Pulse circuit

Details Pulse circuit Pulse circuit

2000-04-28

2002-08-06

Sherry, Michael J. (Department: 2829)

Electricity: measuring and testing

Fault detecting in electric circuits and of electric components

Of individual circuit component or element

C324S765010, C324S676000, C324S678000, C333S020000

Reexamination Certificate

active

06429674

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention is in the general field of pulse circuitry and, more particularly, relates to measurement of the effects of a pulse applied to a device under test.
2. Description of the Prior Art
Electrostatic Discharge (ESD) has long been recognized as a threat to integrated circuit devices. Whether generated through triboelectrification, induction or conduction, the discharge that occurs can contain several hundred nanojoules of energy and generate 3,000 V, enough to destroy all but the most robust semiconductor devices.
Large steps have been taken by the semiconductor industry towards reducing the ESD environment during fabrication in regard to parts handling techniques and in grounding technology and practices. The industry has also sought to address the problem by including ESD protective circuitry in the overall circuit design.
Efforts in regard to the latter approach have been hindered by the difficulty in testing such protective circuits, given the short time, high-current characteristics of ESD events (typically tens of nanoseconds and amps of current). Traditional laboratory equipment was incapable of simulating such electrical pulses, or accurately measuring the resultant circuit damage caused thereby. Without such feedback, it was difficulto assess the effectiveness of protective circuits as well as to develop improved circuits.
In a paper delivered at a 1985 ESD Symposium by Maloney and Khurana of Intel Corporation “Transmission Line Pulsing Techniques for Circuit Modeling,” a different approach was proposed. Instead of attempting to simulate the complex waveforms of human body models, Maloney and Khurana proposed use of simple square pulses obtained from a charged transmission line. Known as Transmission Line Pulses (TLP), their use on the same time and current scale as ESD events was intended to provide a simple pulsed curve tracing system to measure the characteristic curve of ESD protection structures. As originally proposed, a diode and resistor placed at the open end of a charged coaxial line to absorb the pulse energy reflected from the Device Under Test (DUT).
Subsequent test equipment has used other methods to reduce reflected pulse energy from the DUT, including the placement of a 500-ohm resistor between the 50-ohm source and the DUT, which provided a constant current source. A 50 or 56-ohm resistor was placed at the end of the 50-ohm transmission line to terminate the part of the test pulse that did not flow through the 500-ohm resistor. In these test system designs, the current and voltage monitors would typically be located directly at the DUT for close-in measurements. U.S. Pat. No. 5,519,327 to Consiglio is an attempt to improve the pulse generator with additional switches, to prevent an incorrectly assumed impulse source occurring as a result of distributed capacitance between discharge switch contacts.
These present methods of pulse generation have been attempted without a clear understanding of the time domain techniques that must be used to produce pulses with sub-nanosecond nanosecond Gaussian-type risetimes. The techniques used for pulse measurements have suffered from similar technology limitations, resulting in measured data of limited accuracy.
For example, the method most often used to absorb reflected pulses places a diode and resistor at the open end of the charged line used to produce the rectangular pulse. Diodes have capacitance across them and are not “perfect” diodes. This capacitance causes measurement errors from imperfect pulses or reflections that are bypassed through the diode capacitance or is an artifact of the majority carrier lifetime.
Other problems are created by the reflections at lower voltages where the diode is not sufficiently forward-biased. The charging and discharging of a transmission line with long connections between the switches and the charged lines produces non-Gaussian, slow risetimes. Additional circuit complexities result from the existence of capacitance between the contacts of a discharge relay, which cause impulse discharges in addition to the desired rectangular pulse. Analyses of such pulse generation are questionable because of the difficulty in building and correctly analyzing time domain circuits.
Use of the series 500-ohm resistor to provide a constant current source and the shunt 50-ohm resistor most often used to minimize reflections from the variable impedance of the DUT are typically assembled with insufficient consideration of lead length (inductance), which distort fast pulse rise times. The commonly used resistors in the 50- to 500-ohm range have high voltage coefficients and are difficult to assemble without adding more distortion to fast pulses. The 500/50-ohm matching system also reduces the pulse amplitude available to the DUT to about 50% of the amplitude that is available by using a matched (coaxial) attenuator to reduce reflections from the DUT. Placing the current and voltage monitors directly at the DUT also adds parasitic inductance, resistance, and capacitance at the measurement location. Optimum compensation for these parasitic elements can not be provided for the very wide bandwidth measurements that is needed for controlled fast pulse rise times in this time domain system. The imperfections at different pulse heights to both the generated pulse and the measurement hardware provide less than optimum test data accuracy throughout the full testing amplitude range.
Without a controlled risetime, the threat to the DUT is under less control. Real Human Body Model (HBM) pulses have a relatively slow risetime of between 2 to at least 10 nanoseconds to an amplitude where “snapback” occurs. Slower rise time pulses allow the DUT snapback to go to significantly higher voltages than when faster risetime pulses are used. Applying higher voltages to the protection elements before the conductive structure of the protection turns on, places a higher voltage on the voltage-sensitive gate oxide. This is a closer simulation to actual HBM threat pulses. It is also a more severe test than when using faster risetime pulses that cause the protection structure to turn on at lower voltages with correspondingly lower threats to the gate oxide.
Consiglio in U.S. Pat. No. 5,519,327 recites an attempt to solve two “problems” with the prior TLP test equipment, and fails to understand that the first problem is, in fact, a non-existent problem. The Consiglio circuit attempts to reduce a short impulse that is claimed to be produced when the switch armature makes contact to the supply voltage contact, and causes a displacement current to flow in the opposite contact of the pulse-forming switch. The capacitance between the terminals of a SPDT switch are shielded by the armature and amount to a maximum of 0.5 pF. Such a discharge into a 50-ohm load has a 25-pS exponential fall time with an unspecified risetime. With a poorly designed transmission line circuit, the 25 pS impulse that might be formed with a poorly chosen switch would be smeared out and reduced to an unmeasurably low amplitude before it ever got to the DUT. Although the circuit disclosed in U.S. Pat. No. 5,519,327 does remove the resistance of the voltage probe from the leakage current measurement path, it is done with terminal
106
of switch S
1
and terminal
152
of switch S
3
. They create a complex generating and reflecting structure from “T” type branched stubs that produce pulse distortions because it creates reflections of the test pulse before the pulse reaches the DUT.
It is also highly unlikely that the invention described in U.S. Pat. No. 5,519,327 actually operates as described, because when its S
1
is closed, a very high DC potential is applied to terminals
122
and
123
of the scope
124
. No oscilloscope can withstand such high voltages and if an undescribed voltage probe is used, its shunt resistance would decrease the voltage on the charge line by the voltage dividing effect of resistor
102
and the voltage probe resistance to ground. It would also begin to discharge the

Affiliated with

Also associated with

Rating

Pulse circuit does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Pulse circuit, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Pulse circuit will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2921378