Question about USB on the minimig

[pixie]

Well it does if you put Windows to minimig.

To be of any use the USB-stack would need to run outside of the emulation in the first place (this is kind of obvious if you think about it: How would you go on about launching the USB-stack if you're booting from a game floppy? You obviously want to be able to use the keyboard and mouse in games...)

[jkonstan]

As before in another thread on MiniMIG, the issue with USB host on the present Minimig PCB would be the lack of CPU MIPs (the slow clock speed of the 68SEC000 CPU) and lack of spare FPGA I/O pins. A seperate CPU on a USB HOST controller such as the FTDI Vinculum USB HOST controller might work becasue it has a seperate CPU of its own running the USB stack although we would have to use a SPI interface to connect to it since we only have 4 spare FPGA I/O pins on J9 Spare I/O header. The speed of a serial SPI connection from 68K/Xilinx_FPGA would most likely limit the bandwidth/transfer rate. Data sheet has Vinculum SPI port max SPI Clock = 10Mhz => 10Mbits/sec Max SPI transfer rate (if no signal integrity issues arise). A FTDI Vinculum USB HOST controller small daughtercard that connected to MiniMIG1 Spare I/O header J9 could be built. We may have enough 68K CPU MIPS to handle this kind of implementation. Would anyone be willing to provide some Amiga driver software for MiniMIG/FTDI Vinculum USB HOST controller to support such a scheme ?

http://www.vinculum.com/

 :-)

[dammy]

Or go with Coldfire instead of a 68K for a TNG Minimig?

Dammy

[freqmax]

Coldfire isn't code compatible asfair.

[little]

Coldfire would be cool for a next generation projet, they have released v5 that runs at a nice 400mhz, add some PowerVR SGX for the graphics and you have some pretty competent hardware.

But for a project with emphasis in running 16-bit software it is not the right solution since it is not 100% compatible with the 68000.

[techie]

downix wrote:
Alright, let's crunch the numbers:

USB is a CPU-driven, not DMA driven, system.  That means for every batch call, the CPU must be utilized.

For a single mouse, you will, at peak, utilize 12Mbps, roughly 1.5MB/sec.  Now, the Amigas RAM speed is only 7MB/sec.  So, for a single, solitairy mouse, you'd be tying up 40% of RAM access.... just imagine adding a keyboard.

Sorry I'm probably just being Thick here but if this is the case then I have to ask; How does the Anaiis USB stack work then???  :-?
Because it claims to be working on a pretty much stock Amiga 2000 with ZorroII Highway USB and 68000 processor (7.14 Mhz).
(http://aminet.net/package/driver/other/anaiis)

Hattig wrote:
If you eventual point in all this is that you want to use a USB gamepad to play Speedball 2, then this is nigh-on impossible.

Yes, you can attach a USB port to a 68000, the rest of the system is just code. However due to the nature of the software part of USB, you would need to run what is probably quite a heavy duty bit of software alongside the game that already probably uses >90% of the 68000's available power. Alongside that you would need to map the USB gamepad input into standard Amiga joystick inputs. And then you'd realize that Amiga games boot from ADFs on the MiniMig, and that there's no way to stick the USB stuff into that.

The only way it could be done would be to have a second CPU that runs the USB stack transparently, and also does the mappings to classic Amiga hardware transparently. Here, of course, you start seeing the level of work that would be required.

Agreed, this does sound like it would undoubtedly be a project in and of itself but that doesn't necessarily mean it's impossible (just improbable :-))

So my next question would be; would the Vinculum solution that jkonstan mentioned make such a project any easier? If so what would this solution do to the overall cost of the minimig?

[Piru]

@pixie

You had a point there somewhere?

[downix]

techie wrote:

downix wrote:
Alright, let's crunch the numbers:

USB is a CPU-driven, not DMA driven, system.  That means for every batch call, the CPU must be utilized.

For a single mouse, you will, at peak, utilize 12Mbps, roughly 1.5MB/sec.  Now, the Amigas RAM speed is only 7MB/sec.  So, for a single, solitairy mouse, you'd be tying up 40% of RAM access.... just imagine adding a keyboard.

Sorry I'm probably just being Thick here but if this is the case then I have to ask; How does the Anaiis USB stack work then???  :-?
Because it claims to be working on a pretty much stock Amiga 2000 with ZorroII Highway USB and 68000 processor (7.14 Mhz).
(http://aminet.net/package/driver/other/anaiis)

 wouldn't know, as the source to both products is quite closed it appears.  Could be that there's an embedded controller.  Could be that the guy went "fsck it" and ate the performance loss.  Don't forget, you too can render Toy Story on your Amiga 1000.... it just would take centuries.

But it does not make things a practical solution.

[JimS]

techie wrote:
So my next question would be; would the Vinculum solution that jkonstan mentioned make such a project any easier? If so what would this solution do to the overall cost of the minimig?

Yes, the Vinculum would make things easier. You can buy an assembled board from them for a reasonable price. They even have one with a MP3 decoder chip for playing music. So all you'd need to do is make a cable..... oh yeah.... and write the software. ;-)

Personally, I'm beginning to more into the "USB is pointless" camp. It would be cheaper to just use the SD cards... after all to use a minimig you'll need to put an SD card interface in your PC....

[Belial6]

I still don't get what the huge focus is for using USB for classic gaming.  I would want to see USB for use from the Workbench.  I'm perfectly happy using a 9-pin joystick and adfs (hopefully, someone will eventually make an open IPF replacement) for gaming.

I want USB to be able to hook up a scanner, or drawing tablet, or mp3 player, or any other device I haven't thought of.  The explanation I keep hearing for why this can't/shouldn't be done is that you can't use it to replace the joysticks and floppy drives for classic gaming.

