Standards converter rebuild.

[FERNSEH]

Here's a project to keep me occupied some time in the near future. I won't be able to heave HMV 900s around so I'll need to do some light work instead.

I use the Aurora standards converters for my 405 sets and for that reason my own build converters have fallen into disuse.
So why not repair them and add some extra mods to improve thier performance?

Check out the horrific 1980s construction techniques. The long board is a line store and the the smaller one is the 405 sync and blanking pulse generator.

DFWB.

[HamishBoxer]

I believe this was in "405 Alive" all tose years ago??

David

[FRANK.C]

Wow David that's impressive. I heard of your converter mentioned is post's before but I know nothing about it. I looks a great bit of kit. Is there many other boards to make it up? This will be an interesting project.

Frank

[ppppenguin]

David has the honour of being the first enthusiast to build a successful 625>405 converter, With late 1980s technology this was not easy. I am and was then an electronics design professional; I would not have regarded it as simple if a client had asked me to design one in the late 1980s.

I congratulate David for having the persistence to make it work.

[Ed_Dinning]

Hi David that should keep you busy while recuperating. I also have PCB design programme to avoid all those tedious tape-ups. The handbook will act as light reading while you are in hospital.

Ed

[evingar]

Great stuff.

The advantage of TTL type technology is that it's easy to rework (no pesky BGA's etc) and most parts are still readily available (in contrast to today technology that seems to become OBO after about a fortnights production !).

I had a go at designing a converter using similar technology, but it was never stable - badly designed PLL's (etc), so I have every admiration for people who made this stuff work properly

My own effort got shelved (and eventually dismantled) when I managed to get a Pineapple PCB and key components after the main project was abandoned. Still got it somewhere - like most folks though I have an Aurora as my "main" 405 source.

[ppppenguin]

Originally Posted by evingar

The advantage of TTL type technology is that it's easy to rework (no pesky BGA's etc) and most parts are still readily available (in contrast to today technology that seems to become OBO after about a fortnights production !).

FPGAs are even easier to rework No soldering at all.

[murphyv310]

Hi.
That is a well constructed board for a prototype. I certainly wasn't around with 405 line sets when you built your first converter and when I got back into the hobby in 2002 I had absolutely no idea so many 405 line sets had been saved nor did I realise the wealth of knowledge about.
To me a total eye opener!

[FERNSEH]

Originally Posted by FRANK.C

Is there many other boards to make it up?

Hi Frank,
There are more assemblies employed in the converter. Three line store boards, line store control PCBs to care take of the routing of the clock pulses and read/write control.

The pictures show another standards converter project, a line store PAL to 405 NTSC converter. I gave up on this one, could not get it to work at all.
DS2011 FIFOs were used, the difficulty with this device is that the reset is common to in and out.
I do know that a forum member had success with this FIFO IC, but it would not work for me.
Hi Jeffery,
The original prototype converter still exists. It was built in 1985. Except for one PCB the whole thing is constructed using veroboard. Forum member System A looks after it for me.

DFWB.

[FRANK.C]

David that's really fascinating stuff. I look forward to reading more when you start your rebuild.

Frank

[System A]

Hi everyone, as some of you may know, I have David"s first generation converter that he built back in the late 80's? I have taken some of air photos of the converter output plus some from the aurora . David"s does not have interpolation as can be seen but I'm sure you will all agree that the pictures are good . David is to be congratulated on this effort. I will let David explain how it works!
Regards,

Gary.

[murphyv310]

Hi Gary.
You have a piece of history there. The first converter to be constructed by an "amatuer" not that David is an amatuer!

[HamishBoxer]

I am certain as stated earlier,this excellent piece of equipment was in 405 Alive many years ago for those of you who remember or were members.

David

[FRANK.C]

That is truly amazing, what a huge achievement.

