Half-dead A4000D, please help!
Christian Link
I wouldn't normally post such a long request for help, but this
friggin' machine is slowly driving me nuts, and before wasting much
money for spare-parts that may turn out not to be what I really
needed, I thought I'd rather ask first, in the hope that some old
AMIGA guru may come up with a useful idea.
The system:
A4000D, Motherboard Revision B (same for the Zorro slot
daughterboard), Super Buster 11, 4*4 MB FastRAM, 1*2 MB ChipRAM, A3640
Rev. 3.1 CPU card (the A3630 shows the same symptoms I'm going to
describe, so don't worry about that). Kickstart 3.1.
Two Caviar 540 MB IDE harddisks, two Chinon FZ357A HD floppies, one
single Zorro card (Buddha IDE controller, although unused at the
moment).
Power Supply: SkyNet, 145 W
Here are the symptoms:
When powered on, it very often takes ages for the harddisk(s) to
start. Once they do, their usual reset-calibration (or whatever this
"dak-dak-dak-dak-rrrrrrrsssst!" is :-) ...) is interrupted many times
by what I take for continuous resets from the computer (the power LED
goes darker, then lightens again).
Even if the harddisks "survive" the reset, all they do is start
booting and freezing after some time. Mostly after a couple seconds,
sometimes only after five minutes. This "freezing" usually looks like
a WB3.1-greyish, empty screen, staying there forever. If you do a cold
boot, the same procedure starts again, only that the chances of an
actual boot process to start are much lowere :-( .
When I enter the Early Startup Menu and move the mouse around, there's
a very high chance that it will freeze.
Occasionally, all I get is an "error color" screen. Sometimes green,
usually yellow.
So much for the symptoms.
Here's how it actually happened
All this started some three or four weeks ago, after many days of
flawless operation. Heck, I haven't even touched the machine for hours
(Just left it on the workbench screen with hardly any program running
in the background!) when it happened.
For me, this looked like a problem with the power supply. So I
desperately looked for a spare supply, and like a miracle, I've found
one on Ebay, and at a ridiculously low price at that.
Well, I installed this new power supply, fired up the machine (which
had been in the closet for one or two weeks), and pronto: It worked!!!
...For something like 15 minutes or so :-( . Then the screen went
grey, and after a cold boot it stayed black, and the usual harddrive
stutter started again. I really got fed-up with this machine, so I put
it in the closet again just in time not to decide to throw it out of
the window, and used my trusty old A2000 for a couple days.
One night, when I felt adventurous, I took the 4000 out again. My
actual plan was to check the underside of the motherboard, i.e.
whether or not there was a short with the computer case. Before I did,
I gave this machine one more try, and couldn't believe my eyes when it
ran absolutely beautifully!
One hour, two hours, eight hours... Well, I thought that maybe the
zorro daughterboard had shorted a couple pins with the harddisk bay.
Indeed the insulating plastic mat did show some damages. So I removed
the board and cut all those solder points down to reasonable sizes.
Closed the machine, installed it under the monitor, let it run again.
Every day... For a whole week... Until, two days ago: BANG! Grey
screen, cold boot, harddisk stutter - well, the usual story :-((( .
And from that point on, the machine never ran for longer than
approximately 5 minutes :-( .
Finally, this is what I've tried and checked (everything without
success):
- Made sure that the insulating plastic under the mainboard was
intact,
- Resoldered the pins of the mainboard power plug,
- Removed and reseated the Super Buster, which, as you certainly know,
the only socketed IC on that damn board :-(,
- Swapped the Power Supply (Remember, now I have two of them!),
- Divided all power consuming devices (boards, harddisks, floppies...)
on both power supplies, so each of them had less stress,
- Tried the mainboard sans everything except for the processor board
and the video plug for visual output, outside the case, hanging in
mid-air, so no shorts against any case could occur,
- tried the mainboard without the Zorro daughterboard,
- tried the mainboard without *any* device connected to it.
Well, that's about all of it.
What I did notice though, is that on both days where the machine
decided to give up, it was comparably hot (35 degrees Celsius in my
room), and in the later case, it had worked for hours, using the
harddisks a lot, etc. But: If it really was a thermal problem, it
would run now, because I've put it into a much cooler room. I doubt if
I have destroyed something irreversibly, as I've seen the same things
happening a couple weeks before, and after that, it *did* run again
(for a week). So it's not very likely it's a thermic issue, is it?
Furthermore, I've noticed that the power supply is only 145 W? That
seems _very_ little to me, as even my smallest PCs came with a 200 W
supply, if not much more (250, 300). But if the powersupply was just
overloaded, why didn't it work with no consumer except for the actual
mainboard connected to it?
Third, the backplane probably wasn't the cause of the error (however
nice this idea with trimming the solder spots was :-( ...), as the
same behaviour also shows if I remove it from the mainboard.
And last: A friend of mine who has a voltmeter checked both power
supplies. Instead of 12V / 5 V, the values displayed were considerably
higher (14 and 6,5, I think), and this regardless of whether there was
a heavy load on the power supply or not. I still don't know if his
voltmeter was just misaligned, but I really don't believe it. Maybe
it's normal (seeing as _both_ supplies showed these values)?
So, if we take into account that all other peripherals aren't the
cause of the problem, and further, that there were obviously no shorts
(at least not against any part of the case; of course you never know
whether there's another short on the actual board(s), maybe), plus the
fact that _two_ different power supplies didn't help (which rules out
the P/S as the source of the error, unless the guy who sold me his
just sold it because he suffered from the same problems, maybe because
it's a typical sign of ageing?), all it comes down to is the actual
board, right?
That's where I am now :-) .
Now, could _anybody_ out there give me further hints as of what to
check next? I really can't believe the board is damaged, as it _did_
work even after the first long-term occurrence of the error! But, as
we'd say in German, I'm finished with my Latin ;-) . And a new board
would be very expensive :-( .
Anybody out there who dares to try an exorcism on this sucker ;-) ?
Seeing as this newsgroup seems very well-frequented (unlike my actual
home, the 8 bit newsgroups), I'm hoping for the best.
Regards,
Chris.
[H]omer
news:u1ptltss2hlktahuh...@4ax.com...
> Hello,
>
> I wouldn't normally post such a long request for help, but this
<---->
> What I did notice though, is that on both days where the machine
> decided to give up, it was comparably hot (35 degrees Celsius in my
> room), and in the later case, it had worked for hours, using the
> harddisks a lot, etc. But: If it really was a thermal problem, it
> would run now, because I've put it into a much cooler room
> Now, could _anybody_ out there give me further hints as of what to
> check next?
Its probably either floating joints or dead components on the mobo.
Put it in for a soak test.
I dunno who does that in the US, but in the UK its Analogic.
[H]omer
Mike Leavitt
- tried the mainboard without *any* device connected to it.
If it weren't for this, I would suspect your HD's or controller, and you
could try the HD's in the A2000 with the Buddah to check, but if you are
saying you don't eventually get the floppy screen &/or can't boot from
floppy with the HD's out, it is probably your motherboard, sorry.
--
Mike Leavitt ac...@lafn.org + team Amiga +
Zipper
Christian Link <C.LinkS...@GMX.NET
> For me, this looked like a problem with the power supply. So I
> desperately looked for a spare supply, and like a miracle, I've
found
> one on Ebay, and at a ridiculously low price at that.
What about the mobo/powersupply plug pin condition, maybe there is a
marginal contact. I had lately problems with a loose HD plug that had
intermittent
contact making surprising crashes.
Z
Marcel DeVoe
> Hello,
> A4000D, Motherboard Revision B (same for the Zorro slot
> daughterboard), Super Buster 11, 4*4 MB FastRAM, 1*2 MB ChipRAM, A3640
> Rev. 3.1 CPU card (the A3630 shows the same symptoms I'm going to
> describe, so don't worry about that). Kickstart 3.1.
[snip]
> Occasionally, all I get is an "error color" screen. Sometimes green,
> usually yellow.
