New SSTV Demodulator.

Forum for discussion of SSTV topics. Slow Scan television (SSTV) is a picture transmission method used mainly by amateur radio operators, to transmit and receive static pictures via radio in monochrome or colour.

Moderators: Dave Moll, Andrew Davie, Steve Anderson

Re: New SSTV Demodulator.

Postby Steve Anderson » Wed Dec 07, 2022 11:49 am

As an aside, and seeing as Klaas mentioned it, here's the Philips datasheet for the EQ80, Nonode = 9 elements, anode, cathode and seven 'grids' Interesting to note the datasheet also mentions use as an audio tube/valve...

EQ80_Philips_PeG.jpg


Perhaps the best site around for tube/valve datasheets and other stuff, especially European versions...

http://www.tubedata.org/
http://www.tubedata.info.

I'm sure many of you knew about those sites anyway....

In case you're wondering...NO! I'm not going to do a tube/valve SSTV demodulator, I'll leave that for the truly mad! (I'm not quite there yet, but well on my way.)

Steve A.

EQ80.pdf
(365.39 KiB) Downloaded 190 times
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5376
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: New SSTV Demodulator.

Postby Steve Anderson » Wed Dec 07, 2022 2:48 pm

A few very minor tweaks to the analogue (Travis) SSTV demodulator, it's in a state now where I'll go ahead and build, then on to the micro and the internal FIR filtering. I'll not post the minor changes until I've run some tests...and written some code.

Here's a link worthwhile noting, it relates to the MFB (Multi-FeedBack) filters used in the demodulator and how to calculate the component values for different frequencies, gain an Q. There is a small Windows application within the item which takes out the drudgery of stabbing lots of calculator buttons. The link is in the first sentence after the heading, "Calculating Component Values".

https://sound-au.com/project63.htm

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5376
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: New SSTV Demodulator.

Postby Klaas Robers » Thu Dec 08, 2022 12:09 am

In the SSTV monitor, that I built in 1973, I had a zero crossing detector, that uses both zero crossings, up going and down going. I never had the feeling that I missed something, when demodulating SSTV signals that came tot me via the SSB X-tal filter of my HW-101 transceiver. And I can't imagine how narrow band noise can add extra zero crossings when the signal to noise ratio is well above 0 dB. For FM demodulation you want to have that any way. The SSTV signal fills the bandwidth of an SSB filter of 2.4 kHz to almost maximum.

We know that FM has a kind of threshold in S/N ratio. If you are below that threshold, there is very heavy and coarse noise. For SSTV you will avoid that situation. Then the picture is too bad, if you can recognise any thing at all.

But I think the question is: can we demodulate the FM signal mainly in a micro computer with as little analogue circuits around. My experiments went wrong when I sampled the limited FM signal, as the precise timing of the zero crossings is important and sampling spoils that precise timing. It adds noise.
User avatar
Klaas Robers
"Gomez!", "Oh Morticia."
 
Posts: 1656
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Re: New SSTV Demodulator.

Postby kd2bd » Thu Dec 08, 2022 7:08 am

One of the problems with zero-crossing detection is that it can translate amplitude noise into phase noise (jitter). For example, a high frequency "hiss" or a low frequency "hum" can modulate the threshold voltage of the zero-crossing detector and introduce phase noise that wasn't present in the original signal. A poorly recorded SSTV signal may suffer from bias or harmonic distortion that might affect the behavior of a zero-crossing detector, whereas a frequency selective discriminator might be better equipped to ignore the distortion completely.
But I think the question is: can we demodulate the FM signal mainly in a micro computer with as little analogue circuits around.

Absolutely! Software Defined Radio (SDR) technology does this all the time. There are simple "receivers" that are nothing more than frequency converters with I and Q outputs that down convert a region of RF to the audio frequency range that can be digitized by the soundcard in a PC. Software running on the PC provides all the selectivity, AGC, noise blanking, and demodulation functions of a full receiver.

