Amiga & Usb

[Retro1234]

First off, Congratulations

I suprised there hasnt generated more interest.

The idea of USB>Serial has come up a number of times, 1 for a replacement to current USB options and also for Amigas like the A500 that dont have clockports or pcmcia.

One of the problems -The idea has been there but writing the software to use the hardware hasnt been done - So well done and the fact you used AMOS I found fun.
How the hell is AMOS able to communicate with FAT partition

Could someone give a full price list of components.

AMOS is not communicating with the fat partition itself.
This thing works like an old style mailbox or FTP-Server.
AMOS sends shell commands compareable to something like msdos commands to the little computer with usb port that is connected to the serial port via nullmodem connection.
There are commands like DIR(list dir), DLD (delete dir), MKD (make dir), CD (change dir), RD (read file), WRF (write file), etc...
It's also possible to connecta USB Printer.
The chip itself has also a parallel port, as it is 5V tolerant it should be possible to work there too.

[desiv]

Since it's 5V, you should be able to pull 5V from an Amiga 1000 on pin 21 to power this thing?

On all the other Amigas, there is 12V on pin 9 you should be able to transform to 5V???

As long as the amp pull isn't too high...

desiv

[rkauer]

Floppy port for power, perhaps?

As USB can provide up to 500 mA on each port, I would not recommand grabbing power of any of that Amiga-ports, as even the Floppyport has only 250 mA on 5V (should be enough for USB-Stick but not for 1,8" or 2,5" HDDs. Better connect to a drive-power-connector inside the Computer.
When connected to AMIGA Printerport a possible mapping would be:
D0-D7 <-> D0-D7
/ACK <- RXF# When high, do not read data from the FIFO. When low, there is data available in the FIFO which can be read by strobing RD# low, then high.
BUSY <- TXE# When high, do not write data into the FIFO. When low, data can be written into the FIFO by strobing WR high, then low.
/STROBE -> RD# Enables the current FIFO data byte on D0...D7 when low. Fetches the next FIFO data byte (if available) from the receive FIFO buffer when RD# goes from high to low
POUT -> WR Writes the data byte on the D0...D7 pins into the transmit FIFO buffer when WR
goes from high to low.

As I never programmed the parallel port on Amiga, i'm not sure if AMOS works, as it is not used as printer port.
The logic fom BUSY posibly needs to be inverted.
POUT has to be programmed as Output!

You have to do that via programming the CIA Register as AmigaDOS only supports output.
CIAA
BFE101 prb Parallel port
BFE301 ddrb Direction for port B (BFE101);1=output (can be in or out)
CIAB
BFD000 pra /DTR /RTS /CD /CTS /DSR SEL POUT BUSY
BFD200 ddra Direction for Port A (BFD000);1 = output (set to 0xFF)
BFDD00 icr CIAB interrupt control register

As /STROBE ist set automatic from CIA-A polarity and timing has to be checked.
The 8520 can be programmed to conditionally generate a level 2 interrupt to the 680x0 whenever the ACK* input goes active. So an IRQ Handler has to be installed, I don't think AMOS can do that. If one of the above 2 Signals with indirect access can not be handled there is one spare I/O pin the SEL pin

[desiv]

Originally Posted by natas

As USB can provide up to 500 mA on each port, I would not recommand grabbing power of any of that Amiga-ports, as even the Floppyport has only 250 mA on 5V

Wow! Good to know! ;-)

desiv

Is someone interested in writing a driver in assembler for optimal performance?

[Counia]

Intensity: 20mA to 40mA with key.
With my 2,5" disc, 300/400mA but can peak 700/800mA when start.

Parallele-port programmation with AMOS is no an problem, because register are fully adressable, like serial-port control line (no funtion in AMOS).

I not remember the speed of this parallele-port, but i can use an litle ASM code for read these limit with frequence-meter.

[rkauer]

Expect a gain of at least 8x in speed compared to the serial port. Normal operation can read around 300 to 500kbp/s.

[Counia]

I think 300 to 500kbps is an optimistic number, already 8x19200 is only 150kbps.

Just my test with this litle loop:
NOT.B D0
MOVE.B D0,$BFE101
plus wait

I no test more speed, my CIA respond up 8kHz, but time for UP-front(0-->5V) is not negligible.
Perhaps 128000bps with goot timing synchro (6x comparate 19200bps serial speed).

Someone know an reliable transfert with parallele-port at this speed (on Amiga of course) ?

[rkauer]

Typical PARNET speed is 23 to 50kB/s (multiply by 8 to see in bits instead of bytes).

Vrooooooooooommmmmmmmmmmmmmm!!!!!!!!!!!!!!!!!!!!!

[Counia]

Ok, i see signal at this speed.
Require good timing for this frequency: on 7,5µs , 4µs is for up-front.

@Counia as a IRQ can be generated for receiving timing should not be a problem in this direction.
I guess for sending Data Mutlitasking could be problem. When you disable it can you still access fdd or hdd? If not i think a ram buffer would be the solution.

[Counia]

IRQ not resolve this fact, at this frequency if you not read with goot timing the dada is indeterminate 0? or 1? it is the hardware limit when front take more time than read-period.
At this frequency, my 7Mhz 68000 wait in loop, therefore multitasking car run too.
Receiving timing in parallele protocol is more easy than serial, because no buffer, just code like you write and both adapt debit without imposed clock, just cross-signal.

Just with 16ko/s, an floppy image transfert do < minute, i think DF0: do more ;-)
It is an good goal, after if no problem 30k, 40k or 50k...

This reflexion for know limit and parallele version gain, with this number i think i will test it now :-)

As the interupt is generated when the data is valid and the next thing you can do is the port access. Or I'm wrong?

[rkauer]

Any news about this development?

[spock-79]

I have just build this circuit, works perfect!!

Thank´s for this hardware and software

[Counia]

NEWS !!!
No update for serial / VDRIVE

BUT parallele version / Vdip2:
http://img31.imageshack.us/i/amigausb.jpg/
http://img337.imageshack.us/i/amigausbasm.jpg/

No portage for AMOS code in parallele-version because very bad transfert time (19200 serial is beter).

Programmation with Ed ; assemblage with this old SEKA and CI with DPaint4 (LOL 100% by Amiga for Amiga).
First test with file = (JOIN x4 ROM), therefor 2097152 octets (2Mo).
Dh0: -> USB 1'26'' [23,8Ko/s]
Ram: -> USB 55s [37,2Ko/s]

No optimised ASM code and only buffer=256.
At this time just 3x CLI command: USB_Init, USB_Dir et USB_Export.
And GuruMeditation some time.

[Counia]

Originally Posted by spock-79

I have just build this circuit, works perfect!!

Thank´s for this hardware and software

Thanks for your message.
Where you buy the VDRIVE ?
What USB-disc you use ?

[rkauer]

The transfer rate doesn't make sense: the speed must be around 8x the maximum serial speed at least!

Did you used the original parallel.device (par.device)?