🚀 go-pugleaf

For  example - is it indicative of a CPU failure or some chip  on 
the motherboard?
-- 
David Longley

In article <821071506snz@longley.demon.co.uk>, David@longley.demon.co.uk 
says...
>
>For  example - is it indicative of a CPU failure or some chip  on 
>the motherboard?
>-- 
>David Longley
>
>Looks that you have a DMA conflict,
>an easy way to find out is type MSD
>at the Dos prompt and look which DMA
>are in use. Maybe you may want to try
>to change the Network card DMA to an Available One.
>>>>>>>>>>>>>>>juliust@jaxnet.com<<<<<<<<<<<<<<<<<<<

David Longley (David@longley.demon.co.uk) spoke thusly:
: After  many  weeks  of configuring a  small  network,  everything 
: seemed to be in place until Friday. ** DMA TEST FAIL !! ***

: Next, when the system. just before the drives are identified,  we 
: got the message  ** DMA TEST FAIL !! ***

This is a BIOS error message, >likely< because the 8253 (or whatever the hell 
the number was) emulating glue logic on the mother is...

bad. It would appear, imho, that your motherboard is toast. Go get a new one. 

:(
--
Home: dremu@emf.net     find me on the PSQ homepage http://www.emf.net/~quam

Work: support@maxis.com     Maxis Software Toys: The SimCity People
Maxis -=- 2121 N California Blvd, Suite 600 -=- Walnut Creek CA 94596-3572
Voice  510.927.3521   FAX 510.927.3581   BBS 510.927.3910

I remember the letter wrinkled in my hand    "I love you always" filled my eyes
I remember a night we walked along the Seine                Riding on the Metro

I remember a feeling coming over me        The soldier turned, then looked away
I remember hating you for loving me                         Riding on the Metro

and of course, the views expressed herein are those solely of the individual
and in no way are meant to express those of Maxis or any of its subsidiaries

David Longley <David@longley.demon.co.uk> wrote:

>After  many  weeks  of configuring a  small  network,  everything 
>seemed to be in place until Friday. ** DMA TEST FAIL !! ***
>
>System 486DX-100
>       heatsink+Fan
>
>Motherboard PCI Bus Mainboard (48 and a BIG 6 on the manual)
>
>The IDE is disabled and I'm running two SCSI hard disks of a  BUS 
>LOGIC  BT545S. All was working fine until we tried a new  network 
>adapter  and  at  one stage of configuring  Windows,  the  SYSTEM 
>directory was scrambled and we got gobbledegook on the screen.
>
>Next, when the system. just before the drives are identified,  we 
>got the message  ** DMA TEST FAIL !! ***
>
>Pulling  the network card and trying again, the system seemed  to 
>work fine till now. Now I can;t get rid of the message at all.
>
>I have tried swapping the SCSI card for another, a Future  Domain 
>16 bit card and with that card the system just stops at the  SCSI 
>ROM signature. The BUS LOGIC card works fine on another system.
>
>Question 1 - where is this message coming from?
>
>             a) CPU
>             b) Motherboard
>
>What can I do about it?

It's a motherboard message, most likely.  Make sure the card is firmly 
seated in the VLB slot, and that bus-mastering is enabled for that slot.

Remove the network card until you get the system up and running again, 
and then make sure that the two cards are not trying to use the same 
resources - IRQ, I/O address, etc.

Jo Kellner
BusLogic Technical Support
(408) 654-0760

In article <4crdvn$fhr@multimaster.buslogic.com>
           techsup@buslogic.com "Jo Kellner" writes:

> David Longley <David@longley.demon.co.uk> wrote:
> 
> >After  many  weeks  of configuring a  small  network,  everything 
> >seemed to be in place until Friday. ** DMA TEST FAIL !! ***
> >
> >System 486DX-100
> >       heatsink+Fan
> >
> >Motherboard PCI Bus Mainboard (48 and a BIG 6 on the manual)
> >
> >The IDE is disabled and I'm running two SCSI hard disks of a  BUS 
> >LOGIC  BT545S. All was working fine until we tried a new  network 
> >adapter  and  at  one stage of configuring  Windows,  the  SYSTEM 
> >directory was scrambled and we got gobbledegook on the screen.
> >
> >Next, when the system. just before the drives are identified,  we 
> >got the message  ** DMA TEST FAIL !! ***
> >
> >Pulling  the network card and trying again, the system seemed  to 
> >work fine till now. Now I can;t get rid of the message at all.
> >
> >I have tried swapping the SCSI card for another, a Future  Domain 
> >16 bit card and with that card the system just stops at the  SCSI 
> >ROM signature. The BUS LOGIC card works fine on another system.
> >
> >Question 1 - where is this message coming from?
> >
> >             a) CPU
> >             b) Motherboard
> >
> >What can I do about it?
> 
> It's a motherboard message, most likely.  Make sure the card is firmly 
> seated in the VLB slot, and that bus-mastering is enabled for that slot.
> 
> Remove the network card until you get the system up and running again, 
> and then make sure that the two cards are not trying to use the same 
> resources - IRQ, I/O address, etc.
> 
> Jo Kellner
> BusLogic Technical Support
> (408) 654-0760
> 
> 
Please note: The motherboard is a 3 PCI and 4 ISA slot board.

a. The Buslogic SCSI-2 card is in the second ISA slot (not a VLB)
b.  If  I put an IDE hard disk in the system, after the  

                     !!!  DMA TEST FAIL !!!

