Issues with ZZ9000 on A4000

amigasquirrel

amigasquirrel

New member
Joined
Jul 6, 2025
Posts
22
Country
UK
Region
Perthshire
Hi there,

Can anyone help with some issues I'm having with the ZZ9000 in RTG screen modes on an A4000D?

Hardware:
A4000D Acill replica PCB v1.0
16mb chip ram onboard
2mb fast ram onboard
SD to IDE converter
BFG9060 CPU card with 68060RC50 at 50mhz
Replica A4000D daughterboard.
ZZ9000 rev 4 firmware v1.13

The only card installed on the daughterboard is the ZZ9000.
All of the custom chips are soldered onto the board and the system is 100% reliable in native screen modes.
Memory tests of onboard chip RAM fast RAM, BFG9060 RAM and ZZ9000 RAM are successful using Amiga Test Kit for a number of repeat passes over many hours.

Software:
Kickstart v3.2.3
AmigaOS v3.2.3
ZZ9000 driver v1.13
P96 v3.6.2 (rtg.library 43.760)
MMUlib v47.10.1

When testing RTG screen modes the test patterns sometimes have various glitches and incorrect pixels.

When using workbench in RTG modes, I had various dots and lines appearing and parts of windows/text appearing where they shouldn't. Text on menu items is not displayed correctly - sometimes blank or incomplete. Windows and icons don't display correctly, various display glitches.

It also crashes frequently, most annoyingly when trying to export the ZZ Top bus test output into Textedit. When this happens the machine freezes and after powering off and on I get a guru 8000 0004 error, then recoverable error with ZZ9000.card complaining it needs firmware (boot.bin) 1.13, then the ZZ9000 driver will fail to load. Restarting a second or third time will usually result in it booting cleanly but display issues remain.

Running a bus test in an RTG mode sometimes shows errors other times there are no errors at all. Running sequential tests sometimes shows errors / no errors / errors despite no changes being made.

Trying to use the network adapter on the ZZ9000 is also erratic, it only works for short periods of time.

Video showing the issues and bus test results are here:
https://drive.google.com/drive/folders/1QOFNR6xxEL3Kps8oOLAwxmBo2wnU5Mdv

The only other thing I have noticed is the termination resistor arrays on the daughterboard get rather warm.

I've also tried using an original 030 CPU card and clean install of OS v3.2.3 with only P96 and the ZZ9000 drivers installed with similar issues.

Thanks,
 
What is your 5v rail measuring during the points when you experience these issues?
 
How old is the ZZ9000 card?
Only asking as the earlier revisions were known to have a faulty power rail and needed a power mod doing to it.

 
Thanks for the replies,

The 5v rail is 4.9v as measured on the ZZ9000 itself.
This is a new revision 4 card so not affected by the power rail issue.

It seems to be more reliable running in Zorro 2 mode, still lots of display glitches however bus tests seem to be better.
It only complains about a zz_perform_memtest_delaywrites... error.
 
Last edited:
What i would do in this situation is
- check if the SD card of the ZZ9000 is fine and rewite it if needed
- try another version of P96 (Aminet or previous IComp version)
- also verify that the ZZ9000 main board has good contacts with the smaller daughterboard.
Don't worry about the network part, is it no good anyway.
 
You mention this

Replica A4000D daughterboard.

What is the Replica Daughterboard you are using?

Also worth a look here

community.mnt.re

ZZ9000 No longer shows up in Kickstart Menu

No, I haven’t heard of this issue appearing over time. With old card models, it was happening sometimes from the beginning, which was mostly due to the card taking longer than Autoconfig to initialize. For USB, you’re right, the current docs are not great. Nowadays, the USB stick is seen like...
community.mnt.re community.mnt.re
 
Last edited:
Just to throw in here. I use a ZZ9000 on an Amiga 4000 (orignal motherboard) and don't have these issues. Almost looks like video ram issue on the ZZ perhaps? IDK.
 
When using P96 v2.0 (from Aminet) the ZZ9000 driver fails to load - can't seem to get this to work at all. If I re-install v3.6.2 the driver loads correctly.

I'm using the following daughterboard https://github.com/arananet/A4000db
I also have a Firebird PCI daughterboard and get more or less the same issues with the ZZ9000 installed and no PCI cards.

