Should the Elk have been a 'full' Beeb?

on-topic Acorn-related news and discussions not covered by the other forums
RobC
Posts: 1736
Joined: Sat Sep 01, 2007 9:41 pm

Re: Should the Elk have been a 'full' Beeb?

Postby RobC » Wed Sep 07, 2016 3:53 pm

ThomasHarte wrote:
RobC wrote:Agreed. I'd love to create an improved video ULA to give 16 proper colours in mode 2 and a larger palette. I have the components ready and finally have the time on my hands but now I've gone and broken my wrist :(

I'm an electronics dunce, and have recently evidenced this elsewhere, but surely the problem isn't the ULA? Red, green and blue come out as digital signals down in the lower right of the schematic, then a bunch of logical components do the resulting mixing to produce video, seemingly in a binary fashion? If you added a brightness pin, or made those ULA outputs analogue, surely you'd also have to change a whole bunch of stuff in the trimmed schematic hastily edited and attached?

Not really - you would ignore that part of the circuit and take the (now analogue) RGB and sync signals out to a monitor by a different connector (or wired direct to the existing RGB output). The issue with the ULA is that it does the serialisation and palette look-up. If you want to do 16 proper colours in mode 2, it needs to be replaced or augmented.

My idea was to use a Brooktree BT450 RAMDAC which has a 16 x 12-bit palette to give 16 colours from a palette of 4096 (16 shades of red, green and blue).

paulb
Posts: 765
Joined: Mon Jan 20, 2014 9:02 pm

Re: Should the Elk have been a 'full' Beeb?

Postby paulb » Wed Sep 07, 2016 4:02 pm

ThomasHarte wrote:
RobC wrote:Agreed. I'd love to create an improved video ULA to give 16 proper colours in mode 2 and a larger palette. I have the components ready and finally have the time on my hands but now I've gone and broken my wrist :(

I'm an electronics dunce, and have recently evidenced this elsewhere, but surely the problem isn't the ULA? Red, green and blue come out as digital signals down in the lower right of the schematic, then a bunch of logical components do the resulting mixing to produce video, seemingly in a binary fashion? If you added a brightness pin, or made those ULA outputs analogue, surely you'd also have to change a whole bunch of stuff in the trimmed schematic hastily edited and attached?


This was one of my "speculation about the ULA" topics: Enhancement: Palette Definition. Corrections welcome! :wink:

User avatar
daveejhitchins
Posts: 3631
Joined: Wed Jun 13, 2012 5:23 pm
Location: Newton Aycliffe, County Durham

Re: Should the Elk have been a 'full' Beeb?

Postby daveejhitchins » Wed Sep 07, 2016 4:39 pm

I'm working from here, and with Dave (Hoglet) to offer flexibility e.g. The normal Electron input outputs would work as normal but with options, provided by connectors on the FPGA board, to have 'improved' video, if required, and other goodies. I'm trying to keep the purists placated e.g. those who just need a replacement ULA - due to their original one dying - Also, offer to the extremists (most of us, probably!) as many bells and whistles as possible - They will however, be limited to what can be achieved with the chosen FPGA - so there will, no doubt, be disappointment for some!

What I don't envisage is a lot of changes/modifications to the Electron itself. Currently I'm working on just removing the ULA socket and replacing with standard sockets that will allow either the original ULA socket to be plugged back in or for the FPGA board to be plugged in - KISS

Dave H :D
Parts: UM6502CE, GAL22V10D, GAL16V8D, AS6C62256A, TC514400AZ, WD1772, R6522, TMS27C512, AT28C256
Products: ARA II, ABR, ATI, AP6, MGC, AP5 . . .
For a price list, contact me at: Retro Hardware AT dave ej hitchins DOT plus DOT com

User avatar
algenon_iii
Posts: 67
Joined: Sat Nov 25, 2006 6:49 pm
Location: Cardiff

Re: Should the Elk have been a 'full' Beeb?

Postby algenon_iii » Sat Sep 10, 2016 11:24 am

ThomasHarte wrote:
algenon_iii wrote:The Teletext colour palette of the BBC/Elk is one of my pet hates (especially compared to the C64 or even Spectrum). IIRC the BBC RGB port never had an intensity pin and the Model A didn't have an RGB port. So just adding colour intensity on this computer could have been a simple opportunity to provide a genuine 16 colour mode/ palette to use / choose from, instead of wasting the 4th colour bit on whether a colour flashes.

