Should there be further models of the Raspberry Pi?

That wasn't an issue for me (and most likely it would have been so for the sort of customers who were prepared to solder a kit of bits together). ISTR using a spare 5v 10 or 5 amp SMPSU bought at my local Gvt Surplus store for less than a fiver, wound up to 7.5 or 8 volts to overcome the 2 volt dropout of the 7805 regulator chip.

An 8080 based version of the ZX80 would have needed an even more expensive PSU (+5v _and_ a +12v supply requirement).

--
Regards, J B Good
Reply to
Johny B Good
Loading thread data ...

Alas I was only 12 when I built mine, and had not at that time accumulated the huge collection of "might come in useful" PSUs that I have now... so I bought the Sinclair one.

Indeed... the more complicated supporting chips etc would have made the whole computer too expensive anyway. I seem to recall the value of the

1K RAM in the ZX80 was about £30 at the time!
--
Cheers, 

John. 
 Click to see the full signature
Reply to
John Rumm

Somewhat ironic that for the Spectrum they went with 4116s so had to generate not only a +12v but also a -5v supply from the 9v input.

Reply to
Guesser

It currently shows as UKP 8.88 (inc delivery) but it is well reported on. If similar adaptors could be suitably inexpensive in bulk they would remove the need for an inbuild VGA.

James

Reply to
James Harris (es)

Hey. I was neither posh nor spoiled.

And, even in my day, O Level Computer Studies was taught in many establishments.

--
(\__/)  M. 
(='.'=) If a man stands in a forest and no woman is around 
 Click to see the full signature
Reply to
Mark

And I was only teasing.

I'd left school before the BBC machines were released - we had an HP blinkenlights mini at high school, which we got to play with in math class.

Reply to
Rob Morley

I've found my RPi runs OK from an 'emergency phone charger' powered by 3 x AA cells, giving its output via a USB port. I have not found out the duration though.`

Brian.

--
______________________________________________________________ 

Brian Carroll, Ripon, North Yorkshire, UK   
 Click to see the full signature
Reply to
Brian Carroll

No offence taken.

I'd also left school before the BBC machine was released, otherwise I could not have afforded to buy one. My secondary school had no computers. I did O Level CS at a technical college.

--
(\__/)  M. 
(='.'=) If a man stands in a forest and no woman is around 
 Click to see the full signature
Reply to
Mark

An 8080 (1974) based version of the of the ZX80 (1980) would have been an 8085 (1977) based version of the of the ZX80.

The 8085 required only +5V.

--
Roberto Waltman 

[ Please reply to the group, 
 Click to see the full signature
Reply to
Roberto Waltman

Having looked at its wikipedia entry, I see it also did away with the

-5v rail (which I'd forgotten to mention) as well. It still had the data multiplexed on the address bus though, so not so easy to interface with standard memory i/o chips (intel made special memory i/o chips that could be directly connected).

It was a popular chip in some small computers and embedded controller applications, however, looking at the Z80's wikipedia entry shows it wasn't anywhere near as popular as the Zilog design. I noticed that the list for the Z80 failed to include Kenwood's TS140 transciever[1].

[1] Handy that they'd chosen the Z80 cpu for this microcontrolled transciever since there was a curious bug in the remote control interface (e.g. you could adjust the RIT up but not down).

I copied the rom code into a file on my homebuilt Transam Tuscan S100 Bus computer and disassembled the code with my own Z80 disassembler program and found a bug in the command interpreter loop counter which looked like it may have just simply been a single bit programming error of the eprom.

Burning a replacement eprom (using, afaicr, a home brewed eprom programmer - I don't think I'd acquired the DataMan programmer at that time) nicely fixed this issue.

--
Regards, J B Good
Reply to
Johny B Good

All it needed was an octal latch. The Z80 was a better option, no doubt. I just wanted to point that the alternative at the time was the

8085, not the 8080.

Right - For about the same cost the Z80 offered a much richer instruction set, a better interrupt architecture, etc. Also, while a standard CP/M OS required only the 8080 instruction set, many application written for CP/M used the extended Z80 set. So using a Z80 in CP/M machines was almost a requirement.

Reply to
Roberto Waltman

I don't think that was true since a lot of software written to take advantage of the Z80 instruction set often included 8080 support and simply tested which cpu was in control of the show and loaded the appropriate modules. This made such application software, essentially, CPU agnostic, maximising its market reach.

--
Regards, J B Good
Reply to
Johny B Good

I have never seen a CP/M application that did that. Maybe they existed, but it was not a common way of handling this. It would essentially mean writing two versions of your program (the software often was written in assembly language in those days to get the most compact code).

Most commercial software just used 8080 instructions. There was some software that only ran on the Z80, mostly public domain stuff. When I wrote software myself, I always used Z80 instructions.

