have you ever withnessed calculators not working anymore due to corrupted memory?

[pier4r]

As far as I know calculators using ram as user memory should last (in terms of write cycles) quite long. Nevertheless it may happen that the ram gets corrupted and the calculator becomes completely unusable or partially usable. The same applies (actually this should happen earlier with an equal workload) to calculators offering ROM as user memory (48,49,50 families at least).

So far it did not happen with any of my few calculators, but it happened with some RAM banks of computers (just 2 or 3 out 50 or more ram banks that I used) and with a tablet (internal flash memory got corrupted in some crucial sectors, it was not possible to flash the tablet anymore).

Since some of you have a lot of calculators, some quite old an well used maybe you have more experience. For example the ones that often flash the rom of a 34s, did you ever have the internal memory corrupted somewhere that let the calculator be not fully working anymore?

[toml_12953]

(05-01-2017 02:03 PM)pier4r Wrote:  As far as I know calculators using ram as user memory should last (in terms of write cycles) quite long. Nevertheless it may happen that the ram gets corrupted and the calculator becomes completely unusable or partially usable. The same applies (actually this should happen earlier with an equal workload) to calculators offering ROM as user memory (48,49,50 families at least).

So far it did not happen with any of my few calculators, but it happened with some RAM banks of computers (just 2 or 3 out 50 or more ram banks that I used) and with a tablet (internal flash memory got corrupted in some crucial sectors, it was not possible to flash the tablet anymore).

Since some of you have a lot of calculators, some quite old an well used maybe you have more experience. For example the ones that often flash the rom of a 34s, did you ever have the internal memory corrupted somewhere that let the calculator be not fully working anymore?

It happens regularly with the HP Prime. A total reset fixes it temporarily but the memory becomes corrupted again after a while.

Tom L

[Dave Britten]

I had the RAM chip in a TI-58C fail almost completely. It would read back zeros for pretty much anything you tried to store.

[pier4r]

(05-01-2017 02:07 PM)toml_12953 Wrote:  It happens regularly with the HP Prime. A total reset fixes it temporarily but the memory becomes corrupted again after a while.

Tom L

What, really? Should be a pain in the ass of a product then. Is it only yours that behaves so or is it a general problem?

@Dave Britten: could I ask you how old/used was the ti-58c ?

[toml_12953]

(05-01-2017 03:24 PM)pier4r Wrote:  

(05-01-2017 02:07 PM)toml_12953 Wrote:  It happens regularly with the HP Prime. A total reset fixes it temporarily but the memory becomes corrupted again after a while.

Tom L

What, really? Should be a pain in the ass of a product then. Is it only yours that behaves so or is it a general problem?

@Dave Britten: could I ask you how old/used was the ti-58c ?

I have three Primes. One hardware A, one hardware C with old keyboard and one hardware C with the new, high-contrast keyboard. They all exhibit the same symptoms.

Tom L

[Dave Britten]

(05-01-2017 03:24 PM)pier4r Wrote:  @Dave Britten: could I ask you how old/used was the ti-58c ?

It happened maybe a year or two ago. I don't know how heavily the previous owner used it, as I got it from ebay around 2008, and finally got around to rebuilding the battery pack in 2014 or so. Not sure if I killed the chip with static or what - the RAM started acting funny not too long after.

Fortunately 58Cs are cheap, so I just found another one on ebay.

[Geoff Quickfall]

Had a 67 that was short memory. Went to line 168 and looped back to 000. Dead RAM. harvested one from a corroded board and back to 224.

Same with an HP 41CV.

Why they were dead, no idea. Wasn't solder joints as that is the first thing I check. Just dead. Static?

Geoff

[pier4r]

(05-01-2017 04:17 PM)toml_12953 Wrote:  I have three Primes. One hardware A, one hardware C with old keyboard and one hardware C with the new, high-contrast keyboard. They all exhibit the same symptoms.

Tom L

Ok, and it means that the program in memory or whatever is saved gets corrupted? Or the computations yields to wrong results?

Is the internal storage affected as well?

That is pretty a drawback to buy a real prime.

[toml_12953]

(05-01-2017 06:29 PM)pier4r Wrote:  

(05-01-2017 04:17 PM)toml_12953 Wrote:  I have three Primes. One hardware A, one hardware C with old keyboard and one hardware C with the new, high-contrast keyboard. They all exhibit the same symptoms.