Maybe your monitor is too good? I'm going to guess that most Electrons, like mine, were connected up via UHF, in which case the colours aren't all that sharp. Contrasting screen shots from my emulator attached — to repeat, because other authors have given the whole area a bad name: it ends up with those results because it performs a composite encoding of the video signal, then performs a composite decoding. No subjectivity applied, anywhere.


That's a possibility as when I had the electron as a kid I was running it off a B&W portable TV so never really noticed it at home, it wasn't till later that I personally started running it on a decent colour display. But even at the time I do remember the BBCs at school which were all connected to the classic Cub monitor (also friend of mine with an electron had a monitor) so I got the colour palette in it's full gory glory.

User avatar
1024MAK
Posts: 6684
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: Should the Elk have been a 'full' Beeb?

Postby 1024MAK » Sat Sep 10, 2016 1:48 pm

With respect to the colours...

There are lots of variables that affect the actual displayed picture of a composite video, UHF connection or even an analogue RGB/SCART connection on a TV / monitor. The actual colour, brightness and contrast is all affected by the user controls, the engineers internal settings inside the TV or monitor and the component variations inside both the TV/monitor and the computer. With UHF, it is also affected by how well tuned in the TV was to the modulators output. With the RGB-SCART leads that people make, or buy from sellers on eBay or similar, the values of the resistors used affect the displayed colours.

As both the Beeb and the Elk output digital TTL RGB output signals on their RGB monitor ports, you get 100% colour, or none. Hence the available colours being the three primary colours of red, green and blue, plus the available mixes of these, yellow, cyan, magenta and white.

With the RGB monitors, like the Cub, most people wanted bright vibrant colours. So that's what they were designed and set-up for.

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

Commie_User
Posts: 872
Joined: Wed Jan 27, 2016 12:50 am

Re: Should the Elk have been a 'full' Beeb?

Postby Commie_User » Sat Sep 10, 2016 1:58 pm

So BBC Micros couldn't be used to generate TV test patterns? That would have been highly ironic.

A onetime friend of mine, in the line of studio building, didn't take long to mod a Vic20 for such a purpose.

User avatar
1024MAK
Posts: 6684
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: Should the Elk have been a 'full' Beeb?

Postby 1024MAK » Sat Sep 10, 2016 2:07 pm

TV test cards, were actually printed cards mounted and fixed, with a camera pointed at them... At least back in the early 1980's. It was only later that purpose built electronic test cards were made.

Keep in mind that a broadcast standard test card has to exceed the specifications of the broadcast signal. The fine lines and the colours used are used by engineers to check the performance of various parts of the analogue system. At the studio, at the transmitter, and in the shop or the viewers home. The test card also enables an engineer to correctly adjust an analogue TV.

So, although the Beeb was used by the BBC in it's programs, including program lead-in count-downs. As far as I am aware, they are not normally used for test images.

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

paulb
Posts: 765
Joined: Mon Jan 20, 2014 9:02 pm

Re: Should the Elk have been a 'full' Beeb?

Postby paulb » Sat Sep 10, 2016 2:26 pm

1024MAK wrote:So, although the Beeb was used by the BBC in it's programs, including program lead-in count-downs. As far as I am aware, they are not normally used for test images.


Yes, the Beeb's specification did not exactly state that it "be able to run the entire technical infrastructure of the BBC in 32K RAM". :lol:

Still, the Beeb did apparently get used in lots of roles within the BBC, anyway, partly due to certain additions to the specification that the engineers thought might be useful. But due to obvious colour depth and resolution issues, it did not do the test card.

Richard Russell
Posts: 163
Joined: Sun Feb 27, 2011 10:35 am

Re: Should the Elk have been a 'full' Beeb?

Postby Richard Russell » Wed Oct 12, 2016 11:26 am

1024MAK wrote:TV test cards, were actually printed cards mounted and fixed, with a camera pointed at them... At least back in the early 1980's. It was only later that purpose built electronic test cards were made.

