I've recently obtained samples of the HMC470A attenuator and I've produced a small PCB for test purposes. The board is double-sided with a solid ground plane on the bottom and plenty of vias connecting to the top side ground. The circuit is roughly the same as the application circuit shown on page 5 of the data sheet, except that I have 100nF couplers at pins 2 and 11 and all ACG pins are decoupled to ground by 100nF caps. I have 10k pull-ups on pins 12 - 16. VDD is decoupled by 100nF and 100pF in parallel. I'm using a 50ohm test system.
My intention is to use the device as an output attenuator on an RF signal generator operating from around 6MHz to 600MHz. My maximum input power across this range is +18dBm, well below the +27dBm maximum for this device.
Performance appears very good for frequencies above 20MHz. However, at lower frequencies, signal distortion becomes apparent.
For example, at 5MHz with no attenuation (pins 12 - 16 all high) signal purity is OK for input powers below +10dBm. However, at +12dBm distortion becomes apparent, and at +16dBm the sinewave is severely distorted. At 1MHz, there is significant distortion as low as +7dBm.
I do not believe my layout is responsible for these problems because the part performs very well at frequencies above 20MHz.
The datasheet suggests that the device can operate down to DC provided the external caps are large enough, so I'm surprised to see these problems.
I'd welcome any suggestions.