picasso iv card and os 3.1.4

[dschallock]

There seems to be something in the startup-sequence of OS 3.1.4 that is conflicting with the video signal pass through of the picasso iv.  When the picasso iv is plugged into the zorro and video slot of the A3000 and A4000 the pass through works immediately (even before any software is installed) unless... (per the manual) you have other graphic card software installed or have additional monitors in the devs: folder beyond the NTSC and PAL monitor files.
I have a new 3.1.4 ROMS in my 3000T and a fresh install of the operating system.  Plug in the picasso iv and the pass through is not working.  However, if I hold down both mouse buttons to initiate the early startup menu, that screen will pass through the picasso iv just fine, and if I select boot with no start up sequence it will boot and the pass through works.  So the picasso iv is working fine.  Something in the standard 3.1.4 startup-sequence or user-startup is conflicting and I can't figure out what it is.
Although P96 is not required to make it work, I decided to install it anyway just to see if that changed anything... and it didn't.
Does anyone have any experience with this conflict and know a solution?
Thanks!

[beller]

Have you tried commenting out the "If exists DEVS:Monitors" section of the Startup-sequence?  Only thing I see in startup that might hang your card.  I'm using an Indivision and I've commented out the section without any problems....hope this helps!

[Gulliver]

For what it is worth, I prepared a prebuilt AmigaOS 3.1.4 for one of our testers that had an A4000D with a Picasso IV.

I did nothing special besides what I would consider a normal OS deployment, and the tester only had to configure the card screen modes, but everything else worked out of the box.

Maybe in your case it is some third party program that is doing some funky stuff and getting in the way.

PS: The OS when it installs does not create a user-startup file. That one is generated by the user or applications.

[dschallock]

Interesting ideas and thoughts.... I appreciate them.

There is one piece of software I did install after the OS that perhaps replaced a library or something that could be creating a conflict.  The 3000T had a progressive peripherals Mercury 040 accelerator in it.  After installing 3.1.4 I got the normal errors telling me to install the proper 68040 library.  I ran the install disk from the PPS for the mercury card.  It worked in terms of bringing to 040 and its on board memry to life, but I don't really know what it installed.  I think I should look into that.

Thank you!

[guest11527]

  It worked in terms of bringing to 040 and its on board memry to life, but I don't really know what it installed.  I think I should look into that.

While not related to the P-IV problem - there is most likely something missing in the user-startup - you may want to upgrade to the MMUlib-driven 68040.library which is still maintained:

http://aminet.net/package/util/libs/MMULib

[dschallock]

Thanks Thomas.

Well, I figured out the problem isn't directly between the picasso iv and 3.1.4  I have isolated it to the 2 libraries that are installed (and required to work) for the Progressive Peripherals Mercury 040 accelerator card.  There are 2 libraries (68040.library and pps040.library) and a command (init040 that is placed in the startup-sequence), without those 2 libraries and the command the accelerator doesn't work and the 040 isn't recognized.  The moment they are in place the pass through of the picasso iv stops working.  Not sure how I could fix that...

[guest11527]

Well, I figured out the problem isn't directly between the picasso iv and 3.1.4  I have isolated it to the 2 libraries that are installed (and required to work) for the Progressive Peripherals Mercury 040 accelerator card.  There are 2 libraries (68040.library and pps040.library) and a command (init040 that is placed in the startup-sequence), without those 2 libraries and the command the accelerator doesn't work and the 040 isn't recognized.  The moment they are in place the pass through of the picasso iv stops working.  Not sure how I could fix that...

This looks like a very bizarre way how to enable a 68040. Well, anyhow, if it would work, I would say "go for it", but apparently, it does not. One problem might be the inclusion and mapping of the P-IV board which is probably not taken into account by this strange init sequence. It might be that this came to be before SetPatch was updated to load the 68040.library itself, but who knows...

As said, what you *may* want to try is to remove (or rename) these two custom libraries, remove "init040" (for whatever it does) and install the MMULib. Maybe then things work, or at least, I would be more able to help you... (-;

[dschallock]

Thomas,
Thank you.  Your MMU libraries solved it!  I should have done that from the beginning but I honestly didn't know about the distribution.  Why isn't that package part of 3.1.4 out of curiosity?

Anyway... I thank you Thomas very much for your code, and your help.

TO SUM UP THIS ISSUE FOR POSTERITY:

There is not an inherent conflict between OS 3.1.4 and the picasso iv and it's pass through of the video signal.  There is a conflict with Progressive Peripheral's Mercury accelerator's 68040.library and the pps040.library when installed in 3.1.4 and the functionality of the picasso iv's functionality.
Instead of installing PPS Mercury install software, all one needs to do is to: 1) physically install the accelerator card. 2) install OS 3.1.4 ROMS and OS 3) install Thomas Richter's MMUlib package found on Aminet http://aminet.net/package/util/libs/MMULib  That package includes a 68040.library that allows the accelerator to work seemingly perfect and the picasso iv video signal works perfectly as well.

Thanks again everyone.
Dan

[guest11527]

Why isn't that package part of 3.1.4 out of curiosity?

Glad to have helped. To answer your question: There are multiple reasons: First, the package does not work with *all* accelerators. Some P5 boards require their own library, and then would not work entirely with this library, nobody knows why as P5 refused to document what they did. Hence, it makes it a bad choice to ship it by default and install it by default, as it may leave customers with non-working installations.

Second, I don't want to drag my own software into this legal dispute of Hyperion vs. Cloanto. It's easiest this way: It's my own copyright.

[dschallock]

I wanted to add an update to this thread in case anyone using similar hardware in the future is reading this:

Although using Thomas' wondering MMUlib install did indeed get the (Amiga 3000 tower - Picasso IV - OS 3.1.4) combo up and running, I kept noticing that many pieces of software I was installing would break the system and I would have to restore the system back.  This was happening pretty frequently and I was pretty sure the problem wasn't the software, the OS, or the picasso IV.

I am lucky in that I own another powerful accelerator, so I decided to swap that PPS Mercury out and install a Cyberstorm MK ii 060 board and do a fresh install of 3.1.4 and start over.

Immediately every and all issues disappeared.  My x-surf/rapid road combo was also not working with the Mercury setup but with the cyberstorm 060 it works perfectly.  The picasso IV is working perfectly and so is every other piece of software I install on top.

One more oddity was with the Mercury installed when I loaded the workbench prefs in 3.1.4 and added a check mark to the "set MWB colors" option it would then cause my system to crash (flashing green power light) whenever I clicked on the CD0: device in Dopus... very strange, I know.  But once I installed the cyberstorm 060 on a fresh install that option in workbench prefs no longer crashed anything.

Just wanted to add this update.  So, with the help of Thomas' MMULib it seems possible to get things working, but there is alot of incompatibilities still present.  If anyone has a 3000 machine and wants to use PPS Mercury accelerator you might have to settle for a much earlier OS release to get intended stability.  Like maybe 2.0?
 

[paul1981]

Out of curiosity, was your Mercury 040 card working correctly with your A3000 before you upgraded to 3.1.4? I'm just wondering whether the card is faulty.