Hi,
We are using a buck boost configuration nearly identical to that seen on pg 4 of AN-1967. Our prototypes proved reliable and we are close to entering the production phase. It was at this time a new and potentially very important customer asked for several drivers to use in their prototype of a new product that involves flashing an array of LEDs very quickly, faster than we ever planned for, about 20KHz. A couple days ago we were contacted by the customer with a problem. When powered up in a continuous on state, the driver is flashing the 9 LED string at about a 40KHz rate! Thinking just that circuit may have had some instability we took a new production ready board and ran in with a six LED string on our bench, We are seeing a well shaped, full on/off LED current pulse at a 22KHz rate with a tightly defined 2us off time. Exploring the cause of such a clean pulse has lead to finding a 2us, 22KHz pulse on the DDRV pin of the chip. We are at a loos to understand how this can be since the dim input is not active, Enable is solidly on, the OVP input is low and Vcc is ok and not noisy. Changing the input voltage does alter the pulsing width. At 30V it is about 300ns, at 20V it is 1.7us at 20KHz, and at 12V is about 8us at 16KHz.. There seems to be a charging mechanism going on but what sets the threshold and how does the signal get out on DDRV?
The board was carefully laid out to have short shielded paths and as much return plane copper as possible. Any help with identifying the cause is greatly appreciated.
Steve in Cincinnati
