Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
C64: Out of memory error in 0
948 views
Skip to first unread message
dre...@easy.com
Nov 7, 2012, 11:58:38 PM11/7/12
to
Hi,
One of my C64s is showing the "Out of memory error in 0. ready." at startup. Assuming that one or more RAM chip is faulty, I would need to determine which one. One of the chips (U12) is socketed so I was able to rule out that one by swapping.
I tried piggybacking each RAM chip without results. None of the chips gets especially hot.
My only clue so far is the character "(" appearing on the screen at times (in spots that should normally be blank). According to Lemon 64 forums, this erroneous character points to RAM chip U10. Is this correct? Are there other tests I could do to solve which chip(s) are faulty?
One of my C64s is showing the "Out of memory error in 0. ready." at startup. Assuming that one or more RAM chip is faulty, I would need to determine which one. One of the chips (U12) is socketed so I was able to rule out that one by swapping.
I tried piggybacking each RAM chip without results. None of the chips gets especially hot.
My only clue so far is the character "(" appearing on the screen at times (in spots that should normally be blank). According to Lemon 64 forums, this erroneous character points to RAM chip U10. Is this correct? Are there other tests I could do to solve which chip(s) are faulty?
Andreas Meerbann
Nov 8, 2012, 7:47:34 AM11/8/12
to
Yes, the "(" instead of a " " indicates that the RAM chip for bit 3 is causing the issue because "(" = 0x28 and " " = 0x20 in PETSCII coding.
The chip for bit 3 might be U10 (don't have a schematic here in the moment but if lemon64 says so it should be correct).
Try to remove this chip, put a socket inside and test with a chip known to be good. The chance is very high that this is fixing the issue.
Happy (de-)soldering!
Andreas
The chip for bit 3 might be U10 (don't have a schematic here in the moment but if lemon64 says so it should be correct).
Try to remove this chip, put a socket inside and test with a chip known to be good. The chance is very high that this is fixing the issue.
Happy (de-)soldering!
Andreas
dre...@easy.com
Nov 8, 2012, 6:49:40 PM11/8/12
to
On Thursday, November 8, 2012 2:47:35 PM UTC+2, Andreas Meerbann wrote:
> Yes, the "(" instead of a " " indicates that the RAM chip for bit 3 is causing the issue because "(" = 0x28 and " " = 0x20 in PETSCII coding. The chip for bit 3 might be U10 (don't have a schematic here in the moment but if lemon64 says so it should be correct). Try to remove this chip, put a socket inside and test with a chip known to be good. The chance is very high that this is fixing the issue. Happy (de-)soldering! Andreas
Thanks, I will look to get U10 replaced. The error first showed up after a long time of inactivity (about 2 years), which seems to be common with the C64 RAM chips.
> Yes, the "(" instead of a " " indicates that the RAM chip for bit 3 is causing the issue because "(" = 0x28 and " " = 0x20 in PETSCII coding. The chip for bit 3 might be U10 (don't have a schematic here in the moment but if lemon64 says so it should be correct). Try to remove this chip, put a socket inside and test with a chip known to be good. The chance is very high that this is fixing the issue. Happy (de-)soldering! Andreas
Clocky
Nov 9, 2012, 7:32:29 PM11/9/12
to
retro stuff.
One thing to watch out for is the power supply, one with ripple will kill
DRAM first and it's a very common problem with the C64 bricks.
If the brick is one you can open up and replace the capacitors on, do it
before it starts killing the other less expendible chips in the C64.
Ray Carlsen
Nov 10, 2012, 3:55:09 PM11/10/12
to
>> Thanks, I will look to get U10 replaced. The error first showed up
>> after a long time of inactivity (about 2 years), which seems to be
>> common with the C64 RAM chips.
>
> It's common with RAM chips of that age, I'm replacing them all the
> time on retro stuff. One thing to watch out for is the power supply,
> one with ripple will kill DRAM first and it's a very common problem
> with the C64 bricks.
>
> If the brick is one you can open up and replace the capacitors on, do
> it before it starts killing the other less expendible chips in the
> C64.
I never saw the original post, but one way to find the bad RAM IC in a
>> after a long time of inactivity (about 2 years), which seems to be
>> common with the C64 RAM chips.