Tom L

Ok, and it means that the program in memory or whatever is saved gets corrupted? Or the computations yields to wrong results?

Is the internal storage affected as well?

That is pretty a drawback to buy a real prime.

The symptom I see is that the calculator reboots when trying to create a new program or editing an old one. After the reboot the new program is indeed in the list so I know the problem is happening after it's created. It's strange but the exact same symptom occurs in all three calcs.

Tom L

[pier4r]

(05-01-2017 05:25 PM)Geoff Quickfall Wrote:  Why they were dead, no idea. Wasn't solder joints as that is the first thing I check. Just dead. Static?

Geoff

I suppose that, aside from static discharges,electrical components lose the precision after a while. I remember that I studied (electronics, basic transistors) that quite a lot of electrons have to move through transistors, so after a while I think they will "erode" the capacity of some of them.

[pier4r]

(05-01-2017 06:33 PM)toml_12953 Wrote:  The symptom I see is that the calculator reboots when trying to create a new program or editing an old one. After the reboot the new program is indeed in the list so I know the problem is happening after it's created. It's strange but the exact same symptom occurs in all three calcs.

Tom L

Hmm, so If I understand you correctly, you cannot tell whether is the RAM or the internal storage that creates the problem, am I mistaken?

Does it happen for every new program?

[Dave Frederickson]

(05-01-2017 02:03 PM)pier4r Wrote:  As far as I know calculators using ram as user memory should last (in terms of write cycles) quite long. Nevertheless it may happen that the ram gets corrupted and the calculator becomes completely unusable or partially usable. The same applies (actually this should happen earlier with an equal workload) to calculators offering ROM as user memory (48,49,50 families at least).

So far it did not happen with any of my few calculators, but it happened with some RAM banks of computers (just 2 or 3 out 50 or more ram banks that I used) and with a tablet (internal flash memory got corrupted in some crucial sectors, it was not possible to flash the tablet anymore).

Since some of you have a lot of calculators, some quite old an well used maybe you have more experience. For example the ones that often flash the rom of a 34s, did you ever have the internal memory corrupted somewhere that let the calculator be not fully working anymore?

When you say "corrupted" do you mean that the memory simply doesn't contain the expected data or that the memory has failed? If the memory fails then, sure, the calculator is either partially or completely unusable. If the data is simply corrupted then some sort of reset should clear the problem.

The 48,49, and 50 use RAM for User Memory. Using ROM for user memory is illogical.

Dave

[pier4r]

Dave with corrupted I mean "consistently reporting wrong (previously stored) data from all or certain memory sectors, even after a reset" or even not working at all.

Why is illogical to use the rom? I mean libraries are stored there (at least for 48,49,50) and if one needs extra space as temporary working space the rom is the only viable way. Surely it is not common (I can barely fill the memory of my 50g), but someone may have done it. Especially the 48 series as far as I know is pretty low on ram.

[PANAMATIK]

(05-02-2017 06:41 AM)pier4r Wrote:  Dave with corrupted I mean "consistently reporting wrong (previously stored) data from all or certain memory sectors, even after a reset" or even not working at all.

Why is illogical to use the rom? I mean libraries are stored there (at least for 48,49,50) and if one needs extra space as temporary working space the rom is the only viable way. Surely it is not common (I can barely fill the memory of my 50g), but someone may have done it. Especially the 48 series as far as I know is pretty low on ram.

ROM is "Read Only Memory" per definition, it cannot be used for changing contents like RAM. But you are right, in todays calculators Flash memory is used instead of genuine ROM memory and it can be used for changing contents but has limited number of write cycles.

To understand the possible answers to your question you have to know about the interaction of Boot memory, Flash memory and battery buffered and non buffered user RAM memory. Modern calculators use all of these memory types. The Boot memory normally is protected flash memory, which cannot be overwritten accidentally.

It is principally possible, that corrupted memory in each of the above type can lead to permanent malfunction. If (and really If) the boot memory contents is still intact then it is possible to recover the calculator by an update.

Wrong battery buffered RAM contents (and even uninitialized RAM if it is not cleared at power up) can lead to a not working calculator. It depends solely on the operating systems initialization procedure. if the operating software is not water proof and depends on the RAM contents at start up, it can easily run into an endlesss loop or else. Performing a Reset procedure should avoid this problem, because it clears all RAM contents and this can be handled by the non water proof operating software (until the memory becomes corrupted again by some complicated user actions).

