Homeroasters - Home Roasting Coffee Community » ALL ABOUT ROASTERS » Other Roasters
Who is here? 1 guest(s)
Page 16 of 16: 1...13141516
 Print Thread
Skywalker roaster mods
Bill grubbe
Top  #376 Print Post
Posted on 07/26/2024 8:23 PM
I just plugged in my computer into the USB port to start use artisan and the transformer blow on the roaster now everything works put the 240 v part of the roaster so I put in a new transformer and it blow to I think I have a low voltage short somewhere anybody has any thoughts about this it would be very appreciated
 
Dan N
Top  #377 Print Post
Posted on 07/27/2024 7:39 AM

Quote

Robotic Kitten wrote:

BTW, decided to build me a replacement board, based on F412 MCU.
Doh, why it wouldn't let me attach a .jpg or .png
https://1drv.ms/i...w?e=sjmXzZ

That is VERY impressive. Any feature upgrades (thermocouple support, etc)? How's the software progress?
 
renatoa
Top  #378 Print Post
Posted on 07/27/2024 12:43 PM

Quote

Bill grubbe wrote:

I just plugged in my computer into the USB port to start use artisan and the transformer blow on the roaster now everything works put the 240 v part of the roaster so I put in a new transformer and it blow to I think I have a low voltage short somewhere anybody has any thoughts about this it would be very appreciated


I would check the cable connections/order, for roaster and also the PC.
Already had bad experience with cables having wrong wire order.
 
billsey
Top  #379 Print Post
Posted on 07/28/2024 8:27 PM
Was the USB port on the computer plugged into the 'notUSB' port on the roaster, or did you plug it into the USB port on an Arduino running the control software for Artisan to use? Either way I can't see anything obvious that would blow a transformer on the roaster. If anything got blown from plugging it in it would be on the power that provided through that port, which is only 5V, or on the computer side of things.
 
renatoa
Top  #380 Print Post
Posted on 07/29/2024 1:29 AM
A wrong made cable can carry the mains line to the PC ground, and from there, via USB shield, into the roaster case.
 
HarryDog
Top  #381 Print Post
Posted on 08/05/2024 6:18 PM
Hello renatoa, I finally came back to implementing your Spy Temp Filter code into Skycommand. It looks to be running and the FILT: command in IDE reports changes.
Just some basic tests in Artisan all the Lamp, Fan, roaster look to work.
No roast yet but the temp looks to be smooth as I was testing 80% filter and 2 Second sample rate. Not sure how fast I should push this?

Anything else I should look at or test as time has gone on?
 
renatoa
Top  #382 Print Post
Posted on 08/06/2024 2:42 AM
It's ok as you set, but you can lower to 70%, for less delay of the reading.
80% is 4 readings delay, 70% is 2 readings.
Checking "Oversampling" is another way of filtering, equivalent to 50% filter, and available to any people, any measurement system.

Quote

Anything else I should look at or test as time has gone on?

The empty machine drive test, as I posted previous page?
 
HarryDog
Top  #383 Print Post
Posted on 08/06/2024 4:46 AM
Thanks will do some more testing later today.

I have also been playing with two Max chips, the Max 31855 K only type module and a probe I tested out side of the roaster worked fine but the one I have inside the roaster that was working with the Phidget does not work for some reason.

Thinking of using the IDE example with fault detection to see if I get some reason why its not working correctly? It does work if I plug it into my hand meter as well.
 
renatoa
Top  #384 Print Post
Posted on 08/06/2024 12:37 PM
What is the purpose of replacing actual temperature system with another performing at most on par ?
MAX: 0.25C resolution, 4 samples per second
NTC: 0.2C resolution, 9 samples per second.
 
HarryDog
Top  #385 Print Post
Posted on 08/06/2024 1:46 PM
Thanks, I'm not sure I ever read any specs on the original code but if using FILT 70% at 1 second sampling what resolution is the NTC at?

I think I see the issue with MAX31855K, once the Skywalker is powered on I get a short to ground fault on the probe. Is this just an issue with a grounded probe? I tried to ground the other probe to the same plate and it still worked. I have a number of K thermocouples that were purchased all over the place so no idea if grounded or not but assume most are? I should be able to test the continuity from the leads to the outer sheath.

At some point I would like to make a 1kg roaster of some type but unsure if I will get there or not.
 
renatoa
Top  #386 Print Post
Posted on 08/07/2024 4:58 AM
Filtering does not affect resolution (too much) For more resolution you need oversampling.
The main function for which I added the filtering was to smooth the spikes, rather than to increase the resolution.

Actually, for Skywalker the resolution is a very fuzzy concept, due to the non linear scale of the sensor.
The canonical approach for a linear ADC is to divide measurement range by the values count.
Example: 250C measurement range, 10 bit ADC => 1024 values, resulting 0.25C degree per ADC step.

