RoastLogger with USB & webcam input

[GreenBean]

The first batch of the RL-HT-CTRL boards have now all been taken. Two are awaiting installation, all the others are installed and working well.

For further information on these boards see the RL-HT-CTRL webpage and the relevant posts in this thread dated between 23 July and 21 August 2013.

I am not planning to make another batch of boards in the near future as it is hoped that JimG will be back in production of the HTC/HTShield boards in the next few months. If, in a few months time, Jim is still not back in production then I will consider making another batch of boards but that would only be worthwhile if at least 4 people had committed to purchase a board so I would ask anyone interested to register their interest by emailing me at the email address included in the feedback section of the RoastLogger website. If Jim is back in production in these few months I will pass the details of those interested on to him.

[GreenBean]

I have been asked if I have plans to add support for the Phidget 1048 temperature sensor to the RoastLogger. I have had a quick look at the Phidget and have decided not to support it at this time as I consider that it does not compare well to a TC4/C on functionality, flexibility or cost. I understand that the TC4 and TC4C are not currently available but, hopefully, they will be back in production in the coming months.

My thoughts on a comparison of the Phidget and TC4/C are that both devices provide temperature readings from up to four thermocouples to similar accuracy. The only advantages the Phidget has is a true USB interface but the advantages of this are somewhat negated by the fact that it still requires a separate driver installed before it will work*. The advantages of the TC4 and TC4C are that they include various interfaces to allow a user to add control to their roaster and to use a standalone LCD display. It will not be practical to add control or LCD interfaces to a Phidget so anyone using these devices who later decides they want this additional functionality would have to purchase a TC4/C to replace the Phidget.

Other advantages of the TC4 and TC4C are that they provide excellent rate of rise data and the firmware is open source allowing it to be customised to meet individual user requirements. The TC4 and TC4C are also considerably cheaper. The price of the Phidget is $100 in the USA but costs £96 ($155) in the UK. JimG sells the TC4 for $59** (even less in kit form) and a TC4C for $89.

* Compare this to the true USB interface in the RL-HT-CTRL board that requires no drivers in any operating system.
** The TC4 requires a separate Arduino, these are available from about £10.

[GreenBean]

A further minor revision to the RoastLogger to version 2.0.3 is available. This revision only involves changes to the template tab to make comparing all the details of any two previous roasts easier. Previously any two roasts could be loaded and compared in the chart which shows the roast profiles and key roast events. Now, in addition to this, the full template log is loaded in the template tab to allow comparison of all the log details with the full current roast log loaded in the log tab.

As with all previous versions of the RoastLogger, any data the user wishes can be entered into the logs and the template for a new log can be edited, as the user wishes, to structure that data.

New template tab layout:


For further information or to download see the following version information and the RoastLogger website.

Version 2.0.3 22/11/2013

Made comparing all the details of any two previous roast logs easier by modifying the template tab. Previously any two roasts could be loaded and compared in the chart which showed the roast profiles and key roast events. Now, in addition to this, all the data in both roast logs is loaded to allow direct comparison of all the details recorded in the logs. The roast log loaded as the template is displayed in the template tab and the roast log loaded as the current roast is displayed in the log tab.

When any previous log is loaded as a template or current roast the cursor is now set to the start of the log. Previously it was set to the end.

[GreenBean]

A further very minor revision to the RoastLogger to version 2.0.4 is available. This revision only involves the logic used to determine when to automatically adjust the roast phase bands in the chart to set the maillard/development transition to the last first crack start temperature.

Previously if users were testing the RoastLogger whilst not performing a real roast they could click the first crack start button at at any temperature and the roast phase bands would adjust accordingly. Now automatic adjustment only takes place if the first crack start temperature is between the drying/maillard and the development end temperatures. This will always be the case in a real roast and avoids adjusting the bands to impossible first crack start temperatures.

For further information or to download see the following version information and the RoastLogger website.

Version 2.0.4 4/12/2013

This revision only involves the automatic adjustment of the roast phase bands in the chart to set the maillard/development transition to the last first crack start temperature. Now automatic adjustment only takes place if the first crack start temperature is between the drying/maillard and the development end temperatures. This will always be the case in a real roast and ensures that the roast phase bands are not adjusted when the user is testing the RoastLogger by clicking the roast event button(s) whilst not performing a real roast.

[GreenBean]

I am very impressed by the work of Clube do Café member, sergio.m, in Rio de Janeiro. He is experimenting with computer control of a Behmor roaster using RoastLogger.

He has built a TC4 like interface to control the heater and fan and is currently experimenting with control via the PID built into the RoastLogger Arduino Controller. He has posted a video of his interface in action and a chart of his first test of the PID control. In this test he has stepped the setpoint up in stages and charted the result. Considering that this is the first attempt, without fine tuning the PID, It looks very promising:



For further details see his posts on Clube do Café.

[GreenBean]

A further very minor update to the RoastLogger to version 2.0.5 is available. This update only involves a few cosmetic changes to the layout of the chart tab and the Arduino Controller action tables.

For further information or to download see the following version information and the RoastLogger website.

Version 2.0.5 10/12/2013

Cosmetic changes to the chart and action tables as follows:

Reduced the margin around the chart to allow slightly larger chart on low resolution screens.

Improved the layout of the action tables. Now, when the user chooses the menu option to limit the table opening to five items, scroll bars are provided allowing the user to scroll to all settings.

Now a new installation defaults to limiting action table heights to show the first 5 items. This can be changed to show the full 10 items per table using the menu option as previously.

Added tooltips to action table headers.

