Retro SSTV anyone?

[Steve Anderson]

As with any test you need to think of your reference. When I said there's no overshoot is that something to do with the modulator used? So to eliminate any doubt is the sketch below viable? This will generate two tones settable anywhere between 1200 and 2300Hz (as a minimum) modulated by a square-wave at a lower frequency. No filters, no A-D or D-A and all the harmonics you could wish for.

In my modulator there is a filter, I want to eliminate any influence it may have on the results I'm obtaining.

It is a chicken and egg situation.

Steve A.

[Steve Anderson]

Going back to the sync-detection bandwidth issue...if you had a choice of two, what would you chose?

Suggestions follow:-

A= +/-70Hz (narrow).
B= +/-140Hz (wide).

...or values around those?

Three are possible 'down the road'.

Steve A.

[Harry Dalek]

As with any test you need to think of your reference. When I said there's no overshoot is that something to do with the modulator used? So to eliminate any doubt is the sketch below viable? This will generate two tones settable anywhere between 1200 and 2300Hz (as a minimum) modulated by a square-wave at a lower frequency. No filters, no A-D or D-A and all the harmonics you could wish for.

In my modulator there is a filter, I want to eliminate any influence it may have on the results I'm obtaining.

It is a chicken and egg situation.

Steve A.

Hi Steve if i am right you are using this 2 tone PLL modulator to test your demodulator to see if it is filtering locking onto the correct frequencies ?
I have to ask why not just a frequency generator or test wav's ?

[Steve Anderson]

Hi Steve if i am right you are using this 2 tone PLL modulator to test your demodulator to see if it is filtering locking onto the correct frequencies ? I have to ask why not just a frequency generator or test wav's ?

Harry, there's no 'locking' involved, no PLLs anywhere, including in the demodulator. I'm just considering using the VCO sections only. I'm simply thinking of using them as a source of unfiltered modulation. There are two major sections to any 4046, the VCO and the phase detectors. Both are totally separate and could be used independently - just like two op-amps in a single package. It's only when you connect certain pins together that it becomes a PLL.

Here, all I need is an oscillator whose frequency is controlled by a voltage, nothing else.

Frequency generator, i.e. tones, done already, that's how the charts above were generated.

Test waveforms or signals. Where do they come from? Who made them? How reliable are they? Even those I made myself I don't trust until proven otherwise. Though the simple tones I uploaded some weeks back I'm happy with. It's rather hard to get them wrong in Audacity.

Anyway, it's a thought and I'll see what the results are. "if you don't ask, you won't get." "If you don't try you'll never win".

Steve A.

[Harry Dalek]

Harry, there's no 'locking' involved, no PLLs anywhere, including in the demodulator. I'm just considering using the VCO sections only. I'm simply thinking of using them as a source of unfiltered modulation. There are two major sections to any 4046, the VCO and the phase detectors. Both are totally separate and could be used independently - just like two op-amps in a single package. It's only when you connect certain pins together that it becomes a PLL.

Here, all I need is an oscillator whose frequency is controlled by a voltage, nothing else.

Frequency generator, i.e. tones, done already, that's how the charts above were generated.

Test waveforms or signals. Where do they come from? Who made them? How reliable are they? Even those I made myself I don't trust until proven otherwise. Though the simple tones I uploaded some weeks back I'm happy with. It's rather hard to get them wrong in Audacity.

Anyway, it's a thought and I'll see what the results are. "if you don't ask, you won't get." "If you don't try you'll never win".
Steve A.

It reminds me when i was making my caveman SSTV testcard generator not sure how close an idea it is to yours in circuit below it to is also without filtering ,i never really thought of it as a demodulator tester but i suppose as with yours working out the wanted bandwidth black to white limits or below black and above white what is the demodulators working parameters ,it would be better with trimmers than fixed resistors i suppose if it was of any use like yours ... Klass and i have to wait and see how you will test your modulator on your new circuit interested to see and learn .

[Steve Anderson]

True, I am for the moment being rather coy as to how I'm putting this demodulator together. As I mentioned before it's probably been done before though I've not seen any reference to it in all the data I have collected over the years. But the concept is probably 'out there' somewhere - if not in documented hardware in some reference. So I claim no originality for it.

But thus far I'm rather pleased with the results, though there is still a way to go yet. As a replacement for the Robot 70 (even my own 'upgraded' version) it already wins hands-down. But I'm not done yet.

Steve A.

[Klaas Robers]

There is a certain bandwidth needed for the sync:
- the sync pulse itself is 5 msec,
- so a train of syncpulses is a square wave of 100 Hz.
- you want to have at least the fundamental = 100 Hz
- so the sync bandwidth should be 200 Hz, 1100 to 1300 Hz,
- and a little bit wider, say 20%
- thus 1080 to 1320 Hz.

Yes I made the sync frequency tunable. When receiving on short wave SSB the frequency is hand tuned. In practice you tune to optimal sync, but if the audio has been recorded, then you have to do with the frequencies that were recorded.

But next to that, sometimes it is better to tune the sync to 1100 Hz, then the interference that can be caused by the video content (side band of 700 Hz bar in the video) can be made minimal. The same counts for the level of sync clipping. Also adjustable.

Be aware that after the rectifier of the sync burst a low pass filter is needed of about 120 Hz to define the optimal position of the start and end of the sync. FM is by origine a sampled system (zero crossings only) and the sync pulse may start and end somewhere in between the samples.

[Klaas Robers]

Steve, for the SSTV FM-modulator I also used no low pass filtering in the video. So your "switched diagram" applies. But in the demodulator (FM-detector) you want to have low pass filtering. That low pass filter introduces ringing = overshoots.

Preferably this is a constant group delay low pass filter, for video of the butterworth type. Butterworth combines reasonable steep cut off with reasonable ringing. The constant group delay ensures that the ringing is distributed equally before and after the edge. This equally distributed ringing gives an impression of extra sharpness.

The constant group delay can be obtained in the analogue world by an all pass filter, in the digital world by a FIR-type transversal filter.

[Steve Anderson]

Steve, for the SSTV FM-modulator I also used no low pass filtering in the video. So your "switched diagram" applies.

The modulator I'm currently using does have an input filter, so I'm trying to eliminate that as part of the puzzle.

So I plan to go with the non-synchronous square-wave modulator I posted as an idea yesterday. I just wondered if it's valid. In a real-world situation there wouldn't be those harmonics/sidebands stretching to infinity, but if this decodes this into a reasonable semblance of the original I'll be happy.

I'm also trying to eliminate the crud that comes out of my PCs in terms of noise, generally based around 48kHz or harmonics thereof, no matter what the sample rate may be. Simple AC97 codec, no reconstruction filter. Probably symptomatic of most PCs, especially low-power devices like laptops/tablets/etc..

I guess I'll have to put some form of filter (I don't want to) at around 10kHz on the output of the PC. There is already a simple RC filter on the input to the demodulator, but it's just just not good enough to get rid of the PC's crap.

Using a battery-powered audio oscillator the output is a straight line, no noise - in fact the 'scopes input pre-amplifier noise dominates at around 500uV p-p.

Steve A.

[Klaas Robers]

Steve, in 1973 when I built the electronics of the SSTV generator, I had the feeling that a filter before the FM audio oscillator was not needed. The carrier frequency is that low, compared to the modulating frequencies, that the oscillator low-pass-filters the video signal of its own. The timing of the half waves is influenced by everything that happens during each half wave, and it is integrated, so there is an automatic low pass action. Even with the 900 Hz bar of the multi burst pattern I have seen no strange effects, although the 900 Hz is fed as a square wave into the modulator (that is the current source that charges the capacitor), just like all the other bar-frequencies.

[Steve Anderson]

OK, let's for the moment forget the anti-aliasing properties I was trying to avoid in my original modulator. My modulator was/is a sampling system. (An A-D and so on...)

Anyway, taking that out of the loop eliminates that variable.

Let's see how things pan out over the next few days...

Steve A.

[Steve Anderson]

Let's see how things pan out over the next few days...

We all know that feeling - when something is too easy, such was the case with this MkII demodulator. The video part went too well, too easy and all done too quickly. Well, I've paid the price with the sync detection bit! Boy what a headache! But it's done. I'm gonna have an early but well-earned beer!

I've used Klaas's figures from the other day of 1080-1320Hz being sync. I'll knock-up the test modulator I posted as an idea a few days back to confirm this. Also test for any mis-triggering with high-level 700Hz modulation.

Steve A.

A quick breadboard of the first 4046 confirms sync-detection is 1081-1322Hz. Near enough! The 700Hz hi-level test I'll do once I've actually built a final board for this which to some degree may clean up the noise problem. But in reality I need to get rid of the PCs as a signal source or at least filter the 48/96kHz crap that comes out of them.

[Harry Dalek]

I gather the circuit your improved demodulator and sync detector are for cases of SSTV SSB reception cases weak signal detection ? a narrow sync bandwidth 300hz or a touch over ...speculating i am
Are you planning to do a real world test on how good it is analog to analog transmit receive not every one has a SSB radio transmitter receiver or would want one Slow scan is very easy to transmit via light LED and receive light sensitive what ever you have, even mechanical sstv sound to speaker to air sound to mic works well very easy to reduce levels for weak reception test all analog bit harder for me any way to do side band same ideas .

[Steve Anderson]

I gather the circuit your improved demodulator and sync detector are for cases of SSTV SSB reception cases weak signal detection ? a narrow sync bandwidth 300hz or a touch over ...

It could be set anywhere almost, I have in mind to do a narrow version at +/-50Hz just for the heck of it. How useful that may be, no idea, but the option is there and can be changed for any sane value.

The next task is providing a modulator that's up to the task.

Steve A.

[Steve Anderson]

Before having a go at an improved modulator I need to shake this MkII demod down, There is some jitter between the logic derived sync and the analouge video and sync output. I think this is going to a fact of life, what impact it has is unknown at this time. Perhaps more correctly is, "To what degree is this tolerable?" This may take some time to unravel.

Part (if not most or even all) of this jitter may be the darn 48/96kHz garbage on the output of the PCs. I know I keep blaming this for causing problems, but I've either got to eliminate it from the PCs output or find some way of generating signals not using these PCs.

As for weak/noisy signals, not really may aim, but the input stage is the same conceptually as many published before. So it'll possibly be similar to other designs in that respect. SSB with its potential mis-tuning is a problem, but within sensible limits that could be overcome. More work is needed in that area but first I'm more interested in its video-recovery abilities without mis-tuning or incorrect modulation frequencies.

One step at a time...

Steve A.