I've reformatted the ZZ9000 SD card and tried various different firmware images (no fast RAM etc) and also have the serial debug output on the ZZ9000 connected up and there isn't anything obviously wrong there.

Interestingly, if I try to use a PCI graphics card with the Firebird daughterboard and same version of P96 (no ZZ9000 installed) I get similar issues. I've tried a couple of different PCI cards and they both show similar graphical issues.

I've replaced U701 and U714 PAL's and confirmed they are programmed correctly so the next thing to do might be to try another Buster as the system works great unless the Zorro bus is involved. The erratic bus test results in ZZTop kind of point to a hardware problem, and results are worse in the more demanding Zorro 3 mode compared to Zorro 2.

In terms of physical connections with daughterboard and cards etc everything looks great as all of these parts are new - it's essentially a new system with the odd new old stock component.

Dale423, if you do a bus test with ZZTop does it show any errors?
 
So that seems to rule out the ZZ9000 and you seem to have a more general issue with the computer.

I have done a Bustest right now and it saus "Done. O errors".
Note that i am still in firmware 1.8 as i had no reason to upgrade a working system so far (i only want scanlines and they won't do them) and i have IComp P96 version from late 2000 (no upgrade either as it works fine).
 
Last edited:
Swapped over Buster and still getting the same results - so it's not that.

Also tried a older ZZ9000 firmware and that does not make much difference either.

I'll see if I can find an earlier version of P96.
 
From all you said, your problem doesn't seem related to the ZZ9000 or even your Edu's bridgeboard.
Since you have ruled out the Buster (better not insert and remove back and forth this one too much to avoid damaging the socket or the precious chip).
I guess your problem can only be located on your mobo, maybe your PSU, or software wise to this specific P96 version if you couldn't try another one.
 
I have a ZZ 9000 in my A2000 Blizzard2060 and getting it working was a real nightmare.

Pointer desapear, graphics corruption, And some really cool effects, like after a reset, the image that was previously on the screen appearing as the background, as if I had taken a screenshot, spectacular!!.

In my case Basically, three things conflicted: the 68060 library, the Picasso version, and even, for some things, the Kickstart version.
It's finally working, but it's truly maddening.

Last but not least, my accelerator doesn't like INT6; I have to set the ZZ9000 jumper to INT2 for the audio and network to work.

Good luck! 😥
 
@amigasquirrel Hi, I've got a similar setup but original Rev B PCB. Buster 11 is fitted.

I will have to check versions of software I am on.

HDMI cable can make a huge difference. Not seen it on my A4000, but saw lots of artifacts on my A1200 PiStorm. So definitely worth trying a different one. I have a short HDMI extension 5cm cable, going into a HDMI to DVI cable, can't remember length.

I am probably on latest Firmware as I like to keep upto date, and i think 1.13 has been out for years.

I use a modern Small Form Factor PC ATX supply.

I remember having one issue and i think it was related to the MMU setup and adding an memory mapping address to an ENV file? It's on their forum. But it was not causing graphical artifacts.

My system seems nice and stable. And I have a switch at the back for 50/100Mhz (not to be operated when system is powered on).
 
Thanks for all the replies, it seems that getting RTG working on any hardware is a bit of an adventure. :)

P96 Version:
I’ve tried v3.6.1 and v3.6.2 but can’t get v2.0 (from Aminet) to work as the ZZ9000 driver fails to load, I’m guessing I need to try an older ZZ9000 driver and firmware.

Power supply:
I’m using a modern SFX-L power supply, with the sense leads connected up. I’ve also tried an old AT power supply and it's still the same.
Supply voltages on the ZZ9000 itself are good, 5v input is 4.9v, and other rails on the card, 3.3v, 1.8v, 1.2v are all good and look clean on a scope.

Interrupt:
Changing from INT2 to INT6 it does not seem to make any difference.

HDMI Cable:
Changing the cable does not make any difference, the behaviour I’m seeing is something internal to the Amiga. If I open a window then for example use the scroll bars to cause glitches to appear, I can move the window around the screen the window will move to the new location with the glitches unaffected.

MMU:
I’m running MMUlib with recommended entries for the BFG9060, however I’ve eliminated this as an issue by booting from a clean install of OS 3.2.3.

Here's a video of it running with the 68030 CPU card on a clean OS 3.2.3 install with only P96 v3.6.2 and ZZ9000 drivers installed, nothing else.

