Yet another standards converter

[Karen O]

Yuletide greetings to you all,

I've built a standards converter! The motivation is not to compete with Darryl (I don't think anything can compete with the Aurora). It has been on my wish list for years to build a converter and the increasing scarcity of hobby useable parts has presented something of an ultimatum. It has become a case of 'now or never'.

My converter uses FIFO memory ICs as line stores and a couple of PICs to manage things. It is working extremely well, despite the fact that I haven't implemented the output frame synchs properly yet.

In common with other 'simpler' designs, it generates a non-standard signal with a 96 microsecond line duration. I don't think there is a TV that can't be made to lock to this strange standard however, I am uncertain what effect this reduced line duration is having on aspect ratio.

For the moment, I have simply implemented System A line timings pro rata, that is, reduced all timings by about 2.8%. That has undoubtedly affected the aspect ratio of the picture on my TV9-90.

My question to all you CRT experts out there is whether we can predict the effect of a slightly higher line frequency on aspect ratio? My intuition is that the line sweep velocity of a TV is much pretty fixed because it is a function of the rate of current build-up in the LOPT, which is a function of voltage applied, which doesn't change with line frequency (?) If that is the case then my 405 line pixel clock should remain at the 'correct' value of eleven seventeenths of the 625 line pixel clock.

If the sweep velocity IS unaltered then the effect of my shorted line duration is that a couple of microseconds are chopped off the end of the line.

I am grateful to Darius Mottaghian by the way. He was most helpful in explaining the interpolation scheme employed in his converter. I am grateful to Jay Oldstuff too for the TV9-90 I'm using for this development.

I'll keep everyone informed of progress and I will of course make the design information available to the community in due course.

[Focus Diode]

Looks great. Well done on your excellent achievement.
Cheers,
Brian

[dominicbeesley]

Looks good Karen! I would have thought that most sets would be able to adjust out the aspect ratio difference with their width control. It might have more of an effect the EHT but probably not by enough to worry about. PAL/Colour TV's have been fed all sorts of not quite right versions of PAL/NTSC/PAL with NTSC timings/PAL but "close enough" line frequency by home computers, videos and games consoles.

[Hybrid tellies]

Well done Karen, a great achievement.

[Karen O]

Thank you all

I'll think about this over the Christmas break but you're right Domonic, it is easily adjusted for on the telly.

An interesting aspect of the strange standard I'm using is that it is triple interlaced! There are 208.33 lines per field. It is tempting to call it 417 lines but it is really 625 lines over 3 fields.

I'm planning to use equalisation pulses, which in this case will be at three times line frequency and one third line pulse width!

My, I do take on some tasks...

While we're on strange standards, my daughter complained that her old game box won't work on new TVs. Turns out that a lot of these old game boxes used non-interlace - presumably 312 lines at 50Hz (288p) and modern TVs turn their nose up at this. So much for progress.

Thanks again for all your kind comments.

[dominicbeesley]

It can be either the non-interlace or the fact that many tuners will lock on to the frequency by tuning to the sound carrier and if that is missing or at the wrong separation will cause bother...

[Courtney Louise]

Very very impressive

[jay_oldstuff]

Wow what a fantastic achievement. Well done on the work so far.

Jay

[Barry Lloyd]

Impressive indeed. Re. the game problem, I seem to remember that a problem that used to arise with early games is that sometimes they had very rudimentary frame sync pulses and the receiver would not lock on to them properly. I don't know if that is also the case here.

Barry

[julie_m]

It's always nice to see something done the hard way!

I've heard tales of viewers on the South coast picking up 819-line transmissions from France, and displaying two pictures side-by-side as though it was a 409.5-line signal. Your 416.667-line pseudo-standard isn't much further out, so a bit of twiddling with the vertical hold should get a lock on.

The Acorn Atom was also renowned for producing a very non-standard video signal.

[Karen O]

I raised the idea of non-standard 405 line video signals in connection with standards conversion a year or two ago. I'm pretty certain now that other converters have used such a departure from standard to simplify their designs. But after the reception my earlier posting got I can understand why designers of these converters have kept quiet about it!

[Jac]

Very impressive results Karen!

Would you care to disclose some more information of it?

Thanks!

Jac

[Karen O]

Hello Jac,

I will put together a PDF of the project when its finished and disseminate, perhaps through the BATC. That shouldn't take too long. The remaining work is not hugely demanding but I do have to make a decision now regarding the 405 line pixel clock because changing that is difficult once the code is done (there's a lot of instruction cycle counting involved!)

For now, I can say that it uses three line stores, each using a 1k FIFO. Each line is digitised to a little over 700 pixels. The pixels are six bit depth and derive from a flash ADC. The interpolation scheme involved is a simple mixing of three lines to produce two and requires that one of the three input lines feature twice in the output lines. This is facilitated by the FIFO's 'retransmit' feature.

Although my converter uses only twelve ICs it lacks a modulator. Currently, I am using a UHF modulator and the 'both buttons down' trick on my TV9-90. But I have an ETF channel 1 modulator waiting to take over!

Best,
Karen

[Jac]

Thanks Karen,

I'm looking forward to that!

Jac

[Karen O]

Well, I've added proper frame synch and had my converter running properly tonight and the results are very pleasing indeed. I have a rock solid picture with no jittering on my TV9-90.

I'm now trying to figure out how I can get my 'scope to trigger at the right times to check the frame synch waveforms

The width needs adjusting, as expected. That adjustment must be internal with the TV9-90 as I can't see any such adjustment on the outside.

Anyway, I'll do a few more checks and then I'll try hooking it up to my Pye FV1C.

[Karen O]

Well, I've had my converter running with my Pye FV1C and I've been watching 'Spearhead from space' (first Pertwee era Dr Who story).

It worked without any persuasion at all. I didn't even have to adjust the line hold, which was something I was expecting I'd have to do given the slightly high line frequency. I had to look twice to make sure I hadn't still got the Aurora connected.

Anyway, I think I can conclude that it works. And now to the write-up...

[Doctor Hue]

Originally Posted by Karen O

Well, I've had my converter running with my Pye FV1C and I've been watching 'Spearhead from Space'

Congratulations.

Ironic that that "Spearhead" was the first Doctor Who transmitted in 625 lines ( though not the first recorded, that was Enemy of the World part 3 I think )

[Silvered_Mica]

Anyone know if the write up was ever completed?

[AC/HL]

If you click on Karen's avatar, there is an option to email her.

[Jac]

Thank you for your article about the standards converter in the BVWS bulletin Karen.

Jac