But Skywalker is using only partial the whole ADC range, for two distinct measurement ranges, as explained the previous page post.
Add to this design quirk the logarithmic nature of the NTC curve, and you have different resolutions along the whole range.
Analysis performed on the data gathered by josh, that was used to build the interpolation formula(s), reveals that we have at least 6 ADC values per C degree (i.e. 0,17C per ADC step resolution) in the range 30-180C, decreasing to 5 ADC values per C degree (i.e. 0,2C per ADC step) in the range 180-210C.

Two examples:
- the range from 145 to 155C, including 150C value, used by many people as the standard dry end, is encoded by the ADC values range from 610 to 671, thus 61 steps.
- the range from 175 to 195C, the most important for FC and development phase, is encoded ADC values from 379 to 490, thus 111 steps, for 20C range.
Edited by renatoa on 09/11/2024 12:53 PM
 
HarryDog
Top  #387 Print Post
Posted on 09/11/2024 12:46 PM

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).


Yes I would like to look at your interrupt code, would like to try and implement it into the code changes I have done so far.
cross fingers
 
billsey
Top  #388 Print Post
Posted on 10/15/2024 6:02 PM
OK, I give up and need some hand holding... When I have SkywalkerSpy 4 loaded in my Arduino and launch Artisan I don't see temperatures, though everything else seems to be normal. I can run a roast using the SW display. When I flash SkyCommand instead, I see the numbers fine but the SW display goes to E3 and I have no control with it, only with Artisan, and I'm not comfortable enough with using Artisan to get anything like as nice of a roast as the automatic mode in the SW. What can I do to diagnose the issue?
 
HarryDog
Top  #389 Print Post
Posted on 10/15/2024 9:54 PM
Well sounds like you are close, check the connections to make sure they look good.
If using a windows system make sure the Com port is set the same as Artisan.

I have used some of the code but have not used the full Spy4 code.
The Skycommand is not designed to work with the SW control panel.

Check out post #196 make sure you set the TC4 5/6 virtual port for fan and heater.
Com port Speed in program is
Serial.begin(115200)
Edited by HarryDog on 10/16/2024 10:09 AM
 
renatoa
Top  #390 Print Post
Posted on 10/16/2024 2:24 AM
Always when in trouble with a new device on Artisan, check Help/Messages, and try a Help/Serial log. Posts the results here for a diagnose.
Please, use text copy-paste. not images, i.e. screen captures.

If I can help assist you via an AnyDesk session... let me know with a PM.
 
renatoa
Top  #391 Print Post
Posted on 11/13/2024 6:15 AM
Do you think that this roaster has exhausted its possibilities for improvement ?
Think again Grin
Sure... anyone is now trying to find everything about the new V2... but not me.

One of the very first initial issues of Skywalker is now gone, I've solved that too... no longer requires anything additional external cooling.
Even more, this mod is so trivial... doesn't have to glue/cut/drilled anything... just remove the back panel and ... but, let's first hear how we arrived here...

Starting from the observation that the external cooling on a fan as we used to do until now, is done with an ascending air current, which is actually a forcing of the natural tendency of the warm air mass to take it up... and that if I put the same fan in reverse, trying to convince the heat to go down, through the perforations of the tray, against natural convection flow, that is, it will take more time than if it blows in the same tray from the bottom up... so I had a flash and reversed the cooling fans in the back, to let air in instead of out!

Nothing else! just unscrewed, and turned them upside down, with the label inside!

In 2 minutes we now have cold beans, in our own skywalker tray, without having to outsource this function.

With a little clarification though!!! The tray must NOT be closed after the drop!!!
Leave it open by 1-2 cm, so that the warm air has somewhere to escape.
If you close it completely, put your hand on the back and you will feel how the air is pushed from the fans to the sides, instead of going inside to cool the beans.

An observation for those who have never opened the back cover: there are many screws, some hidden in those fan protection cages, and its removal will take longer than reversing the fans themselves.

Pictures needed ? Grin
 
billsey
Top  #392 Print Post
Posted on 11/17/2024 1:22 PM
Yes, pictures would help a lot with the steps needed to remove the back cover, less so with the actual fan mod, since that is pretty well explained already.
 
renatoa
Top  #393 Print Post
Posted on 11/17/2024 3:26 PM
Fortunately, there is a maintenance video that includes the back panel removal.
Check here, starting with 5:05



The cooling fans are visible at 6:42
 
renatoa
Top  #394 Print Post
Posted on 11/17/2024 3:32 PM
After removing back panel, the cooling fans remains attached by only two screws, highlighted in the red ellipse.
Unscrew those, and flip the fans the way they are looking in the second picture, i.e. label inside.
A slightly different routing of the wires might be necessary.
~~~
renatoa attached the following images:
skw_cooling_mod.jpeg image_2024-11-17_233053618.png
Edited by renatoa on 11/18/2024 5:23 AM
 
Page 16 of 16: 1...13141516
Jump to Forum: