Building a Standards Converter

[`FRANK.C]

Hi Semir
As Jeffrey has said the Hedghog hardware base would not be suitable for HD.
I was at a time considering HDMI input and thought that something like this might be useful for experimentation to preform HDMI decoding to LVDS.
For 405 lines apart from the convenience of being able to use HDMI I think any other improvement over composite would probably be small.  819 lines would be a different matter.

Frank

[ppppenguin]

As we know, the modulator chips used in Hedghog (and Aurora) are obsolete and the grey market suppliers are getting down to the floor sweepings. I don't see any hope of a new chip doing anything similar as analogue RF inputs to TV are also obsolete. So we will need a new approach:

1: External modulator. Several traditional designs have been published. Not multichannel.

2: Direct Digital Synthesis (DDS). Frank and I have done limited experiments. Making a Ch1 signal directly in digits isn't hard. Needs a more modern FPGA and a good high speed DAC. Direct to Band III isn't feasible. A Band I signal can be heterodyned up to Band III but you also a get a mirror signal which may not matter.

3: DDS at IF. Perhaps in the 15 to 20MHz range. As above but heterodyned up to all channels. Reduces the DAC and FPGA requirements a little.

4: VSB may be possible quite easily using the Weaver or 3rd method for SSB generation.

5: Sound also fairly simple. Added to vision in digits before DAC.

Any RF/DSP/DDS experts here? I know some of the basics - certainly no trouble to make a basic DDS for DSB vision plus AM sound. But I'm not an RF expert by any standard.

[`FRANK.C]

Hi Jeffrey
It is nice to have all 13 channels but I wonder is there many that actually use Band III, I rarely do.

Band I should be able to be done using a FPGA and DDS. which should suffice most.

I would like to get back to the FPGA modulator but it is out of reach at the moment. The next thing I need to do with it is to get to grips with the constraints file.

I know that Semir is building a System A, 13 channel PLL modulator. I am looking forward to seeing it when it is finished.

Frank

[ppppenguin]

Would be possible to do an analogue modulator (need something better than the traditional 1496 chips) and synthesise the carriers in the FPGA that's doing the video conversion. The PLLs in most modern and fairly modern FPGAs will do 200MHz+ with ease. In what is now a pretty old and near obsolete Xilinx (Spartan 3A) I run one of the PLLs at 192MHz which is nowhere near its maximum speed.

[Terry]

- Perhaps a different output chip as the current one has an LPF with 9MHz cut off (the French 819 line system has a vision bandwidth of 14Mhz)

Not quite so high as that, Semir!

14MHz is the channel width for system E; vision bandwidth is 10.4MHz.

However, I believe that some manufacturers skimped on the bandwidth in their designs so, if you can find the alignment instructions for this Sony, it would be interesting to know the actual bandwidth it was designed for!

You might find that a 9MHz LPF is perfectly adequate, after all!

[ppppenguin]

The problem in the Hedghog (and Aurora) isn't the output. This is a simple R-2R ladder DAC preceded by an oversampler to minimise the need for output filtering. If you programmed a frequency sweep test into the FPGA it would show the characteristics of the DAC and filter. I'm sure it would be good enough for Sytem E.

At present the pixel clock rates are 13.5MHz for the 625 side and about 8.7MHz for the 405 side. I'm not sure what clock rate would be used for an 819 output but I would estimate about 17MHz. This gives a theroretical BW of 8.5MHz. The limitation would be the 625 line input. 5MHz on a 625 signal corresponds to about 6.6MHz on 819.

The 1280 active pixels of a modern 720p system would be enough to give full BW 819. Not quite enough lines but by a very small margin. Remember that 819 is total number of lines, 720 is active lines. System E had about 770 active lines. Obviously any 1080 HD standard would be amply good enough.

[peter scott]

Generating full bandwidth 819 line signals in a standards converter sounds like a difficult task when compared with that using a PC graphics card and VLC player.

Peter

[ppppenguin]

Peter, you're right of course. Still need a modulator. Don't know how wideband the devices in the Hedghog and Aurora are.

[Semir_DE]

Generating full bandwidth 819 line signals in a standards converter sounds like a difficult task when compared with that using a PC graphics card and VLC player.

Peter

Hi Peter,

Thanks for reminding me of this possibility This is indeed an option I will be looking into. Since 819 lines is closer to some PC resolutions maybe even one of my old Laptops with Intel graphics chips can be tweaked to this scan rate. Once I have completed my Modulator project I will look into that.

[Semir_DE]

Not quite so high as that, Semir!

14MHz is the channel width for system E; vision bandwidth is 10.4MHz.

However, I believe that some manufacturers skimped on the bandwidth in their designs so, if you can find the alignment instructions for this Sony, it would be interesting to know the actual bandwidth it was designed for!

You might find that a 9MHz LPF is perfectly adequate, after all!

Hi Terry,

you are right of course I mixed up the RF and vision bandwidths (it was late when I wrote that post   ). Considering that most available 819 line TVs seemed to have cut some corners when it came to IF bandwidths I also think that 9Mhz is probably more than enough and as Jeffrey pointed out the output circuit is probably the smallest problem to deal with in this scenario.

Once I get the TV, hopefully in one piece, I will modify one of my modulators to output the correct frequencies for 819 line systems and give the PC generated signal approach a try. All of this is a learning process for me and I am planing to experiment a little.

Semir