New - Amiga Romulator for A500 - A600

[Ledfoot]

Only at AoTL AFAIK, and they've been out of the A500 model for a while now.

[dalek]

Originally Posted by seb132

Are these wonders available as open source projects ?
Or only sold by AOTL ?
Would be good to have another partner in Europe and/or in Australia because shipping costs from US to overseas became ridiculous during Trump...
Last question, will we see Romulators for 32 bits Amiga ?

cmorley has a flash eprom emulator over on the exxos forums that works really well. His shipping is reasonable to Australia. I'm super happy with mine. Slightly different product to the romulator but gets the end result. They can be linked for systems that require 2 eproms.


https://www.exxoshost.co.uk/forum/vi...hp?f=44&t=2018
https://exxoshost.co.uk/forum/viewto...eprom&start=70

[Ledfoot]

Ok, so couple things...

My MB is an A2000 Rev 6.1.

Took the A2620 out, installed the Romulator. It SEEMS to boot up KS 1.0-1.3 just fine, as well as DiagRom. I say seems because when I then put my Buddha IDE in and tried to boot up the 1.3 HD image I had on a SD card I get guru's during the boot process. Remove the Romulator and put the physical 1.3 chip back in and it boots just fine with no error. (Roms from a new legal copy of Cloanto's Amiga Forever 8).

If I try any of the 2.x or later ROM images in the Romulator I just get a red screen.

I unfortunately don't have any physical rom chips for 2.x or higher, just the 1.3 that was in the machine when I got it. I don't know if the issues I'm seeing are a factor of my MB (bad trace?), some oddity of the REV 6.1 board, or what...

[indigolemon]

Hmmm. Does the 6.1 have the KS Socket pinout issues the rev5 500 has?

[Ledfoot]

Originally Posted by indigolemon

Hmmm. Does the 6.1 have the KS Socket pinout issues the rev5 500 has?

No idea... I'm wondering if that's the case though.. I thought they resolved those in Rev 6 though...

[Zippy Zapp]

Originally Posted by indigolemon

Hmmm. Does the 6.1 have the KS Socket pinout issues the rev5 500 has?

I do not believe so. The 6.1 rev was newer then the A500 Rev 5 and came with 8372 Agnus. It is one of the later revisions so I would not think it has that issue. Did other 2000s have this issue because I didn't think the did but don't really know?

It seems maybe the ROMulator has a few conflicts? Perhaps they can be fixed. You can always revert to an old school kernel switcher or one of the newer open source ones that use higher capacity EPROMs to store multiple kernels...

[cdh]

Some time ago, I came across a note from Dave Haynie posted online about the A2000 which said, in part:
Rev 6.1 corrects the clock noise.
Rev 6.2 addresses the expansion bus noise.

I don't know if 6.1 expansion bus noise is the cause of your issues between the Romulator and the Buddha.

Red screen from Kickstart essentially means that the CPU was able to execute code from the Kickstart ROM, but that a checksum of the ROM failed. Could it be that the ROM checksum code was added after 1.3?

[Ledfoot]

Originally Posted by cdh

Some time ago, I came across a note from Dave Haynie posted online about the A2000 which said, in part:
Rev 6.1 corrects the clock noise.
Rev 6.2 addresses the expansion bus noise.

I don't know if 6.1 expansion bus noise is the cause of your issues between the Romulator and the Buddha.

Red screen from Kickstart essentially means that the CPU was able to execute code from the Kickstart ROM, but that a checksum of the ROM failed. Could it be that the ROM checksum code was added after 1.3?

Yeah, that's about all the info I've been able to find on it. There's apparently changes that can be done to bring it up to a higher rev but I can't find a definitive list anywhere to know what I need to do or what parts, which is annoying, and I don't even want to bother if it won't fix this kickstart issue.

Tempted to order a 3.1 or 3.1.4 ROM chip to see if it's the Romulator or Rom images I'm trying to use or if it's some sort of hardware issue.

Looking through the A2000 Rev 6 schematics, there's A0-A17 (so 18) lines going into the ROM, so 2^18 = 262144 addresses, so I don't see how a 512KB ROM would even work in these boards, it would seem to need 1 more address line to have enough addresses for 512KB. Am I missing something??? Or is it because it's a 16 bits at each address so it's effectively 512KB of 8bit data (ie, high/low packing) ?

[cdh]

Ledfoot, your calculations are correct as to address lines and the number of unique addresses which can be driven. The minor detail you are missing is that this is a 16-bit part. You'll notice in the schematics that Amiga A1-A18 are connected to the chip. It's not in BYTE mode, so each fetched address is 2 bytes. So that's how you get to 512KB ROM.