[snip: again]
> Finally, this is what I've tried and checked (everything without
> success):
[snip: and again]
You've tried everything except one thing, at least you didn't state that
you have. And even if you say you have, I would suggest you try it again..
Take the simm for the chip ram out, the one closest to the zorro riser
board, and take a clean and dry paper towel and rub the coating on the
contacts until they are bright and shiny, then put it back. You might also
do that for the other simms, but the chip simm is the most important.
Unless you have soldered-in chip simms (this is true for A4000's with
SMD 030's on the motherboard and in A4000T's), try that then get back
to us.
> Furthermore, I've noticed that the power supply is only 145 W? That
> seems _very_ little to me, as even my smallest PCs came with a 200 W
> supply, if not much more (250, 300). But if the powersupply was just
> overloaded, why didn't it work with no consumer except for the actual
> mainboard connected to it?
With my A4000 and original 151 watt supply, I have four hard drives, a
CDRW, two floppy drives with a Catweasel controller and 3 zorro boards in
this desktop A4000. All running with it's current setup for about 2 years
now.
> And last: A friend of mine who has a voltmeter checked both power
> supplies. Instead of 12V / 5 V, the values displayed were considerably
> higher (14 and 6,5, I think), and this regardless of whether there was
> a heavy load on the power supply or not. I still don't know if his
> voltmeter was just misaligned, but I really don't believe it. Maybe
> it's normal (seeing as _both_ supplies showed these values)?
His voltmeter is off or you really didn't have the supply loaded
(motherboard powered).
[snip: snip: snip: snip: AND snip]
Sheesh! You writing a novel or something? ;-)
--
Marcel J. DeVoe - mde...@shore.net - Team *AMIGA*
A4091scsi CV64 96 megs CDRW M1764-17" Catweasel FUSION/Emplant
A4000/060 CyberStorm MKII overclocked 66mhz - see "How to Overclock!"
and "DIY A4000 Tower for $45" @ http://www.shore.net/~mdevoe
Marcel DeVoe
> Christian Link <C.LinkS...@GMX.NET
I agree. It also might explain why his friend's meter registers such
high voltages. Although he did say that both supplies registered the
same high voltages which might rule that out.
I think I'll stick with the chip ram solution. ;-)
Christian Link
many thanks for all your quick replies!
I wish I could report that I got my machine working again, because
when I tried it half an hour ago (removing the ChipRAM SIMM, clean and
reinsert it), it _seemed_ to work. For a couple seconds, and a couple
retries. After that it was yellow screen time again :-( .
To answer the respective hypotheses:
On Wed, 25 Jul 2001 22:47:53 +0100, "[H]omer" <nos...@here.shall.com>
wrote:
>Its probably either floating joints or dead components on the mobo.
>Put it in for a soak test.
>I dunno who does that in the US, but in the UK its Analogic.
thanks, Homer, but seeing as I'm neither US nor UK, I can neither send
it in to Analogic nor do I know the term "soak test". Does it involve
putting the whole board into a board of contact liquid for aqua
destillata? I wonder who could do this in Germany, but actually, the
board doesn't really look dusty or corroded at all. This is because
this machine has not been used for more than seven years, so its fan
couldn't suck too much dust into the case :-) .
Or is a "soak test" something completely different?
On 25 Jul 2001 18:31:44 -0700, "Mike Leavitt" <ac...@lafn.org> wrote:
> - tried the mainboard without *any* device connected to it.
>
>If it weren't for this, I would suspect your HD's or controller, and you
>could try the HD's in the A2000 with the Buddah to check, but if you are
>saying you don't eventually get the floppy screen &/or can't boot from
>floppy with the HD's out, it is probably your motherboard, sorry.
I do _not_ like to hear what you say, Mike ;-) !
I can say, though, that the harddrives are perfectly okay. I just
cannot connect them to the Buddha at all, which is part of my problems
that came with this "A4000 tragedy": They are not compatible with the
Buddha (just like my 1.6 GB WD Caviar drives aren't)! The Buddha would
try to reset them and probably read the RDB, then crash (repeatedly).
I've contacted Jens Schönfeld who has developed the hardware, and he
has confirmed this (blaming WD for it, and I have no reason to doubt
his words).
The problem deriving from that is that I have not backed-up the
current data from these harddisks; otherwise I could more or less
happily work with my A2000 until the 4000 can be fixed. And at the
moment, the 4000 is the only machine I have that could read the data.
There are still two possible solutions for this particular problem:
Try to mount the harddisk under LINUX and copy the data over to my PC.
Afterwards, copy them over to the 2000 via AMIGA2PC. Sounds okay to
me, but I don't have LINUX installed. I'm hoping for a fellow
comp.sys.cbm guy who may send me a boot disk with the required filing
system in order to mount the harddisk and copy the stuff, hence saving
me a temporary LINUX installation.
The other solution: I've recently bought an Apollo 2000-AT controller
on EBay, mainly because it was so cheap. It may turn out to be a lucky
accident, because _maybe_ it will cope with these caviar harddisks
that caused the Buddha so much trouble. In this case, I hope I can
read the data on my A2000. This, however, has two catches again which
I would sooner or later have mentioned in another thread, so I may as
well ask it now: The Buddha ads claim that the Buddha was able to
"even" read harddisks that were formatted with a different controller.
I know this may be a problem with SCSI, but I wouldn't have suspected
this to be one with IDE as well. So, can I expect to run into new
trouble again if I try to use these harddisks with that "new"
controller? They were originally installed and formatted on the
A4000's internal controller, under Kick/WB 3.1.
Well, the other catch is that it has still not arrived in my mail, and
it comes without warranty at that :-( .
Anyway, to get back on topic: The harddisks are okay. The problems
would occur with a completely different IDE drive (the Seagate 130 MB
harddisk that originally came with the machine), too.
On Thu, 26 Jul 2001 08:40:12 +0300, "Zipper"
<pekka....@elimaki.fi> wrote:
>What about the mobo/powersupply plug pin condition, maybe there is a
>marginal contact. I had lately problems with a loose HD plug that had
>intermittent
>contact making surprising crashes.
This would be nice ;-), but please remember I've tried it with _two_
power supplies, so the plug on the supply itself is probably not the
cause of the error. Its motherboard counterpart, however... Well...
What more can I do but resolder its six pins on the underside of the
board :-( ?
Now for the final one, and then I'll stop boring you for a while ;-) :
On Thu, 26 Jul 2001 09:41:39 GMT, Marcel DeVoe <mde...@shore.net>
wrote:
>You've tried everything except one thing, at least you didn't state that
>you have. And even if you say you have, I would suggest you try it again..
>
>Take the simm for the chip ram out, the one closest to the zorro riser
>board, and take a clean and dry paper towel and rub the coating on the
>contacts until they are bright and shiny, then put it back. You might also
>do that for the other simms, but the chip simm is the most important.
Yes, *this* was a really good idea. In fact, I did this when the
machine crashed the first time some weeks ago, but after writing this
posting yesterday, I remembered that I did in fact not try it this
time.
So I did like you told me, and indeed found out that the RAM in bank 0
has become lose. One of the two clips that held it in place has bent
outside too far, and when I touched it, it finally broke (Well, from
looking at various documents on the Web, this seems to be a regular
problem for clumsy people like me :-( ...). Anyway, I decided to cry
about that later, removed all RAMs, cleaned the ChipMem strip and
re-inserted it.
And as I've mentioned in the beginning of this posting: It worked - a
couple seconds... Notice, though, that I didn't have any monitor
connected (I was temporarily storing this machine in a cool room to
rule out thermic problems, and I had no monitor there), so I had to
rely on the harddisk noises and floppy clicking (the latter, however
annoying it is in most cases, really proved to be a nice acoustic
watchdog ;-) !).
After some more tries (on/off/on...) it did *not* work anymore. I
repeated the whole procedure, but it didn't get any better. As hard as
it is to believe, especially considering that the machine did not run
for longer than maybe two minutes altogether, and *open* at that, it
really seems to be something thermal, doesn't it? But I don't want to
lead anybody on the wrong track; I'm no techie :-( .