So, my question is, given that we have already established that the processor speed is not an issue due to solutions with there own built in processor, what prevents USB from working from Workbench?

[techie]

JimS wrote:

techie wrote:
So my next question would be; would the Vinculum solution that jkonstan mentioned make such a project any easier? If so what would this solution do to the overall cost of the minimig?

Yes, the Vinculum would make things easier. You can buy an assembled board from them for a reasonable price. They even have one with a MP3 decoder chip for playing music. So all you'd need to do is make a cable..... oh yeah.... and write the software. ;-)

I'm sure it wouldn't be just plug'n play but it could be an asset.

Personally, I'm beginning to more into the "USB is pointless" camp. It would be cheaper to just use the SD cards... after all to use a minimig you'll need to put an SD card interface in your PC....

Which is probably right if all you wanted out of USB is access to storage devices but if you also wanted to use USB joysticks and mice then access to SD cards doesn't really help :roll:



More and more it sounds to me like there are really two groups of people who are asking for USB 1) You have people who want access to USB Mass Storage devices and other high tech gadgetry and 2)You have people who just want access to USB joysticks, mice, keyboards. Now I think it's safe to say that it is looking more and more unlikely that the first group will ever see this functionality without redesigning the MiniMig completely but what about the second group?

What if we only focused on making USB joysticks, mice (2x), and keyboards compatible with the minimig for classic gaming? Would that help reduce the complexity of this sort of project and would that be enough to satisfy most people when it comes to USB?  

 

[little]

So, my question is, given that we have already established that the processor speed is not an issue due to solutions with there own built in processor, what prevents USB from working from Workbench?

Hear, hear!

If people really want to use their USB joysticks with 16bit era machines then the solution is to make an usb to de-9, this could also of interest users of msx, commodore 64, sega genesis, etc. You can already connect some USB keyboards to the ps/2 interface with an included adaptor.

I think the best use for an USB interface is to connect cd-roms and there will be people interested in scanners, external hard disks, etc.

As a side query, does anybody know if the PIC controller could be used for the USB interface or if using a more powerful/bigger PIC would allow such functionality? This would have the side benefit of allowing transparent use by games of USB joysticks :-p an also would keep the cost and simplicity of the design.

[techie]

Belial6 wrote:
I still don't get what the huge focus is for using USB for classic gaming.  I would want to see USB for use from the Workbench.  I'm perfectly happy using a 9-pin joystick and adfs (hopefully, someone will eventually make an open IPF replacement) for gaming.

I want USB to be able to hook up a scanner, or drawing tablet, or mp3 player, or any other device I haven't thought of.  The explanation I keep hearing for why this can't/shouldn't be done is that you can't use it to replace the joysticks and floppy drives for classic gaming.

So, my question is, given that we have already established that the processor speed is not an issue due to solutions with there own built in processor, what prevents USB from working from Workbench?

I guess the focus on USB for classic gaming is partially because people like to have choices when it comes to input devices. I for one would like to have the option of using either a classic joystick or a USB gamepad to play different Amiga games.

There's also the fact that classic controllers simply won't last forever. I have at least two joystick in my apartment which have simply stopped working on me because the cords got a little mangled during my last move (the box they were in got crunched a bit :S). So having access to USB joysticks and mice would give me a lot more options.

I also like the idea of being able to use two mice at the same time in games like lemmings and marble madness without having to resort to using my Amiga 1000 angle mouse + a PS/2 mouse :-(  

[Hattig]

More and more it sounds to me like there are really two groups of people who are asking for USB 1) You have people who want access to USB Mass Storage devices and other high tech gadgetry and 2)You have people who just want access to USB joysticks, mice, keyboards. Now I think it's safe to say that it is looking more and more unlikely that the first group will ever see this functionality without redesigning the MiniMig completely but what about the second group?

The whole point is that (2) is really difficult, requires a lot of extra hardware and why not just use PS2 keyboards, mice and 9-pin joysticks?

(1) is a lot more possible, if you just want to run USB within Workbench. It might be slow given the 68k, but once you have the software stack, you can attach whatever USB devices you want - including keyboards, mice, gamepads (they just won't work in a non-AmigaOS environment).

[techie]

Hattig wrote:

The whole point is that (2) is really difficult, requires a lot of extra hardware and why not just use PS2 keyboards, mice and 9-pin joysticks?

Like I said before, partly because 9-pin joysticks won't last forever and also because it doesn't let you use multiple PS/2 mice for those few games that have two player mice support.

I get the whole PS/2 keyboards and mice argument and I kind of agree since there are already adapters that let you hookup USB mice and keyboards to PS/2 but it still doesn't solve the joystick problem.

(1) is a lot more possible, if you just want to run USB within Workbench. It might be slow given the 68k, but once you have the software stack, you can attach whatever USB devices you want - including keyboards, mice, gamepads (they just won't work in a non-AmigaOS environment).

Wait doesn't this solution STILL require "a lot of extra hardware" to implement as well? Wouldn't this also require a higher end CPU to be integrated to really get anything out of all those high end peripherals?

Sorry it still sounds to me like option two would be easier (albeit a lot of work) to implement in the current minimig design.

little wrote:

If people really want to use their USB joysticks with 16bit era machines then the solution is to make an usb to de-9, this could also of interest users of msx, commodore 64, sega genesis, etc.

While I would still prefer an on board solution this type of adapter would probably be enough to satisfy people like me as long is it could be used with both USB joysticks and mice.

[dammy]

Coldfire isn't code compatible asfair.

Last I heard, they had libs out for a work around.

Dammy