Welcome, Guest. Please login or register.

Author Topic: ZIII RTG lockup in A3000 - solved!!!  (Read 2267 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline joeksterTopic starter

  • Full Member
  • ***
  • Join Date: May 2007
  • Posts: 203
    • Show only replies by joekster
ZIII RTG lockup in A3000 - solved!!!
« on: February 24, 2009, 02:22:31 PM »
I have an a3000/040 running os 3.9. I am trying to get either a gvp spectrum 24/28 or a phase5 cybervision 3d working with picasso96. All seems to work, except that the system will lockup after 1-10 minutes while running p96speed. With the gvp card, if I set it to ZII mode, everything will run perfectly for hours... Because of this, I suspect that the issue is with my buster 7. Is it possible to run either of the cards in ZIII mode, but disable DMA (some config file perhaps)? If not, is there any way to run the cv3d in ZII mode? Could the issue be something other than my buster revision?

Thanks.
 

Offline Crumb

  • Hero Member
  • *****
  • Join Date: Mar 2002
  • Posts: 1786
  • Country: 00
    • Show only replies by Crumb
    • http://cuaz.sourceforge.net
Re: ZIII RTG lockup in A3000 - buster issue???
« Reply #1 on: February 24, 2009, 02:29:18 PM »
If you use a 3640 you may have lots of problems. Check out that your A3640 is the version 3.1. Upgrading to Buster11 will be useful to make your system work better with the 3640 but gfx cards don't use DMA.
The only spanish amiga news web page/club: Club de Usuarios de Amiga de Zaragoza (CUAZ)
 

Offline joeksterTopic starter

  • Full Member
  • ***
  • Join Date: May 2007
  • Posts: 203
    • Show only replies by joekster
Re: ZIII RTG lockup in A3000 - buster issue???
« Reply #2 on: February 24, 2009, 02:30:57 PM »
I'm using a pp&s mercury 040. Way better than the 3640!
 

Offline countzero

  • Hero Member
  • *****
  • Join Date: Mar 2005
  • Posts: 1938
    • Show only replies by countzero
    • http://blog.coze.org
Re: ZIII RTG lockup in A3000 - buster issue???
« Reply #3 on: February 24, 2009, 02:36:00 PM »
buster 7 ? I thought buster 7 doesn't even support Z3. you have to get 11 or 9 at least.
I believe in mt. Fuji
 

Offline jlariv8957

  • Full Member
  • ***
  • Join Date: Feb 2005
  • Posts: 112
    • Show only replies by jlariv8957
Re: ZIII RTG lockup in A3000 - buster issue???
« Reply #4 on: February 24, 2009, 02:41:23 PM »
Buster 7 support ZIII but have serious bug(s) that didn't allow really using ZIII cards (didn't remember what kind of problems).
 

Offline Crumb

  • Hero Member
  • *****
  • Join Date: Mar 2002
  • Posts: 1786
  • Country: 00
    • Show only replies by Crumb
    • http://cuaz.sourceforge.net
Re: ZIII RTG lockup in A3000 - buster issue???
« Reply #5 on: February 24, 2009, 04:57:01 PM »
@joekster

Buster11 will always be a nice addition. Pure Z3-only CV64 is nicer for an A3000 but cv3d is OK too.

BTW, have you checked that your A3000 isn't overheating? have you tried running the tests with the case open?

IIRC CybergraphX allows running the CV3d in ZorroII mode using the right monitor.
The only spanish amiga news web page/club: Club de Usuarios de Amiga de Zaragoza (CUAZ)
 

Offline joeksterTopic starter

  • Full Member
  • ***
  • Join Date: May 2007
  • Posts: 203
    • Show only replies by joekster
Re: ZIII RTG lockup in A3000 - buster issue???
« Reply #6 on: February 24, 2009, 06:50:06 PM »
I'm pretty sure that the a3000 is not overheating as when it is in ZIII mode, it will hang in less than 10min, but simply changing the jumper to ZII mode allows it to run for > 2hours.

Is there an authoritative answer to the minimum buster level for these cards? I checked the manuals and they "reccomend" a buster-11, whatever "reccomend" means...

I tried going to the CybergraphX web site to try a demo, but the site appears to be down... Are they still around? is the software still commercial?
 

