Elektroda.com
Elektroda.com
X

Search our partners

Find the latest content on electronic components. Datasheets.com
Elektroda.com

Will the Sun clear the EPROM?

TechEkspert 11886 35
This content has been translated » The original version can be found here
  • Will the Sun clear the EPROM?
    A long time ago in devices read-only data was stored in EPROMs. These types of memories were electrically programmed and erased with UV radiation. Erasing was possible thanks to a quartz window in the housing (there were also cheaper one-time programmable memories without a window in the housing).

    EPROMs could be found on PC motherboards (BIOS), network cards (booting program), and automation and control devices (they stored the program executed by the microcontroller).
    Most hobbyists had access to or had an EPROM programmer (eg Willem), but very rarely someone had an EPROM eraser. Erasers were equipped with UV-C light sources, tightly closed in an EPROM cassette (UVC is harmful to humans - especially the skin and eyes), better models had a simple timer system that turned off EPROM exposure.


    How did hobbyists deal with EPROM erasing?
    Usually they used the available UVC source, e.g. in the form of a mercury lamp, UV-C germicidal fluorescent lamp or the so-called quartz watches. With a little practice and caution, this "free" method managed to erase the EPROM content. Reading the topics Erasing EPROM memory with UV eyelet and Homemade EPROM eraser you can find surprising suggestions for use pepper shakers and migomat , flash and the sun .

    Will the Sun clear the EPROM?

    Will solar radiation erase EPROM?
    In the last heat wave, I decided to check in practice how long it will take to erase EPROM with sunlight. In the 27C256 Microchip memory specification, I found information that the dose of 15Ws / cm2 should be applied and that a light source with a wavelength of 2537 angstroms (253.7nm UV-C) should be used, with an intensity of 12000uW, the irradiation should last 20 minutes. According to the specifications, the data storage capacity is > 200 years however, I have happened to run into automation devices where the EPROM content has degraded after ~ 20 years.
    UVC radiation is quite well damped by the atmosphere and it is difficult to determine what time of exposure will be needed. According to Wikipedia The sun can erase the EPROM in a week, while a fluorescent lamp (the usual one shining with visible light) in three years. Not without reason on EPROMsprinted paper or aluminum stickers were placed, which, apart from the informative function, blocked the access of UV light that degraded the memory content.


    Link


    For the experiment, EPROMs of unknown origin were obtained, looking at the content of one of the dice, you can see that it came from a device with some language settings.
    Will the Sun clear the EPROM?
    The 27C256 memory was placed on the roof of the building for two weeks. The memory has a capacity of 32KB, which is exactly 32,768 bytes, which gives 262144 bits, ? 262144 = 512, i.e. we can visualize the memory content as a 512x512 pixel image. Bits 0 are marked as a blue pixel, bits 1 as a red pixel, and bits deleted by the sun (0-> 1) as yellow pixels. Below is the "appearance" of the memory before exposure to sunlight and after two weeks of exposure.
    Will the Sun clear the EPROM? Will the Sun clear the EPROM?
    You can see that most of the memory has been deleted, so the method works, however, is quite unpredictable and depends on the current weather. I had a few more EPROMs, I wonder how the memory erasing goes over time. The EPROM was placed on the eastern wall of the building, while the content was read periodically using Arduino Mini and saved on the SD card. A simple code and two '4040 counters made it possible to address all 27C256 memory cells and save the content in 512 byte blocks to subsequent files on the micro SD card. The system was powered by 5V, therefore a 3.3V level converter and stabilizer was used for the SD card.
    Will the Sun clear the EPROM?
    Code for Arduino that reads data from EPROM cyclically:
    Code: c
    Log in, to see the code


    Subsequent dumps of EPROM contents were changed with a simple script in Python to PNG files with a resolution of 512x512 pixels showing the process of memory erasing by solar radiation. The script was run in Python 3.5 using the library to create PNG files PILLOW 3.3.0.

    Python code for converting EPROM data to 512x512 pixel PNG files:
    [syntax=python]
    from PIL import Image
    EPROM_SIZE=32768
    SIZE_X=512
    SIZE_Y=512
    FILES=20

    #base image
    base_file=open("000", "rb")
    chunk = base_file.read(EPROM_SIZE)
    if chunk:
    print("Base memory image saved form file 000.")
    base_image = bytearray(chunk)
    else:
    print("Base memory image saving form file 000 err!")
    base_file.close()


    file_list=open("file_list.txt", "w")
    err_list=open("err_list.txt", "w")
    changes_list=open("changes_list.txt", "w")

    #files processing
    n=0
    e=0
    for f in range(0,FILES):
    in_file=open('{:03d}'.format(f), "rb");
    tmp=n
    print("processing file: "+'{:03d}'.format(f))
    chunk = in_file.read(EPROM_SIZE)
    while chunk:
    raw_image = bytearray(chunk)

    imgSize = (SIZE_X,SIZE_Y)
    img=Image.new('RGB', imgSize, color=0)