After I've drageed the case back into my room to check for the screen
colours, I can now say it's yellow, usually. If I pull the ChipMem
completely, it's green of course. But that's probably because the
presence of ChipMem is checked first during the self test?
What's interesting, though, is that even with the ChipMem (and all
other memory) completely removed, the old error shows: Harddisk
stutter during the reset, PowerLED blinking (without a constant
rhythm, mind you!). So if you had ChipMem contact problems in mind, I
think we can rule this out now, too, or can't we?
There's one more thing I've probably forgotten to mention yesterday,
which may have been somewhat important with regard to the Zorro board:
I've also taken out the Super Buster, cleaned its contacts as much as
I could, and re-inserted it. Needless to say, the results didn't
change at all.
>Unless you have soldered-in chip simms (this is true for A4000's with
>SMD 030's on the motherboard and in A4000T's), try that then get back
>to us.
No, I have one of the earlier boards. But this procedure has revealed
yet another problem: Two broken PS/2 sockets (Please don't shout at me
;-) - I know I should have been more careful :-( ...). Since I don't
have a master's degree for (de-)soldering, would it be possible only
to use banks 2 and/or 4? Or do they have to be filled without any gap.
From what I've read, it should be possible to use two 8- or one 16 MB
SIMM, shouldn't it? If not, well...
>With my A4000 and original 151 watt supply, I have four hard drives, a
>CDRW, two floppy drives with a Catweasel controller and 3 zorro boards in
>this desktop A4000. All running with it's current setup for about 2 years
>now.
That's strange, as I remember that even my old PC wasn't able to
provide four HDDs, six cards, four EDO SIMMs, and a mere Pentium
processor. I admit I don't know much about the more recent Motorola
stuff, but I assume they consumed much more power than the newer
Pentiums. Because, in the P-machine mentioned, I had a 200W power
supply, which obviously was way too weak.
>His voltmeter is off or you really didn't have the supply loaded
>(motherboard powered).
It was loaded as much as we could load it :-) . But I assume that this
"off" theory is quite realistic.
>[snip: snip: snip: snip: AND snip]
>
>Sheesh! You writing a novel or something? ;-)
Yes, a tragic drama, to be more precise. I don't know how it will
finish, but so far its working title is "How my ungraceful A4000 who I
bought so many new toys for simply gave me the finger, made me buy a
spare power supply I never used and even made my clumsiness ruin its
fragile PS/2 sockets in less than six weeks!"
You guys will definitely get credited ;-) !
Greetings, and thanks again for all your ideas so far!
Chris
slu...@earthlink.net
> thanks, Homer, but seeing as I'm neither US nor UK, I can neither send
> it in to Analogic nor do I know the term "soak test". Does it involve
> putting the whole board into a board of contact liquid for aqua
> destillata?
Having worked in a QA department for a videogame company, I got to
know all sorts of debugging lingo. Yay me!
Soak test refers to leaving the application in question (or in this
case, the machine in question) running overnight. So just turn your
machine on, maybe run a few thing on it, and then go to bed. Well,
actually you'll probably want to start soaking it before dinner. When
you wake up in the morning, check up on the machine and see if it's
still working. =)
Keep doing this for every change you make to the machine, such as
when you leave a bit of hardware unplugged to test if that might be
the culprit.
Christian Link
On Thu, 26 Jul 2001 21:22:59 GMT, slu...@earthlink.net wrote:
>> thanks, Homer, but seeing as I'm neither US nor UK, I can neither send
>> it in to Analogic nor do I know the term "soak test". Does it involve
> Soak test refers to leaving the application in question (or in this
>case, the machine in question) running overnight. So just turn your
>machine on, maybe run a few thing on it, and then go to bed. Well,
>actually you'll probably want to start soaking it before dinner. When
>you wake up in the morning, check up on the machine and see if it's
>still working. =)
Ah, so it is a burn-in test, only that you don't run it once after
aquiring new hardware, but more or less regularly (i.e. even in case
of strange errors) - a difference that's quite marginal, I think.
Hmmm, but this is what I *do* with all my vintage (duck :-) ...)
hardware. Just in this case, where I suspected the power supply first,
and where the error showed very quickly all the time, I didn't want to
do it. I'm no techie, so I really have a hell of respect for power
supplies. I mean, I wouldn't want to be waked by a big "Puff" and have
to search for a fire extinguished while still half asleep, you know
;-) ?
Meanwhile, I think I have found the problem, though. I just don't want
to post it to the group yet, because a) I've thought too often that
the error was gone, only to start crying a couple days later again,
making a complete fool out of myself for bothering people again and
again, b) _if_ this was the mistake (and I hope it was, as I have
already bought a replacement part on EBay, and it wasn't that cheap
:-( ...), it was really DAMN stupid of me, and so embarassing that I
still have to search for the diplomatic words that will keep the
people who bothered answering to my desperate cries for help from
sending the fire squad to my home :-( ...
Anyway, thanks for the explanation!
Greetings,
Chris.
Milan BB
Christian Link wrote:
>
Just one idea - when you run out of them :) - check for electrolytes
(or electrolytic capacitors) - my A4K had problem with one on 3640
CPU card - drived me nuts - untill found which one,....
Hope you get your machine working in no time...
Regards
--
Milan Belic
mil...@europe.com
Please remove NOSPAM when replying !!!
Marcel DeVoe
> when I tried it half an hour ago (removing the ChipRAM SIMM, clean and
> reinsert it), it _seemed_ to work. For a couple seconds, and a couple
> retries. After that it was yellow screen time again :-( .
Yellow screen, eh? Here's one cause of that I found this on Warren Block's
A4K hardware Guide:
http://wonkity.com/~wblock/a4000hard/bkprob.html
"Backplane Problems
After adding or removing expansion cards, system no longer boots, displays
yellow screen. Solution: too-long resistor and capacitor leads on solder
side of backplane daughterboard are bent and shorting together. Remove the
backplane and trim leads."
> This would be nice ;-), but please remember I've tried it with _two_
> power supplies, so the plug on the supply itself is probably not the
> cause of the error. Its motherboard counterpart, however... Well...
> What more can I do but resolder its six pins on the underside of the
> board :-( ?
Well, why not? If you have a loose contact on the pins then they can get
quite hot and even burn the board.
[snip]
> No, I have one of the earlier boards. But this procedure has revealed
> yet another problem: Two broken PS/2 sockets (Please don't shout at me
> ;-) - I know I should have been more careful :-( ...). Since I don't
> have a master's degree for (de-)soldering, would it be possible only
> to use banks 2 and/or 4? Or do they have to be filled without any gap.
> From what I've read, it should be possible to use two 8- or one 16 MB
> SIMM, shouldn't it? If not, well...
Two 8 megs will work (check Warren Block's A4K web page for which banks)
but I don't think one 16 will.
BTW, just in case your chip ram "might" be bad, you can use an 8 meg simm
in it's place to check.
>>With my A4000 and original 151 watt supply, I have four hard drives, a
>>CDRW, two floppy drives with a Catweasel controller and 3 zorro boards in
>>this desktop A4000. All running with it's current setup for about 2 years
>>now.
> That's strange, as I remember that even my old PC wasn't able to
> provide four HDDs, six cards, four EDO SIMMs, and a mere Pentium
> processor. I admit I don't know much about the more recent Motorola
> stuff, but I assume they consumed much more power than the newer
> Pentiums. Because, in the P-machine mentioned, I had a 200W power
> supply, which obviously was way too weak.
Yes, if I were to have my old 040 CPU that might put more a of strain as
that got hot enough to fry eggs on. But my 060 is cool enough to put my
finger on, even overclocked.
Speaking of which, did you make sure the CPU board was *fully* seated?
The board is quite flexible and can bow up in the middle on that long
socket. You might have a look under the board while seating it to make
sure it is flush.
