New Spectrum Next

[lforos]

I'm not sure if this is the right branch for this, but anyway...

On Kickstarter they have published a project to revive none other than the mythical Spectrum. Its name is ZX Spectrum Next.

Aimed at running original games (and those that can continue to be developed in the pages of retrocomputing) and with a very similar feel to the Spectrum+ (in appearance, keyboard, etc.). They maintain some key aspects of the architecture, such as the use of the Z80 as a processor or Mic and Ear inputs and outputs for cassette input, but adding more memory (512kb), more modern video outputs (RGB, VGA, HDMI), SD slot, etc. Even a Wifi option.

Among the participants in the project is Rick Dickinson who designed part of the ZX80, ZX81, ZX Spectrum and Spectrum+.

At the moment it has accumulated more than 400,000 pounds of the 250,000 requested to execute the project and there is still almost a month of funding left. The first units are expected to be delivered in January 2018.

Certainly, Spectrum nostalgics (like myself) will be excited...

[whoololon]

Let's see if this one makes it to the stores, not like Spectrum Vega...

[Yorus]

So far the best project of this type, although in some ways less ambitious, has been the ZX-Uno. Spectrum Next seemed to want to go too far, especially because of the extra cost and the complications that the keyboard issue brings them, but I believe (I'm speaking from memory, and this is not my strong point) that they already have a prototype of the board that can be mounted inside the casing of a rubber duck.

[Namiga]

Well, I think that wanting to "revive" an old glory is always good. But what they propose is a kind of Frankenstein, strange and with steroids...
Let me explain.
The keyboard is more like the QL's (and I like that, so many hours with the rubber one burned me out). The exterior aspect is also more like the QL.
At the hardware level, I think they don't tell the whole truth. They talk about resurrecting a classic using a Z80... but where is it?? They talk about FPGA (programmable array), so it's an emulation. In their own FAQ they talk about whether they could "emulate" other retro classics but that's not in their plans... yet.

Let's see... I'm not complaining. But it was my first computer and I remember it with special affection. It's a classic. You can't adapt it to new times. It was what it was... and that's it.

Do you want to surf from a Speccy?.. please...

The only thing I see sense in is the Scandoubler, to be able to use a VGA monitor (something like that was also used with the Amiga). But it's for pure functionality. From there to putting an HDMI... wow...

I'll follow it with interest, but before that I'd go for another Amiga (My cybernetic love) than spend my money on that... monstrosity (with affection)

[Yorus]

I don't have much idea of how an FPGA works, but the creators of the ZX-Uno say the same thing, that it's not an emulation, not software emulation at least.

What doesn't convince me is that they're trying to make a new Spectrum with new possibilities, and if they were to do it I would prefer it to be with real components, not another one behaving like them, but I imagine it's more viable this way. But I'm referring above all to the fact that it will be a machine without a future, because there won't be software (or only four things will come out) that can take advantage of it. These evolutions should have come out in their time as a relay of the original Spectrum, something that the Sam Coupé tried but which arrived when the 16 bits had already taken over the market. In Russia they did come out with clones with more possibilities like the Pentagon or Scorpion, and they even have games.

I stick with the ZX-Uno, which for just over 60€ allows us to use the Spectrum with the DivIDE and other integrated accessories and also can behave like many other machines, so we can have our precious machines saved and use a gadget the size of a Raspberry.

[cobito]

An FPGA is an array of programmable transistors. That means it can be wired to create logic gates as one wants. If you program an FPGA to make an and gate, internally within the FPGA there is physically an and gate; it's not a program that behaves like the gate but a gate itself.

To wire an FPGA, a programming language like vhdl or verilog is used. These languages may look similar in syntax to software programming languages like C but they are very different. They are not sequential, that is, they are not executed one line after another. The only thing they do is wire the logic of the FPGA. They are nothing more than a tool to modify the hardware. It is not software that runs inside the hardware, it is software that defines the hardware: real hardware, not emulated hardware.

In fact you can program a kaby lake with vhdl, send it to have the ASICs made and have physical processors that have been created based on that code. The only thing the code does is say which gates are connected to which gates.

I don't know if I made myself clear.

[Yorus]

Very good explanation @cobito, within my almost non-existent knowledge of electronics.

I have been looking for something about programmable transistors to see if I could understand how the "program" was "recorded" in them. I mean, just like in a magnetic disk, we know that parts of a layer are polarized to be zeros and ones, I couldn't quite understand what makes these matrices of transistors operate in one way or another. Perhaps with the programmer, in some way, by applying certain currents, they change the properties of the silicon at the appropriate points so that each transistor or set of these changes its way of operating?

Because it sounds to me like a programmer is used in the style of Eprom programmers, although in the latter case, it is software that is recorded.

[cobito]

@Yorus Well, the truth is that I don't know the inner workings of FPGAs. It's a world I'd like to get into seriously when I have some time because it intrigues me. I think it's the natural next step after playing around with things like Arduino and Raspberry Pi for a while.

A few months ago I published a post related to this topic focused on CPU design. It talks more about logic than hardware but I think it's interesting.

[Namiga]

@cobito said in New Spectrum Next:

An FPGA is a programmable transistor array. That means it can be wired to create logic gates as you want. If you program an FPGA to make an AND gate, internally within the FPGA there is physically an AND gate; it's not a program that behaves like the gate but a gate itself.

To wire an FPGA, a programming language like VHDL or Verilog is used. These languages may look similar in syntax to software programming languages like C but they are very different. They are not sequential, that is, they are not executed line by line. The only thing they do is wire the logic of the FPGA. They are just a tool to modify the hardware. It's not software that runs inside the hardware, it's software that defines the hardware: real hardware, not emulated hardware.

In fact, you can program a Kaby Lake with VHDL, send it to have the ASIC made, and have physical processors that have been created based on that code. The only thing the code does is say which gates are connected to which gates.

I don't know if I explained myself.

You explain yourself perfectly.
What I was referring to is that to implement that hard, a soft is used... and the one who implements it "can" not know everything that needs to be known about the hard to be programmed... and may not "recreate" exactly the functioning of the original hard... only what may interest them.
That's why I refer to a soft emulation, and my assertion may be incorrect.

The reality is that I don't see much sense in spending time and money on it, if there are already perfectly functional software emulators.

In the style of cars... a fantastic concept car, in this case a Concept Sinclair