Troubleshooting an unfamiliar machine ...

[macto]

So I'm new to the Amiga.  My new toy is an Amiga 2000 with a GVP Combo 030 and SupraRAM 2000 card.  Until recently I thought that the "random" crashes were the product of a user who is unfamiliar with a system pushing it in unusual ways.  I'm beginning to think otherwise.

But the question is, how do I diagnose the problem?  I'm not very familiar with the machine, so just pulling the GVP card and hoping that the problem will go away is not of much use because I wouldn't know how to use the machine in order to reproduce the problem.  So can somebody suggest something which would use a significant chunk of the machines resources when I pull the card.  The software would have to fit on two 720 kB MS-DOS formatted floppies since removing the controller means removing my hard drive.

In terms of software, I have Workbench 1.3 and 2.0.4 (I have Amiga OS 3.9, but it sounds like it needs a 68020 which would disappear when I pull the GVP card).  The Amiga has a 3.1 ROM.  Unfortunately, I don't Workbench 3.1.

Thank-you for any advice.

[QuikSanz]

Hi,

How old and large (small) is your HD? If all I mentioned before had no effect, try to reseat all the chips on the MOBO. Do you get any error messages?

Chris

Click on my name to see whats in my system. As of last sunday I will either have to repair or replace my PSU.
I'm ready for more WATTS.

[QuikSanz]

Hi,

Almost forgot, Have you installed both Boing Bags? With only 1Mb of chip make sure in Prefs/WBPrefs/ Graphics mem = other, not chip.

Chris

[Lemmink]

In general it is no Problem running older WBs on a newer Kickstart. Problem comes when a programm that is functional under 2.x but also supports additional features of 3.x detect a 3.1 ROM. It could assume that the WB on top is a least 3.1 too and it starts to pull Tricks meant for 3.1 and up wich a 2.x WB can`t handle.
It will sure be helpfull if you spezify the crashes you accounter a little bit: When does it happen, what says the error message (or is it just a GURU), what programm crashed asf.

[Noster]

Hi

I must agree Lemmink. You should use a Workbench according to your Kickstart ROMs, i.e. try to install 3.9 and see if the problems are still present.
And please give detailed informations about the errors occuring.

Noster

[Castellen]

Hardware wise, you can reduce the machine to the most simple usable state possible to eliminate expansions causing the problem.

For software, remove all non standard workbench programs (like MCP, MagicMenu, etc) if you're using any and run workbench in it's most basic form.  If you still have problems, it could be hardware.

Find out if the crashes are random, or only happen after a specific sequence of events, or after a certain time.

As mentioned before, reseat all socketed devices on the motherboard.  Remove essential expansions like the CPU board and clean edge connectors and sockets with proper gold contact cleaning spray.

Monitor power supply 5V rails for dropouts or excessive noise.

Flex the motherboard while it's running and see if light mechanical stress causes problems (usually hilighting contact problems).

If possible, try swapping memory modules for known good ones, or continuously run MemCheck (from Aminet) for a few hours to soak test all memory addresses.

Dunno, could be lots of things...  

[macto]

QuikSanz: Reseating the chips seemed to do the job, but I'm going to wait a few days before reinstalling the SIMMs in the GVP card and SupraRAM card.  This is just to be sure, seeming as the errors are semi-random.

One curious thing though.  The chip at U301 (an 8520A) had pin 40 bent and pin 20 looked odd.  I noticed this when I pulled it out, so I presume that the previous owner left it like that.  Should I be replacing the UART?  I haven't been having trouble with the serial port thus far.

Oh, and I'm sure your A2000 will look much more impressive when I know what half of that stuff is.  

On the other comments: I was running a clean install of OS 3.9 and had crashes.  Ditto for 2.0.4, but they didn't seem as common.  I even saw errors with 1.3, which was available on the SupraRAM and GVP diagnostic disks.

As for the nature of the errors, the system would run fine for a while, then any number of things can happen.  Sometimes a guru style error will pop up (8000000n and 8100000n).  Some times the mouse would be frozen and the keyboard unresponsive, including .  Sometimes there would be corruption of the image on the screen, may that be random or a blank block which just floated there.

A broad range of programs would cause or see problems.  The Fonts control panel in OS 3.9 was a consistent source of errors.  The GVP RAM test was another.  In both cases, the machine would have to be running for half an hour or so before things went nuts.  Formatting a partition or even copying a large number of files could cause problems.  Demos would have problems, but I presume that is normal.

Castellen: it sounds like you have a number of good suggestions, but I don't have the equipment to look for noise on the power supply, or replacement RAM modules for the GVP.

Thank-you all.  Hopefully reseating the chips solve the problem.  If not, there is a more complete description of the problem.