Some people with sensitive fingers and don't really press down hard enough
on the board because of the soldered pins sticking up on the other upper
side of the board from the socket. I used a rubber handle of a screwdriver
to press down with. Check the pins in the sockets to make sure they are
all straight and even.
Is your battery okay? They can leak and etch the board and cause similar
problems. Or is it charging okay? Check the voltage on it. It should be
about 3.6 volts whether the computer is on or off (slightly higher on).
Zipper
Milan BB <milanb...@europe.com> kirjoitti
viestissä:3B6090ED...@europe.com...
> Just one idea - when you run out of them :) - check for electrolytes
> (or electrolytic capacitors) - my A4K had problem with one on 3640
> CPU card - drived me nuts - untill found which one,....
Or check every solder joint on the mobo...
Happily it is only Eyetech monitor switcher which pops into life for
me
when resoldered - 3 times so far (short clipped chip legs)...
Z
David Myers
Z>
Z> What about the mobo/powersupply plug pin condition, maybe there is
Z> a marginal contact. I had lately problems with a loose HD plug
Z> that had intermittent contact making surprising crashes.
Z>
I fixed a friend's A4000 that had a loose 5V socket on the power
connector. Although it was soldered OK onto the motherboard, the
socket is a split tubular contact that mates with the pin in the plug
from the power supply and was was making poor contact. The metal it
was made of had lost it's springiness from the heat generated from
the current flowing through it. You can often even see the
difference between it and the other sockets if you look carefully.
Try bending it a bit tighter.
Regards,
David Myers
www.starnet.com.au/davem/amiga.html
Remove numerals from email address to reply.
--
------------------------------------------------------------------------
Posted with Amiga NewsRog
------------------------------------------------------------------------
Christian Link
On Fri, 27 Jul 2001 06:43:54 GMT, Marcel DeVoe <mde...@shore.net>
wrote:
>Yellow screen, eh? Here's one cause of that I found this on Warren Block's
>A4K hardware Guide:
>
>http://wonkity.com/~wblock/a4000hard/bkprob.html
>
>"Backplane Problems
[...]
Sob... That was _not_ it :-( . You probably have overlooked it because
I've talked too much again, hehe... But this is what I've written in
my initial posting:
"One hour, two hours, eight hours... Well, I thought that maybe the
zorro daughterboard had shorted a couple pins with the harddisk bay.
Indeed the insulating plastic mat did show some damages. So I removed
the board and cut all those solder points down to reasonable sizes."
I take it what you guys call a "backplane" is this separate board with
the zorro slots, isn't it?
Well, it wasn't the reason. Although I do admit that the insulating
plastik was _very_ worn and the screws I've used for fixing the
harddives were sticking away from the cage very far. Still...
But speaking of Warren Block, this guy really has done a lot for me
with his A4000 guide. At least he reinforced my spirit each night I
was close to throwing this machine out of the Window. Actually, I was
going to write him a short note of appreciation, but it seems he has
had various (meanwhile invalid) email adresses since. Anybody knows
his current, or whether he's reading this newsgroup? Well, I just
wanted to say "Thanks, Warren!" in case he does ;-) .
>Two 8 megs will work (check Warren Block's A4K web page for which banks)
>but I don't think one 16 will.
Alright, I think I could find two 8 MB PS/2 SIMMs, too. The problem is
(which I found out by trial and error) that you have to use the banks
in the correct order (i.e. if you don't use bank #1, using bank #2
won't help), and unfortunately bank #1 *is* one of the broken ones.
Well, so far they manage to stick there the way they are, but I'm
afraid I won't be able to save myself the work of replacing the
sockets soon :-( .
>BTW, just in case your chip ram "might" be bad, you can use an 8 meg simm
>in it's place to check.
I see _you_ have had Warren's works as night lecture as well ;-)))
...?
>Yes, if I were to have my old 040 CPU that might put more a of strain as
>that got hot enough to fry eggs on. But my 060 is cool enough to put my
>finger on, even overclocked.
>
>Speaking of which, did you make sure the CPU board was *fully* seated?
SOOOB...
Yeah, please, more salt, man! Rub it in really deep, please. I deserve
it :-( ...
Okay, seriously now... I'm not going to say "That was it", because I
have thought that too often during the last six weeks. Still, it looks
as if I found the error - me :-( .
Sorry for having stirred up all that noise for a comparable trivialty,
but...: It was/is the A3640 board. It really seems to be...
See, loudmouth that I am, I *expressively* told you guys (when
describing my system in the first posting) that you shouldn't worry
about the CPU card, as the machine wouldn't run with the A3630,
either...
Yes...
...but I forgot that during that test, I've forgotten to set the two
INT/EXT jumpers back to their original (030) settings. No wonder it
didn't work...
Now I've put it back after adjusting the jumpers to their correct
settings, and indeed my A4000 is runs for 24 hours now - and in a very
hot room at that (open case, though). I've swapped the boards some
hours ago, and the A4000 crashed immediately, so it is quite safe to
assume that the A3640 was the culprit...
WAAAHHH!!!
See, I've checked everything at least twice or "thrice", but *never*
repeated testing the CPU board. And since I've made this damn jumper
mistake during the only test I did with the CPU card, I was relying on
a totally wrong examination.
I'm an idiot. I will kill myself. Promised. But feel free to humilate
me first - you have every right to do so ;-))) ...
By the way: I've resoldered most points I could access on the A3640
because it obviously is only a contact or thermic problem. But to no
avail. Meanwhile, I've "ordered" yet another one on EBay (and I do
hope it's > Rev 3.0, as the auction text didn't tell :-( ...). I'm fed
up with this crap. REALLY fed-up. Should I ever manage to get this
A3640 working again (meaning I'd have two), it will be put into my
A3000T-40 (Yes, I'm proud to say I found one today - maybe fate's
compensation for all the unnecessary trouble lately?) which is missing
its CPU card altogether but has this nice 3000T-040 case, so it _must_
get its A3640 sooner or later, anyway. Guess when it's time to
refurbish this machine, you guys will hear from me, so you better run
quickly if you see a topic smelling like "A3000T problems, please
help!" ;-) .
>Is your battery okay? They can leak and etch the board and cause similar
>problems. Or is it charging okay? Check the voltage on it. It should be
>about 3.6 volts whether the computer is on or off (slightly higher on).
Yup, the battery is in absolutely perfect condition. I cut it out
after it stopped working some years ago - _before_ it could leak, and
finally replaced it with a brandnew one only two weeks ago. Hey, I
even managed to _socket_ this rechargable battery, which I'm a
"little" bit proud of, as I haven't heard of such a thing done to a
standard rechargable battery before. Not that it was that much of a
great deed, but it saves me the "And what if it leaks in the
meantime?" fears once and forever, I hope.
Anyway, guys, I think this topic's over now. I'm the one to blame for
his idioty, but maybe your comprehensive list of possible error
sources and the respective suggestions will be of some use for the
next guy with a defective A4000 when he searches the usenet archives.
Let me add, that I am totally stunned: I thought "we people at
comp.*.cbm" were cool and helpful and everything, but you were really
*great* (Okay, but then again, there's more of you than of us, hehe)!
Thanks for all your time! It was appreciated.
Greetings,
Chris.
Christian Link
<milanb...@europe.com> wrote:
>Just one idea - when you run out of them :) - check for electrolytes
>(or electrolytic capacitors) - my A4K had problem with one on 3640
>CPU card - drived me nuts - untill found which one,....
As you can see from my other posting, it _was_ the A3640. So you may
have hit the right spot. Unfortunately I am no techie, so I cannot
really say what is an ElCo and what isn't. Older parts, yes, but the
parts used on the A3640 look quite strange to me. Do you mean the
small, silver-coloured parts tha look cylindric? If I have counted
right, there are five of them on my board, all with a "c"+number
(Capacitor, right?). And indeed there is one that catched my eye
yesterday when I tried to resolder a couple of the joints. It's C105,
directly above the oscillator. One of the two contacts looks somewhat
"dirty", if not corroded, but I'm not sure.
