New microdrive internal replacement

[Peter]

That part, is a Clive philosophy matter, affordable, low cost, easy and efficient.

The movie "Micro men" quotes Clive Sinclair: "Practicality, simplicity and elegance are the pillars of my vision."

I'd find this elegant, if it was smaller and a single board device. Instead of a three boards, of which one is external.
Ever considered to solder the RP2040 directly and to to get rid of the external electronics, making it small and actually internal as the title says?

For a minute, I even considered to do that myself, for fullsize SD card, so it is not endangered to get lost in the microdrive slot.
Your circuitry would in principle allow it. And it would save further costs, as this project seems so focussed on pricing.
Unfortunately, microdrive storage capacity, while fine for the Speccy, has always been too small for the QL.
So I'd buy such a version, if someone else does the job, but am not sufficiently motivated to design it myself.

[Peter]

Found the point in the video where Sir Clive says it:
https://youtu.be/XXBxV6-zamM?t=3251
And he added "Elegance above all"

[stephen_usher]

But ... not now. I encourage you to take this idea and implement it. For sure some people would love it. Why not?

It already exists, for the vDrive (or anything else that can pull high a line). The LED light triggers a transistor which turns on the ATTiny85 (which costs pence) and outputs to a speaker and has a capacitor to smooth the power to the chip. You could hook it up the the QL speaker if you wish.

https://youtu.be/A9s6XKbfVUY?si=ys22XzmF4OwtDZDB[/youtube]

See: https://github.com/stephen-usher/vDrive-Sound-Board

[Popopo]

That part, is a Clive philosophy matter, affordable, low cost, easy and efficient.

The movie "Micro men" quotes Clive Sinclair: "Practicality, simplicity and elegance are the pillars of my vision."

I've watched it. Time ago...
Exactly what we did with this device nothing complicate that requieres compatibility with 3rd parties or whatever... it is compatibly with original QL, the rest... is not our cup of tea.
And of course, it is elegant, at least for us. very very elegant.

I'd find this elegant, if it was smaller and a single board device. Instead of a three boards, of which one is external.
Ever considered to solder the RP2040 directly and to to get rid of the external electronics, making it small and actually internal as the title says?

Don't get messed or confused. only 2 boards. With your POV even the QL is not elegant because it has regulators, many chips... well... then a FPGA solution to emulate an original QL is your solution.
We could solder directly on the main board the RP2040 components or part of them, but what? where is the point for it? to do it less affordable for people with not high knowledge? that is not what we want. Everyone must be able to built it up and with low cost.
The title says what it is. And this is perfect already awesome and beautiful, just what it must be.

For a minute, I even considered to do that myself, for fullsize SD card, so it is not endangered to get lost in the microdrive slot.
Your circuitry would in principle allow it. And it would save further costs, as this project seems so focussed on pricing.
Unfortunately, microdrive storage capacity, while fine for the Speccy, has always been too small for the QL.
So I'd buy such a version, if someone else does the job, but am not sufficiently motivated to design it myself.

Remember the same principle that you mentioned before... simplicity. We don't want to make it complicate. Loading images (MDV or another format) as Gotek idea. That's all. Hard disk is another matter, and it could implement somehow something like that, but that imply to modify also the main QL board, and we don't want to touch this piece or art, at least... do it as less as possible.
This design is wonderful as it is. Just to manipulate it, how to use it... as simple as a disk but with modern and cheap component.
Also it is possible to create a case (I Won't do it by now) or recycle MDV original case to house it. And that is possible thanks to the design that in versions 1.4 and 2.0 are even more simple than previous one.

Your POV about the capacity of storage files, I like it and could be interesting to code something about it (keeping the same hardware). But maybe you are missing something... Hard Disk or massive storage devices, are actually files, lot of files located by directories (special files) and sub-directories. That is the same concepts than to have a storage with images files. The two main differences is from where you manage the access to one folder or another, and how you open a "directory/image of MDV" to read or access to its files. As simply as that.

Engineering solutions for everyone, that is what we do. That is what Sinclair did.

[Popopo]

But ... not now. I encourage you to take this idea and implement it. For sure some people would love it. Why not?

