interface idea's

[Jeff_HxC2001]

I think the software generated Luma output of SDiskEmul is a great improvement over the models with an LCD display. The BOM reduction and reduction in board complexity results in a product with price point with almost mass appeal.

I think it is a good design.

I have heard that it's video combining can lower the video quality?

I see it was never supposed to directly replace your floppy drive (unlike my suggestion). It allows you to keep your internal FDD (on B: ?) which I can see would be advantageous.

The videos show that it is slightly slow processing the disk images, I guess this is a limitation of the MCU they are using. Perhaps it can use pre-processed images for a speed increase?

alphanumerical LCD (HD44780 compatible) are not as expensive as you seem think. the BOM difference between LCD and video solution is : remove the LCD and add an Video Sync Separator chip.
(if you want make overlay display, you need an Video Sync Separator chip, like in the sdiskemul)
Remember that it's an very very small market... So if you make an floppy emulator, it's better that it can go on any computer/device (i mean an emulator that can be plugged indifferently on computer with (or without!) TV/composite PAL- NTSC or VGA video output )

i think that the cheapest way is to use an 2 digits led displays, like this chinese floppy emulator: http://qhsfd.com/product.asp )

Here somes videos of my standalone version of my floppy emulator (the HxC Floppy Emulator), using an LCD display. (with fast loading )
http://fr.youtube.com/watch?v=0arlGguYEts (cpc atari st amiga)
http://fr.youtube.com/watch?v=KtazXt_Kc4s (amiga)

[Jeff_HxC2001]

How it works on SDiskEmul is via the extra joystick port that's part of the device.

I've took a look at the SDiskEmul, and if the schematics and FW/SW are free, then there is no need to develop almost the same device (only if the new device

would not use the menu with images displayed as overlay on the TV)...

It is a limitation when you want to do more than gaming, e.q. development or using different data disks for music

software/art packages/etc.etc/etc.

I agree the scheme you propose is entirely sufficient for ROm kiddiez style use

Hmm, I haven't thought about that one... Well, then for the better image selection there are at least 3 choices:
- the overlay menu on TV like SDiskEmul
- small 2 (or 4) line LCD display on the device
- serial terminal on other computer (i.e. PC)

Jookie

Here the actual list of Homebrew emulators:

- the overlay menu on TV like SDiskEmul

The SDiskEmul : http://atariamiga.free.fr/sdiskemul_tech.php (Atari ST/ Amiga / CPC)

TFE (Tolga Floppy Emulator) : http://www.commodore.gen.tr/forum/index.php?topic=2278 (Atari ST/ Amiga / many others later)

- small 2 (or 4) line LCD display on the device

the Standalone HxC Floppy Emulator (2*16 alphanumerical lcd): http://jeanfrancoisdelnero.free.fr/flop ... pyemulator
(Atari ST/ Amiga / CPC / Korg DSS-1 / Oric / PC / Super Wildcard / MSX2 / many others later )

the SIO2SD (2*16 alphanumerical lcd): http://sio2sd.gucio.pl/index_en.html some pictures: http://www.yaronet.com/posts.php?sl=&s=104117 (Atari 800)

the 1541-III (with a nice grapical lcd): http://jderogee.tripod.com/project1541.htm (C64)

the Megadrive (with a nice grapical lcd): http://www.sensi.org/~tnt23/megadrive/ (Atari ST/ Amiga / Spectrum / PC / probably many others)

- serial terminal on other computer (i.e. PC)

The SVD : http://www.thesvd.com/SVD/index.php (TRS-80, Heathkit, Apple 2, TI99/4a)

the USB HxC Floppy Emulator : http://jeanfrancoisdelnero.free.fr/flop ... pyemulator
(Atari ST/ Amiga / CPC / Korg DSS-1 / Oric / PC / Super Wildcard / MSX2 / many others later )


And finaly the 1541 Ultimate who use something near to the alexh idea ! http://www.1541ultimate.net/content/ (C64)

maybe i forgot some emulators...

i think that most of these emulators can support the menu disk trick. It's an firmware problem, not a design problem.

