CM_ET/BT LCDs Display

[Terracotta-6]

CM_ET/BT LCDs Display Issue Test conditions: Same foreground and background curves. After pressing keyboard keys T and U and moving the mouse, the LCDs display incorrect values. The correct display should be CM_ET/BT LCDs 00 /00.

[MAKOMO]

Good catch! The recorded CM ET/BT readings only make sense only if the profiles (incl. the same background profile used during recording) are aligned the same way as during the recording. The readings are intended only for display during roasting, but still recorded along all other values. The running LCDs feature displays all readings as they were recorded in the display. There is no recomputation happening for any of those readings. It is a generic mechanism applied to all readings. Thus there is not much that can be done here.

[Terracotta-6]

Curve Matching 功能在 Artisan 中的用途包括：

The Curve Matching function in Artisan is used for:

实时烘焙指导：帮助烘焙师跟踪烘焙进程
Real-Time Roasting Guidance: Helping roasters track roast progress
烘焙重现：通过匹配历史曲线来重现特定的烘焙结果
Roast Reproduction: Reproducing specific roast results by matching historical curves
质量控制：确保烘焙一致性
Quality Control: Ensuring roast consistency
预测烘焙终点：基于相似曲线预测烘焙结果
Predicting Roast Endpoints: Forecasting roast outcomes based on similar curves
鲁棒性改进的重要性
Importance of Robustness Enhancement
在曲线匹配中实现鲁棒性处理特别重要，因为：温度传感器可能出现间歇性故障
Robustness handling in Curve Matching is especially critical because:Temperature sensors may experience intermittent failures
电磁干扰可能产生异常读数
Electromagnetic interference can cause anomalous readings
通信问题可能导致数据丢失
Communication issues may lead to data loss
烘焙过程中不能因为数据问题中断匹配分析
Matching analysis must not be interrupted by data issues during roasting
通过严格处理原始数据和鲁棒性处理断线及非数字读数，Curve Matching 功能可以在各种不利条件下依然提供可靠的匹配结果，帮助烘焙师完成高质量的烘焙过程。
By rigorously processing raw data and applying robust handling of disconnections and non-numeric readings, the Curve Matching function can deliver reliable matching results under various adverse conditions, enabling roasters to achieve high-quality roasts.

My personal take—feel free to flag inaccuracies and correct me!

https://www.youtube.com/watch?v=iUOO6tkkCbc

https://artisan-roasterscope.blogspot.com/2019/05/how-close.html

[MAKOMO]

The curve matching is foremost a quality measure. How well does the current (foreground) profile matches the template (background) profile. There was a long time an offline (calculated after the roast) measure displayed as x-axis label (in one of the modes; you need to right-click that label a view times to spot it). The new extra device types do now calculate that measure during roasting. This gives the operator a numeric assessment how well he is duplicating the roast in addition to the visual clue provided by the drawing of the curves. This numeric value can also be displayed in large LCDs to an audience during competitions, which is the motivation to add it. This two measures might divert in cases as the offline measure is computed by profiles aligned to DROP from DRY END, while the online measure is assuming profiles (re-)aligned by DRY END, averaging the results from DRY END until the current measurement. Basing both on raw measurement instead of smoothed curves makes both computation to resulting in closer results with the drawback that the noise level of the sensors sets a hard minimum in the achievable result. This all is experimental as most of what you find in Artisan.