Both the tests let us know about the core losses and copper losses in single phase transformer, then why is sumpner's test preferred over "open circuit and short circuit test" ?
The Sumpner's test(back to back test) is the very practical, convenient, efficient and minimum power consumption without actual loading(heat run(Temp. rise) test for Hrs.) method to find regulation and efficiency of large power transformer, used in practice. Only needed are two identical power transformers, and is very conveniently used by large power transformer manufectorers, compare to O.C&S.C test. It can be compered with the back to back test, used for D.C machines(Hopkinson's test), power consumption would be equal to no load power consumption of machines/equipments. Two transformers can be tested at a time.
The Sumpner's test(back to back test) is the very practical, convenient, efficient and minimum power consumption without actual loading(heat run(Temp. rise) test for Hrs.) method to find regulation and efficiency of large power transformer, used in practice. Only needed are two identical power transformers, and is very conveniently used by large power transformer manufectorers, compare to O.C&S.C test. It can be compered with the back to back test, used for D.C machines(Hopkinson's test), power consumption would be equal to no load power consumption of machines/equipments. Two transformers can be tested at a time.
To find the temperature rise of transformer at rated load(current), for continuous rateing, 'Heat run test' is carried out. In the Sumpner's test rated current in secondary winding(connected back to back) of transformers in parallel is, achieved to rated current by applying little voltage, using the variable voltage source( of secondary side) instead of actual loading. Please refer any A.C machines book like of author M.G.Say.......
It seems to be a really clever way of simultaneously testing for open and short circuit conditions, and heating under load, as Bhupendra says. The load current is injected separately, in the secondary, to test multiple load conditions, while the scheme is simultaneously testing for copper losses and iron losses. Very clever.
The main drawback seems to be that you need two identical transformers, to obtain accurate results.
Sumpner's test simulates closely the actual full load test when both iron losses
and copper losses are present at the same time without actual full load loading .Many standards specify full load test to confirm the real temperature rise (hot spot) because depending on type of cooling (forced / natural) the cooling is difficult to model .It is the same in dry type transformers also.Usually identical transformers will be available in smaller distribution transformers production and so this type of type test is preferable. Also as indicated by others full load test is
not practicable for large transformers .Temperature rise with rated load current in short circuit is used in testing production runs after type test.Open circuit loss is used in efficiency calculation and also induced voltage test for checking the insulation rating.If you want to see how the Sumpner.s test behaves when transformers are not identical you may like to simulate in Pspice in a very simple manner.It is a good mini project.