I was afraid of that...and was an issue I have on my punch list to check out when I get my modules.
I'll know for sure when I see the circuit behavior on an oscilloscope. But I'm fairly certain the quick flash is from the short start-up interval for programmed microcontroller IC chips. This interval is typically measured in the milliseconds or a fraction of a second. Basically it's like the boot-up time for a PC/laptop where power is being consumed but no useful work is accomplished.
As the diagram above attempts to illustrate, following the application of 5V power to the module, the microcontroller takes a fraction of a second to "boot up" during which time it has not figured out that it is a delay-off controller so the output is essentially off. It then figures out that its purpose in life is a delay-off widget and immediately turns on its output for 10 seconds. Since the output is obviously OFF when power is OFF, this short start-up delay from the application of power to the output turning ON is not noticed.
But when this output goes thru the inverter (2N3904), this short start-up time flashes the LED.
OK. That was a long-winded explanation that does little to correct the behavior.
I'll need to see the circuit behavior in action to understand the parameters of the start-up timing. It can probably be masked and my objective would certainly be to do so in a simple/inexpensive way.