Just in case I may feel courageous enough to give it a try: How did
you fix yours? It seems to be "SMD-glued" to the board, doesn't it? Is
heating the respective contact ("+") enough?
Greetings,
Chris.
[H]omer
<slu...@earthlink.net> wrote in message
news:3B608C1B...@earthlink.net...
> Christian Link wrote:
>
> > thanks, Homer, but seeing as I'm neither US nor UK, I can neither send
Nehmen Sie bitte meine aufrichtigen Entschuldigungen an.
>> ... nor do I know the term "soak test". Does it involve
> > putting the whole board into a board of contact liquid for aqua
> > destillata?
Aahhhh class :)
Er ... no. A soak test is an ambiguous and commonly misused term
with several similar meanings - all related to "testing".
There is a "Burn-In", which is not really a test as such, but an initial
"breaking in" period for IC components to allow the initial
application of energy/heat/electrical current to be applied gradually.
This prevents initially overloading the components, so they don't
burn out upon initial "real life" use. It is also a "test" in a way, so that
if there *is* a problem, it may manifest itself during the burn-in.
Often you will see the phrase "Burn-In 24Hrs" on a motherboard
CMOS/BIOS chip for example.
Then there is a floating joints test. Both tests are performed with the
same principle in mind. When things become hot, they expand. If a
soldered/solid state component is making a floating contact then
when it moves (due to expansion/contraction caused by a change in
temperature) it may break contact, thus malfunctioning. Old or
misused components exhibit this behaviour. Floating or loose
contacts can be caused by exposure (oxidation) or fatigue
(physical stress).
There is the component level multimeter test, which checks for
specific voltages/resistance/capacitance/power etc at certain
discreet parts of the circuit board. The manufacturer/licensed
repair centre will be aware of the correct values at certain "check
points" and if there are any showing a discrepancy of "> +/- X%"
(where X is a known value) then there is a fault. If and how it can
be corrected, is up to the technician.
Then there is the LSI (large scale integration) logic test. This involves
removing individual components (chips) and testing the logic of their
gate arrays using dedicated hardware - something that looks a bit
like a PIC processor/programming tool. This kind of thing is usually
way beyond the scope of a repair centre, but I have seen it done in
circumstances where the component in question was a
deleted/unavailable part. Initially this kind of thing is done by the
manufacturer as part of quality control. Successfully diagnosing a
LSI level logic error is of little use however, all you can do is chuck
the component in the bin, but at least you'll have an answer.
When I use the term "soak test", I'm using it as a generic term
meaning (D) all of the above.
> Having worked in a QA department for a videogame company, I got to
> know all sorts of debugging lingo. Yay me!
>
> Soak test refers to leaving the application in question (or in this
> case, the machine in question) running overnight. So just turn your
> machine on, maybe run a few thing on it, and then go to bed. Well,
> actually you'll probably want to start soaking it before dinner. When
> you wake up in the morning, check up on the machine and see if it's
> still working. =)
And yes, "soak test" is also a term sometimes applied to continuous
software "stress testing"/debugging/benchmarking. Not used
commonly in that context today thou.
In general, early startup problems are bound to be hardware
related, with the two possible exceptions of incorrect settings in
various CMOS/BIOS chips and illogically structured media
descriptors (Buggy filesystems/corrupt partition tables/bootsectors)
Well that's my 22 cents worth.
[H]omer.
SG
> Hello guys,
>
> many thanks for all your quick replies!
>
> I wish I could report that I got my machine working again, because
> when I tried it half an hour ago (removing the ChipRAM SIMM, clean and
> reinsert it), it _seemed_ to work. For a couple seconds, and a couple
> retries. After that it was yellow screen time again :-( .
You said before the screen was yellow and sometimes green. Green
means bad RAM and nothing else. Try another SIMM.
--
Steve.
Angelo Vigliotti
news:1416.608T1000T136...@iolinc.net...
> Gene Heskett sends Greetings to Milan BB;
>
> >Christian Link wrote:
> >>
>
> >Just one idea - when you run out of them :) - check for electrolytes
> >(or electrolytic capacitors) - my A4K had problem with one on 3640
> >CPU card - drived me nuts - untill found which one,....
>
> >Hope you get your machine working in no time...
>
> capacitors installed correctly! To a unit, they've all been installed
> bass ackwards. Thanks to the low voltage, only 5 volts, most of those
> caps have lasted for many years, but with much reduced capacitance
> after the first week or so under power. And a few have shorted out,
> taking the machine down as the psu will do a very quick shutdown when
> they do short.
>
> The shorted ones I tried to replace, then found they worked as well just
> by removing the bad one (or 2), so I haven't been terribly religious
> about replacements of late.
>
> Cheers, Gene
Hi Gene!
Your above post got me curious so I took a close look at
the two 3640's I have. WOA! how the hell did they get ALL of them
bass ackwards? Caps sitting backwards in the device feeder during
FAB?
Everyone, to a match had there colored plus side opposite the
silkscreen +. Now is it possible the silkscreen is backwards
on the boards and not the components? hmmmm ;-)
Too weird.
James Vigliotti
Angelo Vigliotti
> Gene Heskett sends Greetings to Angelo Vigliotti;
> >> capacitors installed correctly! To a unit, they've all been installed
> >> bass ackwards. Thanks to the low voltage, only 5 volts, most of those
> >> caps have lasted for many years, but with much reduced capacitance
> >> after the first week or so under power. And a few have shorted out,
> >> taking the machine down as the psu will do a very quick shutdown when
> >> they do short.
> >>
> >> The shorted ones I tried to replace, then found they worked as well
just
> >> by removing the bad one (or 2), so I haven't been terribly religious
> >> about replacements of late.
> >>
> >> Cheers, Gene
>
> >Hi Gene!
> > Your above post got me curious so I took a close look at
> >the two 3640's I have. WOA! how the hell did they get ALL of them
> >bass ackwards? Caps sitting backwards in the device feeder during
> >FAB?
>
> >Everyone, to a match had there colored plus side opposite the
> >silkscreen +. Now is it possible the silkscreen is backwards
> >on the boards and not the components? hmmmm ;-)
> >Too weird.
>
> > James Vigliotti
>
> the ever handy Wavetek DMM confirmed that the silk screening was correct
> in every place I managed to get the probes to reach, which FWIW, isn't
> all of them. But when you find everything you check backwards, and
> you've probably checked 8 out of 12 or so, then the conclusion is
> obvious. The 12 is a fictitious number of course as I don't recall how
> many there are on a 3640 and I don't have a handy 4k/040 machine to
> check here at home.
>
> In such a case, the 'infant mortality', meaning those that blew the tops
> off in the first 10 minutes of having power applied backwards *should*
> have gotten someones attention in burn-in in Westchester, but their
> noses must have been burnt out by then. And yes, I've seen such caps,
> obviously the blowup was very old by the time I saw it, and the machine
> was still pumping lightwave animations out by the frame when something
> else caused me to open the hood and check the oil. AIR at the time it
> was a dying Rodime 100 meg drive or something similar. FWIW, I removed
> the cap and was gonna replace it later when I got some, but 10 years
> later that spot on the baord is still empty, and the machine is still
> pumping along as a secondary CG yet today.
>
> Cheers, Gene
> --
> Gene Heskett, CET, UHK |Amiga A2k Zeus040, 70MB ram, 31 gigs
Here's a longshot, but maybe the caps are just taps to ground for a
better signal plane? Longshot but since yours continues to run even
with the odd cap not even there this seems plausible.
Just recently got an Xcalibur and a quad doubler fixed up
in 2 4000-D's that I will be selling. Must be quite a bit
of software out there that really can't benefit from an 060
because either local 040 fastram on the XCAL or the
Quaddoubler (50Mhz 040) give an 060 a run for it's money
and even outperforms it in some cases. Lightwave is one
which of course was never optimized beyond 040.
James Vigliotti
Christian Link
I lost track of the many interesting passages that this "flakey
3640"/"backward-soldered capacitors" provided me with, so I'll save me
the fat quotings ;-) .
