David Robinson's System A modulator

[`FRANK.C]

Hi All
This is my attempt at building  David Robinson's System A modulator which covers all 13 Channels. I found the original threads on the UKVRR which are here and here

Both Audio and Vision modulators are the same as David's. The only change that I made was to simplify the audio input as I don't need the separation between left and right channels that David did. There was some debate about whether the 3.5 MHz filter was necessary or not. I decided to keep it, the coil the David used is no longer available, I used a 'Mini Mod' transformer in it's place and it works quite well.

I have made some changes to the Microcontroller side of things. I am using a 18F1330 Microcontroller. The I2C buses for the two modulators now share a common clock. The microcontroler produces a very basic test card and an audio tone of approximately 300Hz which are fed to the modulators via switched phono plugs. When a Video or Audio plug is inserted into a Phono socket the test card or audio tone is disconnected.

I made a PCB for it which made assembly of the SMD's much easier. The PCB is double sided with one side used solely as a ground plane. The PCB was produced using the laser printer method which left the copper pitted but it is still acceptable.

Frank

[ppppenguin]

This is essenitally identical to the modulator used in the Aurora SCRF. Darryl used an external master clock feed (from the FPGA) rather than a pair of 4MHz xtals. A CPU or FPGA based controller is needed to set the channel. The Aurora SCRF uses its FPGA, Frank has used a PIC. It's probably worth having a 3.5MHz sound trap ahead of the vision modulator.

Well done for a nice neat PCB. If I was doing a "one off" I'd probably do it dead bug style on a bit of copper clad board.

PS: Am I right in thinking that it's your article about a PIC based testcard generator in the new Bulletin? That's a lot of result for minimal resources. If I was doing something like that I think I'd store each pattern as a literal on an SD card and read it with a CPU or CPLD. Ample capacity for loads of patterns and redily programmable. I think Karen Orton has done something like this. It's also the basis of the Richard Russell generator and the ImoGen. The difference is that now you can get a lot of fast flash for virtually no money.

[Murphyv310]

Excellent work there Frank.
I'd have a go at this, I've also your Test Card Generator to finish off.
I suppose I'm very lucky with signal sources for 405 in many different guises.
Here goes.
Philips Pattern/modulator generator, does PAL, NTSC, SECAM in all variants and systems. Continuous Digital frequency programming from 33Mhz to 870Mhz, works as 405 line vision modulator.
Multistandard Aurora converter.
SCRF 405 line converters x3,.
Imogen Test card generator
RT Russell Test card generator
Frank C Test card generator (WIP)
ITA Channel 11 test card generator and all peripheries.

So I'm well served for signals but I'd not say no to another modulator though.

[`FRANK.C]

Hi Jeffrey
I peered inside my Aurora to see what chip it was using as I remembered reading somewhere (probably in one of your posts) that it used a programmable modulator. It was when I went googling for information on the MC44BS373 that I found David's thread on the UKVRR and discovered that he had already all the hard work done.

While I was trying to get everything to work together I did build it dead bug style on three bits of stripboard and the surrounding area. It was a real pain trying to solder strands of wire onto the SM chips legs, much easier to plop them onto a PCB and and run the iron along each leg. I use DesignSpark which can produce a PCB quite quickly

That article is one I wrote. The PIC I used had the largest memory (96k) that I could find, there was lager 3 Volt ones but I wanted to stick to 5 volts.
A SD card would probably be an easier/better option but I am afraid SD cards, CPU's and CPLD's are way over my head.

Hi Trevor
That's an impressive arsenal of 405 signals.
The firmware on your PIC may need updating as it is an older version that has an error in it.
Each field on both test cards has an extra line, giving 407 lines in total. This didn't show up in my first builds as the resonators I was using were sufficiently out in frequency (high) to keep the field rate close to 50 Hz. However If the resonator is at the correct 6 MHz or lower a Hum bar will be displayed on the Test cards. If needed the corrected Firmware can be downloaded from here.

Frank

[ppppenguin]

Frank, CPUs aren't over your head. You're using one - it's called a PIC. The RTR and ImoGen generators are basically counters that step through the contents of a memory. Most of the hard work is preparing the data to put in the memory. That's done on your PC or Mac.

[`FRANK.C]

Hi Jeffery
I misunderstood, I was thinking of something much faster and more powerful.
I think it is also worth mentioning Jon Evans Test card generator, that I think works in a similar way to the RTR.
It's one that I would like to build at some stage.
Jon having made the EPROM images available as a download makes it doable.

Frank

[ppppenguin]

This sort of image can be re-used in any sort of memory, provided that it's accessed with the same clock rate as the original.

If I was designing something like this, I'd probably go for a significantly higher clock rate, at least 27MHz for 625. This simplifies the filter needed after the DAC while giving good frequency response. At 54MHz even a very accurate filter becomes trivial for a 6MHz bandwidth. But isn't this a waste of memory? Assuming the memory is fast enough (I think modern SD cards are amply fast) then at 54MHz each picture needs 54MHz/25Hz or just over 2Mbytes. Since 32GB SD cards are cheap and even bigger ones are readily available this is hardly a problem. Even if we allocate 4MB per frame to simplify addressing then a 32GB card would hold over 4000 frames or almost 3 minutes of moving pictures.

[`FRANK.C]

I was asked for the files of this project so I thought I would make a few "improvements" first.
In the first version the switching of the test card and test tone was done by switches on the phono sockets. This I found inconvenient as if the test card was wanted on its own a phono plug needed to be in the audio socket to turn off the tone. So in this new version I have used toggle switches to select the test card and tone. There is now a audio gain control on the front panel.
I have also changed the test card to the PM5544 style one.
The files can be found here.

Oh and I named it Vol as in Vol(e)

Frank

[ppppenguin]

Looks very professional now. The knowledge that Frank has gained by designing the Hedghog has been applied here too.

[`FRANK.C]

Thanks Jeffrey

The test card in the modulator looks similar to the one in Hedghog but there is a big difference in writing them.

The test card in the modulator is wrote in assembler as one file and is over 4000 lines.
The number of instruction cycles between every video transition must be calculated and counted. A mistake in calculating or counting results as an on screen defect.
The pixel frequency was chosen to make 1/1.5/1.875/2.5/3 MHz gratings possible.

The Hedghog test card is wrote in VHDL and is under 400 lines.
No need to count the counters in the FPGA does all that for you.
Practically any frequency grating within the 3 MHz bandwidth is possible. I do remember having to slightly adjust the gating frequency's to remove some shading from them. Maybe caused by the jitter from the DTO.


Frank