Method an apparatus for obtaining reflow oven settings for soldering a PCB
AssigneeSiemens Corporate Research, Inc.
InventorAngelo R. Marcantonio
Life of this patent
Estimating the term…
Abstract
An artificial neural network is trained to recognize inputted thermal and physical features of a printed circuit board, for providing settings for a reflow oven for obtaining acceptable soldering of the printed circuit board.
Claims6
An artificial intelligence system for processing feature data associated with individual printed circuit boards (PCB) within a family of printed circuit boards having similar but not identical combinations of characteristics, for producing as output data settings for a reflow oven for soldering one of said PCB's, said system comprising:
- a neural network;
- and
- means for training said neural network to receive identifying features of each individual PCB of said family of PCB's, and produce as outputs belt speed and temperature zone settings for said reflow oven, for obtaining acceptable soldering for a selected individual PCB.
A system for automatically determining settings for a reflow oven for soldering a printed circuit board (PCB) comprising:
- a microprocessor;
- programming means for installing a neural network program into said microprocessor;
- a reflow oven including a plurality of temperature zones, and a motor driven conveyor belt for moving a PCB through said plurality of temperature zones;
- means for determining thermal characteristics for a representative plurality of different samples of printed circuit boards typical of a family of printed circuit boards that may be run through said reflow oven, whereby for each sample PCB, certain thermal characteristics are selected and used in combination with identifiable physical characteristics for each sample PCB, for obtaining identifying and distinguishing input features between each sample PCB;
- means for determining for each of said plurality of sample PCB boards output features including the belt speed, and temperature settings for said temperature zones, respectively, of said oven for obtaining acceptable soldering of said sample PCB boards;
- and
- means for training said neural network by sequentially and iteratively applying via said microprocessor said input features to input nodes of said neural network, and said output features to output nodes of said neural network, respectively for each sample PCB, whereafter said neural network is operable in a recall mode for receiving features of a non-sample PCB, for providing as outputs from said neural network a desired belt speed, and temperature settings for each zone of said oven, for setting up said oven to solder said non-sample PCB.
A controllable oven for soldering a printed circuit board, comprising:
- artificial neural network means coupled to said oven and being trained to recognize inputted thermal and physical characteristics of a printed circuit board;
- and
- means for causing said artificial neural network means to provide settings including those for a plurality of temperature zones for said oven for obtaining acceptable soldering of said printed circuit board.
Dependencies are inferred from the claim text.