[Noster]

Hi

@macto

> The chip at U301 (an 8520A) had pin 40 bent and pin 20 looked odd.
> I noticed this when I pulled it out, so I presume that the previous owner left it like that.
> Should I be replacing the UART? I haven't been having trouble with the serial port thus far.

OF CAUSE REPLACE IT, this might be the defect of your system that leads to the crashes!

The 8520 do not only serve the parallel and serial ports, they also contain the systems clocks, several control-lines for the floppy-disk drives and other stuff.
They also generate interrupts! If any of these CIAs is bad, the whole system may be disturbed.

Noster

[macto]

Okay, the following site labels the pin 40 on u301 (CIA-B) as PPOUT.  Any idea what this means?  What ever it is connected to should be floating at this point.

http://www.amiga-stuff.com/hardware/8520.html

I think I have a few of these lying around from my previous attempts to get an Amiga.  Here's hoping that the CIA wasn't the problem in those cases.  (I presume these are fairly interchangable.)

[Noster]

Hi

> Okay, the following site labels the pin 40 on u301 (CIA-B) as PPOUT.

You mean POUT ? It is connected to the paper-out signal-pin of the parallel interface

Ok, if that signal is missing that should not lead to your problems, but the CIAs are very "unstable", i.e. they are IC's that die very often in an Amiga (have already replaced three of them) so you should replace it nevertheless.

> I presume these are fairly interchangable.

Yes they are. As Paula they have never been changed really and are interchangeable.

Noster

[Castellen]

Pin 40 isn't too major, but pin 20 is the 5V supply to the CIA if I remember correctly.
If this fails, then you will have major problems.  The system will not function without both CIAs working correctly.

You can get replacement parts from an A500 or A3000 as well.  If you use small pliers to straighten the pin, it'll probably be OK.  Remember to use anti static precautions.

I'd be more concerned about those cheap arsed sockets C= decided to use.  If the device pulls out quite easily, it suggests the socket contacts are weak, and possibly intermittent.

You can remove the socket and solder the IC directly into the board, or replace the socket with a machine pin type, which are far more reliable than those dual wipe ones.

As a last ditch effort, you can solder the ICs into the exsisting sockets.  Just flow solder between the IC legs and socket contacts.

[Morley]

One thing that can cause occasional guru's, are viri...

They are so rare in the Amiga world now, that most people don't bother thinking about them at all. If you have tested lots of old disks on your A2000, chances are one or several of them were infected and you transferred the virus to your hdd. As most viri are rather out-of-date, most don't work at all or they cause major lock-ups, such as guru's.

Try a decent virus-scanner to check this out before you start resoldering all your IC's  :-)

[macto]

On virii: if you can suggest a way to create a clean boot floppy for OS 3.9, I'll bypass the virus scan for now and do another clean install.  The situation over here is that I have three versions of the OS, one of which is on CD-ROM; a batch of driver/diagnostic diskettes, which I can download fresh copies of; and a few things from Aminet, which I can download fresh copies of.

I replaced the CIA with a pair from an Amiga 1000.  There was no change in the condition of the machine.  At this point, I'm going to pull out the GVP card to see if that critter is causing all of the trouble.  Then I'll consider mainboard problems, and replacing the sockets if need be.  Otherwise I risk damaging something which isn't broken.  I dug around my box of Amiga goodies and found a DataFlyer which will give me a hard drive to play with.  Reducing the machine to an 8 MHz 68000 may be painful, but I'll just keep reminding myself that it is only to figure out where the problem is.

Again, thank-you for all of the advice.  Hopefully I can get to the bottom of this soon.

[QuikSanz]

@ macto,

Do you have the docs for that accel card? You may want to check jumper settings.If not let me know and I'll check mine, I've kept it as a spare.

Chris

[macto]

I have two sets of documentation, which conflict in a couple of cases!  Okay, the second is just a sheet detailing the jumpers for a revision 3 board.  Does a part number 500025-03 suggest a revision 3 to you?  It does to me.

The differences between the board and the sheet are: CN15 (was on 2-3, should be on 1-2, listed as reserved, now on 1-2), J12 (was open, should be shorted, selects between autoconfig and extended RAM, left open), CN8 (was on 2-3, should be open (!), listed as reserved, is now open).  There is another thing worth noting, the card has a 68030 rather than a 68EC030.

My A2000 currently has the DataFlyer IDE and DataFlyer RAM.   WB 2.0.4 is being installed.  Hopefully that'll tell me whether it is the mainboard or GVP board which is causing the troubles.  At this point I don't know whether to hope that the 2000 or the GVP board is toast.  The 2000 would probably be easier to fix, but a 2000 is still a 2000 without the 68030.