Atari ST hard disk facts, things to know

[ParanoidLittleMan]

When writing 'NOT using DMA' should specify what DMA .  Plus, how it works internally is not really interesting for Atari users.

I guess that those newer microcontrollers are way faster than what is in STs . Should be able to be fast enough without their DMA.

Actually, what should be harder part is serial-parallel and parallel-serial conversions. Because SD cards have serial data transfer.

Let's see:  Atari DMA chip is clocked 8 MHz. Data bus of it is 8 bits. It can transfer in DMA mode 1 byte in min 4 cycles of DMA chip - that gives 2 MB/sec peak transfer rate - with fastest adapters. And in serial it means 16 Mbits/sec . That would mean something like min 32 MHz clock,  special section to doing it with HW logic instead with SW . 

So, I ask what is clock of that  STM32F103 ?     Hmm - it starts with STM - funny. Or it is how programmed version is marked ? What is marking of unprogrammed ?

 

On 11/17/2020 at 4:51 PM, Chri O. said:

I'm bumping this post because I need some HD solution for my Atari 1040 STe.

This ACSI2STM looks interesting, perhaps I can port this to Teensy 4.1. with 3.3v <> 5v level shifter. (need the speed of native (4 bit SDIO) micro SD card port access)

 

I am also looking into the HD IDE interface for Atari ST/E based on the GAL20V8 IC (Generic Array Logic) solution but I am complete noob at programming GAL chip. Well with the GAL IC the main issue is they are obsolete so I am forced to use Atmel AFT22V10C and using WinCupl as the editor and I got plenty of syntax errors. Wish there was simplified BOOLEAN-EQUATIONS version with out the TOS ROM switching.?

 

Sorry, I missed some posts in this thread - that took over a year actually.  So, how is it currently with GAL logic conversion ?

I have GAL files for IDE - first one is from 1992 (When GALs went in sale), and it has added logic for TOS 2.06 in ST - of course, because I did not know then to solve IDE autoboot with other TOS versions . Plus 2.06 was attractive, and to add that they sold it for 200 DEM in Germany (2 programmed EPROMs) .   Other one is later design, for Mega STE, without data line buffers. There is 2.06 support, and it is not mandatory to use it.  Since adding it costs nothing, sure it is there .

Now should go with such IDE interface what solves timing problems - otherwise it will not work with plenty of CF cards.

http://atari.8bitchip.info/astideTP.html

 

[TGB1718]

7 hours ago, ParanoidLittleMan said:

So, I ask what is clock of that  STM32F103 ?

I believe 72MHz

[Chri O.]

On 7/13/2022 at 3:14 AM, ParanoidLittleMan said:

Sorry, I missed some posts in this thread - that took over a year actually.  So, how is it currently with GAL logic conversion ?

I gave up on that idea and just used Teensy 4.1 microcontroller. ?

 

Teensy® 4.1 Development Board Specifications:

 ARM Cortex-M7 at 600 MHz
 Float point math unit, 64 & 32 bits
 7936K Flash, 1024K RAM (512K tightly coupled), 4K EEPROM (emulated)
 QSPI memory expansion, locations for 2 extra RAM or Flash chips
 USB device 480 Mbit/sec & USB host 480 Mbit/sec
 55 digital input/output pins, 35 PWM output pins
 18 analog input pins
 8 serial, 3 SPI, 3 I2C ports
 2 I2S/TDM and 1 S/PDIF digital audio port
 3 CAN Bus (1 with CAN FD)
 1 SDIO (4 bit) native SD Card port
 Ethernet 10/100 Mbit with DP83825 PHY
 32 general purpose DMA channels
 Cryptographic Acceleration & Random Number Generator
 RTC for date/time
 Programmable FlexIO
 Pixel Processing Pipeline
 Peripheral cross triggering
 Power On/Off management

 

[ParanoidLittleMan]

15 hours ago, Chri O. said:

I gave up on that idea and just used Teensy 4.1 microcontroller. ?

....

 

Hmm.. - that GAL logic was for IDE adapter + TOS 2.06 support in STs .  You used than monster Teensy 4.1 for IDE adapter for ST ?

I see GALs 16V..  , 20V still at AliExpress, and prices are low - like 5x GAL16V8D about 4 euros .

I ordered few years ago some quantity of older versions 16V8B - they were old manufacturing, but all it worked well.

What is price of that Teensy microcontroller ?

[masteries]

On 7/13/2022 at 4:15 PM, TGB1718 said:

I believe 72MHz

Yes, and 50 MHz for I/O  (it´s the maximum switching frecuency).

 

As well it contains a 4 instruction memory in the processor, where instructions are waiting to be executed; such a tiny cache.

Probably its the reason that a STM32 is capable to read ACSI port without use IRQs.

 

Current performance is around 1 MB/s or over.

 

 

While Teensy solution is excellent, if memory serves me well, Teensy solution reaches 1,7 MB/s

But Teensy is a bit expensive compared to the humble STM32F103

 

 

 

[TGB1718]

2 hours ago, masteries said:

While Teensy solution is excellent, if memory serves me well, Teensy solution reaches 1,7 MB/s

But Teensy is a bit expensive compared to the humble STM32F103

I went for the STM32 because of the price of a Teesey, up till now, I've not managed to get it working

I think I have the latest version of the software.

 

The on board LED flashes to indicate activity, but no SCSI devices are found.

I'm using ICD Pro, I have tried the demo versions of PPDriver and HDDriver again both fail

to see any SCSI devices (HDDdriver just crashes anyway, terrible piece of software, even crashes with my "real" hard disk only attached).

 

I'm going to dismantle it and re-wire completely at some point

[masteries]

23 minutes ago, TGB1718 said:

I went for the STM32 because of the price of a Teesey, up till now, I've not managed to get it working

I think I have the latest version of the software.