[simbo]

well we have sort of been here before
in the last rant I made about some sort of internal adapter
however taken from my ideas about this
this is what I want to see
I am going to add one myself anyway so ill need some pals for a joint project
as for designs I have the basics down in proteus vsm for the main internal board
I have a spare stfm v1.04tos 1mb
I want to add to this and the ste I have
some sort of expanded buffered bus
not just to break out address and data
but to serialise it and take some of the handshake pins like LD from the shifter
and read write drive ready etc...
to a better but mounted above the original dma socket port add in internal and external ....
NOW !!!
if we can all have a discussion and maybe a pin list
perhaps we can agree on some sort of added port like this
for instance the dam dma port socket issue for one... add a plug in ide card easier
internal ide on an internal purpose port etc...
added sound card outputs in raw data form or etc.. Atari to pc dedicated adapter... usb project etc ....}}
the dma port just doesn't have enough signals and data i/o pins and is crap and hard to find
add its better to add a second one rather than change the one there
adding a second dma+ port is just a matter of ease.....
answers in schematics form is best left till a pin list is agreed
port post below in this thread will be a updated list by each users name column
added on as soon as they agree with themselves as a choice
then the best default setup chosen
then a steady state will emerge where everyone I am sure will agree
what do you say????
obviously some soldering would need done
and internal screened ribbon would do
I am sure this would make jookie and the other builders happy
and maybe some st machines a bit more portY
I think using a 36way idc or similar maybe 50way {added later or custom }
however its easily confused with a floppy or scsi ide drive etc if the machine is sold on
better to use a standard connector that is easy to find and readily available
not used for such perifs
OR simply ... serialise the whole port to two pins and use a custom Atari stream port method
running at around 8Mhz say via a custom usb2 style ftdi chip core...{easy...}
an internal board
connected behind the floppy via a shielded ribbon to the dma and etc points
and the internal main ATARI UA board has
some buffers and one mcu ic and an ftdi chip
and outputs some standard protocol like say
control signals 5 bytes width: one byte is 8 signals out one byte is 8 signals in so 5 bytes give 40 signals {plenty}
data signals 2 bytes width: { 16bits total }
address signals 2 bytes width: {16bits total }
dma data bidir 2 bytes width: {rx tx 8bits each }
that makes a packet of 11bytes
ATARI UA to ITM1 transfer can also be auto set in the mcu to ignore the address and data streams
and also trim the control word to dma only use etc as needed
ITM1 to ATARI UA is just dma data and control bytes signals to Atari
the ITM's should have 1 mcu for decoding any controls data power control and reset control
and a serial register to get at wide data for drive interfaces hdd smart video sound etc diverse uses
.,. and data monitors or other....
the ATARI UA interfaced to the ITM's via a 5 pin connector {even the same connector can be reused }
+5v RX TX CLK GND
to a small outside ITM converter that has already spaces for ide, scsi, smart, usb, {maybe} etc and its own decoders setup
you can simply plug another second board internal to the Atari
and give them some sort of hierarchy {easy via dma lun} {auto lun per socket used internal external {easy I think}
or add up to 5 of these as separate logical unit numbers controlled by dma
the main board could have a name like
ATARI UA {universal adapter}
and the pods called ITM's {intonatious transfer module's}
all the software should be open source for the cores and any drivers needed {not needed}
don't forget !!!! because you have access to break out your own set of custom signals
then this way you'll get access to shifter data AT THE SAME TIME AS DMA
so you can add a really NICE video card project using a better d-a setup maybe an fpga clip in board
as long as we have the signals outside the Atari
I am happy to help because space is tight in there
its best the keep all addin's outside via a better port thing
than heaps of heath Robinson plugs to the poor supported dma plug and i/o capacity
or use a blitter hogging and quite data dodge ppera ide add in
or the latest fpga ide idea I find a daft {strange and totally mostly unexpandable } project
where this idea
uses one internal board up to 4 external ports and 1 internal
{if you use dma already via hosts then 4- amount of hosts in use - internal port if used} is the max
external ports the ATARI UA could host {5 in total in Atari land !!}
this way you could have an ITM internal
and capture and video and sound etc ITM's external
this way I can do what I want
and give just a common set of i/o's for dma + and the rest is up to an individual
making the units auto detect stuff is really !!!! easy
and serial I find so easy to do these days
maybe some degree of working out the timings first to see what the data transfer clock should be
all transfers in Atari bus take place at 8mhz / 16
so 11bytes max rx would need 11X 8mhz{125ns per clocked bit} / 16 or around 5.5Mhz{181ns per clocked bit} sync clock transfers per 11bytes sent in RT {packet data}
a single 40mhz mcu has a single cycle time of around 1ns-10ns per ins depending what one
some are much faster single cycle types even using senix 100Mhz chip you'll get 100ps resolution
processing a serial byte takes as much time as a single bit of an i/o
so it is Feasible

