Browse Prior Art Database

Method of Charge Pump Defect Detection

IP.com Disclosure Number: IPCOM000244412D
Publication Date: 2015-Dec-09
Document File: 2 page(s) / 45K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method of defect detection in conjunction with semiconductor charge pumps. Charge pumps are commonly used for deriving voltage rails on semiconductors.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

Method of Charge Pump Defect Detection

Disclosed is a method of defect detection in conjunction with semiconductor charge pumps. Charge pumps are commonly used for deriving voltage rails on semiconductors. The charge pump, itself, is instantiated within the semiconductor circuits on the same piece of silicon as the functional logic such as a processor, memory chip, or ASIC, for example. Since these voltages are derived on chip, they usually do not have C4 or BGA access. This, in turn, prevents direct leakage-based measurements. It is commonly known in the industry that many process defects on a chip increase leakage current, and early detection of these unusual leakages can help protect the end customer and improve reliability, availability, and serviceability (RAS), especially since these defects can degrade over time and may fail at a customer. What is needed is a method to detect leakage paths on derived voltages in order to proactively identify defects before they are fatal in the field and take appropriate action.

    A charge pump basically operates by "pumping" electric charge from one charge source node (typically a power supply) to another charge sink node (often call a generated or derived supply). "Pumping" electrical charge involves clocking or cycling charge through passive electrical elements like capacitors (or inductors) and using the energy storage capabilities of these passive elements to move charge from one electrical node (the source) to another electrical node (the sink) while enabling a usable energy and voltage difference between these source and sink nodes. The rate of "pumping" (or clocking) of the charge pump, in terms of "pumps per second", is directly related to the amount of charge (or energy) being called for from the charge pump's load. This "pump rate" (in pumps/second) can be monitored by registers or counters connected
to the charge pump.

    As process defects can go undetected through all of the test gates (wafer, component, system) initially, only to fail later in the field. These fails have often been caused by increasing defect-based current leakage (due to electromigration) from the on-chip derived power supplies that are driven from on-chip charge pumps What is needed is a method to detect these defects before they de...