Ideas for v2.1

[tejeez]

Let's collect some ideas for the next batch of PCBs here.

I'd rather not maintain a separate firmware for v2 and v2.1, so it should be mostly pinout compatible with v2. If something is changed in the pinout, the existing v2 boards can be modified to work with latest firmware.

I decided to not focus that much on mechanical design yet, so I think v2.1 will just use the same enclosure and similar construction as v2.

List of some ideas:

• Fix enclosure mounting hole positions.
• Remove the BJT microphone preamplifier and try using the internal opamp in EFR32. This would reduce component count and allow adjusting microphone gain from the firmware. This should be tested first by modifying v2 boards.
• Should we still use ETC4-1-2 as the balun for EFR32 matching or find a replacement from JLCPCB?
• The class D audio output stage does not have proper filtering. I thought voice coil inductance of a speaker might be enough but haven't actually measured it. Some people, however, have added a headphone connector. How bad is it to continuously feed a ≈1 V 192 kHz tone to headphones?
• https://github.com/tejeez/geckokapula/issues/2
• Make unfiltered RF signals easier to access. EFR32 can tune wider than expected so it would be nice to allow some experimentation. Maybe add some 0-ohm jumpers to bypass filters.
• How about replacing the 3.3 V LDO with a switching regulator for better battery life and less heat? The regulated voltage could be also reduced a bit (say, 3.0 V), so that it maintains regulation even with a bit less battery voltage.

[tejeez]

This could work as the balun: https://jlcpcb.com/partdetail/MurataElectronics-DXP18BN5014TL/C152552

Or two of these connected together in a similar way as a 4:1 balun: https://lcsc.com/product-detail/Balun_Sunlord-Sunlord-BW21S7511A01TF_C138776.html These even specify power and RMS current handling. Looks like they should handle the PA DC bias current just fine.

[tejeez]

* Fix enclosure mounting hole positions.

I decided to not design v2.1 for any specific enclosure, so this is not an issue anymore.

* Remove the BJT microphone preamplifier and try using the internal opamp in EFR32. This would reduce component count and allow adjusting microphone gain from the firmware. This should be tested first by modifying v2 boards.

This is now done.

* Should we still use [ETC4-1-2](https://www.partco.fi/fi/tarjouseraet/3741-tarmuu-etc4-1-2.html) as the balun for EFR32 matching or find a replacement from JLCPCB?

Balun are BW21S7511A01TF now.

* The class D audio output stage does not have proper filtering. I thought voice coil inductance of a speaker might be enough but haven't actually measured it. Some people, however, have added a headphone connector. How bad is it to continuously feed a ≈1 V 192 kHz tone to headphones?

I decided to use a class AB amplifier driven from a DAC instead.

* [Reserve a pin for clock output #2](https://github.com/tejeez/geckokapula/issues/2)

Done

* Make unfiltered RF signals easier to access. EFR32 can tune wider than expected so it would be nice to allow some experimentation. Maybe add some 0-ohm jumpers to bypass filters.

Done

* How about replacing the 3.3 V LDO with a switching regulator for better battery life and less heat? The regulated voltage could be also reduced a bit (say, 3.0 V), so that it maintains regulation even with a bit less battery voltage.

Now that audio amplifier is powered directly from battery voltage, consumption from 3.3 V is not that much. I kept a simple 3.3 V LDO there. EFR32 DAC has its own internal reference too, so even if 3.3 V line stops being regulated when battery goes below 3.3 V, it should not affect audio that much.