Source code released under GPL license is now on my site:
formatting link
To unpack: tar -zxvf pwfax-0.1.tgz
cd pwfax-0.1
Follow instructions in INSTALL, More info in README
I decided to add Goertzel algo for detection of the start and stop tones, as mentioned here:
formatting link
The whole thing is purely experimental at this point. So YMMV.
I get pretty decent decoding also if lots and lots of noise is present, just let my digital recorder[1] run all night recording from the Tecsun SSB receiver at 8038 kHz, and extracted about 12 to 15 maps, most of these without needing correction.
Deliberately only used xlib, this makes the program very small in code size. Maybe or maybe not I will add some mouse callbacks from the X display later.
[1]
formatting link
Pretty nice thing, line input, straight into the Tecsun receiver output.
Interesting! I did this about 20 years ago with much more mundane eqipment. I used an R-392 vacuum tube receiver, A CV-57, which is a narrow band 2nd receiver used for converting RTTY FSK signals to TTY current loop, but I picked off the linear output from the discriminator and fed it into an ADC, fed that into a uVAX and decoded the sync pulse.
On a sunny day (Sat, 24 May 2014 21:39:23 -0500) it happened Jon Elson wrote in :
Yes, a narrow wand FM should work. I should try rtl_sdr (well actually already did) and rtl_fm with a RF converter some day. But the RF converter I have is not really selective.. This Tecsun radio is pretty good. On the boat in the middle of the ocean in a storm (when you need those charts most) you do not want anything complicated. Recording away from the computah and other electronic equipment gives a much cleaner signal, and computah cannot be on all the time (power), so letting the small radio and recorder run unattended, and then in a short time later process the (several hours) recording that is the way to go. It runs great unattended.
Yes I had a Trio tube short wave SSB receiver in the sixties, it drifted, you had to constantly adjust the BFO. Would be useless. These modern radios are true miracles, just type in the frequency and it stays there. Also I have about 50 or 60 weather fax frequencies worldwide in its memory (it can hold 600 frequencies). Just pick one. for the area where you are. Now that brings me to celestial navigation, you know, sextants and so, GPS can break down, or be switched off. Context!
Soundcards typically have a sample clock tant runs independant of the CPU and the interface bus. feeding the sampled data into system RAM via DMA (or isochronous transfer on USB).
There's buckets of latency and some jitter in the arrival time of the packets into RAM, but the samples themselves are as evenly spaced. as the clock on the sound card can manage.
On a sunny day (29 May 2014 07:07:56 GMT) it happened Jasen Betts wrote in :
That is one answer i contemplated giving to Robert, but my brain locked, as it is not complete. It is probably true for 'soundcard' as one you plug in. But for example in the Raspberry Pi _I THINK__ and I am not 100% sure, one of the Broadcom chips PLLs is used as audio clock. The Raspberry can be used as a great programmable signal generator from about 61 Hz to 250 MHz! My signal generator...
formatting link
There ARE jitter situations, and remedies, I copied some of the Broadcom datasheet into the source, see that for jitter specs, or look up: BCM2835-ARM-Peripherals.pdf For integer division that jitter should be zero, so YMMV.
Anyways, Raspi only has audio out...
Any task switching interrupt related delays can also be fixed with a simple FIFO, I do that here on the Raspi for DVB-S video transmission:
formatting link
The symbol rate is set by that same PLL I just mentioned, saving on an other crystal, and making software controlled symbolrate possible So a 'universal' reply is not easy, and mentioning all or rather many possibilities does not fit the scope of this newsgroup I think.
On a sunny day (Thu, 29 May 2014 21:42:02 -0700) it happened Robert Baer wrote in :
No you dont[1].
That has nothing to do with it.
Yes it is.
DeCent, scent, dee-cent whatever. be more specific, write some code. Did you know RELATIVITY comes into play too? If the thing moves close to light speed THAN things happen. And do not forget Doppler, if you move away from your spee-cars all frquencies are no longer matching the song you whistle, you will have to re-tune your mouth. and while you move closer to have a look at the hardware, frequencies will go up. This is serious stuff any 1 interested in true High Fiddle-letty shoot pay a tension too.
[1] Just use the NMI to send the next audio sample
When connecting a receiver to a computer, you really should use galvanic isolation to keep the receiver ground separate from the dirty computer ground (especially if the radio does not have a real ground electrode).
Suitable isolation transformers can be found from old telephones and modems. Adding some ferrites on the audio cable close to the sound card will also help.
In this way, long audio cables can be used, so the receiver can be moved far away from the PC, to prevent direct radiation from the computer to the built in antenna.
On a sunny day (Fri, 30 May 2014 09:52:02 +0300) it happened snipped-for-privacy@downunder.com wrote in :
Audio transformers can have several hundred pF between primary and secundarty, and are a short at HF.
And why bother, my system, that you snipped, recording on a simple battery powered solid state recorder and then playing back and processing the data in a few minutes, works much better - and uses way less power than having a PC running with all sort of cables for several hours, has less interference issues, does not require exotic transformers, and actually WORKS. Got about 8 maps from the last received file.
While ADCs with proper sample clock and sufficient large FIFO can easily hide any CPU or OS jitter issues in _one_directionl_ (streaming) applications, provided the data sink has a higher _average_ throughput than the ADC.
The situation with your Raspi i.e. a DAC application is a quite a bit different, especially if there is no feedback (flow control) to the data source (e.g. some disk)..
Remember the heated debate in various "high end" hi-fi rags a few decades ago, when the first external audio DACs were connected to the CD transport with the SPDIF cable. Since the CD transport might run the disk a bit too high/low speed, generating SPDIF frames at say
44101 frames/s, but if the DAC used a fixed 44.1 sample clock, one alternative would be to drop a sample every second, which might be audible. The other alternative would be to skip the DAC sample rate crystal and use a VCO/PLL to lock into the actual incoming SPDIF frame rate, with a minimal increase in audio pitch.
Since it was desirable to also support 32 and 48 kHz sample rates in the DAC, so why not use a large frequency rate PLL to lock to any SPDIF frame rate between 32 and 48 kHz ?
Unfortunately a wide input frequency band PLL tend to have a large phase noise, which is translated to DAC sample time jitter, which some claimed to be audible in some cases.
The problem was here that there was no feedback path to the disk transport and was unthinkable in those days to use Megabyte FIFOs to compensate the frequency error for an hour of playing time.
Similar issues are with USB connected DACs, which are then forced to use a PLL to generate the DAC clock.
In an integrated CD transport+DAC, no such problems existed, since if there was too much data coming from the disk, just slow down it a bit (via feedback) or if everything was generated from the same crystal with a slight frequency error (20 ppm), accept this very small increase in audio pitch.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.