We have implemented the AD9913 in a design very simular to that of the AD9913 evalution board.
Connections to AD9913 chip:
- Reference clock: 25 MHz Crystal Oscillator
- Seperate 1,8V power supply for analog and digital pins.
- RSET = 4,64K
- A out and B out terminated in 200 Ohm resister to analog ground
- Reset controlled by uC
- Power Down input shorted to digital ground
- IO - Update input controlled by uC
- SER/PAR input shorted to digital ground
- CS input controlled by uC
- CLK input controlled by uC
- parallel adress lines inputs unused and shorted to digital ground
- Internal PLL set to 250 MHz
We are using the AD9913 in a single tone mode, generating a 3MHz output. Programming the device is no problem and the output waveform looks fine.
However the device seems to spontaneously reset sometimes.
Most of the time this means the 25 MHz reference clock stops and the output waveform stops.
Other times there is still some kind of oscillation present at the output but it looks very spurious, impossible to trigger on.(What is very strange is that in this situation there is no reference clock present, but there is still some sort of output signal)
A few times the 3 MHz output simply dropped to 84 kHZ.
By performing a reset (active high) and reprogramming the device we can restart the device.
We can repeat this behavior by generating a small ESD puls in the neigbourhood of the device, this pulse does not interfere with other digital components near the AD9913, such as a µC and memory..
We have tested 3 boards so far, all with the exact same behaviour.
Have you seen this behaviour before? I must emphasize that the ESD pulse is very small.
Any help is appreciated!