[simbo]

see post above for explination of the added port idea



simbos list so far ill add other edited lists as time goes on below
and we can reach a common ground

ok here we go dma first

Pin 1-pin7 = d0-d7 dma port pins 1-8
Pin 8 HDRQ dma port pin 19
Pin 9 RW dma port pin 18
Pin10 A0 dmaport pin 16
Pin11 HDACK dmaport pin 14
Pin12 XHDINT dmaport pin 10
Pin13 HDCS dmaport pin 9
...... other dma chip and perifs signals
Pin 20 16 bit serial address stream {maybe add other bits as avalible}
Pin 21 16 bit data streem {maybe add other bits as avalible}
Pin 22 stream sync clock {16bit pulse{for hold on a serial parallel converter or i/o on a micro}
Pin 23 DTAC
Pin 24 ROM signals as a serial stream
Pin 25 soft Reset signals {one wire to the middle pin reset button{ste}
Pin 26 Hard reset power up etc {optional requires an extra little board between the psu and mainboard}
Pin 27 -30 other cpu signals as required etc time will tell
Pin 33 - other etc .. not sure
Pin 34 - Fused 12v @350mA {no floppy internal used at same time}
Pin 35 - Fused 5v @625mA {no floppy internal used at the same time }
Pin 36 ground {hehe}

this ends the first pin list i thought of its not the best but a great idea
i have selected two ic's that are high speed for both data and address streaming
they give a serial stream from 16bits and a clock
and the hold data word and chips clock can use the 8mhz clock /16 etc as approp easily enough
...
...

[ppera]

Almost OT:

Who made that this thread be constantly on the top of HW section?
I have some ideas about... but come on... this is really forced push of something hmmm... not perspective.
I suggest that instead such moves some people here start to think how to write better readable posts

[Desty]

And yes, I am part of a minority - and I'm proud of this Why stick with a boring mainstream when you can distinguish yourself from the others?

Do what you please, this is a free world but I hope I made some points clear....it is discussion and not a fundamental no go. Just me disagreeing doesn't necessarily mean it is entirely unrealistic....

It seems like you're imagining some complex scenarios where the 'obvious' device design would be a bit cumbersome for you.
But as a programmer, surely you understand that 'crippling generality' can destroy software projects (and presumably hardware ones too), where a design is overarchitected to suit too many people's idiosyncratic needs instead of forcing them to use the one simple, limited-scope system (which is of course easier for the designers, and possible for the users because as intelligent beings we adapt easily).

Why not support the simple design that suits the vast majority of normal use by most users, rather than expecting it to fit your most complex needs?

[simonsunnyboy]

I have no "complex" needs, I'm just looking for somethign which is a true floppydrive replacement and not a readonly loader.
And I do not turn this project down..I just made clear that I do not agree with the limitations inherent in this design sketch.
Anyone is free to buy what suits him.
I bought a Satandisk too, although turned down for exact technical reasons by others...so in this case I'm just on the other side of the fence and objecting this time.

[rocket-dog]

...... forcing them to use the one simple, limited-scope system......

Spot on! Doesn't this remind everyone of a little OS called TOS.......!