[Ledfoot]

Originally Posted by cdh

Ledfoot, your calculations are correct as to address lines and the number of unique addresses which can be driven. The minor detail you are missing is that this is a 16-bit part. You'll notice in the schematics that Amiga A1-A18 are connected to the chip. It's not in BYTE mode, so each fetched address is 2 bytes. So that's how you get to 512KB ROM.

Yeah that's what I figured afterwards...

Ok, well, I've ordered a 1.3/3.1 switcher with physical chip. I give up with this thing. If that works then I think the Romulator is having some sort of issue.. The system seems utterly rock solid with the 1.3 chip in... Anything I try in the Romulator either red screens immediately, or will guru during boot up of Workbench... I BRIEFLY got 2.04 to work but it guru'ed during the WB boot up. Then I couldn't get it to start up again after that... Always a red screen... Sigh.

[mkstr]

You mention it seems to work fine until you put the IDE board in. This makes me think it's a bus loading issue, which could point to this expansion bus noise problem someone mentioned.

Have you got a serial adapter hooked up to it? If you can go on the serial console it would be interesting if you could plug in the IDE board and then do "halt" and then run "memtest" repeatedly. I would like to see if something is messing up the STM32's ability to program the SRAM. The SRAM itself is 5V powered and should work reliably with the Amiga side.

I could potentially add some more aggressive verification during loading and send you a firmware update. You would need to hook up a USB cable to install this.

Mike

[Ledfoot]

Originally Posted by mkstr

You mention it seems to work fine until you put the IDE board in. This makes me think it's a bus loading issue, which could point to this expansion bus noise problem someone mentioned.

Have you got a serial adapter hooked up to it? If you can go on the serial console it would be interesting if you could plug in the IDE board and then do "halt" and then run "memtest" repeatedly. I would like to see if something is messing up the STM32's ability to program the SRAM. The SRAM itself is 5V powered and should work reliably with the Amiga side.

I could potentially add some more aggressive verification during loading and send you a firmware update. You would need to hook up a USB cable to install this.

Mike

No, I was able to get a 1.3 KS working and WB 1.3 booting up off floppy with the Romulator when I stripped the machine down to nothing. But when I put the Buddha IDE in and tried to boot up my 1.3 WB image off that it guru'ed and crashed.

I've tried every KS image I could find (everything off the Cloanto Amiga Forever 8, and a bunch of other images I've had kicking around from various sources, as well as a legal 3.1.4 from Hyperion... Most immediate Red screen with or without the Buddha IDE installed. A few would just hang with a black screen and do nothing (had to power cycle to get back out).

I've done continuity checks between the CPU and the KS socket - all the address and data lines check out, so no bad traces that I can tell.

I even went so far as to change out 68000 chips - the one that came in this 2500 was one that they listed in the revision history as causing some signaling issues. Swapped it out with one from an A500, that had a different revision number on it. No difference in how it worked though.

At this point I'm giving up. Same setup with the original dedicated 1.3 KS chip in, everything works perfectly and super solid. And even if we do get the Romulator working, I'd still have to figure out how to get it working with the A2620 card...

If you can tell me what to hook up and how I'm willing to help look into things a bit until my 1.3/3.1 KS switcher arrives in a few weeks. I have a USB/serial adaptor that I use for programming gotek drives. I assume that's the thing you're talking about to hook up to the Romulator to check the serial console??

[mkstr]

Instructions for hooking up the serial cable further up.

If you do the "halt" command, which will hold the machine in reset, then run "memtest" a few times this should show up any problems in loading the ROM image into the SRAM.

[Ledfoot]

Ok, hooked it up... Did the halt, then the memtest, it seems to cause it to reset to the menu and then the timeout kicks in and it tries to load in the last selected rom. In this case it tried to load in "amiga-os-310-a600.rom" size 524288 kickstart file that I got off the Amiga Forever 8 DVD.

Here's the output when it did that:

CLI> memtest
1388535206.000 main.c(640): Starting SRAM memory test...
1388535206.000 main.c(652): Verifying...
1388535206.000 main.c(675): Done
CL1388535206.000 main.c(486): Detected user request for ROM menu
1388535206.000 usb.c(187): Disabling USB device mode
I> 1388535206.000 main.c(190): System Volume Information 0 [DIR]
1388535206.000 main.c(190): amiga-os-070.rom 262144
1388535206.000 main.c(190): amiga-os-100.rom 262144
1388535207.000 main.c(190): amiga-os-110-ntsc.rom 262144
1388535207.000 main.c(190): amiga-os-120.rom 262144
1388535207.000 main.c(190): amiga-os-130.rom 262144
1388535207.000 main.c(190): amiga-os-204.rom 524288
1388535207.000 main.c(190): amiga-os-205-a600.rom 524288
1388535207.000 main.c(190): amiga-os-310-a600.rom 524288
1388535207.000 main.c(190): amiga-os-314.rom 524288
1388535207.000 main.c(190): diagrom-1_2_1.bin 524288
1388535207.000 main.c(190): kickstart v31-a.rom 524288
1388535207.000 main.c(190): kickstart v31-b.rom 524293
1388535207.000 main.c(190): kickstart v31-b2.rom 524288
1388535207.000 main.c(190): kickstart v204.rom 524288
1388535207.000 main.c(190): kickstart v204-b.rom 528509
1388535207.000 main.c(190): kickstart v204-c.rom 524288
1388535207.000 main.c(417): Card read OK
1388535207.000 main.c(323): default = amiga-os-310-a600.rom
1388535207.000 main.c(330): Found default rom 'amiga-os-310-a600.rom' in slot 7
1388535207.000 main.c(323): delay = 5
1388535207.000 main.c(323): update_default = true
1388535207.000 main.c(323): play_music = true
1388535212.000 main.c(436): Loading rom 7 : amiga-os-310-a600.rom
1388535212.000 main.c(440): Updating saved default
1388535212.000 main.c(253): Image is <= 512K
1388535212.000 main.c(271): Image is an Amiga ROM
1388535212.000 main.c(314): Read 524288 bytes
1388535213.000 usb.c(176): Switching to USB device mode

At that point I just get the Red screen over and over and over.

halt and memtest again just do the same thing.

Same with any of the larger roms. The majority will immediately red screen. I can SOMETIMES get the 2.04 or 2.05 roms to load but then the machine will guru during boot up, or will get stuck in a loop where it'll flash the gray and white and black screens, pause a second and then cycle through them again, over and over.

With the 1.3 rom it will not boot consistently.. I'll try to boot off the WB floppy, it'll start booting and then I'll get a Guru #00000009.002198F6 over and over...

It doesn't do this at all with the original 1.3 rom in - everything works fine - I can boot off the WB floppy all the way to workbench, or if I have the Buddha IDE in and have a formatted SD with WB on it it'll boot right up with it.



At this point I'm out of ideas...

[dwfa]

Hi - anyone know if AotL will have the romulators for A2000 coming back or are they gone for good?

[kipper2k]

Hi, Yes, they will have some more soon!

[R4M]

There are some 68020 accelerator boards with loads of FastRAM, IDE and a MapROM feature for around 140$ (or below 80 if you DIY). Those do almost everything the Romulator can, and loads more. If the Romulator is in short supply or hard to find in your area, you could consider on of those.

The MapROM feature requires the computer to start, before you can load arbitrary ROMs into the MapROM. Therefore you can not use that solution for diagnosing serious hardware problems using a DiagROM, when the computer refuses to boot. But other than that, I do not see any advantage of the Romulator.

[dwfa]

Hi do you have a link to this solution below so I can get more info. I will search, but if you have something already I would appreciate it. Thx ...

Originally Posted by R4M

There are some 68020 accelerator boards with loads of FastRAM, IDE and a MapROM feature for around 140$ (or below 80 if you DIY). Those do almost everything the Romulator can, and loads more. If the Romulator is in short supply or hard to find in your area, you could consider on of those.

The MapROM feature requires the computer to start, before you can load arbitrary ROMs into the MapROM. Therefore you can not use that solution for diagnosing serious hardware problems using a DiagROM, when the computer refuses to boot. But other than that, I do not see any advantage of the Romulator.

[R4M]

I recently built

https://gitlab.com/MHeinrichs/68EC020-TK

That's a great card and completely open source. With 9,5 MB FastRAM, 1MB at $A0 and write protectable 512K at $E0, $F0 and $F8 and really fast autoboot IDE (i get 8.5MB/s with my IDE2SD adapter).
Last time I checked, there are people who sell these cards at amibay. It's also not too hard to solder it yourself. But some of the pitch sizes are really small. And you need a Xilinx programmer for the CPLD and an eprom programmer for the autoboot rom.

[mkstr]

Originally Posted by Ledfoot

It doesn't do this at all with the original 1.3 rom in - everything works fine - I can boot off the WB floppy all the way to workbench, or if I have the Buddha IDE in and have a formatted SD with WB on it it'll boot right up with it.

I forgot about halt triggering the menu, but the memtest did pass before that happened indicating that the SRAM is being programmed properly. It kind of looks like a timing issue with this particular machine (between the SRAM and the motherboard), but it seems to work in the vast majority of machines so I'm not sure what could be different here.

How does the diagrom image behave?

Mike