This technology has grown to the point where modern receivers can digitize RF from the antenna connection and provide all the needed gain, selectivity, and demodulation processes of a complete receiver using DSP software techniques. The only "cost" in doing this (besides the energy to power the CPU) is a slight time delay before the incoming signal is finally heard in the speaker.

Chris pointed out an on-line SDR that was capable of receiving SSTV signals from the ISS. There are many more all around the globe that cover various segments in the HF spectrum: http://websdr.org/

And if anyone ever needs to reverse engineer a pre-existing multi-feedback bandpass filter, this site can determine the gain, center frequency, and Q, among other items of interest: http://sim.okawa-denshi.jp/en/OPttool.php


73 de John, KD2BD
User avatar
kd2bd
Mad Scientist
 
Posts: 65
Joined: Wed Sep 28, 2016 5:47 am
Location: New Jersey

Re: New SSTV Demodulator.

Postby Steve Anderson » Thu Dec 08, 2022 6:35 pm

Klaas Robers wrote:: can we demodulate the FM signal mainly in a micro computer with as little analogue circuits around.

Well, as John mentioned, "Yes...", my posting of the 4th this month in this thread shows how little analogue circuitry is required. However this means a zero-crossing technique is almost forced upon you. So far the consensus within this thread seems to be there are better ways to do this requiring more external analogue processing...unless I've missed something.

In this case I'm not under any commercial pressure to reduce components (cost), board space or assembly time/labour. The mentioned analogue (Travis) demodulator, A-D, FIR filter, D-A and write processor could easily fit on a small board of 92x55mm[1], using DIL chips and human-friendly component sizes. i.e. no SMD. The outputs would be a baseband (demodulated) SSTV signal for an analogue display device and/or an oscilloscope for waveform monitoring as well as the serial SPI data for use with the following up-convertor to 625 or 800x600 VGA.

Steve A.

[1] I am hoping to get the whole thing on a single board of 92x110mm instead of two smaller boards - maybe not including the power supply though.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5376
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: New SSTV Demodulator.

Postby Klaas Robers » Thu Dec 08, 2022 8:12 pm

After my previous experiments, explained somewhere else on this board, I had the idea to make an SSTV front end by limiting the FM signal, send it through an exclusive or and feed it to a PIC. The other input of the exclusive or is connected to an output of the PIC. Then one of the timers in the PIC should be restarted on the negative edge of the input, inverts the exclusive or and the elapsed time could be converted (look up table) to frequency. This gives also the possibility to exclude too long and too short elapsed times.

It will give a stream of numbers, that indicate the value of the frequency at that moment. However the moments of change are coupled to the zero crossings of the input signal, and not synchonous to an internal processing clock. But then.....

What would happen in a FIR-filter unsynchronous with the changing numbers? I can't oversee that problem......



Later...... My problems with the software demodulator some years ago, can be found here:
viewtopic.php?f=24&t=2351&start=90
User avatar
Klaas Robers
"Gomez!", "Oh Morticia."
 
Posts: 1656
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Re: New SSTV Demodulator.

Postby Steve Anderson » Thu Dec 08, 2022 8:59 pm

The input A-D needs to be synchronous with the following FIR filter. In this case 1:1, but it can be different if other sub-systems need a higher sample rate, 2:1, 3:1 and so on...but not here. None of what follows is hard-defined and may (will) change - keep that in mind.

The micro I plan to use is a PIC18F14K50 with a clock of 48MHz resulting in a single-cycle instruction rate of 12MHz or 83.3ns per instruction. The 16-bit by 16-bit signed multiply routine takes 2.83us. The FIR filter requires 31 of these multiply routines for the coefficient weighting and resultant addition, total = 87.7us.

While the micro is performing the coefficient multiplication the A-D starts a new conversion, almost multi-tasking! So the throughput is surpisingly high for such a simple device.

