MIL- STD-883F 2004 TEST METHOD STANDARD MICROCIRCUITS.pdf - 第140页

MIL-STD-883F METHOD 1021.2 15 November 1991 2 1.2 Int erfer ences . There ar e several inter ferenc es that need to be cons idered when t his t est pr ocedure i s appli ed. These incl ude: a. Total dose damage. Devic es …

MIL-STD-883F

METHOD 1021.2

15 November 1991

METHOD 1021.2

DOSE RATE UPSET TESTING OF DIGITAL MICROCIRCUITS

1. PURPOSE

. This test procedure defines the requirements for testing the response of packaged digital integrated

circuits to pulsed ionizing radiation. A flash x-ray or linear accelerator is used as a source of pulses of ionizing radiation.

The response may include transient output signals, changes in the state of internal storage elements, and transient current

surges at inputs, outputs, and power supply connections. The dose rate at which logic or change-of-state errors first occur

is of particular interest in many applications.

1.1 Definitions

. Definitions of terms used in this procedure are given below:

a. Dose rate threshold for upset. The dose rate which causes either:

(1) A transient output upset for which the change in output voltage of an operating digital integrated circuit goes

either above or below (as appropriate) specified logic levels (see 3.2 on transient voltage criteria), and the

circuit spontaneously recovers to its preirradiation condition after the radiation pulse subsides, or

(2) A stored data or logic state upset for which there is a change in the state of one or more internal memory or

logic elements that does not

recover spontaneously after the radiation pulse. However, the circuit can be

restored to its preirradiation condition by applying the same sequence of logic signals to its inputs that were

previously used to establish the preirradiation condition, or

(3) A dynamic upset which results in a change in the expected output or stored test pattern of a device that is

functionally operating during the time it is irradiated. The upset response may depend on the precise time

relationship between the radiation pulse and the operating cycle of the device. For operations requiring

many clock signals, it may be necessary to use a wide radiation pulse.

b. Dose rate. Energy absorbed per unit time per unit mass by a given material from the radiation field to which it is

exposed.

c. Combinational logic circuit. A digital logic circuit with the property that its output state is solely determined by the

logic signals at its inputs. Combinational logic circuits contain no internal storage elements. Examples of

combinational circuits include gates, multiplexers, and decoders.

d. Sequential logic circuit. A digital logic circuit with the property that its output state at a given time depends on the

sequence and time relationship of logic signals that were previously applied to its inputs. Sequential logic circuits

contain internal storage elements. Examples of sequential logic circuits include memories, shift registers,

counters, and flip-flops.

e. State vector. A state vector completely specifies the logic condition of all elements within a logic circuit. For

combinational circuits the state vector includes the logic signals that are applied to all inputs; for sequential

circuits the state vector must also include the sequence and time relationship of all input signals (this may include

many clock cycles).

MIL-STD-883F

METHOD 1021.2

15 November 1991

1.2 Interferences

. There are several interferences that need to be considered when this test procedure is applied. These

include:

a. Total dose damage. Devices may be permanently damaged by total dose. This limits the number of radiation

pulses that can be applied during transient upset testing. The total dose sensitivity depends on fabrication

techniques and device technology. MOS devices are especially sensitive to total dose damage. Newer bipolar

devices with oxide-isolated sidewalls may also be affected by low levels of total dose. The maximum total dose to

which devices are exposed must not exceed 20 percent of the typical total dose failure level of the specific part

type.

b. Steps between successive radiation levels. The size of the steps between successive radiation levels limits the

accuracy with which the dose rate upset threshold is determined. Cost considerations and total dose damage

limit the number of radiation levels that can be used to test a particular device.

c. Latchup. Some types of integrated circuits may be driven into a latchup condition by transient radiation. If latchup

occurs, the device will not function properly until power is temporarily removed and reapplied. Permanent damage

may also occur, primarily due to the large amount of localized heating that results. Although latchup is an

important transient response mechanism, this procedure does not apply to devices in which latchup occurs.

Functional testing after irradiation is required to detect internal changes of state, and this will also detect latchup.

However, if latchup occurs it will usually not be possible to restore normal operation without first interrupting the

power supply.

d. Limited number of state vectors. Cost, testing time, and total dose damage usually make it necessary to restrict

upset testing to a small number of state vectors. These state vectors must include the most sensitive conditions

in order to avoid misleading results. An analysis is required to select the state vectors used for radiation testing to

make sure that circuit and geometrical factors that affect the upset response are taken into account (see 3.1).

2. APPARATUS

. Before testing can be done, the state vectors must be selected for radiation testing. This requires a

logic diagram of the test device. The apparatus used for testing shall consist of the radiation source, dosimetry equipment,

a test circuit board, line drivers, cables, and electrical test instrumentation to measure the transient response, provide bias,

and perform functional tests. Adequate precautions shall be observed to obtain an electrical measurement system with

ample shielding, satisfactory grounding, and low noise from electrical interference or from the radiation environment.

2.1 Radiation source

. The radiation source used in this test shall be either a flash x-ray machine (FXR) used in the

photon mode or a linear accelerator (LINAC) used in the electron beam mode. The LINAC beam energy shall be greater

than 10 MeV. The radiation source shall provide a uniform (within 20 percent) radiation level across the area where the

device and the dosimeter will be placed. The radiation pulse width for narrow pulse measurements shall be between 10 and

50 ns. For narrow pulse measurements either a LINAC or FXR may be used. Wide pulse measurements (typically 1 - 10

µs) shall be performed with a LINAC. The pulse width for LINAC irradiations shall be specified. The dose rate at the

location of the device under test shall be adjustable between 10

and 10

rads(Si)/s (or as required) for narrow pulse

measurements and between 10

and 10

rads(Si)/s (or as required) for wide pulse measurements. Unless otherwise

specified, a test device exposed to a total dose that exceeds 20 percent of the total dose failure level shall be considered as

destructively tested and shall be removed from the lot (see 1.2a).

2.2 Dosimetry equipment

. Dosimetry equipment must include a system for measuring total dose, such as a

thermoluminescent dosimeter (TLD) or calorimeter, a pulse shape monitor, and an active dosimeter that allows the dose rate

to be determined from electronic measurements, e.g., a p-i-n detector, Faraday cup, secondary emission monitor, or current

transformer.

MIL-STD-883F

METHOD 1021.2

15 November 1991

2.3 Test circuit

. The test circuit shall contain the device under test, wiring, and auxiliary components as required. It shall

allow for the application of power and bias voltages or pulses at the device inputs to establish the state vector. Power

supply stiffening capacitors shall be included which keep the power supply voltage from changing more than 10 percent of

its specified value during and after the radiation pulse. They should be placed as close to the device under test as possible,

but should not be exposed to the direct radiation beam. Provision shall be made for monitoring specified outputs.

Capacitive loading of the test circuit must be sufficiently low to avoid interference with the measurement of short-duration

transient signals. Generally a line driver is required at device outputs to reduce capacitive loading. Line drivers must have

sufficient risetime, linearity, and dynamic range to drive terminated cables with the full output logic level. The test circuit

shall not affect the measured output response over the range of expected dose rates and shall not exhibit permanent

changes in electrical characteristics at the expected accumulated doses. It must be shielded from the radiation to a

sufficient level to meet these criteria.

Test circuit materials and components shall not cause attenuation or scattering which will perturb the uniformity of the beam

at the test device position (see 2.1 for uniformity). The device under test shall be oriented so that its surface is

perpendicular to the radiation beam.

2.4 Cabling

. Cabling shall be provided to connect the test circuit board, located in the radiation field, to the test

instrumentation located in the instrumentation area. Coaxial cables, terminated in their characteristic impedance, shall be

used for all input and output signals. Double shielded cables, triax, zipper tubing or other additional shielding may be

required to reduce noise to acceptable levels.

2.5 Transient signal measurement

. Oscilloscopes or transient digitizers are required to measure transient output

voltages, the power supply current and the dosimeter outputs. The risetime of the measuring instrumentation shall be less

than 10 ns for pulse widths greater than 33 ns or less than 30 percent of the radiation pulse width for pulse widths less than

33 ns.

2.6 Functional testing

. Equipment is also required for functional testing of devices immediately after the radiation pulse in

the radiation test fixture. This equipment must contain sources to drive inputs with specified patterns, and comparison

circuitry to determine that the correct output patterns result. This equipment may consist of logic analyzers, custom circuitry,

or commercial integrated circuit test systems. However, it must be capable of functioning through long cables, and must

also be compatible with the line drivers used at the outputs of the device in the test circuit.

2.7 General purpose test equipment

. Power supplies, voltmeters, pulse generators, and other basic test equipment that

is required for testing is general purpose test equipment. This equipment must be capable of meeting the test requirements

and should be periodically calibrated in accordance with ANSI/NCSL Z540-1 or equivalent.

3. PROCEDURE

. An outline of the procedure is as follows: a) determine the state vectors (or sequence of test vectors

for a dynamic test) in which the device will be irradiated; b) following the test plan, set up the test fixture, functional test

equipment, and transient measurement equipment; c) set up and calibrate the radiation source; d) perform a noise check on

the instrumentation; and e) test devices at a sequence of radiation pulses, determining the transient response at specified

dose rates. The dose rate upset level can be determined by measuring the transient response at several dose rates, using

successive approximation to determine the radiation level for dose rate upset.

3.1 State vector selection

. Two approaches can be used to select the state vectors in which a device is to be irradiated:

a. Multiple output logic states. Partition the circuit into functional blocks. Determine the logic path for each output,

and identify similar internal functions. For example, a 4-bit counter can be separated into control, internal flip-flop,

and output logic cells. Four identical logic paths exist, corresponding to each of the four bits. Determine the total

number of unique output logic state combinations, and test the circuit in each of these states. For the counter

example this results in 16 combinations so that the upset must be determined for each of these 16 state vectors.