Wednesday, August 31, 2011

Some things that need doing

There are a number of opportunities to improve parts of Ham Radio. This posts collects some random thoughts along those lines
  • Automatic HF beam forming. The idea is to create multiple antenna patterns that put nulls in the direction of interference. The algorithm can make multiple beams simultaneously. Extend the algorithm to put a null in the direction of static crashes, too. The multichannel Softrock is built and working. The next step requires implementing a test case in software. The SVD math has me stumped at the moment. Sometime it work, sometimes not.
  • Faster JT-65b timing. The default is one minute intervals but with strong signals, the timing could speed up. Several approaches might work. Short sequences that break and listen to estimate signal strength. Or, take advantage of knowing own stations runs high power, if the other guy has a strong signal, we must have one as well.
  • Wide band staring demodulator for FSK-441. EME stations can work without spotting assistance using MAP65. But Meteor Scatter mostly needs a schedule to work a station. Operating procedures for random contacts, tail ending, and spotting don't work very easily. Develop a staring receiver that demodulates all the MS bursts across the MS band. Develop operating procedures that have stations pointing at selected common points in space.
  • Use JT65b or FSK-441 as an alternative to CW. Stations sometimes switch to CW to work if SSB doesn't get through. For example in the VHF contests, W2SZ runs 1500W and stations calling often run 10 to 100W. The could benefit from a box that transmits call signs, grid square, and QSL using a digital protocol. JT65B works for weak signals in noise, FSK-441 could cut through interference from SSB. The big gun would run a JT-65b and FSK-441 decoder on the Rx output all the time.
  • Change contest scoring for the microwave bands. The ARRL contest rules give a lot of extra points for making any kind of contact on the microwave bands, probably as part of an effort to get people on the bands and to develop equipment for the bands. But we don't advance the state of the art with some kludged up CW radio for 5760 MHz. Those bands have good point to point data link capability and good local area network capability, but they will never be a rag chewer band or a random contacts band. The contest rules should support advancing the technology with more appropriate type of contacts. How about getting points for putting up a microwave link between two stations and maintaining it?

Tuesday, August 30, 2011

Solid State Amp Controller out to fab


The SSAC board went out to Sunstone for fabrication and will arrive about Sept 6. This board contains two channels of RF sampling, Relay controller, and 50V, 50A FET switch. This all fits on a 2.35 x 2.8 inch board. The board uses the Arduino Shield interface.

The 50A potential current flows through some traces that do not have enough width if you believe design guidelines. 1 oz. Copper has about 500 microOhms per square, which dissipates 1.25 Watts per square at 50 Amps. Most of the connections are less than 1 square, but that just means they're small. This board, and the FET transistor will mount to SilPad thermally conductive sheet to conduct heat away from the board.

Saturday, July 23, 2011

Solid State Amplifier Controller or Transistor Control Board

Freescale has released a new VHF/UHF transistors that has excellent ruggedness, reasonable matching, and high power output, the MRFE6VP61. Every amplifier needs a control board, so this is my attempt.

This controller design consists of a PCB with he following major blocks
  • 20 x 4 character LCD status and metering display
  • Analog meter drive
  • Drain Voltage solid state power switch
  • Drain current measurement
  • Forward and reverse power measurement
  • Relay sequencing for 28V input, output, and tower mounted coax relays
  • 28V to 5V power supply
  • Arduino Uno microcontroller socket
The initial design implements the controller in a box with the following features
  • On/Off switch which functions as fault reset
  • 20x4 character metering and status display
  • Analog meter for RF Power Out
  • 50V 50 Amp Input and output connectors using 75 Amp Powerpoles
  • EMI Filtered AC input
  • 28V power supply for Relays and SSAC
  • External Key input, ground to enable
  • External relay connector for Input, Output, and Tower Mounted Relays
  • Current limited switched 28V relay outputs
The Arduino Uno processor has a very easy to use programming interface that support good I/O, display, and fast response to events. The code resides on my SSAC_Arduino public svn.

The schematic and layout files reside on SSAC_Eagle svn repository.

The parts datasheet files, amplifier design documents, etc. reside on the SSA_Controller svn repository.

I wish these could all reside in one place, but apparently not. SVN doesn't use links, probably for good reason. Maybe, I can get Arduino and Eagle to use them...we'll see.

The SSA Controller implements a "shield" in Arduino terms that mates to the Ardiuno. The SSAC mounts to the metal enclosure to dump a few watts of heat and the Arduino mates to that.

Wednesday, March 30, 2011

This image of the finished four channel HF receiver comes from a scanner. They have good depth of field and maintain good focus ever with significant component height. The original image was scanned at 600 dpi. I don't know what blogspot will do to it.

The SMA input connector (for 7 MHz?!) go out the back side. This will go on the lid of a 4x5 die cast box.

The input network covers the 40M band.

Sunday, March 27, 2011

Rx_4 Update

The first two channels appear to work about the same...still need to wind the input bandpass filter inductors for the remaining two channels.

Did you know that Minicircuits marks their case with the dot on Pin 6? I didn't...until checking the data sheet more carefully.

QSD_EN is really QSD_EN/ ... which is inverted from where it needs to be to have PTT/ mute the receiver.

Four Channel HF Softrock

Blind beamforming uses Singular Value Decomposition to form beams automatically that separate multiple overlapping signals. This sounds like a good idea for the HF bands. The link goes to one paper, there's a lot more, almost all of it over my head mathematically.

I'm in the process of playing with some simulation code. K1LT sent me some DVD of data streams captured from his 4 antenna linear array for 160M. The initial results seem promising, but more development requires my own multiantenna array.

I designed a variant of the Softrock Ensemble II receiver with 4 channels using Eagle schematic and board layout tools and upload the results regularly to my Subversion repository.
  • http://www.wa1hco.net/Rx_HF_Multichannel/
Here's links to .pdf files
The /trunk/ directory contains the results of debugging and will be the V2 design if it gets built. The /branches/V1/ directory contain the exact files that went into the circuit card fab.

Current status is that the USB interface and Si570 control appears to work under usbsoftrock software. Quisk recognizes the card, but tuning doesn't resulting changing frequencies. I'm just starting the process of debugging the RF signal flow, QSD mixer and analog amplifier.