[simbo]

Almost OT:
Who made that this thread be constantly on the top of HW section?
I have some ideas about... but come on... this is really forced push of something hmmm... not perspective.
I suggest that instead such moves some people here start to think how to write better readable posts

ppera
any user can make a sticky.... post im sure,,, or can ask one is made this way...

i made this a sticky becouse it is an interesting post many people will maybe get some time to comment on
i personaly for one and all that... would like to have and to add other interfaces in other forms to my older machines

and its a good idea i thought to add a sticky in a more perminant form
used to promote interface ideas
setups and uses
..

[simbo]

The main problem with using usb devices with ataris
ie atari as a usb device is that you need to use your own usb id unique to your interface
and they cost a packet to buy a set {10000 american per 256keys}


thats it in a nutshell

[alexh]

ie atari as a usb device

Unusual thing to want to do but hey...

is that you need to use your own usb id unique to your interface and they cost a packet to buy a set {10,000 american per 256keys}

The chip you use to interface the Atari to the USB bus will come with a vendor ID and you can set the product ID fields to anything you like in software or sometimes with an I2C EEPROM.

As for costing $10k, AFAIK it only costs $4000 for a USB Vendor ID. But as I have said, you don't need one as the interface chip you use will already have one.

As for buying 256 keys... either we have our wires crossed and you are talking about something other than USB Vendor ID's or you are mistaken

[simbo]

i think alex your talking at cross purposes mixing vendors of ic's software and licence
with providers of mcu

if i were to buy a set of ready made chips for usb
i would not be able to write into there code
other than adding there own id's and code and some form of librarys for inclussion
in no way does any bare mcu chip come with a vendor id embeded
unless you licence it from the makers of the ic's or 3rd party tool they support

maybe something from ftdi it has a locked vendor id becouse they are the vendor
however there hardware is imature and moslty unusable
as is many makers of ready made interface ic's like the company your frequent

a single vendor key id costs from a company who makes interface ic's and supplies there librarys or some 3rd party about $4000 for a 'site licence'
however to licence atari as a unique device of its own
you need your own set of id's and they cost a packet >$10,000 for a block of 256 'unique' id's
notice the use here and above of this word 'unique'

and this is what i ment

im sorry you missunderstood what i was intending to say....

[simbo]

further

we as retro guys should !DEMAND! a free allocation of say 256 id's not pids ~{public ids}
but uid's for retro machines

or occupy and squat and take them by force !!!! for sure the authorities would give way im sure at least be resonable
with our plight

i dont see why not
we simply get together with other retro mobs and demand it or else !!!


its that simple

gota show yer spiky metal or they walk all over you

my ham licence says i am legaly allowed to occupy certain bands in certain zones of the spectrum 0-~

this should also therefore apply to usb allocations as zone of ham and retro interest !!!

[Nikolas]

Abaut that: http://www.usb.org/developers/vendor/ that cost way too much, and once device is out, dev ppl stop to support those devices just coz of buisness math dont work
Im shure you ppl know this allrdy that usb devices moustly sux. Bad hw drivers, bad hw/sw device responding, bad software anyway.

But usb is wery handy.

[alexh]

if i were to buy a set of ready made chips for USB I would not be able to write into their code

Depends what you mean by "code" but if we are talking about the same thing this is not true. Most USB chips are fully customisable through programmable registers.

other than adding my own id's and code and some form of libraries for inclusion

Isn't that what we are talking about?

in no way does any bare mcu chip come with a vendor id embedded unless you licence it from the makers of the ic's or 3rd party tool they support

If you buy a USB device controller chip it will come with a VENDOR ID (VID) register which can be programmed to any value with the CPU it is attached to. You need to buy a unique VID number from the USB group and they cost $4000 as I pointed out.

maybe something from ftdi has a locked VENDOR ID because they are the vendor

I do not know their parts, but our USB Device chips (and cores) have a fully programmable VENDOR ID register which can be set to anything during initialisation in software. I dunno if ours come with a default value in the VID register which you can use without paying USB, possibly.

hardware is immature and mostly unusable as is that of many makers of ready made interface ic's like the company your frequent

The Oxsemi (formally Transdimension) USB parts have been in production for many, many years. They are considered rock solid in the industry and are used in everything from mobile phones, set-top-boxes, PDA's, etc.

