Indivision AGA MK3 for A1200

Quote from Jens

Timm, it appears to be board- and accelerator specific. mbruines reported that the old ACA1233 works fine, while the ACA1233n doesn't, and it's worse on one main board than on another. If you want to replicate the problem, you should use the ACA1233n-55. IHave sent one of these accelerators to Peter as well, hoping that the problem shows on his machine.

I can confirm that this demo does have problems with an ACA1233n-55 here, and I saw black screens and the OSD several times. Also there were interruptions in the sound. Just for the record, when watching this demo on an ACA1233n-55 with native video/audio, it's not free of a plethora of other problems, and it crashes early.

Quote from Timm

I can confirm that this demo does have problems with an ACA1233n-55 here, and I saw black screens and the OSD several times. Also there were interruptions in the sound. Just for the record, when watching this demo on an ACA1233n-55 with native video/audio, it's not free of a plethora of other problems, and it crashes early.

At my place the demo it self runs fine on a 030, but crashes on a stock a1200 after a minute or 10 (the issues start after the text intro when the fast switching of images start, till the crash it looks perfect without accelerator).

I think this difference in crashing on a 030 comes from the whdload version used. Did you run the diskversion? Dont think its llowed to upload the exact whdload version(?) so i will try to find the disk version,, must have it some where.

Hardwired is fixed for faster amigas in the whdload version i have. But the issue is less prominent in this demo and there fore harder to reproduce.

I didnt test the sound over HDMI since i prefer my dedicated amplifier, but i did test DVI mode on/off.

Enigma demo goes crazy when displaying the “raytraced fractal” animation loop.
I’m starting to get disheartened with my purchase now.

Just my word here.

I don't have the ACA1233n-55 but with demo's Jesus, Hardwired and Enigma no black screens, no crashes. Everything runs fine.

A1200 KS3.1.4 Rev 2B

Blizzard 1230/4 no FPU, 64MB Ram

Indivision AGA MK3

WHDLoad 18.5

Hi All,

I was going to post this info to a new thread, but it may fit better here because my problem appears to be aca1233n-55 related.

I’m experiencing constant blackouts in workbench in vanilla PAL Hi-Res mode using the Indivision AGA MK3 with its default settings.

The problem seems to happen once the Amiga warms up. When the machine is booted from cold, all seems fine for around 15 mins, then the blackouts start and are reproducible by simply moving a window or entering text.

The instability miraculously goes away if I reduce the aca1233n-55 frequency to speed step 3 (27.78MHz) or less using acatool. Hope this helps someone.

Amiga

- A1200 mainboard REV 1D.1

- aca1233n-55

- Indivision AGA MK3 with latest firmware (20200824) and config tool (1.5)

- AmigaKit clock module (fitted to Amiga’s mainboard clockport)

- Kickstart/Workbench Hyperion 3.1.4.1

- Compact Flash HDD

- modern commercial PSU (also tried original commodore supply with, similar results)

- Dell U2415 monitor (HDMI input used)

Chipset

- (see photo)

Other things that have helped

- updating to latest AGA MK3 firmware

- Turing off graffiti emulation

- removing E123C, E125C (this seems to have delayed the onset of blackouts but not solved them)

I’ve also tried experimenting with CCKL Pull-up, CCKL Capacitance, PCLK Resistor (all modes) but this did not appear to make a difference (or I didn’t find the right combo perhaps).

Any chance people can try Enigma demo? Specifically the “Raytraced Fractal” scene (green balls loop anim).

Thanks.

I can try the demos too if it helps, as I also have a 1233n. Which specific versions of which demos do you need testing out?

More than happy to try out Enigma, but the disk I have only works on my A500. Where did you get a fixed one that works on the A1200? My A1200 has the Indivision AGA MK3 as well as a GVP 1230 accelerator and CF card as its HDD.

Quote from dmac

- modern commercial PSU

Please specify - there are some PSUs on the market that are sold as "new and great", but most of the ones I know are based on MeanWell chassis, which are not suitable for a heavily expanded Amiga.

Quote from Jens

Please specify - there are some PSUs on the market that are sold as "new and great", but most of the ones I know are based on MeanWell chassis, which are not suitable for a heavily expanded Amiga.