Frank

[FERNSEH]

Hi Gary,
The old converter still delivers an excellent pictures. The converter is the result of work started in the autumn of 1984. An early attempt at 625-405 conversion was readied by January 1985 but it wasn't all that good, so it was back to the drawing board, so to speak. In the 1980s the shop was very busy, I was working long hours repairing TVs along with a booming TV and audio sales dept. so the project was held up. The BVWS meeting in November '85 was coming up so I had to pull my finger out and set about to build a viable converter to show at that event.
What you see is an simple converter without any interpolation of the dropped lines, all 220 of them. There should be a serious loss of vertical detail in the converted picture, but in practise the results can be quite acceptable. Bits and pieces of the interpolation line store and control boards still exist in the workshop somewhere.
I still like building projects with Veroboard, I did notice Maplins still sell it.

Way back in the eighties I remember an assistant in a local electronics parts shop saying to me: "are you that bloke who still builds things with TTL"
Surely TTL had not gone out of favour by then? Anyway what's wrong with TTL?

Gary has to thanked for saving this early converter. That was in 1990. The remains of it were lying about on my metal working bench and it was going to be used as a source of parts, TTL parts of course.

The last picture taken by Gary shows the reconstructed converter.
In the background and on the left is the David Looser designed vision and sound modulator, it's the original Television August 1984 circuit. On the right side is a Farnell SMPSU. Excellent unit.
Next are the Veroboard constructed line store boards. Uses Hitachi HM6164 S-Rams. To their right and again made on veroboard is the line store switching unit.

Assembled on metal backing plates is the 405 line sync and blanking board.
On PCBs are the 625 timing and clock pulse generator. The other one that is just visible is the 625 input and 405 output processing.

One Forum member has expressed an interest in building a replica. There could be an problem however, some of those seventies and eighties chips are becoming scarce now. Some redesign of the original circuits may be necessary.

DFWB.

[davegsm82]

Hi David, its nice to see you on the forum, I hope you are doing well.

RE: your last post, was that "Local Electronics Parts Shop" Aitken Bro's Electronics by any chance? It may surprise you to know that one of the Assistants who worked there now works in the Freeman hospital Dialysis dept, although I don't know if he was there as far back as the 80's.

I hope i'm not going too far OT on the subject of TTL, but a similar experience befell me at my last workplace when I was asked to design a drive system for a totem pole capacitively coupled 30kW power converter, working on a Phase shift design for the power output control.

I designed the whole thing using a mix of 555's, LM324's, a 741, CMOS counters, Logic and I think there was even a couple of TTL Chips in there too (all on breadboard, of course!). After some teething problems with the IGBT's we eventually got it working 100% and achieved a respectable 95% efficiency while proving that the thing was absolutely bulletproof, it NEVER blew an IGBT in the time I was running it. Then they decided to move to an FPGA at which point I moved away from the project, from the day they installed it I don't believe it ever ran again. Simply turning on the 600v supplies would send shards of power transistors all over the factory.

Turns out the FPGA was too susceptible to noise, by a long shot. The noise from the transistors switching was enough to lock the FPGA causing shoot-through with catastrophic consequences. This made me question a few things, like, WHY does everything have to be done in a uProcessor? Logic is logical and doesn't break easily, its cheap and readily available too. I'd created a system which worked and one day I found it laying in the bin but not for long, i took it home, then guess what? they asked where it was along with all the schematics i'd drawn! needless to say, all of that was conveniently 'missplaced'.

I miss the days of REAL electronics.

Dave.

[ppppenguin]

Originally Posted by davegsm82

I miss the days of REAL electronics.

There are many different ways to design a standards converter. I have written about this in several articles over the years all of which can be found on my website. These range from an almost entirely analogue design to the Aurora which is based largely on an FPGA. None of these is right or wrong but they all represent appropriate use of the available technolgies to solve the problems.

David Boynes' largely TTL converter used parts available to the amateur at reasonable cost at that time to achieve a result that could otherwise only have been done with optical conversion. A professional design of that period would likely have used similar technologies though small PLDs were available which would have simplified things. A very high volume product might have used a fully custom chip.

Fast forward to the early part of the 21st century. Darryl uses low cost FPGAs to produce the Aurora SCRF. The price is right for a large number of enthusiasts to work with 405 line TVs without worrying about the cost and availability of signals.