>
> It's common with RAM chips of that age, I'm replacing them all the
> time on retro stuff. One thing to watch out for is the power supply,
> one with ripple will kill DRAM first and it's a very common problem
> with the C64 bricks.
>
> If the brick is one you can open up and replace the capacitors on, do
> it before it starts killing the other less expendible chips in the
> C64.
computer that comes up with less than the normal 38911 bytes free is to
"piggyback" a known good IC over each board RAM, one at a time (power
off, of course), and see if the bytes free changes at power up. Change
that IC. The piggyback test chip must be oriented the same way as the
board chip, and all pins of the test chip must contact the board chip
pins. Since there is only 5 volts there, you can use your fingers to
hold the chip in place for the test. Note that shorted RAM (usually
caused by a bad power supply) will produce a blank screen and the
piggyback test will not work for that. A suspected power supply should
be tested under full -artificial- load (not a computer) to prevent
repeated damage to the computer.
Ray
Message has been deleted
dre...@easy.com
Nov 10, 2012, 9:34:36 PM11/10/12
to
On Saturday, November 10, 2012 10:55:10 PM UTC+2, Ray Carlsen wrote:
> >> Thanks, I will look to get U10 replaced. The error first showed up >> after a long time of inactivity (about 2 years), which seems to be >> common with the C64 RAM chips. > > It's common with RAM chips of that age, I'm replacing them all the > time on retro stuff. One thing to watch out for is the power supply, > one with ripple will kill DRAM first and it's a very common problem > with the C64 bricks. > > If the brick is one you can open up and replace the capacitors on, do > it before it starts killing the other less expendible chips in the > C64. I never saw the original post, but one way to find the bad RAM IC in a computer that comes up with less than the normal 38911 bytes free is to "piggyback" a known good IC over each board RAM, one at a time (power off, of course), and see if the bytes free changes at power up. Change that IC. The piggyback test chip must be oriented the same way as the board chip, and all pins of the test chip must contact the board chip pins. Since there is only 5 volts there, you can use your fingers to hold the chip in place for the test. Note that shorted RAM (usually caused by a bad power supply) will produce a blank screen and the piggyback test will not work for that. A suspected power supply should be tested under full -artificial- load (not a computer) to prevent repeated damage to the computer. Ray
Hi,
> >> Thanks, I will look to get U10 replaced. The error first showed up >> after a long time of inactivity (about 2 years), which seems to be >> common with the C64 RAM chips. > > It's common with RAM chips of that age, I'm replacing them all the > time on retro stuff. One thing to watch out for is the power supply, > one with ripple will kill DRAM first and it's a very common problem > with the C64 bricks. > > If the brick is one you can open up and replace the capacitors on, do > it before it starts killing the other less expendible chips in the > C64. I never saw the original post, but one way to find the bad RAM IC in a computer that comes up with less than the normal 38911 bytes free is to "piggyback" a known good IC over each board RAM, one at a time (power off, of course), and see if the bytes free changes at power up. Change that IC. The piggyback test chip must be oriented the same way as the board chip, and all pins of the test chip must contact the board chip pins. Since there is only 5 volts there, you can use your fingers to hold the chip in place for the test. Note that shorted RAM (usually caused by a bad power supply) will produce a blank screen and the piggyback test will not work for that. A suspected power supply should be tested under full -artificial- load (not a computer) to prevent repeated damage to the computer. Ray
The computer is just showing an out of memory error at startup, so piggybacking should in theory work. However, I tried doing it with each RAM chip (including U10, which is the suspect here), and the same error still appeared each time. I even tried piggybacking multiple RAM chips at a time in case more than one has failed.
Does this mean replacing them won't work either? If I did the piggybacking wrong (some pins didn't make contact or so), would I know? Would it just produce a blank screen then?
I don't get the normal startup screen, so don't know how many bytes free it would show.
Dombo
Nov 11, 2012, 5:40:48 AM11/11/12
to
Op 11-Nov-12 3:31, grs...@easy.com schreef:
case. If the outputs are not floating the faulty chip outputs will
conflict with the outputs of the piggybacked chip. In that case it is
indeterminate which chip 'wins'.
> However, I tried doing it with each RAM chip (including U10, which is the suspect here), and the same error still appeared each time. Does it mean replacing them won't work either? If I did the piggybacking wrong (some pins didn't make contact or so), would I know? Would it just produce a blank screen then?