    CH=0
    ERR=0
    for y in range(0,SIZE_Y-1):
    for x in range(0, SIZE_X-1):
    bit_val=(raw_image[((y*SIZE_X)+x)//8])&(1

    Cool! Ranking DIY
    Can you write similar article? Send message to me and you will get SD card 64GB.
    About Author
    TechEkspert
    Editor
    Offline 
    TechEkspert wrote 3266 posts with rating 2550, helped 11 times. Been with us since 2014 year.
  • #2
    krru
    Level 32  
    You asked a weird question at the end, because although hardware damage caused by changing the EPROM content does occur, it is usually difficult to pinpoint the cause. This is a long process (you noticed it yourself), it is difficult to damage the memory contents with one 'flash' of some radiation (then the cause would be obvious). While it is easy to cut off the memory from visible light, for example, some X-ray devices nearby (in the sense of a neighboring building) may work when the user of the memory does not even know about them.
    And one more important difference - you checked the erasing process, and you did not check the damaging process - that is, how long will the memory last not until completely wiped, but until the first content change occurs. It is definitely a much smaller dose of radiation.
    One cool aspect can be seen in the picture presented - the memory center has been erased quite effectively, but the extreme upper and lower parts remained uncleaned. Probably it is the result of the shadow cast by the casing on the structure as the sun moves across the sky.
  • #3
    TechEkspert
    Editor
    That's right, it is usually difficult to establish the influence of some radiation as the cause of a failure. Hence my question, I am curious if someone went beyond the area of assumptions and managed to clearly establish such a cause of the fault.
    The problem of the influence of ionizing radiation occurs, for example, in the case of devices working in space, some manufacturers provide systems "radiation hardened" which are used, for example, in satellites.
    The influence of ionizing radiation can also be observed in the case of the CCD matrix, you can find descriptions of attempts to use the camera in a smartphone for such purposes: http://www.pocketmagic.net/ccd-radiation-detectors

    In the second "online" trial, the first reading errors appeared after two weeks (sometimes single "0" bits were read as "1"), the effect intensified over time, you can also see the moment (after almost three weeks) when an avalanche of memory deletion occurs. In both trials, EPROMs with a different structure appearance were used, the first EPROM had a larger, more square structure, the second memory had a smaller, more rectangular structure (seen in the photo with three memories).
    The time of deletion was probably influenced by the weather, as well as the "past" of the acquired memories. For a repeatable experiment, the same memories programmed just before the test should be used. This experiment was not laboratory and you can afford such an approach, the use of EPROMs with their own "secret" was more interesting ;) .

    In the topic mentioned in the material, there is also information about the method of programming "shallow" and "deep" EPROM depending on the time of the programming impulse, this may also affect the erase time and resistance to degradation.
    Interestingly, in the first "offline" attempt, if I removed the EPROM from the roof a few days later, the whole thing would probably be deleted and it would be impossible to notice the blue areas at the beginning and end of the memory.

    As for the shadow cast by the casing, it is visible on the thumbnail of the YT movie, it is surprising how external factors can influence the results of the experiment in an unpredictable way. During the "online" trial, I saw that depending on the time of day the structure was lit to varying degrees.
  • #4
    Jacekser
    Level 21  
    At the moment, it would probably be more useful to gain experience with the durability of writing common FLASH memories which are an extension of the EEPROM memory. Of course, they do not have a window and the need for UV exposure to erase, but I wonder how resistant they are to penetrating environmental radiation. In my practice I have met with spontaneous (?) erasure of this type of memory. EPROM UV memory I happened to meet damaged, but only for electrical reasons (e.g. lack of grounding of a high-power device). As for the noticeable phenomena related to damage to electronic equipment, it happened to me mainly TV series. different with identical damage to the same transistor (BU508 - horizontal deflection control) at about the same time. When the Internet era reigned, because I am interested in space and astronomy, I associated it with info on the activity of the Sun.
  • #5
    TechEkspert
    Editor
    Great idea, I would like to read on elektroda.pl about the results of the test with EEPROM and FLASH memories that have been exposed to X-rays or some form of ionizing radiation that would reach the structure in the housing.
    Unfortunately, I do not have access to such sources and the experience to perform such tests in a safe manner.

