When I was in the beeb it was a clear ball with sea painted black* and a bulb inside. A bent mirror behind it and a small strip lamp behind the BBC NI text underneath. A big old B&W camera pointed at it (Maybe a 405 Orthicon converted to 625?). Colour added somewhere else. It lived in corner of a small Radio continuity "cubical" (it would be misleading to call it a studio but larger than a Broom cupboard).
One of the mechanical models would be easy to reprogram. I imagine the electronic cow has very little program code and most of the storage is a bitmap. In any case if the Ident was text a search through the "dumped" EPROM would find it and as long as any replacement is exactly padded with spaces to same storage no code change is needed. So with an EPROM programmer and spare EPROMs the same type, changing the Ident is trival and even changing the image might not be hard. There are only a small number of ways then that people stored images. Any non-compressed image type is pretty easy to unravel the format of.
(* or clear sea and black land, can't clearly remember)
Hello,
As a fellow owner of one of these bovine beasts its great to see another one being brought back to life.
To refresh my memory on a few points I powered mine up and had the world spinning away on the tv while I pondered what to write. I didn't get very far as a loud crack drew my attention to the acrid grey smoke that was wafting up from the COW!! so, it looks like I get to write the first COW fault report on here.
The smoke maker turned out to be a 0.047uF RIFA class X capacitor across the mains input terminals of the "commercial" switchmode psu which is housed inside the BBC PS4/11 PSU module. There is also one between Neutral and Earth ( which should be class Y cap) the case of which looks a bit "crazed" ie small cracks forming.
These were both snipped out and the psu seems happy without them untill the replacements arrive. There are also 2 more RIFA class X caps this time 0.022uF in parallel between Neutral and Earth on the output of the EMI/RFI filter can, so for future reliability I will change these for the correct class Y types as well.
My crate is number 7 and has BBC1 with switchable Northern Ireland as the caption but also came with a set of eproms (currently installed) for BBC1 with switchable Cefax 888
I will dig up some more info and post soon
Neil
First, as you point out it needs a mixed sync input. I use a sigma Black/Burst generator. The clock generator is referenced to sync so without it the system timings are all out and the "sequence" data (the rotating map overlay) is not read from the eprom cards.
At the moment I use one of the COW's RGB outputs together with sync from one of the BB gens outputs fed into a SCART lead via a little "bodge box" where a AA cell provides the SCART pin 16 RGB switching voltage. I have a Kramer encoder tucked away somewhere but have not tried to do a pal output yet.
With the crate powered up you should have all 4 leds on the psu lit (the -12v led on mine was dry jointed and intermittant)
On the UN23S/163 card (red handle) the red led should be out, if not check the 2 toggle switches are up and the blue rotary pre-set is set to 0,4 or 8
On the bottom card rack the R/H group of 4 cards known as the "fixed memory" should all have their yellow leds lit.
The L/H group of 10 cards - the "sequence memory" should have one led at a time lit and this steps from card to card as the rotating map data is read.
Hopefully if all this is correct you should have an output.
The CEEFAX switch toggles between the plain "BBC1" logo and "BBC1" logo with "CEFAX 888" underneath or, instead of "CEEFAX" the National or Reginal Ident eg "East" "Northern Ireland" "Wales" etc.
The multipin CONTROL socket on the rear panel only uses 3 pins, the top left and right are for a remote "CEEFAX" switch the bottom left is chassis earth presumably for cable screen.
To select weather front panel or remote CEEFAX switch (or neither) are used there are jumpers on the UN26/65 card
More info to come soon
Neil
A bit more COW information for the UN23S/163 card (red handle)
The red led is lit when the top toggle or rotary switch are not "normal"
The top toggle switch has 3 positions :-
UP - NORMAL. The map data is read seqentially from the 10 sequence data rom cards
CENTRE - HALT. The world stops turning at whatever field is currently being read from seq data
DOWN - STEP. Each operation of the switch steps the map sequence forward 1 field (toggle is spring biased back to centre position when released)
The bottom toggle switch has 2 positions
UP - NORMAL
DOWN - reset map sequence to field 1 (toggle is spring biased back to UP position when released)
The blue 10 position rotary switch :-
Posn 0,4,8 - Normal
Posn 1,5,9 - Test patern 1 "blue disc"
Posn 2,6 - Test patern 2 "gold sphere"
Posn 3,7 - Test patern 3
After moving the switch you may need to toggle the top or bottom switch to display the test picture
Neil
Hmm... the "map" is perhaps a 1 bit mask between the gold and blue orb images?
