[aprssig] slightly OT talking about 9k6

Charles Blackburn ai4ri at cfl.rr.com
Sat May 7 22:13:21 CDT 2016

Sounds like an interesting idea, I would however be more inclined to use a gpio for PTT purely and simply because then you can have handshaking on the serial line, and also then control more than one ptt at the same time.

I'll have to read through this more when I have a little more sanity than I do at the minute (read ex wifre and kids lol)


-----Original Message-----
From: aprssig [mailto:aprssig-bounces at tapr.org] On Behalf Of Ross Whenmouth via aprssig
Sent: Saturday, May 7, 2016 18:46
To: aprssig at tapr.org
Subject: Re: [aprssig] slightly OT talking about 9k6


I am working on the design for a "Pi-hat" audio card for the Raspberry pi, specifically for amateur radio use (including APRS and 9k6 packet). 
I am targeting ~ 4Hz to ~ 90 kHz audio input bandwidth (DC - 90 kHz output), compatibility with 600 ohm balanced interfaces (and unbalanced interfaces too), low QRM, complete galvanic isolation and good common mode rejection.

The plan at this stage is to use 2x INA137/SSM2143 line receivers, 2x
DRV134/SSM2142 line drivers and an AK4556VT 24bit 192kHz stereo codec with galvanic isolation between the Raspberry Pi and the codec/line transceivers. At this stage I am looking at using something like the
Si8641/ISO7841 to galvanically isolate the I2S digital audio stream, with opto-couplers for the 2x PTT outputs.

The expected transmit audio out response goes all the way down to DC, but the receive audio input rolls off below about 4 Hz - according to the plot on page 8 of http://sss-mag.com/pdf/gmsk_tut.pdf it looks like a 4 Hz rolloff is unlikely to be a problem for K9NG/G3RUH 9k6 packet/APRS. The RX chain, all the way to the codec ADC is DC coupled, however, if the digital DC-block filter in the codec is disabled, the codec can only be operated as an I2S slave (I plan to operate the codec as a master with its own 24.5760 MHz clock, because when the Raspberry pi is operated as an I2S master, its I2S clock output has poor phase noise). Yes, I could choose a different codec, but the AK4556VT also seems to be very easy to interface with, and hand-solderability (only
just!) at 192 kHz sampling rate are not such a common combination.

I am considering attaching an STM32F103C6Tx/STM32F103C8Tx to the I2S bus and Raspberry pi GPIO on the Raspberry pi side of the galvanic isolation barrier, and using this MCU to control the PTT optocouplers. The idea is to be able to software configure PTT watchdog timeout (short for APRS, long for WISPR, off for a DMR repeater, etc), and for PTT control by either I2S triggered VOX, or Raspberry Pi GPIO (eg use RTS for the Raspberry Pi's uart AMA0).

Ross Whenmouth
aprssig mailing list
aprssig at tapr.org

More information about the aprssig mailing list