ad9910/ 9954 modules

[phenol]

i recently bought the above two modules on aliexpress.
haven't been able to test them yet as i have no supporting documentation.
the seller is YourCee. i have asked them to send whatever they have, but still...
does anyone have schematics?
 

[iMo]

Looks like the datasheet schematics implementation, nothing special there on the pcb except the DDS chip itself (you have to download its datasheet first and read)..
Follow the signal's labels at the headers. You would need a microcontroller to setup the DDS (perhaps you may find the ready drivers for popular ones), and you will get signals out of the SMAs.
Only thing you would need for the setup of the DDS is the frequency of the crystal oscillator there on the pcb (you will find the freq written on its case).

[phenol]

thanks, iMo
looks like it, but I'd like to know the output filter component values.
The on-board oscillator is a 40MHz cmos output thing overdriving the ref input with a 3.3v square wave.  this one will go in the bin. I have the original 9910 AD demo board for which i have a direct drive 1GHz reference source and a small controller handling basic cw tone generation

[iMo]

Yup, the external 1GHz source is good for better spectral purity. I would expect 400MHz low pass on both channels, for initial testing purposes. What real part values have been populated there is a different story, of course..
You will get it when you start to sweep with the output frequency.. 

[phenol]

i figured it out and while it works, the ref clock sensitivity is some 10db worse than AD's demo board. not sure if they populated it with subpar chips but upon closer inspection it becomes obvious that the board is poorly laid out with decoupling caps placed far from the power/gnd pins. the chip designers have thoughtfully located gnd pins next to every Dvdd/Avdd/clk_vdd pin but the board designer failed to make proper use of them

[phenol]

i soldered 0402 decoupling caps very close to all avdd pins but it didn't improve anything.  input sensitivity is nowhere near the claimed 50mv p-p figure in the datasheet. either those chips are scrap or the ref clk inputs are shot thanks to blasting them with 3.3v swing from the onboard 40mhz oscillator. the max swing there shall not exceed 1000mv p-p as per the datasheet.
the board has provision for a pi attenuator from the oscillator to the ref input but it's jumpered with a 0r resistor instead and a 20p cap to gnd... i cant help but wonder why