My experience is that piggybacking rarely works. The only way to be
certain is to remove the suspect chip and replace it with with a known
good one. While you are at it I recommend you put in a socket so you
won't have to desolder it ever again (the PCB won't survive that). Based
on the symptoms you can work out which chip is most likely the culprit.
> I don't get the normal startup screen, so don't know how many bytes free it would show.
On a C64 that would be 38911.
> On Saturday, November 10, 2012 10:55:10 PM UTC+2, Ray Carlsen wrote:
>>>> Thanks, I will look to get U10 replaced. The error first showed up >> after a long time of inactivity (about 2 years), which seems to be >> common with the C64 RAM chips. > > It's common with RAM chips of that age, I'm replacing them all the > time on retro stuff. One thing to watch out for is the power supply, > one with ripple will kill DRAM first and it's a very common problem > with the C64 bricks. > > If the brick is one you can open up and replace the capacitors on, do > it before it starts killing the other less expendible chips in the > C64. I never saw the original post, but one way to find the bad RAM IC in a computer that comes up with less than the normal 38911 bytes free is to "piggyback" a known good IC over each board RAM, one at a time (power off, of course), and see if the bytes free changes at power up. Change that IC. The piggyback test chip must be oriented the same way as the board chip, and all pins of the test chip must contact the board chip pins. Since
there is only 5 volts there, you can use your fingers to hold the chip in place for the test. Note that shorted RAM (usually caused by a bad power supply) will produce a blank screen and the piggyback test will not work for that. A suspected power supply should be tested under full -artificial- load (not a computer) to prevent repeated damage to the computer. Ray
>
there is only 5 volts there, you can use your fingers to hold the chip in place for the test. Note that shorted RAM (usually caused by a bad power supply) will produce a blank screen and the piggyback test will not work for that. A suspected power supply should be tested under full -artificial- load (not a computer) to prevent repeated damage to the computer. Ray
>
> Hi,
>
> The computer is just showing an out of memory error at startup, so piggybacking should in theory work.
Only if the outputs of the faulty chip are floating, which is rarely the
>
> The computer is just showing an out of memory error at startup, so piggybacking should in theory work.
case. If the outputs are not floating the faulty chip outputs will
conflict with the outputs of the piggybacked chip. In that case it is
indeterminate which chip 'wins'.
> However, I tried doing it with each RAM chip (including U10, which is the suspect here), and the same error still appeared each time. Does it mean replacing them won't work either? If I did the piggybacking wrong (some pins didn't make contact or so), would I know? Would it just produce a blank screen then?
My experience is that piggybacking rarely works. The only way to be
certain is to remove the suspect chip and replace it with with a known
good one. While you are at it I recommend you put in a socket so you
won't have to desolder it ever again (the PCB won't survive that). Based
on the symptoms you can work out which chip is most likely the culprit.
> I don't get the normal startup screen, so don't know how many bytes free it would show.
Clocky
Nov 11, 2012, 6:14:41 AM11/11/12
to
voltage wise under load but it's the ripple that kills DRAM (and other chips
in time).
Clocky
Nov 11, 2012, 6:19:29 AM11/11/12
to
hold an address line either open or closed. Piggybacking won't work with
those faults present.
The only way to be
> certain is to remove the suspect chip and replace it with with a known
> good one. While you are at it I recommend you put in a socket so you
> won't have to desolder it ever again (the PCB won't survive that).
> Based on the symptoms you can work out which chip is most likely the
> culprit.
It could also be a problem with the PLA or DRAM select support chip.
dre...@easy.com
Nov 11, 2012, 7:44:56 AM11/11/12
to
My plan is to replace U10 and PLA with sockets. U10 because it is the suspected faulty RAM chip and PLA because of an altogether different symptom that the machine has had since 2000 - unable to write "through" the ROM, resulting in some programs not working.
dre...@easy.com
Nov 11, 2012, 8:02:59 AM11/11/12
to
Which chip is this "DRAM select support chip"?
Clocky
Nov 11, 2012, 4:04:02 PM11/11/12
to
works. Piggybacking might find a dud DRAM but it could just as easily not
show up a dud RAM chip. Whilst it's easy to point to a DRAM failure, have
you tried any cartridges and do they work?