It already exists, for the vDrive (or anything else that can pull high a line). The LED light triggers a transistor which turns on the ATTiny85 (which costs pence) and outputs to a speaker and has a capacitor to smooth the power to the chip. You could hook it up the the QL speaker if you wish.

https://youtu.be/A9s6XKbfVUY?si=ys22XzmF4OwtDZDB[/youtube]

See: https://github.com/stephen-usher/vDrive-Sound-Board

Nice,
You could do it even easier.
You could re-code part of the firmware of RBP to get the output for the local speaker through one GPIO (there are enough free to do it) and put the footprint in same Driver PCB (of the Microdrive replacement).
About to use the general speaker (QL), ... think about it twice. If you like... go ahead, but that measn you need to retouch things in the original QL Board... and could get conflicts with software and other devices trying to access to the same speaker. If you want to emulate the sound of a MDV... would be it better to built it dissociated from main system? that would give you an extra point to do not get a conflict with other systems and rise compatibly and simplicity,

[Peter]

Don't get messed or confused. only 2 boards.

I hoped to be able to count to three The RP2040 development board, the board for the MDV area and the external board.

With your POV even the QL is not elegant because it has regulators, many chips... well... then a FPGA solution to emulate an original QL is your solution.

Indeed, I find the QL mainboard not very clever and Sinclair style, compared to the ingenious Spectrum design. But for different reasons.
I wasn't advocating FPGA for microdrive emu. It is amazing what can be done with the cheap RP2040.
Side note: FPGA logic implements hardware, just as a PLD, there is no emulation as many think.

We could solder directly on the main board the RP2040 components or part of them, but what? where is the point for it?

Allow an internal solution, have a more compact board despite a larger card socket, a single board instead of three, and save costs for a machine soldered batch.
I was just sending feedback what I would like & buy. If you prefer the way it is right now, fine. It's a hobby after all, have fun! Nice video by the way!

Hard Disk or massive storage devices, are actually files, lot of files located by directories (special files) and sub-directories. That is the same concepts than to have a storage with images files.

Of course not. E.g. think about using C68, SMSQ/E, many files open in parallel, long files...

[tofro]

Indeed, I find the QL mainboard not very clever and Sinclair style, compared to the ingenious Spectrum design. But for different reasons.

The ZX81 was a masterpiece of "engineering the cheap", using ingeniously clever tricks to save cost.

The Spectrum was still clever, but way more "standard"

Seen from that viewpoint, the QL indeed is boringly off-the-shelf (that's overstating facts a bit, I admit ). The IPC data link and the microdrives are some reminiscences on former "glory of simplification", but the rest is a fairly standard 68k system. That's probably not a result of loss of cleverness of the Sinclair engineering teams over time, but simply the fact that more modern CPUs had grown too complex to be "tricked" into simplification.

[stephen_usher]

Well, the QL is really an 8 bit machine with an 8/16/32 bit processor shoehorned in.

Remember, it was going to have a Z80 until relatively late in the design stage.

[Peter]

Seen from that viewpoint, the QL indeed is boringly off-the-shelf (that's overstating facts a bit, I admit ). The IPC data link and the microdrives are some reminiscences on former "glory of simplification", but the rest is a fairly standard 68k system.

QL mainboard has two 8 bit wide banks for both ROM and DRAM anyway, so they could at least have made a standard 16 bit 68000 system.
Instead, they crippled everything with the 68008, not even saving significant costs. (IIRC some sources say the 68000 would even have been a little cheaper.)

And while the data link to the IPC was simple, the IPC itself achieved almost nothing. The Speccy could decode the keyboard matrix completely without IPC and I have heard more complex sounds from the Speccy than from the QL. A waste of board space and costs.

We have to face it, our beloved QL was a clumsy hardware design. Only the good looking case, the OS, and the fact that it was 68K at all, made it worthwhile.

That's probably not a result of loss of cleverness of the Sinclair engineering teams over time, but simply the fact that more modern CPUs had grown too complex to be "tricked" into simplification.

I have designed with both Z80 and 68K. Up to 68020 (with reduced bus size) I never found the latter more complex hardware-wise.
Personally, I think the cause was Richard Altwasser leaving Sinclair, plus the time pressure.

Remember, it was going to have a Z80 until relatively late in the design stage.

That might explain why they didn't make use of having ROM and DRAM ready for 16 bit.

[tofro]

That's probably not a result of loss of cleverness of the Sinclair engineering teams over time, but simply the fact that more modern CPUs had grown too complex to be "tricked" into simplification.

I have designed with both Z80 and 68K. Up to 68020 (with reduced bus size) I never found the latter more complex hardware-wise.
Personally, I think the cause was Richard Altwasser leaving Sinclair, plus the time pressure.

Well, that might well be the case. But: You simply can't implement some very "clever" stuff like the ZX Spectrum's "floating bus" or the ZX81 video circuitry with a 68008, at least not that I would know.