Downloaded from https://www.everyspec.com AMCP 706-10 is amrbiguous. It is convenient to classify increasing stress, For example, if extremnelyr step-stressing into three categ~ories: accurate voltage steps are desired, a stepping switch might be used with a voltage divider. 1. Large ste~ps in whivii the steps are otherwise, a slow motor might be used to presumed high enough and long enough so turn a multiturn potentiometer, that, for a given tttep the, damage accumul- ated at all previous steps Is negligible. Less testing time is the major advintage of step-stress tests over constant-stress tests. A 2.
S~mall steps in which\the steps aire direct comparison of the methods requires small enough so that in thel.analysis one can an assumption of a theory of cumulative presume with neglligible arrorithrt the severity dainage (see par. 7-6), In the area of natal V level Is steadily increasing. *i is then Juut fatigue, there are many thenries of cumu- tht, propes~sive-step case-,in ptogressivo-stress lative damage, In elcatronics,'s simwple linear the stress incroases at a consta0i rate. model Is assumed most, often because of its simplicity and the 'lack of knowledge of the ~ 3.
Medilum steps tor whiich the assump- actual processes, tdons for neither small not large steps are valid, T1e cumulative damage at previous A linetar model of camolatlve damoqe is aII steps mnust be taken into account, but the gross approximation. In some aircumnstsnces, stepE are not small enough that the saverity it consistently underestimates and in other level can bts oonsidcreci to be continuously circurnitancem, consistently overestimates the increating. corroct resulits, Repardleas of these defician- cles, It offers Ilic uadvantages c'tbeing tvact- In addition, the following torminology IN able, easily remembered, anti widely used, used:
large/stop-stresu, medium/ttep-streso, So uou the linear model unless you know of and sinall/sttp-stress, The &Izt designations ., some other which is better, But remember 4 are niot ab~olute, but iire relative to the kind the arbitrariness of any nssumption. of arueiysis that must tbe petformad. An Important parameter in step-strrsi Largo/step-stross tests are analyzed as if testing Is the ratio of' severity step size to the they were cLonntant-stress tests being run at time Rt ecccl level.
This controls the rate of the se'verity levol of the last step, Parts that Increase of' the stress severity and Is the pars- arm ),try expensive or otherwime difficult to moter that is varied when running several acquire or test often ait treated in thi.s way, tests on a particular population of Itensa. Often a sample of only one is used. It Is wise to consider the resulls ms "ballpark" For some kinds of Items, the maximl-lin figures, since the necessary assumption that ustiful. severity level will be exmiedeoJ before the effects of previous steps are niegligible Is the device tailt In the propor mnode. Pot likely to be. in error.
Preliminary tests often example, on thermally stresned trmirsistors are run In this wsy and are. followed by more there are bametlmes autectk, points where comprehensive met of teat& later on, maeltin~g occurs and the tranuistor essaistislly Smal/stp-sresingis aalyed ii he ame coames to be a trattdstor, If this happens, Smal/stp-srasingIN nalzedin he ame the v'ate of increasling tho stvess severity way as progreusivv-stressing, and, In fact, musut be dectoased. The slope of the steps by definition, there is really no distinction duting the course of the testR also can be between them.
In many cases, there may bo 2flodified 'The severity lovel limits (I.e., the an economic advantage to choosing either level where ti-i dlevice cee*dm to function in very small step increments or a continuoualy Its usual manneor) ar~ tau iuripcrtatt 'Airniation ..................... ............... IU. .. .. .. .. ............. JW1* *.,"Aw