I do concede they are used far less these days as most devices contain SoCs (system on a chip) which have USB embedded cores (which we also sell).

9 years in production is hardly immature.

a single vendor key id costs (from a company who makes interface ic's and supplies there librarys or some 3rd party) about $4000 for a 'site licence'

Site licence? What are you talking about?

You can buy a VENDOR ID (16-bit value) from USB for $4000. This is your companies 'unique' ID. You use this with all your products. There is a second ID with USB which is the PRODUCT ID (PID). This is another 16-bit field for which there is NO license. Once you own a VENDOR ID you are free to have 65536 uniquely different products.

however to licence "Atari" as a unique device of its own you need your own set of id's and they cost a packet >$10,000 for a block of 256 'unique' id's notice the use here and above of this word 'unique' and this is what I ment

This is all NOT true.

How can you buy 256 unique VID's? VID is only a 16-bit field! USB would have sold out of unique VID's in no-time if they sold then 256 at a time!

we as retro guys should !DEMAND! a free allocation of say 256 id's not pids ~{public ids} but uid's for retro machines

AFAIK (and I am no expert) there is no such thing as UID in USB nomenclature.

PID in USB nomenclature is not {PUBLIC ID} it stands for {PRODUCT ID}

[simbo]

you miss the point totaly alex and launch into a rant about usb if you wish to shell out 4000 quid
becouse when i posted the post i did this price was 10000 and the price was reduced
and your cross thoughts and fragments of a misconstrued text i posted in responce to your last rant

i simply pointed out
1} there should be a listed set of vids specificaly dedicated to retro users to avoid landfill
2} unique id's for this DONT exist public id strings are of no use becouse any old project can just step on your toes
3} the usb org should service this part of there device by makers listing free to retro i/o device makers who choose to register under a specific set or sets of specific retro vid's and further to promote a green theme and approch

if you read the whole set of posts i made you would see this

it is a fact if you use specific ic's the makers now want you to use there own tools
and buy a block from them to also use there libs and features some features on usb ic's are worth paying
this for... i dont want to pay for it or $4000 to make a device as a user of a retro machine

i feel we have a right to demand a block from the 'usb org' wtf do they think they are to deny this type of table
if you use some companies vid's then there is a legal possibility they can simply forclose on the deal or in other ways you end up loosing rights etc... na not good enough

to buy anything from the u/s usualy costs X2 anyway and in the last year i think the £ lost 43% of its value againt the $ so in real terms its 10,000 + costs compaired to what it was last year... so maybe it now costs 20000£$'s in realterm 'value' to buy the same 16bit block of vids....


now instead of fighting if you can arrange for some free sample ic's for me to use some sort of api in c++ for interfaces i will add your companies ic's and supported accepted vids etc to comvdd and my other proteus vsm projects for serial and usb comvdd v3 is in progress to add usb buffered inputs for all functions
i need a micro with at least 16 32-bit a-d
i plan to use several ic's to allow 128 pots adjustment in realtime 32bit possibly 64bit pots 32bit displayed steps
rotary controller banks can be added using LS165's
just now i use 16 slider's
but i built a 16 rotary controller cashed bmp

so really i need an ic that provides 64bit a-d X128 or 32 is ok
runs at about 100mhz is nice...

[alexh]

when i posted the post i did this price was 10000 and the price was reduced

Ooo... lets go back to say... July 2006...

http://web.archive.org/web/200607192009 ... rs_landing

And you can see it is $4000 Do you always make up stuff when you don't know it?

i dont want to pay for it or $4000 to make a device as a user of a retro machine

As I tried to tell you... you don't have to. If you use someone else's IC it will come with a VID (which you can change in software if you like) and you set the PID yourself, together with the ID string. Chances of you choosing a number for which your host already has a driver for is slim. Plus I would think that endpoint enumeration would give you a working device anyhow. (But dont quote me on that)

i feel we have a right to demand a block from the 'usb org' wtf do they think they are to deny this type of table

I hope this was a sarcastic joke? You sound like a nutter. "Jihad on USB.org!!"

in the last year i think the £ lost 43% of its value against the $ so in real terms its 10,000 + costs compared to what it was last year...