The drive boots (bit I see no SCSI devices of course).

c.  At  this  point  tye ONLY cards in the  system  are  the  Oak 
Technology VGA (256K) card and the BUS LOGIC BT545S card.

d. If I run PC TOOLS SI or Checkit. the motherboard passes as doe 
sthe CPU. The DMA controller passes both.


One  other oddityt is that the characters go blue sometimes,  but 
not  all  the screen. Although there is no CD ROM on  the  system 
now, this line triggers the colour change:

C:\SBPRO\SBP-SET /M:12 /VOC:12 /CD:12 /FM:12 /LINE:12

I'm  very  worried  about the DMA failure  and  no  SCSI  devices 
appearing. I have tried clearing CMOS, that makes no difference.

What should I do?
-- 
David Longley

Just in case anyone else comes across this peculiar problem...

System Board PCI Bus Mainboard 486
       3 PCI
       4 ISA
       Supports P24T, DX100
       Write Back Cache

Intel DX-100
16MB 60ns non parity RAM
Bus Logic 545S SCSI card with default settings
Oak Technology VGA Card (256K)
IBM 540 MB SCSI HD
SEAGATE 400 MB SCSI HD
1.4MB Floppy

The  *** DMA TEST FAIL !! *** message appeared just at the  point 
that the SCSI hard was about to show what was on the SCSI bus.

I  ended  up  clearing CMOS, which did nothing. I  ran  PC  TOOLS 
SI.EXE  and  tested  the  DMA controller  and  CPU  and  rest  of 
motherboard.  All  passed.  Then tried  the  VGA.  Something  odd 
happened  on  the colour bar display, almost as  if  the  display 
wasn't  synchronising. So, I tried another video card,  a  Compaq 
ISA VGA card (1988). On power on, there were two slow bleeps, but 
no DMA fail message. When I put the OAK TECHNOLOGY card back into 
the  last ISA slot, I just got a white grid on boot up.  Changing 
the card to the next slot, and there was no DMA FAIL message.

Now  it  *may*  be  dirty  ISA slots I  guess.  It  *may*  be  an 
intermittent  fault  in the DMA controller, or maybe  some  other 
intermittent  motherboard problem. Anyone ever seen this sort  of 
thing before? Why should a video card contribute to a DMA failure 
at the point the SCSI card was scanning its bus?

-- 
David Longley

The problem came back. 

I  put  the  Oak card back in, and after  2  hours,  Windows  for 
Workgroups had crashed. Looking at the had disk, several files in 
the WINDOWS directory had been corrupted e.g.

PROTOCOL.INI  ---> PRO\OCOL.INI

And some other directories too. This happened the first time too.

I put a COMPAQ 386/20s 256K VGA card in its place and am  leaving 
that   running  PC  TOOLS  diagnostics  on  the  CPU,  RAM,   and 
Motherboard..

Anyone seen this before?
-- 
David Longley

On Tue, 09 Jan 96 10:14:57 GMT, David Longley
<David@longley.demon.co.uk> allegedly wrote:


>Anyone seen this before?

Yes, I have had a <lot> of trouble with early Oak cards and
drivers.Bin it and try another card brand.I always thought <I> had my
share of problems with hardware.You seem to be making my life seem
plain sailing , at the moment. <g>
                   Regards, 
                           Chris Wilson.
                (Gatesgarth Racing Developments)
                          Gatley,Cheshire,England.

In article <30f922f0.19386626@nuntius.u-net.net>
           chris@gategart.u-net.com "Chris Wilson" writes:

> On Tue, 09 Jan 96 10:14:57 GMT, David Longley
> <David@longley.demon.co.uk> allegedly wrote:
> 
> 
> >Anyone seen this before?
> 
> Yes, I have had a <lot> of trouble with early Oak cards and
> drivers.Bin it and try another card brand.I always thought <I> had my
> share of problems with hardware.You seem to be making my life seem
> plain sailing , at the moment. <g>
>                    Regards, 
>                            Chris Wilson.
>                 (Gatesgarth Racing Developments)
>                           Gatley,Cheshire,England.
> 
Well,  I have been using a fair mixture of systems over the  past 
three months:

386/20DX ISA
386/25SX ISA
2 x P120 Endeavours
P133 Endeavour
486DX4-100 Mainboard PCI Bus 486
Ambra DX2-66
P24T-S VLB
486DX-33 VLB
486SX-25 VLB

The main problem has been the DX100- PCI Mainboard 486.

It's  now got a Compaq 386/20s 256K VGA card in as a Stealth  PCI 
2MB  VRAM  produces  *nothing* on the  screen,  not  even  setup. 

Clearing  CMS  and specifying that VGA is in PCI  just  does  not 
help.  Going back to the Compaq ISA card and putting a  PCI  SCSI 
card  in  the machine (AHA2940) just results in  the  SCSI  light 
staying on....

Has  anyone else had trouble with video cards in PCI  slots  this 
way?


-- 
David Longley