I take it that these electrolytic capacitors are these small,
silver-coloured, cylindric things that actually look like a knob with
one half (seen from above) being painted black, right? Yes, I know
what you think: "Why does this nerd dare to solder on this board if he
cannot even identify a capacitor" ;-))) ... But the ones I knew from
older devices looked different. Still, the ones I have in mind have
"C" part number printed on the board, and there are five of six of
them on the A3640, I think.
From what you wrote, I figured that the black half has to point into
the "+" direction (or my whole A3640 is oriented the wrong way,
too...). Well, in the case of my board, they all do. But definitely
not in my actual 4000 :-( . Hopefully you, Gene, are right and due to
the low voltage it won't do any harm, although I think I lately found
an article on the net, where some technically knowledgeable guy made a
real science out of what can/will happen if one leaves them installed
this wrong way. Anyway...
You were, however, right with your assumption about floating joints.
I've given this board a last try yesterday and with my clumsy fingers
and limited experience with soldering tried to "re-glue" all these
five or six "knobs" back to the board. Not really a clean job if you
use a normal soldering iron and haven't got very calm hands, but -
shit, it seems to work! Put the A3640 in, powered on the machine, and
it ran for hours!
Now I don't want to sound over-optimistic, as this board has worked
for one week after the first fault, before just showing the old
symptoms again. It may be that I will cry about this damn A3640 in a
couple days again, I fear. But maybe it really were some lose contacts
with these capacitors, as some of them really looked somewhat
corroded.
In the meantime, I've bought myself a different A3640 on EBay. Maybe -
assuming my old one will now continue to work - this was done for
nothing and I've wasted my money, but then again, I think I could fit
it into my A3000T-40 later, which came "naked". So I would have bought
it anyway (Well, *if* the board in that 3000 is still okay, which I
haven't tested so far).
I have one question left regarding these electrolyte capacitors, just
to prepare myself for the next problems ;-) :
>caps have lasted for many years, but with much reduced capacitance
>after the first week or so under power. And a few have shorted out,
That's what caught my eye in your very first posting. Capacitors store
energy, hm? I take it they aren't made for storing it a long time, but
is there a slight possibility that all my 3640-related problems
(especially with the board working after laying in the corner for a
week or so, and failing again after another week in use) are due to
the capacitors having some flake and have to slowly discharge for a
while before the board would run again (which may explain while it did
after that "pause" outside the computer!)? I know this sounds
absolutely stupid, but I just want to make sure. I'm no electronics
genius, you know ;-) ?
>The shorted ones I tried to replace, then found they worked as well just
>by removing the bad one (or 2), so I haven't been terribly religious
>about replacements of late.
You mean you removed them without replacing them? What did you do with
the two contact points they were soldered to? Connect them or left
them open? I mean, as a very last resort I may consider this one, too,
in case I got a spare board. But I don't think I would be able to
remove and resolder them backwards again. The "solder points" are
really too small for me, so I would rather melt the whole part instead
of the metal ;-) .
Anyway, your information has helped me a lot. Maybe this already was
it (resoldering the points). I really hope so, but after the last
weeks, i.e. after saying "Wow, now I finally found the problem and
fixed it" for at least three times, I'm not so optimistic anymore...
Greetings,
Chris.
Tim
<C.LinkS...@GMX.NET> wrote:
> I take it that these electrolytic capacitors are these small,
> silver-coloured, cylindric things that actually look like a knob with
> one half (seen from above) being painted black, right?
Capacitors come in a variety of wildly different casings. Buy a cheap
electronics magazine, and you'll see pictures of many of them. That's
easier than trying to describe them all.
Polarised caps (the ones that have to be used a certain way around),
usually have some indication of which lead is negative.
> From what you wrote, I figured that the black half has to point into
> the "+" direction
Depends on the cap, but often it's the negative lead that's marked.
Put in backwards, then can short out, explode (if there's enough
current), or just plain not work.
Some devices will apparently work fine with capacitor faults, some
won't, you may not be immediately aware of any problems.
>> caps have lasted for many years, but with much reduced capacitance
>> after the first week or so under power. And a few have shorted out,
> That's what caught my eye in your very first posting. Capacitors store
> energy, hm? I take it they aren't made for storing it a long time,
Capacitors store a charge. Usually only for a short time period, but
some devices (clocks, etc., that use very little current) can use them
to keep them running during a power failure.
The time they'll hold a charge depends on their value (larger values
can hold more charge for longer), on how little current you draw from
them, and how long it takes for the current to leak away internally.
>> The shorted ones I tried to replace, then found they worked as well just
>> by removing the bad one (or 2), so I haven't been terribly religious
>> about replacements of late.
> You mean you removed them without replacing them? What did you do with
> the two contact points they were soldered to? Connect them or left
> them open? I mean, as a very last resort I may consider this one, too,
> in case I got a spare board. But I don't think I would be able to
> remove and resolder them backwards again.
The circuit was designed to have them in place. Don't short out
capacitors (including removing one, and shorting together the
contacts), you'll be putting current through things that shouldn't
have it (possibly in large amounts). You may get away with removing
some caps, and leaving the contents open, but don't count on it.
> The "solder points" are really too small for me, so I would rather
> melt the whole part instead of the metal ;-) .
Any half decent tech could solder a board for you, even if they're not
competent enough to understand the circuitry on it. It's just metal
work.
james
[snip]
> Speaking of which, did you make sure the CPU board was *fully* seated?
>
> The board is quite flexible and can bow up in the middle on that long
> socket. You might have a look under the board while seating it to make
> sure it is flush.
>
> Some people with sensitive fingers and don't really press down hard enough
> on the board because of the soldered pins sticking up on the other upper
> side of the board from the socket. I used a rubber handle of a screwdriver
> to press down with. Check the pins in the sockets to make sure they are
> all straight and even.
When seating my A3000 daughterboard I was having a similar problem.
The motherboard was flexing downdard towards the middle, between the
two card connectors. This made the card not seat fully in the conn-
ectors.
My solution? An old, old one. When the A3000 was a new machine
people suggested cutting two soft (non-abrasive) pencil erasers to
the correct height and putting them under the inside ends of the
card connectors. Thus, the motherboard is supported when pressing
the daughterboard down and the card seats fully. It still works!
;v)
BTW, the right height is so that the motherboard is sitting on the
erasers and the metal motherboard supports. You don't want the
erasers exerting upwards pressure on the motherboard when the
retaining screws are replaced.
Marcel DeVoe
> On Fri, 27 Jul 2001 06:43:54 GMT, Marcel DeVoe <mde...@shore.net>
> wrote:
>>Two 8 megs will work (check Warren Block's A4K web page for which banks)
>>but I don't think one 16 will.
> Alright, I think I could find two 8 MB PS/2 SIMMs, too. The problem is
> (which I found out by trial and error) that you have to use the banks
> in the correct order (i.e. if you don't use bank #1, using bank #2
> won't help), and unfortunately bank #1 *is* one of the broken ones.
> Well, so far they manage to stick there the way they are, but I'm
> afraid I won't be able to save myself the work of replacing the
> sockets soon :-( .
Some people have reported success with "hot glue". ;-)
> Sorry for having stirred up all that noise for a comparable trivialty,
> but...: It was/is the A3640 board. It really seems to be...
> ...but I forgot that during that test, I've forgotten to set the two
> INT/EXT jumpers back to their original (030) settings. No wonder it
> didn't work...
> Now I've put it back after adjusting the jumpers to their correct
> settings, and indeed my A4000 is runs for 24 hours now - and in a very
> hot room at that (open case, though). I've swapped the boards some
> hours ago, and the A4000 crashed immediately, so it is quite safe to
> assume that the A3640 was the culprit...
> WAAAHHH!!!
[snip]
> See, I've checked everything at least twice or "thrice", but *never*
> repeated testing the CPU board. And since I've made this damn jumper
> mistake during the only test I did with the CPU card, I was relying on
> a totally wrong examination.