My day job is a Railway signalling tech, so all the info about the COW was gleaned from t'interweb - mostly the eng.info articles and Andrew Wisemans COW page. The rest was trial and error and try all combinations of switches and see what happens :-))
I have made a start at backing up all 428 27128 eproms to image files as the roms are getting on for 28 years old now and it will only take 1 duff rom to ruin the perfect animation sequence.
One thing to be wary of, quite a few of the eprom window labels had fallen off exposing the eprom to potentially damaging light and the originals were only paper so not the best UV blocking material to start with.
As I read each batch of roms I re-label them with Avery L6008 silver heavy duty labels.
It will be interesting to see which region the wireless museums COW turns out to be. Maybe if a few more COW owners read this we might be able to get a "regional ident rom image" exchange going so you can have a bit of COW variety.
Incidentally does the museum COW have the bottom rack front cover panel ? Mine is missing and I'd love to get a good quality pic or scan to get a replacement made up
Here's an extract from Andrew Wisemans site:- http://625.uk.com/tv_logos/bbc1_85.htm
"Now we can begin to see how the globe player works. The transparent blue sphere representing the ocean is the background image. The processor uses the map data to decide for each pixel whether to leave it as the sea or overlay a pixel from the golden sphere for the land or simply plot a black pixel to represent the land on the other side of the world"
More detailed description see here http://www.bbceng.info/Designs/designs_ ... _world.htm
Regards to all
Neil
Logically they should be TIFF, one version of which is very similar to Fax transmission. It would lovely if they are essentially the data part of a TIFF file
Bearing in mind the realtively limited computing resources used to generate the COW data I suspect that the compression algorithms were pretty simple.
That's why I'd suspect TIFF (or very similar) as it was well known from Fax and computationally easy to unpack.
TIFF itself is somewhat later than the RLE used in Fax images
TIFF was originally created as an attempt to get desktop scanner vendors of the mid-1980s to agree on a common scanned image file format, rather than have each company promote its own proprietary format.
Actually "TIFF" is somewhat a red herring, but what is likely that it uses "CCITT Group 3 1-Dimensional Modified Huffman RLE" (which is used by Fax and the most basic TIFF row based format)
http://www.netnam.vn/unescocourse/computervision/104.ht m"> http://web.archive.org/web/200206281953 ... on/104.htm
If this is the case, then some simple programming in VB using a 3rd party TIFF library will allow creation of a program to import/export the ROM images to/from Baseline Row encoded TIFF images created / edited in a photo editor.
... Annoyingly TCC compresses to 136967 bytes, a little over 128K which would be a nice size ...
Interestingly, I've just reduced a copy of TCC to 378 lines (in Paint!) and got a PNG image of 139K!
However, I wonder if your TCC has location arrowheads? Early versions didn't and you are not centering it in a slide scanner, so could you get the size down by omitting the arrowheads, thus permitting you to store just one line of data and repeating it as required to create the top and bottom borders?
Probably not, but it was just a thought ...
When all else fails, read the instructions
I know we're mixing up 2 subjects here but for my own TCC compression problem
Use the standard Modified Huffmann I linked to. It's hardly any 8051 code to expand. As it's using a symbol table it will give more compression than your re-invented RLE.
Hello all,
Glad to see that there is so much interest in bovine globes. I will press on with the massive eprom backup task and am happy to pass the images on to interested parties as long as you ONLY use them for private educational studies and do not re-distribute. Remember The data and encoded images are still BBC copyright.
Likewise if anyone has access to a cow I would love a image of the regional specific roms from the "BC" and "BY" "fixed" memory boards. This is in theory 16 eproms which hold the switchable sub caption
However there is a bit of a rub with the main "BBC1" caption being a bit smaller on some of the regions which might mean more of the eproms are region specific. When I find the DVD I did of some experiments I'll put up a grab to show what I mean
The memory boards can, according to the eng.inf article take 27128 27256 or 27512 ( when they finally become available ) so it might be possible to reduce the rom count. There are dip switches on the boards to set the address range but, apart from noting down their current positions I haven't had a play yet.
Regards
Neil
Quick post tonight
Here are the screen grabs from a quick video I did whilst having my first play with my COW. This is the result of having the Northern Ireland luma (BY)and London chroma(BC) "background" "fixed memory" cards in the COW together. As you can see the "BBC1" is smaller on the NI ident than the London one. the theory was that only 16 roms needed to be unique to each region, 8 on the background luma card and 8 on the background chroma card this can be seen in the pic of the London (BC) card I posted previously - the custom roms being the 888 labelled ones.