In fact the electronic version of Test Card F entered service in 1984 I think, so perhaps earlier than you thought. Before that high-quality slides would have been used (Test Card F appeared in 1967, and previous testcards would have been slides for some while too).

Richard.

B3_B3_B3
Posts: 67
Joined: Sat Apr 08, 2017 9:42 pm

Re: Should the Elk have been a 'full' Beeb?

Postby B3_B3_B3 » Mon Apr 10, 2017 10:02 pm

At the time I was surprised that the Electron was 32K when the spectrum was 48K, and when the Beeb (which I had) was viewed as memory challenged, especially in 20K graphics modes:

but assuming that more than 32K memory was infeasible
(even if other savings had been made: e.g. drop the monitor port/connectors perhaps leaving RGB/Composite signals available via edge connector only, on grounds most electrons would only be connected to a TV, ULA simplification etc):

Perhaps Acorn could have fitted a faster 3 or 4 Mhz 65(C?)02 so that the much slower 4bit wide RAM access was countered by the faster ROM access.

On a different note, some thoughts on the ULA:

was the BBC SN76489, sound chip expensive? If not would fitting it and thus simplifying the ULA circuitry by removing from it the need for a sound generation section have saved much cost?

With no real mode 7 teletext, could mode 6 (and perhaps 3?) not been made into proper text modes (i.e. with 1K or 2K respective video space requirements)?
On a electron perhaps a 'proper' mode 6 text display needing only 1K could implement the teletext upper char-set codes (160-255) and other char differences, so that rough monochrome teletext emulation was possible (control codes shown as spaces but do nothing).

If Mode 0 was troublesome (because ULA working at its max) perhaps they could have dropped it (I never used it....)
For school children doing mode 0 programming projects at home, perhaps mode 0 could be simulated by mode 4 but with, perhaps 4 column by 8 row characters to simulate 80 columns: allowing code screen layout to be tested but with less legible text .

As suggested up thread, if modes 2, 1 and 0 caused so much trouble and slow down with 4 bit wide RAM, perhaps just skip them (and map to nearest 10K mode) . I wonder how much cost the reduced ULA complexity would save...

As mentioned up thread, surely one or two Atari style D joystick connectors would have suited the intended for school /games market at low cost: presumably could have been read via Adval (or osbyte 80) unless a real BBC style ADC-based analogue joystick interface was plugged into the rear expansion bus.

paulb
Posts: 765
Joined: Mon Jan 20, 2014 9:02 pm

Re: Should the Elk have been a 'full' Beeb?

Postby paulb » Mon Apr 10, 2017 11:23 pm

B3_B3_B3 wrote:was the BBC SN76489, sound chip expensive? If not would fitting it and thus simplifying the ULA circuitry by removing from it the need for a sound generation section have saved much cost?


I think that the sound is pretty much the only thing that I can regard as inexcusable on the Electron. When you get a game that is rather nice on the Electron, the sound is always going to be the thing that lets the side down.

Other things are more excusable. The slowdown in higher-bandwidth graphics modes is understandable in the context of the cheap memory requirement, and it isn't as if those modes are completely unusable as a result. Mode 7 was useful for saving memory on the Beeb, but not having it on the Electron wasn't exactly a show-stopper. Things that used the extra memory and not graphics, like big text adventures for instance, could have been released on disk for the Electron had there been enough demand (and publisher interest). (Double-height text and flashing colours in a somewhat different character set is all very well, but some kind of attribute colour mode would have been good enough. See davidb's palette experiments for an idea of what might have been possible, that also extending to graphics.)

But baking in a simple tone and noise generator into the ULA, to more or less compete nose-to-nose with the Spectrum, was cost-cutting taken too far. Furthermore, it appears that you can't even get expansions to output via the Electron's speaker, although my own experiments in this area are not very sophisticated, and the speaker circuit might not really support decent audio, anyway. So, improving the sound output requires adding more hardware than is necessary on the Beeb.

Welcome to Stardot, by the way!

B3_B3_B3
Posts: 67
Joined: Sat Apr 08, 2017 9:42 pm

Re: Should the Elk have been a 'full' Beeb?

Postby B3_B3_B3 » Tue Apr 11, 2017 9:48 am

Thanks.