 

The on board LED flashes to indicate activity, but no SCSI devices are found.

I'm using ICD Pro, I have tried the demo versions of PPDriver and HDDriver again both fail

to see any SCSI devices (HDDdriver just crashes anyway, terrible piece of software, even crashes with my "real" hard disk only attached).

 

I'm going to dismantle it and re-wire completely at some point

 

Its your STM32 an authentic STM32F103C8?; clones such DKS, CH32 and misslabeled as original ones cannot support the Timer 4 features used by v2 and v3 of the software.

 

Also, use the +5V and GND pinout for powering, the micro USB introduces many incompatibilities with power sources,

 

 

 

[TGB1718]

24 minutes ago, masteries said:

Its your STM32 an authentic STM32F103C8?; clones such DKS, CH32 and mislabelled as original ones cannot support the Timer 4 features used by v2 and v3 of the software.

 

How can you tell if it's a clone or not ?

[ojanhk]

You can first check the chip itself to see what’s written on it.  If it’s not STM32F103, you have a clone. If the chip is properly marked, you might still have a clone, in which case I would advise using the following test program : https://mecrisp-stellaris-folkdoc.sourceforge.io/bluepill-diagnostics-v1.6.html

 

Hope this helps. 

[TGB1718]

1 hour ago, ojanhk said:

Hope this helps.

Thanks, downloaded and programmed the ST32 using an STLINK with the diagnostics, unfortunately

mine doesn't create a "COM" port when the USB is connected, so I can't see the results

 

I've looked for drivers for the USB COM port and could only find one (from ST.com) which again didn't seem to create a port when connecting the device.

[ojanhk]

Did you try using an USB to Serial adapter and connecting it to pins A9 et A10 ? I haven't used it since troubleshooting my Internal ACSI2STM adapter for MegaSTE, but I don't recall using the USB port as a serial port. I always use a USB to Serial(TTL) adapter for programming and debugging. Also, did you switch back the jumpers after programming (I forgot a few times)? 

[TGB1718]

3 minutes ago, ojanhk said:

Did you try using an USB to Serial adapter and connecting it to pins A9 et A10

Oops, forgot that, will try that later ?

[TGB1718]

21 hours ago, ojanhk said:

Did you try using an USB to Serial adapter and connecting it to pins A9 et A10 ?

Unfortunately that doesn't work, the software .pdf says the serial uses the USB connector (I've never had any success with serial on that connector).

Looking at the source code it is using the USB port

Can't change the code and recompile as I think it's written in FORTH ⁉️

 

However a lot of searching on Google and I came across something that says Fake STM32's fail the basic "blink" program.

Guess what, mine fail the blink program ? so looks like I got some fakes.

 

Will try get a "real" one and rebuild

 

@ojanhk thanks for letting me know about this, been banging my head against a wall with this for ages.

[ParanoidLittleMan]

Are those fake actually really manufactured by someone else, or they are just those originals which did not went thru quality control, went to someone who should to dispatch them to some recycling or trash disposal, but ... you know where they went at end - in World where money is most important.

Because I see more and more of such things, not only in electronic.

Politicians, governments should lead war against such people, not war for territory, control over people.  Here in my country some sale people gained enormous moneys in few years. While people complains about crap quality, short lifetime over decades, and it just goes worse and worse. And worst what can happen if some are catched - some financial punishment. Although, I really do not remember hearing that anyone was punished seriously for selling crap. And it seems not only in Eastern Eur.  In my case it went so far that I really hate to go shopping. All it is well known, and that those up talking about: very bad times come, people use less electricity, gas, etc.

We going in anarchy. It's already worse in many things than in movie Idiocracy.

Sorry for being off topic - not me started with it.

[TGB1718]

55 minutes ago, ParanoidLittleMan said:

Are those fake actually really manufactured by someone else, or they are just those originals which did not went thru quality control,

From what I've read, most are Chinese fakes, although I suppose some could have fallen out the back door after failing quality control. 

[ParanoidLittleMan]

Well, I will add something really interesting to this talk about fake, crap ware on sale:

So, yesterday wire leading to 3.5 mm connector with 4 contacts just pulled out from small case where mic. and button are - without any excessive force, I just took it out from my shirt pocket. And I bought it 1 month ago.  Yes, it was very cheap - about 2 Euros, and I actually did not want to buy it because low price and because was not clear has it mike. Most visible was text 'super bass' . But in that shop was no other head set good for my phone - only those without mic, or with micro USB conn.

So, I asked sale person to open it, and I saw that there is mic and button.  OK, buying .

So, after that detaching I decided to combine that wire with good 3.5 mm connector with better head set (better sound) part - where there was problem near connector with wires (they break mostly there). As said, putting there 3.5 mm connector what can buy separately is out of question - they are utter crap.

So, I removed insulation from cable end - and what the Hell ? Only 3 wires instead 4. OK, I must've cut forth. More removing, and still 3 wires ... So, I opened that small case with mic and button, and here is what saw:

Yep, left and right channel are soldered together - red and blue wires. So 1 line for 2 channels. And that's 1 wire. Other 2 are: GND (not colored copper) and mic/button common one (as usual).  I noticed that sound is strange, and bass is poor. Well, used it mostly for phoning in city noise.

How can such stupidity happen at all ? How much they saved with that 80 cm wire with 1 (tiny) line less ?  Maybe they just had such cable a lot, so use it ... I need to go back in that shop and report cheat - on box it writes:

'MUSIC' 'SUPER BASS' 'Headset Stereo'   LOL .

And what is worst in all this: I expecting instead measures against such manufacturers, sale firms and related  that we will have more and more of it. 

Edited July 17, 2022 by ParanoidLittleMan