Ghpsdr3 protocols

From HPSDRwiki
Revision as of 00:28, 11 August 2010 by WA8YWQ (Talk | contribs) (Initial translation from .pdf)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

server HPSDR hardware server

The HPSDR server handles communications with the HPSDR hardware and is designed to work with multiple receivers either on a single Mercury card or multiple Mercury cards. It demultiplexes the I/Q samples into separate streams for each receiver. The hardware server listens on port 11000 for TCP connections from clients. The client sends commands to the server over the TCP connection. IQ data is sent over UDP.

  • Commands to server
    • attach rx
    attach client to receiver rx
    The receiver number rx is an integer from 0 to 3
    example: attach 0
    • detach rx
    detach client from receiver rx
    example: detach 0
    • frequency f
    Set attached receiver frequency to the integer frequency f (Hz)
    example: frequency 7056000
    • start iq port
    start sending I & Q samples for the attached receiver to the UDP port specified
    example: start iq 13000
    • start bandscope port
    start sending bandscope samples to the UDP port specified
    example start bandscope 12500
    • stop iq
    stop sending I & Q data to the attached receiver
    • stop bandscope
    stop sending bandscope data

The server sends a response to the client after parsing each command. The response is either OK or ERROR. Some responses include a short message.

A typical exchange would be:

ClientServer
attach 0OK 96000
set frequency 7056000OK
start iq 13000OK
set frequency 7056100OK
stop iqOK
detach 0OK

The I/Q samples are accumulated in the server until there are 1024 I/Q samples to send. The samples are sent in 512 byte UDP packets, to the port specified in the 'start iq' message, with 12 byte header and up to 500 bytes of I/Q samples.

The UDP packet header contains:

  • 8 byte sequence number
  • 2 byte offset within this set of samples
  • 2 byte length of samples

Note that there are currently no commands to control setting the number of receivers, preamp, dither, random, samplerate or clock sources. As this was developed to support multiple receivers, I decided to not allow clients to change these as they would be changed for all receivers. They are all controlled by command line arguments when running the server.

server command line options

  • --receivers <1..4>
  • --samplerate [48000|96000|192000]
  • --dither [off|on]
  • --random [off|on]
  • --preamp [off|on]
  • --10mhzsource [atlas|penelope|mercury]
  • --122.88mhzsource [penelope|mercury]
  • --micsource [janus|penelope]
  • --class [E|other]

Example:

./server --receivers 4 --samplerate 96000 --dither on --10mhzsource atlas

Note: As an experiment, there is also a softrockserver that supports the same protocol with the restriction that there is only one receiver and it uses the USB/si570 interface to control the frequency. The I/Q samples are taken from an Audio device specified on the command line.

dspserver

dspserver is a client to the hardware server. In turn, it is the server for another client, typically a GUI, such as jmonitor. It provides a dsp processing service using DttSP. Typically one dspserver is started for each receiver. The dsp server listens for client TCP connections on port 8000+rx. i.e. rx0 listens on 8000, rx1 listens on 8001, etc. The client sends commands to the dspserver over the TCP connection, and dspserver sends data to its client on the same port. The first byte of each packet sent to the client identifies the type of data in the packet.

  • 0 spectrum data
  • 1 audio data
  • Commands to dspserver
    • getSpectrum points
    Request points of spectrum data. The value sent for each point is an 8 bit unsigned integer. If the calculated spectrum contains more points than the client requests, adjacent points are combined, and the highest value in the group is sent.
    Example: getSpectrum 480
    • setFrequency f
    Set the receiver frequency to f, in Hz
    example: setFrequency 7056000
    • setMode mode
    mode is an integer specifying the type of demodulation
      • 0 - LSB
      • 1 – USB
      • 2 – DSB
      • 3 – CWL
      • 4 – CWH
      • 5 – FM
      • 6 – AM
      • 7 - DIGU
      • 8 - SPEC
      • 9 - DIGL
      • 10 - SAM
      • 11 – DRM
    example: setMode 0
    • setFilter low high
    Set filter cutoff frequencies
    For LSB, both frequencies are negative
    For USB, both frequencies are positive
    For AM or DSP, typically low is negative and high is positive
    example: setFilter 150 3450
    • setAGC n
    Select AGC time
      • 1 - LONG
      • 2 - SLOW
      • 3 - MEDIUM
      • 4 - FAST
    example: setAGC 3
    • setNR n
    Set noise reduction
      • 0 - off
      • 1 - on
    • setNB n
    Set noise blanker
      • 0 - off
      • 1 - on
    • setANF n
    Set automatic noise filter
      • 0 - off
      • 1 - on
    • setRXOutputGain gain
    Set receiver audio output level to the value gain
    The value is a floating point number in the range 0 to 1.
    example: setRXOutputGain 0.5
    • startAudioStream buffer size
    start sending audio data.
    buffer size is optional, and is an integer, typically 480.
    If omitted, a value of 480 is used.
    • stopAudioStream
    stop sending the audio data.

data packet format Each packet has a 48 byte header.

Byte Function
0 Packet type: 0=spectrum data; 1=audio data
1-31 Reserved
32-39 Sample rate
40-47 Meter reading
48-end spectrum or audio data

Audio data is sent as 1 channel 8-bit aLaw samples at 8000 samples per second.

Sample rate and Meter reading numbers are in ASCII characters.