True. But it never was $10,000 and I've proved it.

so maybe it now costs 20000£$'s in realterm 'value' to buy the same 16bit block of vids....

I tried to explain before, you do not normally buy a "block". You buy one VID for your company and then you are free to make up PIDs (each a different product) as you see fit!

now instead of fighting

Hey you are the one talking nads!

you can arrange for some free sample ic's for me to use

Sure. PM me your address and I'll look around for an old devkit.

some sort of api in c++ for interfaces

It is not my area but AFAIK you get an example driver for HID and Bulk (which is what most people wanna do) in ANSI C for embedded processors. All you need is an 8-bit SRAM type bus to connect the chip to (Address, Data, R/W & CS).

i will add your companies ic's and supported accepted vids etc to comvdd and my other proteus vsm projects for serial and usb comvdd v3 is in progress to add usb buffered inputs for all functions

I have no idea what any of that means.

[simbo]

well sorry im NOT a foo fighter m8t

take it onboard please....!

[alexh]

??

[simbo]

http://www.sumzi.com/en/articles/16/5617.html

here is a nice usb host/slave controller

might be handy for atari users

[simbo]

hi

here is a cool idea

why not reuse the floppy port pre controller

as the dma chip its safe to assume has no actual transaction pin timing set over and above that used for dma pin to pin etc for floppy transfers as the dma chip is otp fpga/flash/lattice in nature
i assume the floppy systems transaction is solo limited by the transactions them selfs and the floppy clocking

so this fully depends on its sections clock and handshaking

perhaps we can use double 8 bit read/ write much like the dma port
to make a second fat dma partition's a/b drives 2 4 8 gb
with a tos patch adding the right sizing /buffers and controller signal changes rotations
we need to have but not essential
the commented source for tos2.06 maybe someone reads this in ex atari and feels sorry for us and posts it to pm or something
this is possible and possibly a good idea for sdcards as a media that more interchangable with pc's

ie we can keep the good points of the floppy and add the sdcard as a more durable replacement

anyone interested to work on this contact me

obviously the first thing to do is set some limits like 1gb 2gb 4gb sizes
and the floppy icon bmp in tos needs adjusted to a sdcard icon
and some sort of hot swap made possible using drive signals
possible reuse of external port and also st level machine etc compatable etc accross platforms

i see this as of use to many machine types .... in future

as i say 'built for comfort not for speed' it best here
im on a little in the idea and it take only one micro plugged into where the floppy controller was and maybe one or two lines from the existing dma asci port support ic's

this is basicaly a win32 sdcard via floppy port

'built for comfort not for speed'

a tip for taking apart atari apps is to use w32dasm in the pc or ida pro
they give 16 bit assembler and string referances at least of nodius

[simbo]

fat 12 floppy stuff

has a max partition size of 34mb

before virtual extendors

so a 34 mb drive in sdard

that the card its self is just a buffer

and the pc/mac is connected at the same time and can alter files
on the remote card

is easy using gemulator parts its easy,,, to create images etc.....

replaces the whole floppy handeling system
in atari st...X

with sdcard {hardies!}

[dexterslab]

tbh, it sounds like a lot of work... but what does it really give you over using an ultrasatan drive?

not to mention a lot of people still need their floppy drive!

[Jookie]

but what does it really give you over using an ultrasatan drive?
not to mention a lot of people still need their floppy drive!

As I understand it, it should be a floppy drive replacement - so you could easily upload floppy disk images and play games, etc. You can't do that with any hard drive, as the games are mostly not modified to work from hard drive.

[dexterslab]

but what does it really give you over using an ultrasatan drive?
not to mention a lot of people still need their floppy drive!

As I understand it, it should be a floppy drive replacement - so you could easily upload floppy disk images and play games, etc. You can't do that with any hard drive, as the games are mostly not modified to work from hard drive.

but thats been done allready, why re-invent the wheel?