> I don't get the normal startup screen, so don't know how many bytes
> free it would show.
dre...@easy.com
Nov 11, 2012, 8:46:59 PM11/11/12
to
On Sunday, November 11, 2012 11:04:01 PM UTC+2, Clocky wrote:
> dre...@easy.com wrote: > On Saturday, November 10, 2012 10:55:10 PM UTC+2, Ray Carlsen wrote: >>>> Thanks, I will look to get U10 replaced. The error first showed up >>>> >> after a long time of inactivity (about 2 years), which seems to >>>> be >> common with the C64 RAM chips. > > It's common with RAM >>>> chips of that age, I'm replacing them all the > time on retro >>>> stuff. One thing to watch out for is the power supply, > one with >>>> ripple will kill DRAM first and it's a very common problem > with >>>> the C64 bricks. > > If the brick is one you can open up and >>>> replace the capacitors on, do > it before it starts killing the >>>> other less expendible chips in the > C64. I never saw the original >>>> post, but one way to find the bad RAM IC in a computer that comes >>>> up with less than the normal 38911 bytes free is to "piggyback" a >>>> known good IC over each board RAM, one at a time (power off, of >>>> course), and see if the bytes free changes at power up. Change >>>> that IC. The piggyback test chip must be oriented the same way as >>>> the board chip, and all pins of the test chip must contact the >>>> board chip pins. Since there is only 5 volts there, you can use >>>> your fingers to hold the chip in place for the test. Note that >>>> shorted RAM (usually caused by a bad power supply) will produce a >>>> blank screen and the piggyback test will not work for that. A >>>> suspected power supply should be tested under full -artificial- >>>> load (not a computer) to prevent repeated damage to the computer. >>>> Ray > > Hi, > > The computer is just showing an out of memory error at startup, so > piggybacking should in theory work. However, I tried doing it with > each RAM chip (including U10, which is the suspect here), and the > same error still appeared each time. I even tried piggybacking > multiple RAM chips at a time in case more than one has failed. > > Does this mean replacing them won't work either? If I did the > piggybacking wrong (some pins didn't make contact or so), would I > know? Would it just produce a blank screen then? > It depends entirely on how the DRAM has failed as to whether piggybacking works. Piggybacking might find a dud DRAM but it could just as easily not show up a dud RAM chip. Whilst it's easy to point to a DRAM failure, have you tried any cartridges and do they work? > I don't get the normal startup screen, so don't know how many bytes > free it would show. 38911
While it would be interesting to test a cartridge, I currently do not own any.
> dre...@easy.com wrote: > On Saturday, November 10, 2012 10:55:10 PM UTC+2, Ray Carlsen wrote: >>>> Thanks, I will look to get U10 replaced. The error first showed up >>>> >> after a long time of inactivity (about 2 years), which seems to >>>> be >> common with the C64 RAM chips. > > It's common with RAM >>>> chips of that age, I'm replacing them all the > time on retro >>>> stuff. One thing to watch out for is the power supply, > one with >>>> ripple will kill DRAM first and it's a very common problem > with >>>> the C64 bricks. > > If the brick is one you can open up and >>>> replace the capacitors on, do > it before it starts killing the >>>> other less expendible chips in the > C64. I never saw the original >>>> post, but one way to find the bad RAM IC in a computer that comes >>>> up with less than the normal 38911 bytes free is to "piggyback" a >>>> known good IC over each board RAM, one at a time (power off, of >>>> course), and see if the bytes free changes at power up. Change >>>> that IC. The piggyback test chip must be oriented the same way as >>>> the board chip, and all pins of the test chip must contact the >>>> board chip pins. Since there is only 5 volts there, you can use >>>> your fingers to hold the chip in place for the test. Note that >>>> shorted RAM (usually caused by a bad power supply) will produce a >>>> blank screen and the piggyback test will not work for that. A >>>> suspected power supply should be tested under full -artificial- >>>> load (not a computer) to prevent repeated damage to the computer. >>>> Ray > > Hi, > > The computer is just showing an out of memory error at startup, so > piggybacking should in theory work. However, I tried doing it with > each RAM chip (including U10, which is the suspect here), and the > same error still appeared each time. I even tried piggybacking > multiple RAM chips at a time in case more than one has failed. > > Does this mean replacing them won't work either? If I did the > piggybacking wrong (some pins didn't make contact or so), would I > know? Would it just produce a blank screen then? > It depends entirely on how the DRAM has failed as to whether piggybacking works. Piggybacking might find a dud DRAM but it could just as easily not show up a dud RAM chip. Whilst it's easy to point to a DRAM failure, have you tried any cartridges and do they work? > I don't get the normal startup screen, so don't know how many bytes > free it would show. 38911
Clocky
Nov 12, 2012, 1:38:53 AM11/12/12
to
Well it depends on the system but support chips like the 74138/139 and
buffers are often used.