> I'm an idiot. I will kill myself. Promised. But feel free to humilate
> me first - you have every right to do so ;-))) ...
> By the way: I've resoldered most points I could access on the A3640
> because it obviously is only a contact or thermic problem. But to no
> avail. Meanwhile, I've "ordered" yet another one on EBay (and I do
> hope it's > Rev 3.0, as the auction text didn't tell :-( ...). I'm fed
> up with this crap. REALLY fed-up. Should I ever manage to get this
> A3640 working again (meaning I'd have two), it will be put into my
> A3000T-40 (Yes, I'm proud to say I found one today - maybe fate's
> compensation for all the unnecessary trouble lately?) which is missing
> its CPU card altogether but has this nice 3000T-040 case, so it _must_
> get its A3640 sooner or later, anyway. Guess when it's time to
> refurbish this machine, you guys will hear from me, so you better run
> quickly if you see a topic smelling like "A3000T problems, please
> help!" ;-) .
Whoo boy! Now I really AM confused. By your ramblings you say the 3640 is
a rev3.0 and is the cause of your crashes? If so, do you have any ZIII
boards? If so again, a 3640 rev 3.0 can truly be a problem.
But there IS a partial fix for it, if you go to Warren Block's web page
again and go to the "Boards" section and read the 3640 reference page;
http://wonkity.com/~wblock/a4000hard/a3640ref.html
I had intermittant crashes myself when I bought my CV64 graphics card
when I still had my 3640 rev3.0 CPU, and that fix cured it.
Christian Link
On Mon, 30 Jul 2001 07:42:17 GMT, Marcel DeVoe <mde...@shore.net>
wrote:
[Broken PS/2 sockets]
>> Well, so far they manage to stick there the way they are, but I'm
>> afraid I won't be able to save myself the work of replacing the
>> sockets soon :-( .
>
>Some people have reported success with "hot glue". ;-)
Hmmm... Why do you put this in quotation marks? Do the Americans/Brits
have a different term for this than "Heißkleber" (I think you're
Dutch, so you probably understand this?)?
Anyway, unless there are different types of "hot glue" which different
"strenghts" or "hardness factors" (whatever _these_ are called ;-)))
...), I didn't succeed. It simply wasn't stable enough, more like hard
bubblegum like we've found under our school d... er, well, you know
what I mean...
>Whoo boy! Now I really AM confused. By your ramblings you say the 3640 is
>a rev3.0 and is the cause of your crashes? If so, do you have any ZIII
No, no, no! I probably put it into too confusing words as usual...
What I have here in my A4000, i.e. the CPU card that probably caused
all this trouble, definitely _is_ a Rev 3.1 card. I just don't know
what the one is that I've just bought on EBay - as the seller didn't
tell, and I didn't have the time to ask him first ("Buy know" at 03:40
in the night...).
Greetings,
Chris.
Marcel DeVoe
> wrote:
> [Broken PS/2 sockets]
>>> Well, so far they manage to stick there the way they are, but I'm
>>> afraid I won't be able to save myself the work of replacing the
>>> sockets soon :-( .
>>
>>Some people have reported success with "hot glue". ;-)
> Hmmm... Why do you put this in quotation marks? Do the Americans/Brits
> have a different term for this than "Heißkleber" (I think you're
> Dutch, so you probably understand this?)?
Nope, US American. ;) It was just quoted because that was a subject of
discussion here about a couple of weeks ago. Some people swear by
it. Others have said to watch out for it.
> Anyway, unless there are different types of "hot glue" which different
> "strenghts" or "hardness factors" (whatever _these_ are called ;-)))
> ...), I didn't succeed. It simply wasn't stable enough, more like hard
> bubblegum like we've found under our school d... er, well, you know
> what I mean...
Hmm, I didn't know they came in different hardnesses. Maybe that's the
reason some have said to stay away from it. Some have even said it
contains a substance that might corrode metal contacts. It could be true
of only certain made glue sticks. I would certainly stay away from that.
But most of the stuff I've personally come across has been very hard.
I even saw someone who had his fishing pole repaired with it.
> What I have here in my A4000, i.e. the CPU card that probably caused
> all this trouble, definitely _is_ a Rev 3.1 card. I just don't know
> what the one is that I've just bought on EBay - as the seller didn't
> tell, and I didn't have the time to ask him first ("Buy know" at 03:40
> in the night...).
Well 3.1 certainly shouldn't have caused any real problems. What about
chip U103 on the motherboard under the CPU card? In Warren Block's
description of the Fastlane SCSI controller board in the Boards section,
it mentions that the board came with a replacement chip for U103 if you
have a rev3.0 3640 CPU.
I'm not 100% sure of this, but I believe that that chip change fixes the
same reversed signal problem that the cut-and-jumper does for rev3.0,
mentioned in his 3640 board section (I should check with Warren on this).
But the problem is if you then go with a rev3.1 board after changing that
chip, that signal problem may be reversed back again. The Fastlane section
says U103 originally is a 74FCT244 clock driver chip but gets replaced
with a 74FCT240 chip provided with the board.
So you might want to make sure it's the original 74FCT244?
SG
> > By the way: I've resoldered most points I could access on the A3640
> > because it obviously is only a contact or thermic problem. But to no
> > avail. Meanwhile, I've "ordered" yet another one on EBay (and I do
> > hope it's > Rev 3.0, as the auction text didn't tell :-( ...). I'm fed
> > up with this crap. REALLY fed-up. Should I ever manage to get this
> > A3640 working again (meaning I'd have two), it will be put into my
> > A3000T-40 (Yes, I'm proud to say I found one today - maybe fate's
> > compensation for all the unnecessary trouble lately?) which is missing
> > its CPU card altogether but has this nice 3000T-040 case, so it _must_
> > get its A3640 sooner or later, anyway. Guess when it's time to
> > refurbish this machine, you guys will hear from me, so you better run
> > quickly if you see a topic smelling like "A3000T problems, please
> > help!" ;-) .
>
> Whoo boy! Now I really AM confused. By your ramblings you say the 3640 is
> a rev3.0 and is the cause of your crashes? If so, do you have any ZIII
> boards? If so again, a 3640 rev 3.0 can truly be a problem.
Green screen no-boots, though?
Thats a bad RAM configuration everytime, in my experience.
> But there IS a partial fix for it, if you go to Warren Block's web page
> again and go to the "Boards" section and read the 3640 reference page;
>
> http://wonkity.com/~wblock/a4000hard/a3640ref.html
>
> I had intermittant crashes myself when I bought my CV64 graphics card
> when I still had my 3640 rev3.0 CPU, and that fix cured it.
>
> --
> Marcel J. DeVoe - mde...@shore.net - Team *AMIGA*
> A4091scsi CV64 96 megs CDRW M1764-17" Catweasel FUSION/Emplant
> A4000/060 CyberStorm MKII overclocked 66mhz - see "How to Overclock!"
> and "DIY A4000 Tower for $45" @ http://www.shore.net/~mdevoe
--
Steve.
Marcel DeVoe
>> > By the way: I've resoldered most points I could access on the A3640
>> > because it obviously is only a contact or thermic problem. But to no
>> > avail. Meanwhile, I've "ordered" yet another one on EBay (and I do
>> > hope it's > Rev 3.0, as the auction text didn't tell :-( ...). I'm fed
>> > up with this crap. REALLY fed-up. Should I ever manage to get this
>> > A3640 working again (meaning I'd have two), it will be put into my
>> > A3000T-40 (Yes, I'm proud to say I found one today - maybe fate's
>> > compensation for all the unnecessary trouble lately?) which is missing
>> > its CPU card altogether but has this nice 3000T-040 case, so it _must_
>> > get its A3640 sooner or later, anyway. Guess when it's time to
>> > refurbish this machine, you guys will hear from me, so you better run
>> > quickly if you see a topic smelling like "A3000T problems, please
>> > help!" ;-) .