However it looks like the kerning (?) of the type face didn't come out right and so all the roms on the "background" card had to be changed - well for NI at least. the NI roms are labelled differently too their code being BCBE and BYBE (Background Chroma (Luma) BBC Eire ???????).
As the background is the blue disc and BBC1 logo created as a quantell paintbox image then saved and compressed using the BBC RLE algorithm I'm guessing the new smaller BBC1 version (being a "new" image)subtlety changed the data enough to require the full set of roms ??
Its getting late and perhaps i'm starting to wibble, but possibly all the regions had the same small BBC1 background roms (apart from the 16 for the region caption) and the London one is the odd one out. Maybe when you get a chance you can get a pic of the roms on the Wireless museums background cards to see how they are labelled (beware flash guns on unlabelled eprom windows though!)
On this image the "CEEFAX" switch is off removing the regional branding
Time for bed now
Neil
Another late night post, just a thought but in theory there is a parallel digital video bus accessible as it goes into the DAC to give the Y V U analog component output. Maybe this could be used to capture a clean digital feed of one loop of the animation which can then be formatted to playback on any device you want.
I'm sure I've seen a parallel digital to SDI converter box (by AJA ???) which would then give me a feed to a Black Magic sdi capture card to get it on disk. I have a BBC PA20/542 656 component to parallel digital to component codec unit that came in the same deal as the COW so I have something to try feeding any potential digital COW output into. Of course I can always feed the analogue component O/P to digibeta and get a sdi output from that but then it adds an (admitably broadcast quality) D/A/D conversion
Regards
Neil
- well for NI at least. the NI roms are labelled differently too their code being BCBE and BYBE (Background Chroma (Luma) BBC Eire ???????)
Not likely! BC BE and BY BE
BE = Belfast.
It was too late at night!
When I worked for NTL, all our sites were identified by four letter codes which appeared on, amongst other things, optical fibres.
A fibre would be labelled at each end with the codes for the source and destination and circuit number at each end, thus SEVE (Circuit No.) BEXL (Circuit No.) would be a fibre from Seven Kings to Bexleyheath.
I only ever saw one variant of this where all the fibres were labelled BELF (Circuit No.) LDRY (Circuit No.) at one end (Belfast to Londonderry) and DRRY (Circuit No.) BELF (Circuit No.) at the other ...
________________________________________________
Isn't it amazing the number of times that you spot the typo after you've pressed the Submit button, then add another bit of gobbledy gook in the process of sorting out the first mess ...?
When all else fails, read the instructions
Oh hell ! Sorry for the geographical faux pass don't suppose I could have got that any more wrong. I blame a long day fighting with Network Rails (non) intelligent infrastructure monitoring system. I'm sure a quick google would have got me the bbc region codes
Well, I actually worked in BBC BE in mid to late 1970s. Communications Engineer in CTA.
Pre-Cow as the "world" was mechanical using a B&W camera in a small Radio Presentation room/booth.
I'd done two big stints at Evesham. I think I was an idiot to leave, but probably loads got made redundant later.
Oh yes! I remember that mechanical model well! There was one on display at the Langham in late 1972 in an exhibition to mark the 50th anniversary of the BBC and I took my then new girl friend to it - we celebrated our Ruby Wedding Anniversary in July!
She was really fascinated by it - probably because you could actually play with the knobs to experiment with all the combinations of foreground and background colours!
When all else fails, read the instructions
You can read an SD card and feed a DAC fast enough just with an 18F PIC I think. But maybe only in mono. It's got HW SPI interface, ability for high speed parallel out and 16 MIPs speed. They are so cheap you could do RGB by having 3 of them? I think the 16 MIPS is too slow for more than mono.
A Raspberry Pi can do it from a USB stick, but only for whatever flavour of composite is "baked in". I find mine impressive running RiscOS (in 1987 on Archimedes there was no Web Browser and TCP/IP like RiscOS has today).
In today's world of cheap tiny shiny gadgets with HW Codecs and HDMI, what we need is an all modes HDMI in to analogue converter (22, 120, 220, 405, 441, 525, 625 819 etc). Even my TV set plays USB sticks (and SD cards and HDD via USB), but does it downsample to the SCART? I must investigate. Interestingly my DVD player can output upsampled and converted to analogue 1440 x 576p or 1440 x 480p from film DVDs on component. Certainly Region 1 DVDs are greatly improved compared to NTSC RGB SCART or Composite to CRT TV (even though it had an NTSC mode).
The other cheap way to read the contents of an SD card at speed would be to copy the contents into a desktop computer memory stick on boot up and then read it from there.
