Chip-by-Chip Measurement Correction Using On-Chip Error Tables
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Integrated circuit devices often exhibit variations in output due to practical limitations in circuit design or architecture, leading to inaccuracies that can affect market acceptance, as users expect devices to function according to published specifications.
Innovation Solution
A method is introduced to generate error correction values, which are stored in a nonvolatile memory as part of an error correction table, allowing for chip-by-chip measurement corrections by calculating the difference between desired and actual outputs, enabling the host processor to correct measurement inaccuracies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If circuit design and architecture are simplified for ease of manufacture, then manufacturing cost and complexity are reduced, but measurement accuracy and output consistency deteriorate
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing error correction values in a lookup table during the manufacturing process. The test system measures actual output values, calculates correction factors, and stores them in non-volatile memory before the integrated circuit is shipped to the customer. This allows the simplified circuit design to achieve accurate measurements through pre-computed corrections rather than through complex hardware design.
Solution Approach 2:
The patent introduces an intermediary element - the error correction lookup table - that mediates between the simplified circuit design and the required measurement accuracy. Instead of relying on complex circuit architecture to achieve precision, the lookup table serves as an intermediate layer that compensates for manufacturing variations and enables accurate measurements from simpler, more cost-effective circuit designs.
2Measurement precision
If individual chip calibration is implemented to improve measurement accuracy, then output precision is improved, but manufacturing time and complexity increase
Solution Approach 1:
The patent performs calibration measurements and generates correction values during the manufacturing process as a preliminary action. The test system automatically measures each chip's output characteristics, calculates the appropriate correction factors, and stores them in the chip's non-volatile memory before the chip is packaged and shipped. This automated preliminary calibration enables individual chip accuracy without requiring time-consuming manual adjustment at the customer site.
Solution Approach 2:
The patent implements self-service by enabling each integrated circuit to automatically apply its own correction values during operation. The microcontroller within the integrated circuit uses the pre-stored lookup table to compensate for its specific manufacturing variations, allowing the chip to self-correct its measurements without external intervention. This eliminates the need for manual calibration services and enables high-volume automated manufacturing.
3Measurement precision
If error correction tables are stored in nonvolatile memory, then measurement accuracy is improved, but device complexity and memory requirements increase
Solution Approach 1:
The patent extracts the complexity of accuracy adjustment from the circuit design and concentrates it in a separate, manageable component - the error correction lookup table stored in non-volatile memory. Instead of increasing circuit complexity through additional analog components or complex digital signal processing, the correction data is extracted and stored as simple lookup tables that can be easily generated during manufacturing and require minimal processing during operation.
Data Source
AI summary
A method for allowing measurement corrections on a chip-by-chip basis. Error correction values are generated responsive to the input value to a circuit of the calibrated integrated circuit chip and to a measured value from the circuit of the calibrated integrated circuit chip. The error correction values are stored within an error correction table within a nonvolatile memory of the integrated circuit chip.