Once that's done the result is bit-reduced (truncated) and sent the the MSSP where it is written into the external RAM. RAM write address tracking and sync detection, both vertical and horizontal is also performed by the same micro.

Some of the timings are getting marginal, if really necessary a 64MHz device might be required, e.g. PIC18F14K22, it is early days after all...

Steve A.

I chose the PIC18F14K50 as I have quite a few in stock (no idea why) and the PIC18F14K22 isn't available until March next year at most suppliers.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5376
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: New SSTV Demodulator.

Postby kd2bd » Fri Dec 09, 2022 3:09 am

For what it's worth, earlier this year I explored the possibility of using a microprocessor-based tone decoder for CW decoding and perhaps 1200 Hz SSTV sync detection. I found this article that describes such a circuit using a PIC12F683 and an application of the Goertzel signal processing algorithm.

Of importance in this application is that the '683 has a 10-bit resolution A/D converter that is used to sample the input signal. The fact that the input signal is handled in a completely linear fashion raises the possibility of being able to reliably detect signals in the presence of noise. The article mentions this feature as being one of the circuits strong selling points, and displays some plotted results.

I didn't have a 12F683, but was able to fit the code into one of my PIC16F88s. I got the tone decoder to work, but found the bandwidth too narrow for either CW or SSTV use. However, the source code can probably be modified and the clock frequency increased beyond 8 MHz to expand the bandwidth and decrease the decoder's response time. But I'm not sure if even a 20 MHz clock (a 5 MHz instruction cycle) would be fast enough to reliably detect a 5 ms wide 1200 Hz tone burst. So, I'm just "putting it out there"... 8)

73 de John, KD2BD
User avatar
kd2bd
Mad Scientist
 
Posts: 65
Joined: Wed Sep 28, 2016 5:47 am
Location: New Jersey

Re: New SSTV Demodulator.

Postby Steve Anderson » Fri Dec 09, 2022 12:23 pm

Interesting...I've only skimmed through the item but the first problem I see, as John notes, is the slow response time, though a massage of the code and a faster micro may help resolve this. If being used only to detect SSTV sync pulses at 1200Hz for 5ms, that's only six cycles, it might need quite a bit of surgery to be reliable.

But it certainly holds promise...I'll read more thoroughly later...

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5376
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: New SSTV Demodulator.

Postby Steve Anderson » Fri Dec 09, 2022 3:35 pm

Just wondering...what's your thoughts on the acceptable FIR passband ripple for this demodulator? So far I've been working with +/-0.5db (1db total) up to the cut-off frequency of 1.0kHz, and -40db once in the stop-band (which can vary in frequency). Sampling rates have varied between 6, 8, 10 and 12kHz.

Update: I'll be using a different micro at first, a) because I have some in stock, b) it runs with a 64MHz clock, c) the PIC18F14K22 is on a long lead-time, March 2023, d) it's a 3.3V device therefore compatible with the serial RAM. Downside, it's 28-pins not 20, no big deal really...
I may use the same micro for the FSTV read processor when I get to the up-convertor stage. This demodulator is a lead-in to the up-convertor.

One thing I also should do is re-visit the DDS SSTV modulator as that also has a Butterworth analogue filter within it...and replace it with something digital.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5376
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: New SSTV Demodulator.

Postby kd2bd » Sat Dec 10, 2022 4:36 am

I wonder if an "Integrate and Dump" filter could be made to work here? With a clock used to sample the discriminator output at regular intervals (say 4151 Hz for 256 samples for a 15 Hz line to 4654 Hz for a 16.66 Hz line), the process would go as follows:

1) Integrate the raw waveform coming out of the Travis discriminator over one of these discrete 4 kHz (pixel) sampling intervals.
2) Latch the integrator's output voltage in a Sample and Hold circuit (LF398).
3) Alert the A/D converter that a new sample has arrived.
4) Reset the integrator and repeat the process over again.