Wrong Flash memory leads always to unpredictible behaviour, if it concerns the operating software. But normally wrong Flash contents will be covered by the bootloader, which checks its checksum before jumping into it.

Wrong Bootloader memory is the worst case. If the bootloader isn't present any more, the operating software will not be started and it inhibits any update. Only special programming tools at the manufacturer can be used to restore the calculator.

Bernhard

[BartDB]

As to the original question I have not had any failed calculators where I could specifically pinpoint RAM problems (e.g. a P50 that would turn on and have funny response to keys - but as it's a single chip implementation no diagnostics were possible) but I do have an old Canon Canola with a suspected problem of the delay-line memory .

(05-02-2017 06:41 AM)pier4r Wrote:  Dave with corrupted I mean "consistently reporting wrong (previously stored) data from all or certain memory sectors, even after a reset" or even not working at all.

Why is illogical to use the rom? I mean libraries are stored there (at least for 48,49,50) and if one needs extra space as temporary working space the rom is the only viable way. Surely it is not common (I can barely fill the memory of my 50g), but someone may have done it. Especially the 48 series as far as I know is pretty low on ram.

You have to be careful quoting "48", the 48gII does have Flash that's available to the user, but the 48S and 48G series had fixed ROM and 32k to 128k RAM, with the designated "X" units that could take additional ROM or RAM cards (I have had a 32k RAM card failing).
.

[pier4r]

Thanks panamatik and bartdb. I knew what is a romantic but normally those are hardly read only in later calculators (or even computer bios) but I did not know the additional details, so thanks for sharing.

For the 48,yes.my fault, I tried to recall from memory but mine got corrupted long ago.

[Dave Frederickson]

(05-02-2017 09:57 AM)pier4r Wrote:  I knew what is a [rom] but normally those are hardly read only in later calculators (or even computer bios) but I did not know the additional details, so thanks for sharing.

Let's be clear about this definition. ROM means Read-Only Memory. Modern computers store their BIOS in flash which is Non-Volatile RAM, or NVRAM, meaning that the contents of memory is retained without power. Your computer's BIOS used to be stored in EPROM, a UV-erasable type of ROM.

Our calculators may have NVRAM, but it shouldn't be referred to as ROM.

Dave

[Nigel (UK)]

(05-01-2017 02:03 PM)pier4r Wrote:  Since some of you have a lot of calculators, some quite old an well used maybe you have more experience. For example the ones that often flash the rom of a 34s, did you ever have the internal memory corrupted somewhere that let the calculator be not fully working anymore?

My experience of calculators with memory problems:

• TI-58C bought nine years ago and used occasionally; its RAM stopped working completely. Replacing with a new RAM chip (still available, amazingly enough) fixed the problem.
• HP-25C - RAM failed. Would still do calculations but not store numbers or run programs. "Fixed" it by adding Panamatik's upgrade chip.
• Sinclair Enterprise Programmable - looks beautiful, does calculations slowly and not very accurately, but doesn't store programs. It looks too nice to risk disassembly!
  

I flashed my WP-34S many hundreds of times (literally) and never had any memory problems with it.

Nigel (UK)

[toml_12953]

(05-01-2017 08:30 PM)pier4r Wrote:  

(05-01-2017 06:33 PM)toml_12953 Wrote:  The symptom I see is that the calculator reboots when trying to create a new program or editing an old one. After the reboot the new program is indeed in the list so I know the problem is happening after it's created. It's strange but the exact same symptom occurs in all three calcs.

Tom L

Hmm, so If I understand you correctly, you cannot tell whether is the RAM or the internal storage that creates the problem, am I mistaken?

Does it happen for every new program?

It seems to be triggered by downloading certain programs. Once it starts happening, it happens for every new program until I do a reset. Somehow the downloaded programs show a problem in storage memory.

Tom L

[pier4r]

(05-02-2017 03:33 PM)Dave Frederickson Wrote:  Let's be clear about this definition. ROM means Read-Only Memory. Modern computers store their BIOS in flash which is Non-Volatile RAM, or NVRAM, meaning that the contents of memory is retained without power. Your computer's BIOS used to be stored in EPROM, a UV-erasable type of ROM.

Our calculators may have NVRAM, but it shouldn't be referred to as ROM.

Dave

Well, I appreciate the correction. Thanks again. (and yes, it makes sense, ROM is a bit misused in common usage)