buffers are often used.
dre...@easy.com
Nov 16, 2012, 10:46:35 AM11/16/12
to
Okay, I replaced U10 with a socket and swapped in a new RAM chip. Result: normal startup screen with 38911 bytes free! Thanks everyone for the help, problem solved.
Clocky
Nov 16, 2012, 5:09:47 PM11/16/12
to
Shaun Bebbington
Nov 27, 2012, 8:22:51 AM11/27/12
to
comp.sys.cbm will always be my friend :-)
I'm having the same problem on a C64A, I think that I need a better PSU.
Is this a common problem with old PSUs, or can I trust a C64C PSU maybe?
Also, would it be worth replacing all of the RAM chips on my machine?
Regards,
Shaun.
I'm having the same problem on a C64A, I think that I need a better PSU.
Is this a common problem with old PSUs, or can I trust a C64C PSU maybe?
Also, would it be worth replacing all of the RAM chips on my machine?
Regards,
Shaun.
Shaun Bebbington
Nov 27, 2012, 8:35:13 AM11/27/12
to
I assume that these chips are compatible:
http://www.ebay.co.uk/itm/MCM4164BP15-Ram-IC-16PIN-DIP-/280929561860?pt=UK_Sound_Vision_Other&hash=item4168b4ed04
Regards,
Shaun.
http://www.ebay.co.uk/itm/MCM4164BP15-Ram-IC-16PIN-DIP-/280929561860?pt=UK_Sound_Vision_Other&hash=item4168b4ed04
Regards,
Shaun.
Clocky
Nov 28, 2012, 5:25:33 AM11/28/12
to
Shaun Bebbington wrote:
> comp.sys.cbm will always be my friend :-)
>
> I'm having the same problem on a C64A, I think that I need a better
> PSU.
>
> Is this a common problem with old PSUs, or can I trust a C64C PSU
> maybe?
>
The white bricks are rubbish. At least the black ones are repairable by
> comp.sys.cbm will always be my friend :-)
>
> I'm having the same problem on a C64A, I think that I need a better
> PSU.
>
> Is this a common problem with old PSUs, or can I trust a C64C PSU
> maybe?
>
replacing the capacitors with good quality new ones. The problem is that
capacitors don't age terribly well, so every power supply of that age is
suspect (even if voltages under load are OK, the ripple is what kills chips)
> Also, would it be worth replacing all of the RAM chips on my machine?
>
sockets but on earlier C64's the more chips you remove the more likely you
are to damage the board.
Ofcourse if you have a desoldering station like I do it's a no-brainer as
removing all the RAM and fitting sockets is pretty easy.
If you are confident and/or have the right tools, go for it.
Clocky
Nov 28, 2012, 5:35:06 AM11/28/12
to
You can also use 41256 DRAM but you need to make a small modification.
http://www.c64-wiki.com/index.php/RAM
Computer Nerd Kev
Nov 29, 2012, 6:20:26 PM11/29/12
to
On 28 Nov 2012, Clocky wrote:
> The white bricks are rubbish. At least the black ones are
> repairable by replacing the capacitors with good quality
> new ones.
And the white ones aren't?
> The white bricks are rubbish. At least the black ones are
> repairable by replacing the capacitors with good quality
> new ones.
--
__ __
#_ < |\| |< _#
Clocky
Nov 29, 2012, 7:43:34 PM11/29/12
to
Computer Nerd Kev wrote:
> On 28 Nov 2012, Clocky wrote:
>
>> The white bricks are rubbish. At least the black ones are
>> repairable by replacing the capacitors with good quality
>> new ones.
>
> And the white ones aren't?
I think the white ones that look like the black ones are, in fact any that
> On 28 Nov 2012, Clocky wrote:
>
>> The white bricks are rubbish. At least the black ones are
>> repairable by replacing the capacitors with good quality
>> new ones.
>
> And the white ones aren't?
can be opened and are not epoxy filled should be repairable ;-)
0 new messages