I still think a proper 1K text mode 6 would have been useful and a cheap improvement at the time..

paulb
Posts: 765
Joined: Mon Jan 20, 2014 9:02 pm

Re: Should the Elk have been a 'full' Beeb?

Postby paulb » Tue Apr 11, 2017 10:51 am

B3_B3_B3 wrote:Thanks.

I still think a proper 1K text mode 6 would have been useful and a cheap improvement at the time..


It is interesting to try and quantify the cost. They could have just kept the different discrete components, but the cost of those components might have been significant once they've all been added up. Also, reading through various accounts of Acorn's history, one gets the impression that people felt that Acorn could have optimised for cost more aggressively, particularly later in the Beeb's life. So, there was presumably disappointment that the B+ and Master series were not cheaper. I read one remark about how the board design was an opportunity to reduce costs that Acorn had not supposedly taken. But this is exactly the kind of thing that they did for the Electron!

B3_B3_B3
Posts: 67
Joined: Sat Apr 08, 2017 9:42 pm

Re: Should the Elk have been a 'full' Beeb?

Postby B3_B3_B3 » Tue Apr 11, 2017 11:08 am

B3_B3_B3 wrote:...I still think a proper 1K text mode 6 would have been useful and a cheap improvement at the time..

Replying, paulb wrote:It is interesting to try and quantify the cost. They could have just kept the different discrete components, ....

I was presuming no extra non-ULA circuitry, just different ULA contents (perhaps at expense of other modes eg mode 0).

User avatar
1024MAK
Posts: 6684
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: Should the Elk have been a 'full' Beeb?

Postby 1024MAK » Tue Apr 11, 2017 2:03 pm

I don't know how "full" the ULA is, but I think trying to shoehorn in a low memory text mode into the ULA as well as keeping everything else would have been a non-starter. More on the normal way of creating a text based display in your Beeb thread.

I too think that the cost cutting resulted in cutting too many things. But ultimately, the Electron was suppose to be an educational machine to complete with the ZX Spectrum (which was also supposed to be an educational machine). Both aimed at home users leaning to use and program in BASIC.

As it happened, the ZX Spectrum became the leader in its market segment very quickly due to its low cost, which made it hard to complete with. Acorn launched the Electron in the run up to Christmas, but problems with production of the ULA resulted in Acorn being unable to keep up with demand. So potential customers went elsewhere. This was Acorns downfall. Even if it had been possible to have better sound, and a low memory text mode, would it have been possible to get large numbers into the shops for the Christmas period? I don't think so.

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

paulb
Posts: 765
Joined: Mon Jan 20, 2014 9:02 pm

Re: Should the Elk have been a 'full' Beeb?

Postby paulb » Tue Apr 11, 2017 2:06 pm

B3_B3_B3 wrote:
B3_B3_B3 wrote:...I still think a proper 1K text mode 6 would have been useful and a cheap improvement at the time..

Replying, paulb wrote:It is interesting to try and quantify the cost. They could have just kept the different discrete components, ....

I was presuming no extra non-ULA circuitry, just different ULA contents (perhaps at expense of other modes eg mode 0).


The thing that allowed the ULA to work at all is arguably its uniformity: modes 4 to 6 need every other RAM access slot during the active display region, modes 0 to 3 need every RAM access slot during that region. Had modes 3 and 6 not been needed, that would have made things simpler, but supporting them merely involved introducing a two display line pause between each character line (as far as I remember).

The different modes are effectively combinations of basic but essential features: data input at those two rates, pixel output at 160/320/640 per line rates, 1/2/4 bits per pixel colour interpretation, palette-to-colour encoding, and that "pause" feature (which admittedly is the odd thing out).

If you dropped mode 0, it might be to simplify the pixel output, and you'd therefore also be dropping mode 3 and only offering a 320 pixel horizontal resolution, which was not good for 80-column users. Or maybe you'd be simplifying the RAM access regime and would therefore also be dropping modes 1 to 3, which means that you would only be supporting four colour screen modes.

So I'm not sure that it would be so easy to remove individual modes. If I had to drop stuff from the ULA it would indeed be other things like the sound generator or even the cassette-related functionality, although I don't have much familiarity with either of those.

