SAA5070

[davidb]

Branching from the Master enhancement speculation topic, here's what I've found about the SAA5070 "LUCY" chip:

• A basic overview can be found in the first result in this search.
• It was used in the STC Executel 3910.
• There's some basic information in the PDF available here, but probably nothing more than in other documents.
• You might still be able to buy them.

This may all be well known, of course.

[BigEd]

This is probably known too: there's a 4 page "Product Brief" in this 1982 Signetics MOS Microprocessor Data Manual.
Edit: Maybe already in the link above which isn't working for me.

This recent thread hints at the possibility an upcoming disassembly of some driver code.

[davidb]

There's some basic information in the PDF available here, but probably nothing more than in other documents.

Actually, this looks like something related to the SAA5070, not that chip itself.

[BigEd]

Just managed to get that PDF - it's 8 pages, from Mullard, describing the VM6500 "Viewdata line terminating unit module" which is a PCB with the SAA5070 and supporting circuitry.

[davidb]

There's also this article in Italian: Decoder Televideo Realizzabili con L'Euro-CCT SAA 5240.

[Pernod]

Thanks both, but nothing useful unfortunately. We need a full datasheet that gives detail on it's internal registers, whereas the brief datasheet only provides an overview and block diagram. We even bought a Mullard databook about a year ago from the mid 80's hoping to contain this datasheet, but it didn't.

[BigEd]

Will disassemblies do the trick? Or at least fill in the most important gaps?

[Pernod]

Will disassemblies do the trick? Or at least fill in the most important gaps?

No, that only tells us what data is being sent to the SAA5070, but gives us no clues as to what the data represents or how the SAA5070 should respond to it.

[davidb]

Another article about it, but no real information.

[BigEd]

This won't help:

(From Everyday-Electronics-1981-07)

[davidb]

Well, if you squint and look really closely...

[hoglet]

Cricklewood Elecronics sold these back-in-the-day, and the appeared in their adverts in ETI from 1983 to 1986:
https://worldradiohistory.com/hd2/IDX-U ... saa5070%22

They actually still sell them now, for a very reasonable price of £8 + VAT
https://www.cricklewoodelectronics.com/SAA5070.html

Worth buying one to play with? Or to decap?

Worth giving Cricklewood a call even?

From the block diagram, it appears to have thirteen 8-bit registers:

It's possible there is some similarity to the SC2681 Dual UART:
https://archive.org/details/Signetics-S ... erDUARTOCR

[BigEd]

Oh, good find, and good thought. It is starting to sound like a decap, photograph, and reverse-engineer is the most likely way to go... only 6000 logic gates, perhaps somewhere in the region of a Z80 for complexity.

(Only thing is, plastic DIP is a bit tricky, I think. Certainly compared to a lidded package or a ceramic DIP.)

[hoglet]

Oh, good find, and good thought. It is starting to sound like a decap, photograph, and reverse-engineer is the most likely way to go...

I think you could also figure out quite a lot by interfacing one to the Beeb 1MHz Bus and experimenting.,,

[BigEd]

Agreed! I'll sub you if you want to buy, say, 3 of them.

[Pernod]

Oh, good find, and good thought. It is starting to sound like a decap, photograph, and reverse-engineer is the most likely way to go... only 6000 logic gates, perhaps somewhere in the region of a Z80 for complexity.

(Only thing is, plastic DIP is a bit tricky, I think. Certainly compared to a lidded package or a ceramic DIP.)

I can arrange the decap, but any reverse-engineering from the photographs is well beyond my capabilities. Do you really think a complete understanding can be gained from this? I know extracting data from a decap is usually successful but I'd expect deriving the logic would be much more complex.

I don't think plastic DIP is an issue as they usually end up much cleaner in an acid bath.

[BigEd]

That's encouraging... yes, I do think reverse engineering can answer all questions, given enough applied effort. What usually happens - understandably - is that only some of the questions are answered, as people run out of enthusiasm, or time, or inclination. But the photographs and vectors stick around for future explorers. This is a very niche part, so one wouldn't expect a lot of people to turn up. For that reason, I'm thinking Dave's suggestion of some more conventional exploration is a much more promising starting point, and it doesn't preclude the other tack.

If there's enough stock to put a couple of parts under acid and to keep a couple back for behavioural analysis, I'm OK to sub for the initial purchase and postage.

[paulb]

Nothing technical here, but I found it interesting that the LUCY chip won a Design Council award:

"LUCY Integrated Circuit by Mullard Ltd", Design, May 1982.

And there's a Mullard advert for it on the facing page. Internet Archive login required for borrowing, although the preview might work otherwise.

[philpem]

I've picked up a Tandata TD1100 - which seems to be an AlphaTantel motherboard with a different case and keyboard. It's sparked a bit of a discussion on the Discord.

Short summary - The LUCY data seems to be in a databook titled: "LSI Circuits for Teletext and Viewdata - the Lucy Generation" (1981), Mullard Ltd., reference M81-0001.

I found a mention of it in the Philips "Technical Review", Electronic Components and Applications, Volume 3, Number 2 but only as "Mullard technical publication M81-0001, 1981".

Searching for that turned up a patent which referenced it - https://patentimages.storage.googleapis ... 1359A1.pdf (line 23 of the main text). There's probably more about LUCY in that document but I haven't read it through yet.