Reply to
Rob

For me, the computer wasn't simply the means to an end (that didn't happen until after I put my first IBM PC together) so I had zero experience of CP/M application software. I'd simply seen references to being able to detect the CPU to allow the software to be compiled for both and for the more compact Z80 version to be loaded whenever that proved possible.

I just simply assumed that this would be a no-brainer way to get the most performance out of a commercial distribution when used on a Z80 powered system without having to make two completely different versions. My error is down to my assuming that the software companies would use this technique to make their products Z80/8080 compatable in such a transparent way as a matter of routine. It seems I stand corrected in this matter.

As did I. The stuff I was writing was, essentially, for my own personal 'amusement' so I had no need to consider the 8080 other than as an old fashioned curiousity.

--
Regards, J B Good
Reply to
Johny B Good

The reality is that the BIOS was written to a specific processor target.

Applications were pure 8080. As was CP/M itself.

Unless you coded an app for your actual hardware.

--
Ineptocracy 

(in-ep-toc?-ra-cy) ? a system of government where the least capable to lead are elected by the least capable of producing, and where the members of society least likely to sustain themselves or succeed, are rewarded with goods and services paid for by the confiscated wealth of a diminishing number of producers.
Reply to
The Natural Philosopher

Well, there may be some cases where a piece of software required a lot of memory block copies and included a Z80 detection (I don't remember if there was a documented and reliable way to do that...) and then used the LDIR instruction instead of a loop. That would be a quite local thing, maybe in a subroutine. It would not even be visible from the outside. To globally take advantage of Z80 instructions, like relative jumps, DJNZ, 16-bit subtracts, IX/IY etc would take the inclusion of two separate binaries of the entire package. And two times the coding effort, deciding on which registers and which tricks to use, etc. I never encountered that, but it may have been there as I did not have

*that* much CP/M software.

My machine was a Z80 as well. In fact it was a TRS80 model 1 that I had heavily reworked to allow it to run standard CP/M, by adding bankswitching that replaced the lower 16K of the address space with RAM. (which I had available from 64K RAM chips) I wrote the BIOS for it in Z80 code.

The same system could also run the typical TRS80 operating systems like TRSDOS, LDOS and NEWDOS/80. Those were overlay-based. When I added a harddisk to my system which was only working under CP/M (because I had not enough docs to know how to extend NEWDOS/80 for it), I wrote a CCP replacement that was using overlays as well, and was written in Z80 code. I named it ZCCP. CCP (the console command processor) normally was 2KB in size and could do almost nothing. My version was a 1KB "resident" part and a 1KB "overlay" loaded for many different commands, so it was now possible to perform many system tasks even while application programs were running, something that the TRS80 operating systems could do as well.

But it was only for personal amusement, I never distributed it. There was an open source project named ZCPR but it used applications in the TPA, not the overlay technique in CCP.

Reply to
Rob

...

Still thinking about this from the viewpoint of what might allow the makers to increase turnover and/or increase the longevity of the product and/or themselves in the face of market changes and development of competing products by other manufacturers. Possible Pi developments are, frankly, not many as the device is very well specified just now but might include:

  • models with inbuilt real-time clock (and its battery)
  • models with inbuilt wireless or similar
  • models with more RAM
  • models with greater power to the USB ports
  • models with more USB ports
  • models with multiple CPUs
  • models with VGA (possibly)

On the VGA point, as people have pointed out, the Pi could be given VGA output by simple addition of a converter lead and they are not too costly.

Regarding the USB ports, once keyboard and mouse have been connected directly all such ports are gone. Yet keyboard and mouse don't need the USB

2 speeds.

As some people want the minimality of the Pi but others want to use it for more-general computing could there be a market for a base station that the Pi could plug in to, one that provided a mains-powered USB hub, more USB ports, power to the Pi and some other gizmos? Such a component could be left on a desk with many leads permanently connected.

Just suggestions, not criticisms. Not everyone will want extra hardware but people's needs vary and people like choice even if at a range of prices. For example, if the foundation announced tomorrow that a new model was coming out with a real-time clock I'm sure some people would be interested even if it was a little more expensive. And if they came out with models that could stack together that would interest still further people.

The computer industry - from CPUs to motherboards to almost any other components - is awash with model variants. Presumably they make so many variants available because their customers want the choice and are happy to pay different prices.

James

Reply to
James Harris

Why not just buy a little Atom PC that has all those things already and has socketed ram and expansion cards though?

Reply to
Guesser

The reason is power consumption. Simple as that. My previuos "little PC" webserver used 17W, the Pi only 3,5W. No more "little PCs" for 24/7 uses here.

Paul

Reply to
Paul Berger

My atom server with 1.5TB of spinning rust uses about 8W,

How much would a 1.5TB pi use?

--
Ineptocracy 

(in-ep-toc?-ra-cy) ? a system of government where the least capable to lead are elected by the least capable of producing, and where the members of society least likely to sustain themselves or succeed, are rewarded with goods and services paid for by the confiscated wealth of a diminishing number of producers.
Reply to
The Natural Philosopher

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.