If a design crashes and burns, whether it's a standards converter or a high power system then that's down to poor design and/or manufacture. It might be that you are trying to achive something that's on the limit of what's possible. Or it may simply be poor choice or utilisation of components.

Boards full of TTL (or RTL, DTL, ECL, CMOS etc) are part of the history of digital design, a subject worthy of study and preservation. They are not a way to new products except perhaps as an exercise in nostalgia. Or the unlikely case where, for whatever reason, they are the only way to get the job done in a way that works. You see this in specialist fields, notably spacecraft, where the severe environment and reliability requirements mean that components used are normally some years behind the leading edge.

[davegsm82]

Originally Posted by ppppenguin

Boards full of TTL (or RTL, DTL, ECL, CMOS etc) are part of the history of digital design, a subject worthy of study and preservation. They are not a way to new products except perhaps as an exercise in nostalgia. Or the unlikely case where, for whatever reason, they are the only way to get the job done in a way that works. You see this in specialist fields, notably spacecraft, where the severe environment and reliability requirements mean that components used are normally some years behind the leading edge.

I understand the points you make, however I still think that there are some jobs that are done BETTER in logic.

The case I made is an example, the Power supply was designed with a certain frequency, switch period etc all set in stone, then the drive circuit was designed around this. This means it will NEVER be required to have any more functionality, so why use £10 worth of uProcessor when 50p's worth of logic and linear actually does the job better?

I think the point i'm trying to make is that it seems like the uP is the 'done' thing, rather than making an assesment of what does the job best. I'm not trying to put a downer on the microprocessor, i'm from the 80's, the computer generation I suppose so I embrace them openly.

Dave.

[FERNSEH]

Originally Posted by davegsm82

RE: your last post, was that "Local Electronics Parts Shop" Aitken Bro's Electronics by any chance? It may surprise you to know that one of the Assistants who worked there now works in the Freeman hospital Dialysis dept, although I don't know if he was there as far back as the 80's.

I miss the days of REAL electronics.

Dave.

Hi Dave,
It was indeed Aitken Bros. I was rather surprised when I was greeted with that statement. However to be fair to them, by 1984 amateur interest in constructing electronic projects was on the wane, if had not gone altogether. Most of the components for my converter projects were sourced from RS and Farnell.

DFWB.

[cmjones01]

Originally Posted by davegsm82

I understand the points you make, however I still think that there are some jobs that are done BETTER in logic.

The case I made is an example, the Power supply was designed with a certain frequency, switch period etc all set in stone, then the drive circuit was designed around this. This means it will NEVER be required to have any more functionality, so why use £10 worth of uProcessor when 50p's worth of logic and linear actually does the job better?

While there will always be duff designs out there, I agree with Jeffrey about the time for discrete logic having long passed. Indeed, a programmable logic device is the modern equivalent of a board full of assorted chips, except that it's cheaper to manufacture and quicker to develop with. And 'programmable logic' (whether it's an FPGA or something smaller) does not have to mean microprocessor.

As an example, I have on my workbench right now a system in which I've used a CPLD containing a few thousand logic gates as the interface between a microprocessor and some safety-critical power switching. The CPLD has all the advantages of a board full of TTL: it doesn't crash, so it's perfect for safety-critical functions, and it's excellent for insulating the real world from the vagaries of the (very complex) software which runs on the microprocessor. It also uses only a few milliwatts of power, and it speeded up the design time by a couple of orders of magnitude because I could quickly modify it as the system was developed and tested. I couldn't have done the design as quickly and cheaply using discrete logic.

However, I strongly agree that the 'put-a-PIC-in-everything' approach shows a lack of imagination and can easily lead to over-complexity and difficulty of thorough testing. Not such a problem for a homebrew project, but a real pain for anything which has to be manufactured in quantity and used by people other than the designer!

And by the way, this stuff is very much relevant to technology preservation. As old TTL parts get harder to obtain, especially obscure ones, the ability to replace them with a general-purpose logic device programmed to do the job is extremely useful, and will keep many vintage pieces of electronics running well in to the future.