8bits4ever announces new MSX

Because the OCM platform only supports PS/2 and nobody’s ever implemented USB protocol support for it (probably because the FPGA is full). The original OCM has USB ports but they are not actually functional. Maybe in the future the MSX-SM will usher in an era of bigger FPGAs so there is more room for these kind of extensions?

Is there actually any FPGA-based retro platform with USB hardware support? I'm under the impression that there isn't. But if there is, it would be great if we could build upon it.

Nowadays FPGA reimplementations are very close to the original chips. V9958 is very close to perfection on OCM and similars. It also has a faster blitter than the original VDP which you can turn on if you wish.

nice

already have a ocm

low price of this one makes me want one though

USB is quite complex to implement in hardware alone. It requires some software or a very complex state machine. It is certainly doable but it is not the simplest way of doing it. Every HID device (mouse/keyboard/joystick) sends a description that allows to interpret the contents of the data it reports. So where in the report to get the information (like mouse dx,dy) and how to interpret it, changes for every device.

I think it would probably be more effective to have a small micro-controller to handle that and report the sanitized data to the FPGA using 2-3 pins. You need some initialization and a flexible parser to interpret the reports and then send it to the FPGA only what is really needed.

PS/2 is super simple by comparison.

Then, for USB support, we need one of these:

1. USB drivers for MSX (probably impossible due to processing requirements)
2. USB support baked in the FPGA (probably impossible due to complexity and FPGA space requirements)
3. A low-cost, low-energy updateable solution involving some microcontroller (sounds complex but doable albeit not any time soon)

So far the prospects don't seem good. I wonder if it will ever happen?

EDIT: I also wonder if it would be any simpler if only USB was supported, with no DB-9 joystick or PS/2 connectors at all.

USB drivers for MSX are possible speed-wise I think, but it would not be compatible with software that reads the key matrix directly through the PPI.

My line of thought would be to implement a mini processor in FPGA, then run a program on that which reads USB serial data and updates the key matrix. This can minimise the amount of FPGA space needed I think, however it seems rather complex to implement, involving building both a tiny CPU as well as the software that runs on it. And since the 1chipMSX FPGA is so full, I wonder still if there’s space for that, though the PS/2 support can be removed of course. That’s why I mentioned the bigger FPGA in MSX-SM earlier, but of course it would be the awesomest if it could be done on our current 1chipMSXes.

@iamweasel2 An active USB-PS/2 adapter works by having a microcontroller inside which runs a program which translates the key events from the USB serial protocol to the PS/2 serial protocol. A passive USB-PS/2 adapter relies on the keyboard itself supporting being able to switch to a PS/2 mode (few do, nowadays).

See this forum topic:

Connecting a USB keyboard to One Chip MSX

Not likely, I've seen a few guys asking Victor Trucco about this and his answer as that USB is not doable...

While I generally do not agree it is not doable as you could do it without an external PHY for USB low speed, but then, I understand it as "not doable" as it would waste lots of cells and would not be complete (just try to browse for FPGA implementations of a FULL USB stack, generally just replay enumeration, no real logic). Leaving the USB protocol to the z80, as far as I understand, is not going to cut it... Either you will get not free CPU time for other tasks and it will be really interrupt intensive, or it is not going to be able to respond back to the device in proper time.

So, I agree with Edevaldo, if someone wanna do it, you better user a cheap micro controller that has USB host functionality built-in, and hack inside the FPGA code to scan that device... Or, like someone else told here, you can use an active USB->PS2 adapter... Or you can have something like MSX-VR where it is a Raspberry PI (or similar microcontroller) emulating everything and it already has USB keyboards on the host system that virtualizes MSX...

Perhaps a hybrid solution like this. It already supports MSX (presumably just MSX1)
https://www.youtube.com/watch?v=e5yPbzD-W-I

Have the ARM CPU handle the external USB devices and such, and the FPGA for running the cores.

I'm ok with a Rookie drive and a active USB-PS/2 adapter.

That's what this system is all about. Everything can be added as needed.
There will be no machine that suits everyone.

This industry is unfair. We noticed this when our brand new, expensive MSX 1986 was simply replaced by the great second generation. For less money.
As soon as you pay for the most wonderful shit at the checkout, it stinks even more from the store shelf.
I absolutely agree with 100 Euro and the size for the desk or self made projects is fantastic.
Of course you can also make up a RasPI, but the feeling is not the same.
I'd rather see a built-in cartridge slot on this model than the external solution. I'd even rather have an all-in-one solution with a built-in power supply.
But with this FPGA nothing will block a self-built MSX laptop.

Great setup, Hamlet!

I thought that you had to go through BIOS to read the keyboard matrix? Correctly written software would use the BIOS routine rather than directly accessing the PPI, I remember in the early days of MSX a Microsoft spokesman said in a UK MSX magazine that you should use the BIOS to access the keyboard, if it was replaced by an infrared keyboard then a serial interface would replace the PPI in this case, and your software would not work, however, if the infrared keyboard was used the BIOS routine would be modified, no need to alter your software if you used the correct BIOS routine.

https://github.com/Konamiman/MSX2-Technical-Handbook/blob/master/md/Chapter5a.md#3-keyboard-interface

The mist board actually has a MAX3420E SPI USB Host controller and a USB 4 port hub. I'm certainly no expert on how the MSX core is implemented, so I can be wrong. But the MAX chip is actually controlled by a AT91SAM7S256 micro-controller. I do not think the FPGA handles any of the USB interfacing.

In any case the FPGA on the Mist is about 3x bigger than the one on the OCM.

There any many MCUs today that include an USB otg port, so this would be a little simpler than the solution on the Mist. The SAM part on the mist can only handle USB device side, not host. One thing that could be done for the OCM is to have the USB MCU + Hub in a separate board that connects to the pins allocated for the USB interface...

That’s the theory. In practice, not having a working keyboard in a bunch of games is not really acceptable for the end user.