SolderSmoke Podcast #157 is available for download.
http://soldersmoke.com/soldersmoke157.mp3
December 27, 2013
-- Santa Report: Rigol Scope at SolderSmoke HQ. Kites at VK3YE.
-- Project updates: Bill's BITX 40/20 build.
-- Peter finishing up BITX40 (in a big box)
-- BITX 17 (5 watts SSB) works JA, ZD7, ZS!
-- The mystery, elegance, and gentleness of phasing SSB
-- Phasing explained in 1970 ARRL SSB book
-- Phasing SSB: From Hallicrafters HT-37 to SDR
-- The SP5AHT Phasing Rig
-- I & Q for you: The Binaural Experience
-- Direct Conversion receivers and Software Defined Radios
-- Simple DC receivers plugging into sound cards
-- The joy of receiver building
-- 144 MHz aircraft bounce (Melbourne to Sydney)
-- VHF Across The Great Australian Bight
HAPPY NEW YEAR TO ALL!
Our book: "SolderSmoke -- Global Adventures in Wireless Electronics" http://soldersmoke.com/book.htm Our coffee mugs, T-Shirts, bumper stickers: http://www.cafepress.com/SolderSmoke Our Book Store: http://astore.amazon.com/contracross-20
Excellent! I've been waiting for the follow up!
ReplyDeleteHappy Holidays to all (and congrats on the new scope Bill!)
W9PDS
Nice to hear you and Peter chat more. I'm eagerly awaiting my copy of Sprat to read about Peter's latest. I thought I'd pass along a supplier that I've used before that's super cheap for those "bulk component" buys. Try http://www.taydaelectronics.com out. You can get 2N3904s for $0.02 a piece. LM386s are somewhere between $.23 and $.36 each. They stock a wide variety of caps, resistors, diodes and the like. They don't have the selection of digikey or mouser, but what they do have, they have really cheap.
ReplyDeleteBill,
ReplyDeleteGood news about your new o-scope.
Since my trusty Hameg packed up I'm considering an actual PURCHASE (as usually people thrust this stuff on me to repair and keep if I can get it working).
So when you have the time please do a soldersmoke review of the Rigol DS 1052E .
73
Tony
I-Q SDR seems like magic but is really quite simple. I hope the following explanation is of some help, but it might be old news to you. Or it might be total gibberish.
ReplyDeleteThe following examples consider the reception of simple carriers. Since all signals can be viewed as the sum of a bunch of carriers (sinusoidal components) at different frequencies, these examples generalize to all types of signals.
The image problem in receivers is well understood. For example, a 7000 kHz oscillator mixed with a received carrier at 7020 kHz yields sum and difference signals of 14020 and 20 kHz. The higher frequency component can be filtered out with a low-pass, leaving the desired 20 kHz baseband signal. Unfortunately, a signal at 6980 kHz sent through the mixer will also appear at 20 kHz, interfering with the desired signal. IFs and sharp crystal filters are typically used to kill the unwanted sideband.
Now consider a direct-conversion I-Q receiver. Its mixer outputs two signals instead of one - I and Q.
In the I-Q mixer, carriers 20 kHz above or below the oscillator frequency (at 6980 or 7020 kHz) will both cause 20 kHz sinusoids to appear on the I and Q outputs of the mixer. We seem to be in the same predicament as before, unable to tell the sidebands apart -- but no! By looking at the phase relationship of the I and Q sinusoids, the SDR software can tell if a carrier piped through the mixer is above or below the oscillator frequency. If there is a signal at 6980 kHz and you only want to receive a signal at 7020, the 6980 signal can be killed.
A carrier 20 kHz above the oscillator frequency results in a 20 kHz I output which leads the 20 kHz Q output by 90 degrees.
A carrier 20 kHz below the oscillator frequency results in a 20 kHz I output which lags the 20 kHz Q output by 90 degrees.
Now, if you command your SDR software to receive signals between 7020 and 7021 kHz, it will do the following:
Since we want to get rid of all frequency components below the oscillator frequency (7000 kHz), it will first filter out all components where I lags Q -- the "negative" frequencies. Now only components where I is leading Q -- the "positive" frequencies -- remain.
Of these remaining signals, it will null out frequencies outside the +20 to +21 kHz band. That 1 kHz band is then shifted to 0 Hz.
Finally, the SDR software throws away the Q signal. The remaining I signal is your mono RX audio.
How the SDR software separates the I-leading-Q and I-lagging-Q components of the signal is a whole other can of worms involving complex numbers. But the basic intuition that USB signals have I leading Q and LSB signals have I lagging Q is much of what you need to know.
An analog is a computer mouse (or a dial on most digitally tuned radios nowadays)
ReplyDeleteMoving the mouse left or rignt just gives a series of pulses, the number indicating how far the mouse moved. But just counting pulses doesn't give direction.
However, there are two sensors, arranged 90degrees apart. So the stream of pulses from each is identical, matching how far the mouse moves, but one lead sensor leads or lags behind the other, depending on which direction the mouse is going. It's that information about ahead or behind that allows the following circuitry to know which direction.
Michael
Michael and W1CJM: Thanks for the thoughtful comments on I and Q. I will study them shortly.
ReplyDeleteTony: I like the Rigol, but go for the 100MHz model.