B3_B3_B3
Posts: 67
Joined: Sat Apr 08, 2017 9:42 pm

Re: Should the Elk have been a 'full' Beeb?

Postby B3_B3_B3 » Tue Apr 11, 2017 2:18 pm

I suggested dropping Mode 0 because Steve Furber mentioned problems with getting it to work on the ULA due to speed: so I thought maybe dropping it might have made the elk available sooner.

Others suggested dropping Modes 0,1,2,3 to simplify the ULA RAM access circuitry (no need to halt CPU) which might also have speeded electron development (and perhaps made a text mode 'affordable').

Surely a machine competing at Spectrum price level needs a cassette interface?

Keeping the BBC sound IC might also have speeded ULA development by removing the need to design and test a sound section.

:)

I

User avatar
1024MAK
Posts: 6684
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: Should the Elk have been a 'full' Beeb?

Postby 1024MAK » Tue Apr 11, 2017 2:23 pm

The sound section though, is not likely to have been that much work compared to the video section...

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

User avatar
pstnotpd
Posts: 392
Joined: Wed Jan 20, 2010 11:05 am

Re: Should the Elk have been a 'full' Beeb?

Postby pstnotpd » Tue Apr 11, 2017 3:46 pm

When I finally got a Beeb some years back I was actually surprised that the game sounds weren't that different from the electron versions I remembered from the 80's. Indeed the sound restrictions were annoying at the time but that was compared to the C64 which was the big one in the Netherlands back then.

In my view the biggest shortfall was the processing speed (compared to a B) which seemed to be trivially taken to very acceptable levels by the elektuur/slogger board. Elektuur was a dutch electronics magazine which started to fiddle with dead cheap German electrons whicht were dumped on the Dutch market.

Second shortfall was the unusable hardware scrolling.

paulb
Posts: 765
Joined: Mon Jan 20, 2014 9:02 pm

Re: Should the Elk have been a 'full' Beeb?

Postby paulb » Tue Apr 11, 2017 4:54 pm

pstnotpd wrote:In my view the biggest shortfall was the processing speed (compared to a B) which seemed to be trivially taken to very acceptable levels by the elektuur/slogger board. Elektuur was a dutch electronics magazine which started to fiddle with dead cheap German electrons whicht were dumped on the Dutch market.


I guess that small RAM chips with eight data lines were fairly cheap - the Slogger Pegasus uses one for its workspace - and they probably became very cheap once Samsung started to ramp up memory production (apparently). It is said that Acorn were aware of the turbo option, but maybe they figured it out too late.

(It's amusing that in the Hermann Hauser interview transcript he makes references to the genius of Furber and Wilson for designing a great instruction set for the ARM, in contrast to IBM doing extensive simulations to figure out what the optimal RISC instruction set should be, and yet the need to effectively cache the lowest pages on the 6502 - those that include the crucial zero page bottleneck - is exactly the kind of outcome you would get from analysing the architecture's performance. Maybe no-one really cared enough about it until it was too late.)

pstnotpd wrote:Second shortfall was the unusable hardware scrolling.


This was what got me interested in the complexity of the ULA to begin with. Was it really impractical to add a few more bits to the screen start register? If so, a lot of potential enhancements fall off the table straight away, I imagine.

User avatar
pstnotpd
Posts: 392
Joined: Wed Jan 20, 2010 11:05 am

Re: Should the Elk have been a 'full' Beeb?

Postby pstnotpd » Tue Apr 11, 2017 5:53 pm

The restrictions of the electron compared to the C64 and Beeb took me on a path of striving for efficient thinking and coding, which has proven quite worthwhile for me up to this day. I still believe that, would I have opted for the C64, I would have stuck to just playing games.

Just something I thought I had to add to this topic....... :D

User avatar
algenon_iii
Posts: 67
Joined: Sat Nov 25, 2006 6:49 pm
Location: Cardiff

Re: Should the Elk have been a 'full' Beeb?

Postby algenon_iii » Tue Apr 11, 2017 9:23 pm

If as has been suggested Mode 0 had been dropped, it might have been an idea to replace it with a halfway house resolution sitting between Mode 1 and 5 that might have been better suited to gaming. Basically a compact 4 colour 30 column x 25 row text / 240x200 resolution mode (using blanking lines at the top and bottom of the screen) taking up around 12k. That would have provided a similar resolution to the ZX Spectrum but without the colour clash.