Assuming this could be realized in practice, the Integrate and Dump filter would provide the optimum amount of filtering for your specified sampling rate. Furthermore, it operates completely in the time domain, so it can't ring or overshoot! Each and every pixel gets filtered independently of one another, without any chance of overlap or crosstalk between them.

Here is an analog representation of this process applied to the processing of digital signals. Note that this illustration is deceptive in that the Integrator and the Sample and Hold are not commanded simultaneously -- the output sample is latched very briefly before the integrator is reset:

img23.png

The "dump switch" that resets the integrator after each sample is latched could be one section of a 4066 CMOS switch (or all 4 section tied in parallel if they're not needed elsewhere):

int+dump.png
int+dump.png (42.56 KiB) Viewed 6959 times

Of course this could all be done in software by running the A/D converter at perhaps 10 or 20x the 4 kHz sampling rate, feeding the result into an accumulator after each sample, storing the result for processing/storage/display, and then resetting the accumulator and repeating the process.

But if you can't, at least the analog hardware is cheap. 8)

73 de John, KD2BD
User avatar
kd2bd
Mad Scientist
 
Posts: 65
Joined: Wed Sep 28, 2016 5:47 am
Location: New Jersey

Re: New SSTV Demodulator.

Postby Steve Anderson » Sat Dec 10, 2022 2:06 pm

kd2bd wrote:...Of course this could all be done in software by running the A/D converter at perhaps 10 or 20x the 4 kHz sampling rate, feeding the result into an accumulator after each sample, storing the result for processing/storage/display, and then resetting the accumulator and repeating the process. 73 de John, KD2BD

This is familiar somehow, from a long time ago, but maybe not applied to SSTV in that case. But in principal it looks feasible.

I'm assuming the output of the Travis discriminator into the A-D would be from IC201d, pin14, i.e. no filtering prior to the A-D (posting of Nov. 20 in this thread). The maximum sampling rate of the internal 10-bit A-D within the PIC devices I have chosen is around 40-70kHz. Or an external version is easy up to 100kHz at 12-bits. (Microchip MCP3201).

The rest of the process would be handled in software which would be easier and faster than a FIR filter. This concept is worth investigation...

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5376
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: New SSTV Demodulator.

Postby Steve Anderson » Sat Dec 10, 2022 9:17 pm

I've just checked my stocks and I have 3 LF398s, so as it's such a simple arrangement I may have a go at that first before going the digital route....however it needs to be digitsed at some point for up-conversion, so it may be a moot point.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5376
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Re: New SSTV Demodulator.

Postby kd2bd » Sun Dec 11, 2022 7:58 am

Steve,

The output from the Travis Discriminator is a bit "messy". An integrator can handle this without an issue. But to protect and preserve its full dynamic range, an A/D converter may benefit from some gentle low-pass filtering.

I'm really looking forward to your results. This could be a "game changer"... :D

73 de John, KD2BD
User avatar
kd2bd
Mad Scientist
 
Posts: 65
Joined: Wed Sep 28, 2016 5:47 am
Location: New Jersey

Re: New SSTV Demodulator.

Postby Steve Anderson » Sun Dec 11, 2022 11:48 am

kd2bd wrote:The output from the Travis Discriminator is a bit "messy". 73 de John, KD2BD

Yes, I did take a look at that point in the circuit with a 'scope of previous versions, I thought, "How does something as simple as a low-pass filter make sense of all that lot?" One could do all the math behind the process, but that would hurt my brain to no advantage. It works, leave it be.

kd2bd wrote:...an A/D converter may benefit from some gentle low-pass filtering.

A LPF may also reduce the chance of aliasing as the output of the discriminator/rectifiers surely contain harmonics/junk above the Nyquist frequency when digitising the signal. This may be the only virtue of the Butterworth filter I used previously.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5376
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

PreviousNext

Return to SSTV

Who is online

Users browsing this forum: No registered users and 1 guest

cron