ET and BT and compared to Skywalker temp calculated by your equation. Skywalker temp is 1C off(low) in room temp already. Switchbot room temp and RTD ET/BT matches within 0.2C.
 
renatoa
When talking about physical units measurement, whatever of them, we have qualifiers as precision, resolution and linearity.
Related, but different meanings. Not all of them are important for whatever unit or measurement scenario.

Such precision is not necessary for roasting coffee, even 10C off is allowed.
That's because the rolling bed of beans means a measurement environment dominated by noise.
Not acoustic or electric noise, but measurement/environmental noise, that is even harder to reproduce and control, or minimize.

In the great industry, for similar scenarios, the approach of measuring something inside process has been abandoned, because is simply not relevant, nor consistent.
For example a limestone furnace, where the pebbles can be equated to beans.
Is far more accurate to measure input heat, output and environmental temperature, and compute the roasting progress using a mathematical model, having known all physical constants.
Instead trying to measure pebbles temperature using whatever sensors...
Please note also that limestone cooking does not help us with a milestone event, as the first crack Grin

What we want from a measurement system for coffee roasting is good linearity and consistency required for watching the progress in the most critical area: the development.
This could happens at 180C for radiant roaster, or near 200C for for hot air roasters... and both are right, even if only one of them is true.
What is important is to have a consistent monitoring of evolution of progress for the next 8-16 degrees during development.
Because this value is a differential of temperatures, it will be at least 10x more precise than the temperature values it is computed.
Edited by renatoa on 07/02/2024 3:22 AM
 
Tomoaki Nakamura

Quote

billsey wrote:

Quote

billsey wrote:

Yeah, that's pretty much what I tried. I get to the Attachment; Browse button, select a file (such as PXL_20240626_201552585.jpg, 354KB), the attachment is registered as C:\fakepath\PXL_20240626_201552585.jpg and when I select the Post Reply button it returns a attachment filetype not allowed, even though .jpg files are in the list of supported file.


OK, let me try again, I reduced the file size to closer to 100KB each. Same error, but one of the files actually worked that time. Weird...



Aha, That is exactly what I am trying to make now but using one male and two female usb 2.0 connectors as a cable. I am encouraged to know someone else is doing the same.
 
billsey

Quote

Tomoaki Nakamura wrote:
Aha, That is exactly what I am trying to make now but using one male and two female usb 2.0 connectors as a cable. I am encouraged to know someone else is doing the same.

If we get a box designed and printed, it will have a USB-A male sticking out one side and sockets on the other side with a USB-A female and a USB-C female. That will sit flush with the rear of the roaster and you will plug the display into the USB-A. If you connect your computer to the USB-C and run Artisan you should be able to control with either and monitor with either, or both. That should be very clean looking...
 
billsey
So, I'm looking at wiring up to the Pro Micro... Connecting to D2 and D3 seems simple enough as does GND. Is Vin really needed? There's a Vcc on the Arduino, but no Vin. Since the Pro Micro is likely running at 3.3V I don't want to connect 5V to it. My thought is that voltage is only used to power the display, and not needed for the data transfers. If so we should be able to use just the three wires...
 
Jump to Forum: