Downloaded from https://www.everyspec.com MIL-HDBK-816 NOTICE 1 5.1.8.8 Optimum use of radiation facilities. Such factors as the number of devices which can be irradiated at one time, the length of time required for the irradiation, and the length of time required to complete post-irradiation measurements should be considered when an irradiation schedule is being planned. The selected irradiation schedule should then be itemized in the test plan. 5.1.8.8.1 Selection of radiation levels. For each type of radiation and for each estimated device failure level, a series of irradiations should be selected so that data is obtained most efficiently.
The lowest level of radiation should be below that which produces any failures. In general the subsequent irradiation increments will be sizeable fractions of the failure level. If significant nonlinear behavior is produced, however, then the location of the nonlinear region must be estimated and the radiation increments should be made smaller than they are outside of it. The highest radiation level should be high enough so that all of the tested parts will fail.
This selection will not only allow the device specification to be written for the maximum capability of the part type, should that be necessary, but will provide information on how close the part is to failure for a given system application. 5.1.8.8.2 Selection of radiation levels and abrupt failure. For step stress measurements and the presence of abrupt failure, the proper selection of radiation stress levels is of particular importance because a bad choice can compromise the measurements. The spacing between stress levels should be as small as feasible.
At the very least, care should be taken to assure that the spacings are small compared to the estimated standard deviation of the failure fluences for the parts to be tested. An estimate of this standard deviation may not be easy to obtain. As a first approximation, data on similar parts may be used. If there is no suitable data, smaller than expected spacing may be necessary for the first few parts and then, if costs can be lowered thereby, the spacing for the rest of the parts can be adjusted according to the measured standard deviation. For a complete characterization, the highest stress level must be large enough to drive all the tested parts to failure.
Very wide stress levels may require an extra large sample size. Less than four stress levels are unacceptable if abrupt failure of devices is considered a possibility. 5.1.8.9 Efficient use of test samples. The costs of radiation testing and, for some part types, the cost of the parts themselves both provide strong reasons for keeping the number of test parts as low as possible. In these circumstances it is essential that the experiments be specifically planned so that the maximum amount of useful information is obtained from the sample size that is available.
Whenever possible, preparations for the actual tests should be conducted on spare and perhaps less expensive parts which are not part of the final sample. 5.1.8.9.1 Selection of exposure sequences. Because most characterization measurements are destructive, a single device is exposed to only one radiation environment and the exposure sequence simply proceeds from low levels to functional or parametric failure. The cases where one device can be used for two radiation environments are those when the dose rate upset threshold measurement, which can be non-damaging, is made first and is followed by neutron, gamma ray, or SEU tests.
In these cases care must be taken to ensure that the accumulated ionizing radiation dose in the dose rate testing has not appreciably changed the characteristics of the device. The use of one device for two radiation environments is not recommended. It should be considered only when sample costs are so high that every possible way of conserving sample size must be used. Time dependent effects should be considered when a series of exposures is being planned. 5.1.9 Radiation exposures. Radiation exposures should be made in accordance with the test plan and should contain at least the information listed in table VIII (see 5.1.8.4).
Initial test system checkout, e.g., of test equipment operation, dosimetry, etc., should be accomplished with parts which are less costly and readily available; one possibility is to use parts of the same type as the sample but selected to a less stringent visual criterion. The use of a commercial equivalent device type can also be considered but such use will need to be validated experimentally. 5.1.9.1 Neutron exposures. Neutron exposures can be made either at fast burst reactors or at water moderated reactors.
In either case, the dosimetry practices at the selected facility should be checked to make sure that a valid 1 MeV displacement damage (Si) equivalent fluence can be obtained (see ASTM E722). Test method 1017 of MIL-STD-883 can be used as a guide for the exposure procedure. Short term (of the order of seconds) annealing effects do not have to be evaluated on QML part types. 5.1.9.2 Ionizing radiation dose (total dose) exposures. Test method 1019 of MIL-STD-883, entitled “Ionizing Radiation (Total Dose) Test Procedure”, specifies Co-60 as the radiation source to be used for ionizing radiation dose testing.
Ionizing radiation testing can also be performed with Cesium-137 sources and with low energy x-ray sources. If a source other than Co-60 is used, correlation measurements must be performed to effect a comparison with Co-60. High energy electron Linacs are not recommended because of the SUPERSEDES PAGE 22 OF MIL-HDBK-816 22