The Sinclair QL Forum

bixio60 wrote:I can accept the idea to have (initially) a PS/2 keyboard, I should have one of them somewhere, but what about having the possibility to manage external usb keyboard and mouse ? I mean something integrated wireless like Logitech system. I am using it at the moment with Mist

The MIST uses a separate Microcontroller for USB keyboard/mouse. One could design something similar and add it to the Q68 extension bus or I²C.
It is unlikely I will do that myself, personally I'm fine with classic PS/2 keyboard+mouse. This way, driver software can run on the QL side, which I find more interesting than dealing with non-QL environments.

That's what I'm saying, Peter:

QL keyboard matrix to PS/2 shouldn't be too difficult to implement for the average µC tinkerer.

Tobias

Just needs a i2c GPIO expander chip and a native driver. No uC required at all!

G

Wow!

Not meaning to be critical - I can only imagine how hard and time-consuming it must be to design something like this! - but things like VGA, PS2 and so on seem to be on their way out. Are HDMI and USB (as on the RasPI) very much harder to implement? I know its chasing a fast-moving target, but going for these later standards would give your design an extra 5-10 years of life?

Per

pjw wrote:Are HDMI and USB (as on the RasPI) very much harder to implement?

Yes, they are. In addition, to build an HDMI port, you need to have a professional license ($$$) from the HDMI consortium - Way out of the picture for Peter, I guess.

With regards to USB: I don't think it's so much a problem with the hardware - In an FPGA it shouldn't be really hard to implement USB support, I would also reckon there'll be open designs for a USB host around already. Here the problem would be the software. Who's going to write proper USB driver support for QDOSMSQ? For USB, you'd also need USB support in the OS, not just in the driver (If you think about the wide variety of stuff you can plug into an USB port you can imagine,I guess). And half-baked USB (like HMI only) has practically no advantages over PS/2.

Hardware standards have grown so far away since the 80es that you simply cannot support all the modern interfaces with reasonable effort in the context of a hobby project anymore. I think Peter is doing exactly the right things if he ever wants to finish Q68.

Tobias

The Spectrum Next managed to get a license for use of HDMI without paying high fees - I am not sure if they paid anything - but it is always worth asking the HDMI foundation and explaining the background of the project.

The other option is to add a header for the Raspberry Pi Zero and people can add one of those to provide the HDMI output (as the Raspberry Pi is itself licensed).

As far as my knowledge goes (nor very far, I must admit) the Spectrum Next uses the same approach for HDMI that Ben Versteeg uses for his HDMI project - According to the Facebook page, It seems to use a Raspberry Pi to circumferent the needed HDMI license.

(Note: Apparently, even when you sell a licensed RPI as part of your device, you need your own license. Ben wrote somewhere he will not be able to supply his board completely built including the RPi for exactly that reason.)

Regards,
Tobias

pjw wrote: Are HDMI and USB (as on the RasPI) very much harder to implement? I know its chasing a fast-moving target, but going for these later standards would give your design an extra 5-10 years of life?

The first Q68 hardware was designed over 8 years ago. So if I started today, I might use a newer component here or there, but still not add USB hardware.

Why? Because the basic idea of the Q68 is to be a strictly native hardware, with 68K CPU, no emulation at all !

That's what differs from emulation under ARM Linux like Raspberry, even from FPGA boards running (limited) USB support on a second processor. The basic idea of the Q68 requires that all hardware must be supported by QL drivers. As Tobias explained, this makes USB unrealistic.

HDMI doesn't contradict the "strictly native hardware" approach, but it tends to make things expensive and harder to build. And there is an additional point: For being the only video output, HDMI might still be "too new". For example, my existing monitors for retro computing are still in 4:3 format and have VGA! I would end up converting HDMI back to VGA by an external converter.

It seems easier to go the opposite way and use mass market VGA-to-HDMI converters, in case the monitor no longer has VGA input. The Q68 has at least VESA standard 1024x768, which is so widespread that every converter should be able to deal with it.

As a last resort, in case PS/2 dies, we could still go "non-native" and tinker a little co-processor board with USB that connects to the Q68 expansion bus. But not now.

Thanks for the info, Tobias, Rich, Peter.

I see there are VGA to HDMI adapters, so if they can be used, there would be no problem. I suppose PS/2 to USB adapters are available too.

My original (admittedly naive) thought was of seeing USB/HDMI merely in terms of connectors, not to try to implement the full Monty. After all how complicated can one make converting a key press to a signal, or a word in memory to a coloured dot somewhere on screen? Behind all the modern jiggery-pokery lies but the archeology of stuff we already know. I remember how shocked I was to discover the banal simplicity of the Windows boot sector (until NT, when I lost track) and the vastly expensive tools you "needed" to fix or change things. Well, I know nothing, so Ill shut up. I was acting under the adage that "Fools challenge; the wise conform. Without fools the world would never change!"

Just saw Peter's message now: I totally understand and accept your philosophy on this matter

pjw wrote:I see there are VGA to HDMI adapters, so if they can be used, there would be no problem. I suppose PS/2 to USB adapters are available too.

The problem with most PS/2 to USB adaptors is, they just provide passive connectivity for mice that have built-in support for both protocols. A pure USB mouse won't work this way.

PS/2 to USB converters that also translate the protocol exist, but seem expensive and not very common, like this one:

https://www.tripplite.com/ps-2-to-usb-c ... ~0DT60002/

So USB for Q68 might indeed require tinkering...