With this setup at least the mouse pointer displays correctly and doesn't disappear. :LOL:
It crashed just after I tried to go through the menus in TextEditor, it’s almost guaranteed to crash if I do that.

The inconsistency makes me think it's an interference/signal integrity issue on one of the data paths, the behaviour is similar on each reboot but never the same.
 
What if you use other libs than MMUlibs? Phase 5 libs shall still work with the BFG card. Latest are available from a1k.org.
 
I contacted MNT support and they suggested to try another SD card first. After doing this the occasional yellow / recoverable error on startup from ZZ9000.card complaining it needs firmware (boot.bin) 1.13, so that problem appears to be solved.

They also suggested trying firmware without the autoboot/usb feature - this made a big difference.

Here are the results of the past few hours of testing:

To eliminate any other potential causes and make fault finding easier, I stripped the system back to a simpler configuration to run some more tests:
I removed the BFG9060 CPU card and replaced with 68EC030 CPU card at 24Mhz
Installed a fresh copy of OS v3.2.3 with only P96 v3.6.2 installed and ZZ9000 drivers v1.13 and no other software/drivers.

Results running firmware zz9000-firmware-1_13-zorro3
Here is a video of the original behaviour:

The machine crashed at the end of this video when trying to browse the menus in TextEditor – it always crashes in RTG modes if I do this.

Results using firmware zz9000-firmware-1_13-zorro3-noauto-nofast
This is a significant improvement; there are still issues with the left side of the display – see video:
ZZ9000_3a_68030_clean_install_zz9000-firmware-1_13-zorro3-noauto-nofast.mp4

The bus test consistently completes with no errors, which is the first time I’ve seen it do that.

I then restart the machine a few times, display looks the same but bustest fails – see video:
ZZ9000_3b_68030_clean_install_zz9000-firmware-1_13-zorro3-noauto-nofast.mp4


Results using firmware zz9000-firmware-1_13-zorro3-noauto
Seems to behave the same as zz9000-firmware-1_13-zorro3-noauto-nofast – see video:
ZZ9000_4a_68030_clean_install_zz9000-firmware-1_13-zorro3-noauto.mp4


And another test after a reboot – see video:
ZZ9000_4b_68030_clean_install_zz9000-firmware-1_13-zorro3-noauto.mp4


So, in summary using a firmware with the autoboot/usb feature is causing most of the issues.
I still have some display artefacts/flipped pixels using a noauto firmware, it is hard to see this in the videos so here are some photos:

640x480_32bit.jpg
https://drive.google.com/file/d/1ZdKHFcrzH3iUaY47PJ95mEEK6e3hs30-/view?usp=drive_link

800x600_32bit.jpg
https://drive.google.com/file/d/1KTX-pf1X-NwuzL5uQH4S4kYvZY7k7ZNx/view?usp=drive_link

1024x768_32bit.jpg
https://drive.google.com/file/d/1Nhz2ZgjyBTYuK48q80LTNHVudT6PCBRB/view?usp=drive_link

1024x768_32bit_closeup.jpg
https://drive.google.com/file/d/1jdAsB-eU8GjCb7s8c3DylMTdzsWNIMLk/view?usp=drive_link

Still not perfect, but much improved and the machine no longer crashes. :)
The screen mode and colour depth does not affect this so it does not seem to relate to how hard the card is being driven.
Have reported this back to MNT support so will see if they can suggest anything else.

I have to say that (when it works!) the display on the ZZ9000 is excellent and incredibly sharp when viewed on a good monitor, what with being all digital over HDMI.
 
You must log in or register to reply here.

Similar threads

arnljot
A1200 spare out of storage - has issues
Replies
2
Views
92
arnljot
arnljot
dougal
A1200 having issues with both PCMCIA/CF cards and SDBox
Replies
15
Views
176
softinn
S
Boing-ball
A4000TX Mainboard owners. Are there any pitfalls with these? Issues that need to be aware of?
Replies
4
Views
456
Boing-ball
Boing-ball
A
Amiga 1200 And ZZ9000 Tower setup
Replies
1
Views
415
amigat
A
sheepz
For Sale A4000 060 ZZ9000 and USB - All original parts included UK based
2
Replies
22
Views
3K
Lord_protector
L
Back
Top Bottom