Offline Zac67

  • Hero Member
  • *****
  • Join Date: Nov 2004
  • Posts: 2890
    • Show only replies by Zac67
Re: ZIII RTG lockup in A3000 - buster issue???
« Reply #7 on: February 24, 2009, 08:27:53 PM »
My A3k has been running a Merlin in Z III mode for 14+ years. No problems apart from some glitches the Merlin has.

A Buster 7 works AOK most of the time, at least if there's only one Z3 card active. What it simply can't provide is Z3 DMA - you'd need a Buster 11 for that (for a single DMA-able card), but from the Deneb FAQ I've learned that an 11 doesn't even help with a stock A3000 since the onboard SCSI already counts as Z3 DMA. So you'd need SCSI on your accelerator card to run one Z3 DMA card.
Further reading: http://www.thule.no/haynie/chips/buster/docs/buster2.pdf

However, RTG cards are non-DMA so it's a completely different story. What board rev is yours? There might be heat and timing related problems with some board and PAL revs. If you have PAL programming facilities at hand, some of the PAL codes are available for testing: http://www.thule.no/haynie/systems/amiga3k/pals/
 

Offline joeksterTopic starter

  • Full Member
  • ***
  • Join Date: May 2007
  • Posts: 203
    • Show only replies by joekster
Re: ZIII RTG lockup in A3000 - buster issue???
« Reply #8 on: February 24, 2009, 11:44:12 PM »
My A3000 is rev 8.9, 35th week of 1991. It origionally came with a buster-06, so maybe it isn't such a late model. I wouldn't mind replacing the socketed pal's, but I have no way of programming them...
 

Offline Zac67

  • Hero Member
  • *****
  • Join Date: Nov 2004
  • Posts: 2890
    • Show only replies by Zac67
Re: ZIII RTG lockup in A3000 - buster issue???
« Reply #9 on: February 25, 2009, 06:20:26 PM »
Quote
Rev 8.9         ****    Shipped to end customers.    ****      
         This rev is just like the rev 9.0 artwork, only some    
         small changes were made. This rev is actually the      
         rev 8 inner layers combined with the rev 9 outer        
         layers. This allowed rom towers to be phased out        
         sooner than otherwise possible.                        
                                                                 
         No rework required.                                    
                                                                 

Sounds pretty good actually...
 

Offline joeksterTopic starter

  • Full Member
  • ***
  • Join Date: May 2007
  • Posts: 203
    • Show only replies by joekster
Re: ZIII RTG lockup in A3000 - solved!!!
« Reply #10 on: February 26, 2009, 02:31:58 AM »
Well, I removed my mercury accelerator and everything worked fine in ZIII mode. So, I started playing with the jumper settings. When I put a jumper on D1, all is good. Looking at the manual it states "DO NOT JUMPER" as the description for D1. Maybe someone has more detailed info about the opperation of this card??? I did notice a 7.6% decrease in memory speed when jumpered... Not such a big deal for me as with the internal cache of the 040 the effective performance penalty is about 1-2%, well worth having ZIII opperation!

Maybe it deserves it's own thread as the mercury is one tricky accelerator?
 

Offline matthey

  • Hero Member
  • *****
  • Join Date: Aug 2007
  • Posts: 1294
    • Show only replies by matthey
Re: ZIII RTG lockup in A3000 - solved!!!
« Reply #11 on: February 26, 2009, 11:53:46 PM »
It's not good to go around setting jumpers that you don't know what they do. Do you have static column zips? If so you should note from http://amiga.resource.cx/exp/mercury ...

"if the motherboard is fitted with static column fast RAM the first chip must be replaced with the supplied page mode ZIP DRAM chip - this solves a bug in Ramsey and allows the board to do burst RAM access"

A newer Ramsey may allow you to get around this. A lot of people say the DMAC has to be upgraded at the same time but I haven't seen problems. The newer Ramsey is available but the newer DMAC's are hard to find and expensive. I have also seen some accelerators have problems with the motherboard SCSI with the newer Ramsey and DMAC combo. Putting the DRAM chip in may slow down motherboard memory significantly as static column memory is faster. This is still way better than slowing down the accelerator memory though.