Thanks for the response. It's a Keelog PSU (edit: Link removed, known-bad PSU!) I bought around the same time as the aca1233n-55 because the Keelog website made it sound amazing. On closer inspection, internally I think it's a meanwell RPT-60B or similar. Weirdly, the PSU case seems to glued tight so I can't easily take the lid off for a closer look!

I know these PSUs - we've made a test-purchase and found that it's a MeanWell chassis. Send it back, claim your money back. This is known-bad, and the vendor has been made aware of this in writing (yep, classic old snail mail).

I've just sat through Phenomena's Enigma demo, Anarchy's Hardcore, and LSD's Jesus on E's (in full, I now have a headache - I can't believe I used to go to raves every weekend!), all runnng from HDD via WHDLoad.

Enigma seems to run 100% fine, with no display or audio dropouts at all.

Hardcore seemed to run 100% fine also.

Jesus on E's runs fine until the last bit of the demo (which I think is the credits?). I didn't see any screen blackouts, or audio dropouts during the main demo part which lasts 30 or so mins. I did see a couple of times when the animations messed up a little but I think was just timings from running on a 030 at 50MHz and the HDD, the little fist logo in the bottom right corner never blinked out at all, nor did the TV lose sync. At the very end of the demo, when a different. non-rave audio track starts, then the screen stayed black but the music was faultless and didn't stutter, and my HDTV didn't report a loss of signal or anything.

For info purposes, my A1200 is a 1d4, recently recapped but with no timing fixes yet, powered by a CA-PSU from IComp. The Indivision AGA Mk3 is set to output 1920x1080@50Hz but with the VSync and Auto-Res both off, as either or both cause my display to fail to detect a signal. Audio is set to 96kHz, and only the CCKPull up resistance is active. It is connected to an HDTV via digital, and I am using the v1.5 Indivision tool.

The ACA1233n is set to 55MHz, pio mode 4+, vbr on and maprom off via the ACATool v2.4. MuFastROM and MuFastZero from the MMULib tools are running via the UserStartup. I don't know if this is optimal at all, it's just what it's set to at the moment in case it matters.

I just noticed something else on my 1d1 board while just playing around a bit.

When the computer is still 'cold' i can get a blackscreen when a big screen change is happening in 4 color mode, cant reproduce this in 256 colors. but till now it happens everytime i switch back to 2 bitplanes and the computer is cold.

Used screen mode to fillup the screen with a folder is hires 640x256. This also happens when opening/closing an application (nearly).screen size.

After some warming up time i cant reproduce this anymore, this is at the same point Jesus on Es stops giving black screens.

This seems to be the oposite of my 1d4 board. On this board i can reproduce this when warmed up in 256 color mode, but till now not in screenmodes with 4 bitplanes or less in workbench.

Btw. Today i had the change to try the indy on a 98" 4K display at the office, the image was very nice , but be aware that some demo's can be bit to much on displays like this..

When errors are occurring, are you seeing SHIRES appear/flashing while you leave the Live Edit OSD on screen?

I too suspect a thermal related link as I’ve seen issues occur badly one day and not so bad on another.

Haven’t ruled out all issues (inc. PSU) but when disabling the ACA1233n it does appear more stable.

BTW LeeiComp and ProdigiousLee = me

Correction to above: I actually have an ACA1233n-new (26MHz).

Currently experimenting with speedstep options.

Quote from LeeiComp

Currently experimenting with speedstep options.

My guess is that at some slower settings, problems reduce. This smells like an address/data hold time issue, and the flicker fixer sampling accesses too late. Like I wrote earlier, the ACA1233n is on it's way to Peter, so he can replicate the problem. That's a necessary step before attempting a fix.

Quote from Jens

My guess is that at some slower settings, problems reduce. This smells like an address/data hold time issue, and the flicker fixer sampling accesses too late. Like I wrote earlier, the ACA1233n is on it's way to Peter, so he can replicate the problem. That's a necessary step before attempting a fix.

Step 0 (13.33MHz) = not tested

Step 1 (16MHz) = no issues

Step 2 (20MHz) = no issues

Step 3 (26MHz) = issues are there

Step 4 (26MHz) = issues are there

I'll put acatool setspeed to 2 in the startup-sequence as a workaround, but I loose the extra speed and IDE acceleration

Just tried slower speeds with the aca1233n on my 1d1, at slower speeds the issue still appears (when cold) Only when switching to the 68020 when cold i dont see any issue.