|
|
NWin determines a test unit's polarity by applying a unipolar pulse and examining the test unit's response. When creating a polarity test an engineer can display a test unit's response by selecting Run on the main test window. Three black horizontal lines indicate the upper limit, nought volts and the lower limit. The programme determines pass or failure from the first excursion, negative or positive, to cross a limit, one of the dotted horizontal lines.
When testing a system it is a useful convention to execute the polarity test on the bass driver by setting Test Unit LF. There will be three possible polarity faults with a two way system: bass driver polarity reversed, tweeter polarity reversed, and, both drivers polarity reversed. If the bass driver is reversed then the test unit will fail the polarity test and because there is phase reversal between the bass driver and tweeter the frequency response will usually have a suck out in the crossover region. It should be possible to set up the frequency response test so that the suck out causes a response failure. If the tweeter is reversed then the test unit will not fail the polarity test but because there is phase reversal between the tweeter and bass driver, and a resultant response suck out it will fail the frequency response test. If both drivers are reversed then the test unit will fail the polarity test but because there is no phase reversal between the drivers, it will pass the frequency response test. With this knowledge, a test operator can diagnose which driver or drivers are at fault. Alternatively, if the test sequence is set up with two polarity tests, one for the bass driver and one for the tweeter, then even this simple analysis is not needed. The operator will know from the test failure summary which unit or units are reversed. The time penalty for an additional test is minimal but it may well need some time from the engineer to fine tune the test parameters so that they are independent of each other. If the mike is some distance from the test unit and the PC is a speedy one then it is just possible that NWin will have stopped looking before the response arrives at the mike. To correct this set a delay so NWin waits before looking. Naturally, it is possible to over correct and to set too long a delay so the response is long gone by the time NWin starts looking. What is more likely on a system or dual driver is that the response's front edge is complex (despite setting Test Unit LF) and it is difficult to get a stable result. A solution is to a make the pass or fail decision on the bass driver's second excursion. By setting a delay you force NWin to look at the later part of a unit's time response. Fairly obviously Drive (V) has to be set so the unit's response crosses a limit or it will be judged a failure. |
|