Why is the differential impedance of RFIP/M and RFOP/M of AD8340 50 Ω nominal? Usually, the differential impedance of the most RF devices is 100 ohm. If AD8340 RFI and RFO are connected to these devices, how to match the differential impedance? Is there any application note for this available?
This is inherent in the design.
Please see the applications sections in the datasheet titled "RF INPUT AND MATCHING" and "RF OUTPUT AND MATCHING". Figure 30 also gives a good visual on how to convert the signaling to single-ended.
Thanks for your reply. I have studied the corresponding section.
But what I am concerned about is how to connect 50ohm differential impedance to 100ohm differential impedance. In my design, all the RF signals are differential signal with 100ohm differential impedance, but not the single-ended signal. so how do I do?
You could connect directly to the 100 Ohms if your system can handle the degradation to the return loss.
Else, I would use a 2:1 transformer to match from 50 Ohm differential to 100 Ohm differential.
Yes, I want to use a 1:2 balun to transfer the diff impedance. In fact, Iwant to know the reason why this device has a 50 diff impedance, while most of other devices have a 100ohm diff impedance. :)
This part was released in 2004. I do not believe 100 Ohm differential was nearly as common then, as it is now. 50 Ohm single-ended was much more common then.
The impedance had to be chosen to be something, and the device architecture is inherently differential. 50 Ohms differential was chosen to to easily interface with 50 Ohm single-ended via a 1:1 balun, given the popularity of 50 Ohms single-ended at the time.
I thought about it again yesterday, and it was still a little strange.：）
If, as you said, 50 ohm single-ended impedance was more popular 10+ years ago, then why not design the RF signal of this chip directly into a 50 ohm single-ended signal, so that even a 1:1 BALUN is not necessary for its connecting to a 50ohm single-ended signal source or next stage payload.
The part not having single-ended interfaces at the input and output could have been for several reasons, I really don't have the specific details.
The part is 50 Ohms differential input & output; that is not changing. So some sort of matching is required if you do not want to degrade the return loss.