Our carrier suppression is about 20 dB. Is there any internal adjustment (initialization) that I can perform to increase performance? Something I am perhaps missing?
- Tx or Rx?
- What software are you using? No-OS or Linux?
It's on the TX side. I am using your No_OS driver, your initialization etc.. That is working great! I just feel I'm missing something.
I moved this (for now) to the no-os support, we can check with Dragos, he can make sure you are turning on the Tx DC correction, and then if needed, we can move it back to the chip support section.
We expect to see about -50dBc carrier rejection at 0dB attenuation.
The issue is most likely due to part configuration. Common problem is that Rx LPF BW and/or test tone frequency for Tx Quad cal are not set correctly.
Our design is using Tx2A_N only. The other Tx outputs are not used.
Tx2A_N only? Tx outputs cannot be used in single ended configuration. Can you share a schematic of your Tx output configuration? This may not be the only issue. Calibration/configuration issue may still be present.
After further review of your docs, we no are design is incorrect and will require a re-spin but, somewhere we thought we read, connect any unused Tx output to a supply. Is that true? We now understand Tx cannot be used as single ended outputs. Any advise you can give us will be appreciated.
Yes, your design will need another iteration. I recommend using a balun with a center tap for Tx output DC bias.
Tx outputs are differential. Unused Tx outputs need to be terminated by connecting them to 1.3V.
Rx inputs have the option of being used as single ended or differential, but Tx ports can be used only differentially.
Also, what should we do with unused outputs?
AD9361 Reference Manual (UG-570)
The Tx outputs are differential and require an external output bias. These outputs need to be biased to 1.3 V DC supply voltage (nominal)
Connected to an antenna, or not -- the requirements are the same: 1.3V termination.
Check out the FMCOMMS2/3 schematics.
Check out TX2B_N, TX2B_P - which are "unused" (but not disconnected) - they are tied to a clean 1.3 V supply.
For I/Q/DC issues, you can try to run the TX Quad Cal with a calibration phase offset value(AD9361 Transmit Quadrature Calibration (Tx Quad Cal)). For example:
ad9361_do_calib_run(ad9361_phy, TX_QUAD_CAL, 13);
Try to do this after after the part was initialized.
If you update to the latest No-OS or Linux AD9361 drivers.
It'll find automatically the best fit Phase Offset Value for the TX QUAD Cal.
We are concerned about board space and would like to span the entire frequency range support by the AD9361 (70 - 6000)
Thanks to you both
I know the wide band board that we have is using Mini Circuits TCM1-63 the RF transformer.
http://www.minicircuits.com/pdfs/TCM1-63AX+.pdf (4.06 x 3.81 x 4.06mm) - which isn't exactly tiny compared to a balun in an 805 footprint.
There just aren't many (any) very wide (70 - 6000 MHz) baluns that I'm aware of in very small footprints. Suggestions welcome.
From a previous response:
Where and how do you set these parameters?
One of the responses was:
Can you give me more details? Rx LPF BW - test tone? How are these set? Is an external test tone required?
If you are using the latest version of the ADI provided driver.
It should take care for the Rx LPF BW, test tones and phase value.
Retrieving data ...