Terrible Fire Accelerators

[plasmab]

Originally Posted by Mathesar

I made some progress on the open source A500 IDE controller I promised I would be working on.

Happy "building wishes" to the thread, I expect to see some 030 powered Amiga 500's soon!

Great work.

[Mr.Flibble]

I'm currently noodling around with a pretty basic kickstart flash rom board - will post when I've got the boards and been able to test it

[RichMK]

Originally Posted by Mathesar

I made some progress on the open source A500 IDE controller I promised I would be working on. I have received the PCB's (from Elecrow BTW) and soldered one up. I made one mistake in the layout so I had to lift one leg on the Cypress chip and install a botch wire on the underside. Nothing serious so far... (fingers crossed ).

The board is a standalone Gayle style IDE controller for the Amiga 500. It reads HDF files from a standard FAT formatted SD card. I hope to program it such that you won't need an _ovr and _int2 patch wires. The board also contains 1Mbyte fast RAM of which I intend to use half a megabyte to load Kickstart 3.1 (which is needed for the Gayle SCSI.device) from the SD card. This way no real 3.1 ROM (which is troublesome on older A500's like mine) is needed. Of course you would need a properly licensed rom file instead.

This project will be open-sourced once finished and would make a nice companion to a TF520. However, I am doing this in my own slow pace. I also tend to work in secret so don't expect any blogs, vlogs or other updates anytime soon

Happy "building wishes" to the thread, I expect to see some 030 powered Amiga 500's soon!

Great work! It amazes me the effort people are still putting into the Amiga after all these years! open sourced is the way

[Tachyon]

Originally Posted by Mathesar

I made some progress on the open source A500 IDE controller I promised I would be working on. I have received the PCB's (from Elecrow BTW) and soldered one up. I made one mistake in the layout so I had to lift one leg on the Cypress chip and install a botch wire on the underside. Nothing serious so far... (fingers crossed ).

The board is a standalone Gayle style IDE controller for the Amiga 500. It reads HDF files from a standard FAT formatted SD card. I hope to program it such that you won't need an _ovr and _int2 patch wires. The board also contains 1Mbyte fast RAM of which I intend to use half a megabyte to load Kickstart 3.1 (which is needed for the Gayle SCSI.device) from the SD card. This way no real 3.1 ROM (which is troublesome on older A500's like mine) is needed. Of course you would need a properly licensed rom file instead.

This project will be open-sourced once finished and would make a nice companion to a TF520. However, I am doing this in my own slow pace. I also tend to work in secret so don't expect any blogs, vlogs or other updates anytime soon

Happy "building wishes" to the thread, I expect to see some 030 powered Amiga 500's soon!

I certainly could be wrong, but my research of late is leading me to believe that it's going the A600 style Gayle emulation route that leads to the necessity for the _INT2 & _OVR jumpers.
That said, I think there's possibly a catch-22 here in that not going that route eliminates the 68k socket option and forces you to a Zorro option anyway.
At the least, I think a device will need it's own Autoconfig ROM to load it's own driver at system initialization, just like any bootable Zorro slot device.
I dunno, just my thoughts based on what I've read so far.

Man, I used to know this stuff inside and out, but it's been a lot of years.

As for working in secret, I can respect that. I do it myself. But there's a lot to be said for having a place to bounce your ideas off others and exchange ideas, information, and hard earned experience. The issue is usually more about finding the right "forum*" to do that, as we discovered yet again, in this thread recently. The point being, that I think lonewolf efforts end up needlessly butting heads against problems that have already been solved, and being held up for lack of information that other's already have. But certainly your call though.

I just find it frustrating that the vintage community has so much of an ingrained mindset that's opposite to the kinds of workflow and collaboration that allowed projects like Linux to essentially takeover the world. When we stop worrying about hoarding IP and ideas and start thinking in terms of getting as meany eyes on a problem as possible, huge progress becomes the norm rather than the pipe dream. But keep in mind, that's a general statement, not a personal admonishment. ;')


*Because I know butthurt will result if I don't clarify, I'm using the term forum here in it's more traditional, general definition, not in the sense it's used today to mean specific online groups.

[wXR]

@Mathesar thank you for being a hero!

[PeteAU]

Originally Posted by Mathesar

I made some progress on the open source A500 IDE controller I promised I would be working on. I have received the PCB's (from Elecrow BTW) and soldered one up. I made one mistake in the layout so I had to lift one leg on the Cypress chip and install a botch wire on the underside. Nothing serious so far... (fingers crossed ).

The board is a standalone Gayle style IDE controller for the Amiga 500. It reads HDF files from a standard FAT formatted SD card. I hope to program it such that you won't need an _ovr and _int2 patch wires. The board also contains 1Mbyte fast RAM of which I intend to use half a megabyte to load Kickstart 3.1 (which is needed for the Gayle SCSI.device) from the SD card. This way no real 3.1 ROM (which is troublesome on older A500's like mine) is needed. Of course you would need a properly licensed rom file instead.