If released in 1983 in time for Christmas the electron could have stood a chance to do well, but by summer 1984 when it finally became available in decent numbers the home computer world had moved on. The Spectrum 48 was £129, C64 was already at £199, and when the CPC 464 was announced around that time it was priced at £249 inc monochrome monitor. So in 1984 I think it would have needed better sound, more colours, 64k and maybe Slogger Turbo level speed to have stood a real chance at the £199 price point.

User avatar
sydney
Posts: 1968
Joined: Wed May 18, 2005 9:09 am
Location: Newcastle upon Tyne

Re: Should the Elk have been a 'full' Beeb?

Postby sydney » Tue Apr 11, 2017 10:14 pm

algenon_iii wrote:..... at the £199 price point.


I wonder how many people actually paid that much for one. If it had been a success and remained at that price I would never have owned one.I'd probably have got a speccy instead because at £199 my parents could not afford it.

User avatar
algenon_iii
Posts: 67
Joined: Sat Nov 25, 2006 6:49 pm
Location: Cardiff

Re: Should the Elk have been a 'full' Beeb?

Postby algenon_iii » Wed Apr 12, 2017 6:43 am

sydney wrote:
algenon_iii wrote:..... at the £199 price point.


I wonder how many people actually paid that much for one. If it had been a success and remained at that price I would never have owned one.I'd probably have got a speccy instead because at £199 my parents could not afford it.


I agree that it only became popular when the price was cut heavily, I'm sure one friend only got his when it was sub-£100 bundle (and was cheaper than the Spectrum / Spectrum+). Its price would have had to have come down anyway even if it hit the shops before Christmas '83, because by the summer '84 the C64 was £199, so it would have been overpriced.

Originally the plan was to sell the electron for £150, if they managed to make that the release price in '83 they'd have been onto a sure-fire winner.

User avatar
danielj
Posts: 5148
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester

Re: Should the Elk have been a 'full' Beeb?

Postby danielj » Wed Apr 12, 2017 7:55 am

I think we paid full price. The thing that sold it iirc was that it used the same programming language as the computers in school and had a proper keyboard. I was 7 and don't remember games coming into it other than that it definitely had them (efmba, snapper, arcadians and hopper spring to mind as the first ones I was aware of), and I personally never felt I missed out in that regard.

If it had further cut modes that the beeb offered (eg 0) it would have lost elements of one of its selling points...

d.

User avatar
algenon_iii
Posts: 67
Joined: Sat Nov 25, 2006 6:49 pm
Location: Cardiff

Re: Should the Elk have been a 'full' Beeb?

Postby algenon_iii » Wed Apr 12, 2017 9:41 am

Forgot to say that I think we got ours when it was full price or nearly full price, purely because it was a cheap BBC 'compatible' computer.

As mentioned before, with hindsight, a stripped down Model A (cheaper psu, no empty chip sockets, no provision for additional ports on board and all on a smaller PCB) with 32K might have been the best approach to take and doable for £200 (but probably not the original £150 target price).

User avatar
1024MAK
Posts: 6684
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...

Re: Should the Elk have been a 'full' Beeb?

Postby 1024MAK » Wed Apr 12, 2017 11:24 am

One thing to add: a cut down BBC with extremely limited on-board expansion and so a PCB half the size. Say a 32k model A equivalent as far as software is conserved (see below) called a A+. And a less powerful and less expensive power supply. It could have been able to have been brought to market a lot quicker than designing a new computer like the Electron.

But it is debatable whether it would have been able to be sold at the price point that they (Acorn) were aiming for. And it is clear that they did not want to disrupt their existing educational markets where they were selling most of the model B machines...

Despite its limitations, the Electron was mainly a good idea, just one that did not work out.