[silron1]

Thanks for that GB - Nice little touch. Not for me but interesting stuff on the Behmor

[GreenBean]

Thanks for that GB - Nice little touch. Not for me but interesting stuff on the Behmor

No problem Ron. I am afraid I have made one more revision to the RoastLogger. The good news is that this should be the last revision you ever have to manually update to.

[GreenBean]

A further update to the RoastLogger to version 2.0.6 is available. I am aware that there have been too many updates to the RoastLogger in the last few months. Hopefully, this will be the last manual update to the RoastLogger you will ever have to make.

This update adds an automatic updater to the RoastLogger to make it easy, in future, to update to the latest version. The updater is easy to use and typically takes less than 20 seconds to update to the latest version. Whilst downloading and installing the latest version it copies all your user settings to the new version. It does not affect your currently installed version in any way so you can try the latest version before deleting your previous installation.

For further information or to download see the following version information and the RoastLogger website.

Version 2.0.6 16/12/2013

Added an automatic updater to the RoastLogger to make it easy, in future, to update to the latest version without affecting your currently installed version. This allows you to try the new version before deleting your previous installation.

The updater is started by clicking the menu option Help > Update RoastLogger. The updater then opens and shows the details of the last three updates to the RoastLogger:



Once you click the Update button the following actions are performed:

1. Downloads the latest version of RoastLogger.zip
2. Creates a new RoastLogger folder alongside (i.e. in the same parent folder as) your existing installation and adds the version number to the new folder name.
3. Extracts/installs the relevant files from RoastLogger.zip to the new folder.
4. Sets the executable permissions for the required files.
5. Copies all your user settings from the previous installation to the new installation.
6. Deletes the RoastLogger.zip file downloaded in 1. as it is no longer required.



The entire process typically takes less than 20 seconds to complete. Obviously internet access is required for this and you may need to adjust your firewall to allow RoastLogger access although this has not been necessary in my tests.

From version 2.0.6 the above is built into RoastLogger and you only need to select the menu option described above to use it. If you wish to use the updater to update to version 2.0.6 you can download RoastLogger-Updater.jar separately, drop it into your existing RoastLogger installation (alongside RoastLogger.jar) and double click it to run the updater.

[GreenBean]

Several people have registered an interest in obtaining an RL-HT-CTRL board in the last few months. As Jim is still not back in production of his HTC/HTShield boards, I am, as promised, considering building another batch of RL-HT-CTRL boards in the near future. All those who have registered an interest have been provided with the relevant details and asked to confirm if they wish to go ahead. If anyone else, who has not yet registered an interest, wishes to obtain a board please let me know as soon as possible and I will send them the relevant information. My email address is listed in the feedback section of the RoastLogger website.

[Edit: added link to RL-HT-CTRL webpage]

[GreenBean]

The experimental PID built into the RoastLogger is being tested by another user. CoffeeSnobs user fg1972 is currently using it to control his Corretto roaster.

Corretto controlled by the PID built into RoastLogger:


See his thread on CoffeeSnobs for details.

He states that "the RoastLoggers PID functionality is working very well with excellent repeatability". For most electrically heated roasters, such as Hottops and the Quest M3, I think the action tables produce better results than a PID. For roasters that react reasonably quickly to a change in heat input the PID may be better in some cases.

For anyone wishing to try the PID some initial notes on its use are available here.

[GreenBean]

The current batch (January 2014) of RL-HT-CTRL boards are now well oversubscribed. Further information on these boards is available in the review by Randy Glass and the relevant posts in this thread dated between 23 July and 21 August 2013. I will consider making another batch of boards in a few months time if JimG is still not back in production of his boards and there is sufficient interest. If you are interested in obtaining a board from the next batch, and have not already registered an interest, then please do so by emailing me at the email address included in the feedback section of the RoastLogger website.

If I decide to build another batch I will contact all those who have registered an interest and provide further details for them to consider if they wish to go ahead.

[GreenBean]

It has now been confirmed that the USB HID communication used by the RL-HT-CTRL works fine on Windows 8. I like the colours so much that I have added an image to prove it.

Arduino Controller running on Windows 8 and connected to a RL-HT-CTRL board:

[greencardigan]

Hi GB

I've been using Artisan with a TC4 in the past but have downloaded the latest version of Roastlogger to try out the PID functionality.

I noticed that the sliders to control the heater and fan power change in 5% increments. This is too coarse for my air roaster. Can I put in a request to make the increments 1% or have a user defined increment?

Also does the PID update the heater once per second?

My TC4 is running a highly modified/enhanced version of JimG's aArtisan code and I just made further modifications to make it compatible with Roastlogger.

I'll try out the PID functionality next time I roast and let you know how it goes.

Brad

[GreenBean]

Hi Brad, welcome to TMC I have followed your posts on other forums with interest. I confirm that the PID included in the Arduino Controller recalculates and updates the heater once per second. The PID calculates to 1% and this is rounded to the nearest 5% when it is applied to the heater slider.

The Arduino Controller originally controlled the heater and fan in 1% increments. JimG requested I changed it to 5% so that it is compatible with his stand alone (or use with RoastLogger) LCD adapter board. We were both happy that 5% increments were more than adequate for control of a coffee roaster.

I would be reluctant to change this as it may cause problems for users of Jim's LCD adapter board. I have never used an air roaster but, as the heater setting is updated by the PID every second, I find it very difficult to believe that you can perceive any difference in the results with 5% or 1% increments. Many PID's have an increment of 100% i.e. they are on or off.

You will see from posts in this thread that others say they are getting good results from the PID. I find, with Hottop roasters, I get much better results using the action tables.