[time-nuts] AD9832 in FE-5680A

[Murray Greenman]

Bert,
Your comment about the AD9832 being a 25MHz part made me think, and I too wonder how it might all work. The data sheet calls it a 25MHz part, and I agree with you that it might be expected that some extra division occurred between the ~60MHz Rb loop and the synthesizer MCLK.

That's not the case however. I don't know how it's achieved, but the FE-5650A and FE-5680A (the versions with AD9832 and user-programmability, I mean) do in fact appear to operate with a ~60MHz MCLK. I support that statement with two facts:

1. The AD9832 in these models is capable of generating moderately good 12MHz output, which it certainly would not with a 25MHz MCLK.
2. The frequency calculations used give you a step size of about 10mHz, and as you will realize, 50MHz/2^32 is about 10mHz. I've used the ~50MHz figure (actually either 50.255MHz or 60MHz) in my software.

Does anyone know how we can resolve this apparent inconsistency?

73,
Murray ZL1BPU