This project will be open-sourced once finished and would make a nice companion to a TF520. However, I am doing this in my own slow pace. I also tend to work in secret so don't expect any blogs, vlogs or other updates anytime soon

Happy "building wishes" to the thread, I expect to see some 030 powered Amiga 500's soon!

Did you write your own SDcard<->IDE software? Sounds like quite a feat.

[plasmab]

Originally Posted by PeteAU

Did you write your own SDcard<->IDE software? Sounds like quite a feat.

Without taking anything away from Mathesar. SDCard to IDE software already exists in the MiST firmware. Also SDCard to 1772 floppy emulation.

[tnt23]

@Mathesar: cool! Looks like you are using KiCad

[thebajaguy]

Originally Posted by Mr.Flibble

Yeah, the A1000 is flipped upside-down iirc, so the even pins are on top and the odd are underneath - need to confirm that though - will have a look for sources shortly.

<br />
<br />
The A1000 edge connector is (effectively) identical to the A500 (top and bottom) but 180 degrees on the horizontal - meaning A500 peripherals on the left facing forward become rearward-facing on the A1000 right.

[Codetsu]

plasmab Wellcome back

[Mathesar]

Originally Posted by plasmab

Without taking anything away from Mathesar. SDCard to IDE software already exists in the MiST firmware. Also SDCard to 1772 floppy emulation.

Exactly! That is the power of open-source. We can build upon what others did before us. I am going to use the code written for the Minimig (MiST probably uses something similair). The only thing I have to do is to intercept all writes to Gayle using the CPLD part of the PSOC and direct that into the Minimig/Mist code.

PS good to see you back!
PS2 Yeah, I am on KiCad.

Posted from my Amiga 3000

Would this accelerator work on an Amiga 2000 ? If so, where would be the best place to get _OVR and _INT2 from ?

Thanks - Patrick

[Zetr0]

@prynhart

Indeed it should, you can get these from either the CPU port itself or just to the rear right hand side - Kipper2k has some information on obtaining these signals from his website.

[roy bates]

you can the signals off the co processor slot

pin 17 is _OVR
pin 19 is _INT2

could probably get it off the custom chips,buster and a cia ,i think,but the co pro socket is nearer

[plasmab]

Originally Posted by roy bates

you can the signals off the co processor slot

pin 17 is _OVR
pin 19 is _INT2

could probably get it off the custom chips,buster and a cia ,i think,but the co pro socket is nearer

Just as a note. As of the latest firmware you don't technically need OVR.

[roy bates]

fair enough

[havok1919]

Finally finished with taxes over here, so I put a little time in on bringing up my TF530. In case it's of use to anyone else I did a short, progressive smoke test as I'm going along:

*no CPU, FPU, or oscillator fitted
*3.3V regulator and all other components on the board loaded
*blank CPLDs

Applying 5.0V DC power from a lab supply showed about ~60mA draw which seemed reasonable. 3.3V rail read spot on.

*Programmed the BUS CPLD

Current consumption up to about 90mA.

*Programmed the RAM CPLD

Current consumption up to about 120mA.

Now I need to go snag a 64 pin DIP header from storage and either find a suitable oscillator or remember where I put my blank programmable ones...

-Clay

[plasmab]

Originally Posted by havok1919

*Programmed the BUS CPLD

Current consumption up to about 90mA.

*Programmed the RAM CPLD

Current consumption up to about 120mA.

Now I need to go snag a 64 pin DIP header from storage and either find a suitable oscillator or remember where I put my blank programmable ones...

Power consumption is a little high but not off the scale. I'd expect about 60mA per without the CPU or FPU. Keep an eye on the clock buffer chip getting hot when you hook up an oscillator as a short on the clock lines can kill the buffer chips.

My feeling is you have a short somewhere. Possibly across the RAM pins somewhere. Could be completely wrong but worth keeping an eye on.

[havok1919]

Originally Posted by plasmab

My feeling is you have a short somewhere. Possibly across the RAM pins somewhere. Could be completely wrong but worth keeping an eye on.

Possible... I gave it a pretty good eyeball-ing though, so I doubt it's a short (the SRAMs alone could easily be over 10mA with the enables floating before the RAM CPLD was configured) and we'll be wasting a fair bit in that LDO with ~1.7V drop across it too. I'll give it a quick peek with the thermal camera just in case.

[havok1919]

I think it's fine-- not seeing any temperature spike on any one of the SRAMs on power-up. Most all the heat is in the regulator and the CPLD's. (Of course this is all un-clocked still so it'll go up dramatically once we get some edges happening...)

(I even set the markers to °C since there's probably more non-USA people reading this than not.) ;-)