>>
>> Whoo boy! Now I really AM confused. By your ramblings you say the 3640 is
>> a rev3.0 and is the cause of your crashes? If so, do you have any ZIII
>> boards? If so again, a 3640 rev 3.0 can truly be a problem.
> Green screen no-boots, though?
> Thats a bad RAM configuration everytime, in my experience.
He said yellow mostly, but anyways...
Green screen for me has been my chip simm needed cleaning... everytime.
Doobrey
> > "strenghts" or "hardness factors" (whatever _these_ are called ;-)))
> > ...), I didn't succeed. It simply wasn't stable enough, more like hard
> > bubblegum like we've found under our school d... er, well, you know
> > what I mean...
>
> Hmm, I didn't know they came in different hardnesses. Maybe that's the
> reason some have said to stay away from it. Some have even said it
> contains a substance that might corrode metal contacts. It could be true
> of only certain made glue sticks. I would certainly stay away from that.
Those gluesticks come in various specs..they melt at different temperatures
depending on what material they were designed to stick to (eg polystyrene
sticks have a lot lower melting point ..but still enough to make you scream
and shout when you get it on your fingers).
Their `hardness` can also change, as some sticks are for permanent fixture,
while others are softer and made only for temporary/removable use.
I dunno about them having
corrosive properties, I just thought they modified the plastic to suit the
temperature. I`ll have to ask our local 3M sales rep when he next visits.
But as for the average user, they`ll have one of them hobby type guns that
only melt at one temperature which use a general purpose glue stick. So
being general purpose, I`d assume it`d be pretty safe on metal, although
it`s
probably a wise idea to do a test on a small scrap of similar metal just to
check.
Christian Link
On Tue, 31 Jul 2001 04:15:55 GMT, Marcel DeVoe <mde...@shore.net>
wrote:
>Nope, US American. ;) It was just quoted because that was a subject of
>discussion here about a couple of weeks ago. Some people swear by
>it. Others have said to watch out for it.
Usually, I swear by it, too, especially for insulating things I'm
really sure I don't have to fiddle around with anymore (it's perfect
for flooding a whole plug case once you know it is working). But for
really fixing things, especially things under pressure in the opposite
direction (like the PS/2 sockets' pins), _my_ hot glue was of no use.
Too soft, really.
>Hmm, I didn't know they came in different hardnesses. Maybe that's the
I've read what Doobrey wrote later, and it sounds reasonable. I, too,
have just such a "hobbyist pistol", and I guess I would burn my
fingers with it if it heated the stuff inside even more than it does.
So the sticks I'm using are probably more the "temporary, removable
glue" tye - lucky for me, so I could remove all that mess I've put on
my PS/2 sockets ;-) .
>reason some have said to stay away from it. Some have even said it
>contains a substance that might corrode metal contacts. It could be true
>of only certain made glue sticks. I would certainly stay away from that.
I made sure I just touched the plastic and no metal at all. Still, no
luck. Anyway. For the moment, the RAM modules would work even
"unfixed", although I wouldn't trust this machine until the sockets
are replaced. But as a temporary solution (and for backing up the
contents of my harddisk), that's okay
>But most of the stuff I've personally come across has been very hard.
>I even saw someone who had his fishing pole repaired with it.
Hmmm... Thinking of it again, this reminds me that I once had sticks
that were milky-white and non-transparent. It really became very hard
later. The ones I'm using now are transparent and really soft. Maybe I
should buy some other sticks indeed, as my pistol could handle these
white sticks, too (without melting in my hand, I mean ;-) ...).
But, anyway, I've just got accustomed to the thought of soldering in
new sockets someday. I just want to make sure my A2000 is properly
installed before that, as this work on the A4000 board may take an
unexperienced "solderer" like me a couple days. If I tried to do it in
one day, it would probably make make me aggressive enough to fire that
board against a wall...
>Well 3.1 certainly shouldn't have caused any real problems. What about
>chip U103 on the motherboard under the CPU card? In Warren Block's
>description of the Fastlane SCSI controller board in the Boards section,
>it mentions that the board came with a replacement chip for U103 if you
>have a rev3.0 3640 CPU.
Um, you seem to have missed a posting of mine: Resoldering the
capacitors on the A3640, like a kind person in this forum suggested,
seems to have done the trick - so far, that is. The A4000 ran for
about two days with the A3640 installed, but, remember, it once ran
for a whole week before giving up again. We'll see. What makes me
somewhat optimistic, though, is that one of the capacitors seemed to
have a very dirty, somewhat corroded "solder point". Maybe that one
was the culprit...
Greetings,
Chris.
Gl...@canit.se
+Christian Link <C.LinkS...@GMX.NET> wrote
>>Some people have reported success with "hot glue". ;-)
>Hmmm... Why do you put this in quotation marks? Do the Americans/Brits
>have a different term for this than "Heißkleber" (I think you're
>Dutch, so you probably understand this?)?
>Anyway, unless there are different types of "hot glue" which different
>"strenghts" or "hardness factors" (whatever _these_ are called ;-)))
>...), I didn't succeed. It simply wasn't stable enough, more like hard
>bubblegum like we've found under our school d... er, well, you know
>what I mean...
My SIMM's in my A4000 is hold in place with hotglue, and have been there
for several years :)
And yes, there are alot of hotglue types, some get very hard, some become
like rubber, some are low temperature (for gluing stuff that doesn't allow
heat) some are high temperature (if you are gluing things that will
become hot), you can get it in different colors and so on, but kind of
any type would do for just gluing SIMM's..
But when hotgluing, use alot.. it's not like superglue (cyanoglue) where
you have to mimimize the gaps, hotglue gladly fill gaps :)
Hotglue is not a glue, it's a religion :)
--
| Apollo fastslot accelerators page - Http://www.canit.se/~glenn/apollo.html |
|----------------------------------------------------------------------------|
| ___ | Email «» Sha...@bay-watch.com |
| / __\ __ | Homepage «» http://www.canit.se/~glenn |
| __ / /__ / /__ ____ ____ __ | IRC «» XT600 @ IRC-net |
| (__/ /_ // / -_) _ ) _ )__) | Amiga - Silicon Graphics - 8bit comps. |
| \___//_/\__/_//_/_//_/ | T h e K i n g d o m o f S w e d e n |
Marcel DeVoe
> But, anyway, I've just got accustomed to the thought of soldering in
> new sockets someday. I just want to make sure my A2000 is properly
> installed before that, as this work on the A4000 board may take an
> unexperienced "solderer" like me a couple days. If I tried to do it in
> one day, it would probably make make me aggressive enough to fire that
> board against a wall...
Make sure you get slanted sockets instead of vertical ones. They may not
leave enough room for your bottom zorro board.
> Um, you seem to have missed a posting of mine: Resoldering the
> capacitors on the A3640, like a kind person in this forum suggested,
> seems to have done the trick - so far, that is. The A4000 ran for
> about two days with the A3640 installed, but, remember, it once ran
> for a whole week before giving up again. We'll see.
I hope you still have the heat sink on the 040. Mine was quite beefy held
on with some mighty strong spring clips. Considering the temperature of
it even with the heat sink, I don't think it would run very long without
one.
> What makes me
> somewhat optimistic, though, is that one of the capacitors seemed to
> have a very dirty, somewhat corroded "solder point". Maybe that one
> was the culprit...
--
John T Maguire
<C.LinkS...@GMX.NET> wrote:
>Hello guys,
>
>many thanks for all your quick replies!
>
>I wish I could report that I got my machine working again, because
>when I tried it half an hour ago (removing the ChipRAM SIMM, clean and
>reinsert it), it _seemed_ to work. For a couple seconds, and a couple
>retries. After that it was yellow screen time again :-( .
>
Chris, if it were me - I'd hack in a PC power supply and see what
happens - at least you'd totally eliminate it being a PS problem. The
how-to should be on Aminet...
John T Maguire, Kennebunkport, Maine
The Maine Webcam Network; http://www.maine-webcams.net
Kennebunkport ScreenSavers; http://www.port-gifts.com