Features of the A+
  • Compact size case but uses a Electron like keyboard so no number pad or separate function keys.
  • 32k RAM (maybe using 32k DRAM chips like Sinclair did).
  • Supports display modes 0 to 6. Upgradable to Mode 7 by fitting a Teletext chip.
  • Maybe have the sound output to the RGB port and modulate it to the TV if this is a lower cost than having a built in audio amplifier and speaker.
  • A single ROM socket containing the ROM with OS and BASIC (32k byte ROM). No other ROM sockets.
  • External ports: RGB video out, TV RF video out, cassette in/out (does not support motor control), AC power in, edge-connector expansion/ROM cartridge port.
  • Apart from the ROM and the socket for the Teletext chip, all chips directly soldered to the board.
  • No provision for any on board expansion of any type apart from the socket for the Teletext chip. Note: no provision for a "user" 6522 VIA.
  • PSU not able to supply power to any expansions - expansions would require a more powerful PSU. If the cassette and sound circuits can be designed to operate from a single rail supply (+5V only), this simplifies the PSU even more. Maybe the PSU is an external unit instead of being build in.

Mark
For a "Complete BBC Games Archive" visit www.bbcmicro.co.uk NOW!
BeebWiki‬ - for answers to many questions...

Commie_User
Posts: 872
Joined: Wed Jan 27, 2016 12:50 am

Re: Should the Elk have been a 'full' Beeb?

Postby Commie_User » Wed Apr 12, 2017 11:38 am

I think you've detailed what I was trying to say in my opening post.

The Electron was an excellent concept but it was wasn't on time, had nowhere to park and was just a crippled version of a bigger machine for people. The Spectrum was already out, so cheap and cheerful was already catered for. The Commodore 64 was beginning to nail down the seriously glossy end of pleasure computing and if you wanted a BBC-type machine, there was already the BBC.

And wanting a BBC-type machine was apparently just a passing fancy for 1983. Still, you could play most of the games on the proper BBC, so those owners must have been laughing.

B3_B3_B3
Posts: 67
Joined: Sat Apr 08, 2017 9:42 pm

Re: Should the Elk have been a 'full' Beeb?

Postby B3_B3_B3 » Wed Apr 12, 2017 12:13 pm

If their idea was a work at home BBC BASIC machine for taking ComputerScience projects in/out of school then surely mode 4,5,6 support was enough.
A 1K mode would seem useful or else no room at home but room at school.....
(on the other hand 28K would be a long load from cassette so...)

mode 0 could be simulated for BASIC with mode 4 and sliced 4 by 8 characters....

I wonder if abandoning 16K/20K modes and the cost free alternate video/CPU access of the BBC for the 'video ULA as master of shared video ram area style' of the Spectrum's ULA might have been simpler and/or cheaper and/or faster to implement?

I wonder if keeping the BBC style case might have been cheaper than designing a new Elk case... even if an external PSU and cheaper key 'switches' /membrane was used...
Changing the keyboard layout from BBC always seemed an odd choice to me.... Cheaper keys yes, moving them around a lot, no? I could understand function key functionionality via a FUNC 'shift' key, instead of separate orange keys though.
Plus one/disc inteface could then have been just have been internal pcbs on ribbon cables...


Steve furber said they never examined any low cost competitor hardware.....
( I always thought the Spectrum design was impressively minimalist perhaps that might have inspired them....)
Criticisms of Elk memory design ignore the fact that the price was the overriding feature.

As an aside The VIC-II CRTC has MOS part numbers(eg 6569): would they have had to sell it to any other company that asked?
I suppose a BBC BASIC cartridge* for a C64(or other low cost unit) might have undermined BBC education sales? *would need ability to save/read BBC cassettes :)

Commie_User
Posts: 872
Joined: Wed Jan 27, 2016 12:50 am

Re: Should the Elk have been a 'full' Beeb?

Postby Commie_User » Wed Apr 12, 2017 1:28 pm

B3_B3_B3 wrote:I suppose a BBC BASIC cartridge* for a C64(or other low cost unit) might have undermined BBC education sales? *would need ability to save/read BBC cassettes :)


I assume the very absence of a BBC cartridge proved its irrelevance for that market. There was Simon's BASIC for those who wanted it, which wasn't many.

And there were too few kids wanting compatible machines for homework anyway, with a too small range of actual compatible BBC-Electron software also. So if the educational market compromised itself in reality, any Commodore incursion wouldn't have been a factor at all.


Return to “general”

Who is online

Users browsing this forum: No registered users and 2 guests