After that there was a recent thread on Vintage Radio - https://www.vintage-radio.net/forum/sho ... ?p=1564491 - where someone mentions that it had an orange cover showing someone holding the latest electronics.

Beyond that, I've found a bit on Prestel in the BT Engineering https://archive.org/details/bte-198204/ ... 8/mode/1up but it talks more about Prestel and only briefly about LUCY.

Hardware

This explained the hardware onboard LUCY, and listed the "Lucy Generation" databook in the references. It's an autodialler, with a 1200/1200 and 1200/75 modem, Kansas City-style tape modem (300 Baud, 1200/2400Hz FSK tones), two IBUS (seems to be a versatile I2C/SPI interface) receiver-transmitters, and two GPIO ports.

Pernod mentioned the LUCY has 13 registers - with the above hardware list, 13 seems about right. Most will probably follow a pattern of a data read/write register and a control/status register. If/when I have a spare moment, I'll see if I can figure out enough from the MAME driver to disassemble the ROM, and whether any of the LUCY register functions are obvious. I'd expect data registers and their associated control bits would be.

Reverse engineering etc

I think the priority should be trying to find the LUCY manual ("LSI Circuits for Teletext and Viewdata - the Lucy Generation") - it might be worth looking into who holds the Mullard company archives, or perhaps talk to the British Library (when their IT is back up and running).

If someone's dead set on having a LUCY SAA5070 decapped and imaged, I can make that happen -- I just need some LUCY chips.

And now I have to figure out why my TD1100 doesn't work - it powers on but displays a screen full of garbage...

[BigEd]

Thanks for the info and links!

[philpem]

No problem, BigEd

If anyone happens to have a copy of that LUCY LSI manual (M81-0001) or the Prestel Terminal Specification, I'd rather like a copy... Although Wikipedia suggests the Prestel spec was generalised into CEPT T/CD 06-01, then into CEPT3, which later became ETSI ETS 300 072 and ETS 300 072 Amendment 1, the latter adding a few more character sets and display modes.
I also found something called the "Viewdata service terminal specification" from the GPO in 1976, which may be the Prestel Terminal Specification.

I also found this little gem, an online viewdata viewer called "vd-view": https://vd-view.azurewebsites.net/
There are a bunch of tech documents on the Help page including ...

• Prestel Bulk Update Technical Specification (seems to be a way for editing terminals to send updates to Prestel in bulk)
• Prestel Telesoftware spec

In the meantime I've found out that my dead TD1100 has the same motherboard as an AlphaTantel, and the one on Binary Dinosaurs had the same problem - caused by the battery rotting out the LUCY and its socket.
I've replaced the LUCY socket already but things haven't improved. On the off-chance the LUCY has succumbed to battery rot, I've ordered a couple from Cricklewood.

While my preference is still for someone to find a copy of the Mullard manual, it does give us one chance to stick a chip under the microscope.

I've also attached the Tandata TD1100 ROM for anyone who's interested in gawking at it ... I've been told it does the same on startup as the AlphaTantel.


Edit to add.. Phone socket wiring

These have a GPO Prestel plug, aka Plug 505.
The mating sockets are the 96A and 96B.

The wiring data is in the Mullard VM6500 Line Terminating Unit Module datasheet:

Code: Select all

Bell contact: plug points 1 to 5
Line loop: plug points 2 to 3
Not used: plug point 4

Helpfully, the wiring diagrams are in the above-linked Viewdata spec, on PDF page 20, as figure 2.

If you've got a lead ending in a BT 431A plug, check the colours match up with this website, then the wiring from GPO socket to BT is as follows...

• Line A - White - GPO socket pin 2
• Line B - Red - GPO socket pin 3

Bell wiring isn't needed - and actually that's not what the 1/5/5A contacts are for.
Per the wiring diagrams: in a Prestel installation, Line B would connect to pins 5 and 3. Switch contact 5A would connect to pin 1. The master socket then connects to pins 1/5A and 2 of the Prestel socket. That means...

• When the Prestel terminal is unplugged, the existing telephone and/or master-socket would be connected to the line via pin 5 and switch contact 5A.
• When the Prestel terminal was plugged in and offline, its relay would connect the existing phone to the line (pins 1 and 5 connected).
• When the Prestel terminal was online, the relay would disconnect the phone and the terminal would have full control over the line.

I think this relay is the "line seize" relay in the VM6500 spec, but I'm not certain.

... And now back to LUCY.

[arg]

the Prestel Terminal Specification, I'd rather like a copy...

I had one, and I've just passed it to John Newcombe with a view to getting it scanned.

[Pernod]

Our search is over, the full SAA5070 (LUCY) datasheet can be found in https://bitsavers.org/components/philip ... ystems.pdf

[BigEd]

Excellent - what a huge PDF - 350th page, marked as page 341

[johnkenyon]

Thanks for the pointer to the LUCY datasheet, and the Viewdata specification.
I've been able to update my disassembly of the AlphaTantel ROM that I parked a couple of years ago, and suss out exactly what the code does.

[philpem]

Thanks for the pointer to the LUCY datasheet, and the Viewdata specification.
I've been able to update my disassembly of the AlphaTantel ROM that I parked a couple of years ago, and suss out exactly what the code does.

Would you be able to share that disassembly? And which ROM version did you disassemble?

I was going to disassemble V1.2 (from my TD1100) this evening, using the information from the MAME emulation and the LUCY datasheet.