    I see that the topic was researched by NASA: https://nepp.nasa.gov/workshops/etw2012/talks/Tuesday/T13_Irom_NVM_Radiation.pdf
    https://www.esa.int/gsp/ACT/doc/EVENTS/memris...rois_2015-04-30%20ResistiveMem%20workshop.pdf
  • #6
    Jacekser
    Level 21  
    You can try in the same way as with this EPROM. Put one FLASH cube written with a homogeneous matrix 010101 ... to solar exposure and not only - generally put "up to the sky" (as with a friend - on the roof), and hide the other identical one at home in a dark box "under the roof". Testing method may be similar. And be patient because we will probably wait a long time for some effect.
  • #7
    deus.ex.machina
    Level 32  
    A few words from me:

    - firstly, the erase efficiency depends on the design of the memory cell (size, geometry, technology, etc.)
    - secondly, on the method of programming (the so-called fast algorithms offer a shorter life than a slow algorithm)
    - thirdly, it is worth remembering that the EPROM leads should be compact (ESD foam is enough - electrically conductive) - while exposing the memory, quite a significant electric charge is generated which, in an extremely unfavorable coincidence, may damage the memory.
  • #8
    Sofeicz
    Level 18  
    To the author on the linguistic issue.
    "Will the EPROM clear the sun?"
    In which language is this?
    It turns out that the Sun will be cleared by EPROM.
    Author, correct it, because it beats the eyes!
  • #9
    TechEkspert
    Editor
    @Sofeicz Nice setback, but not mine because as the author of this material I gave it a title "EPROM memory vs the sun" (which can be checked, for example, on FB elektroda.pl https://www.facebook.com/elektroda/). Maybe the original title wasn't perfect, but it was definitely better than what you saw. I'm changing to "Will the Sun clear the EPROM" it will be something in the middle ...
    Thanks for the info.

    PS
    In addition, the changed topic you noticed has now found its way to the main page of elektroda.pl and to e-mail notifications :D we will wait in some time should it change.
  • #10
    Jacekser
    Level 21  
    Sorry for the light OT, but I remembered such a fact. In the 1980s I worked on a digital terminal program. The program was usually placed in EPROM dice. As it was problematic then to obtain components (green currency), components from KDL were selected (internships know what's going on). The chips from Big Brother were the most interesting. Apart from the fact that they had a 2.5 mm raster, they either could not be deleted after programming or could not be programmed at all. There were rumors that these versions are resistant to atomic attack and only a near explosion of a neutron bomb can delete it . :wink:
  • #11
    TechEkspert
    Editor
    I wonder what error in creating the structure caused such effects, i.e. either the inability to introduce / hold the charge for the gate, or the inability to remove the charge. Wrong type of material / technological process?
  • #12
    Jacekser
    Level 21  
    They can be watched on the net (USSR EPROM) And how much gold they can make for jewelry.
  • #13
    deus.ex.machina
    Level 32  
    Jacekser wrote:
    The chips from Big Brother were the most interesting. Apart from the fact that they had a 2.5mm raster, they either could not be deleted after programming or could not be programmed at all. There were rumors that these versions are resistant to atomic attack and only a close explosion of a neutron bomb can delete it . :wink:


    Hehehe, I confirm, I have a few Soviet EPROMs and yes, they were very hard to delete (several hours of exposure), as for programming - it depends on how they were programmed - 2KiB ( ;) 573??2 - 2716 counterparts actually after a few attempts it was OK ...
    As for the raster - the so-called metric inch (25mm), it was not a problem in the case of DIL 24/28 pin - it was becoming a problem in the case of DIL 40.
  • #14
    cb656
    Level 10  
    Quote:
    Have you ever had any damage to your electronics caused by UV, X or other ionizing radiation?

    Yes, yes. At the time, Eltra had a problem with the new product, it was about the turret just put into production. Massive information was poured into the service that the towers were behaving strangely - the volume changes automatically, there were device shutdowns or switching radio stations, signal sources, etc.
    After analyzing the problem in the production service cell, it turned out that it was to blame ... the sun and other sources of infrared radiation, for example a light bulb shining closely in a lamp on a desk.
    The pilot sensor, located on the main plate of the tower, after being illuminated by the sun or a light bulb, generated a random noise of pulses, among which sequences that triggered this or that appeared randomly. As if someone were pressing buttons on a remote control.
    The problem was solved by the method of color filters in front of the sensor eye, selectively
    transmitting a selected infrared band.
    It did not eliminate the phenomenon 100%, but it significantly limited it.
    PS It was the end of the 90's of the last century.
  • #15
    CMS
    Administrator of HydePark
    As if I have a lot of EPROMs with "eyelets". They come from printers (like probably the one from the tests). I can give some for rehearsals. They can be programmed, or I can delete them, because I have a professional "eraser".

    Best regards.
    CMS
  • #16
    Hektar Zahler
    Level 34  
    krru wrote:
    Although hardware damage caused by changing the EPROM content does occur, it is usually difficult to pinpoint the cause.
    Remember that not only UV destroys EPROM, because interference is definitely more important !!!
    This is a parallel rail.
    There have been failures many times before, when someone who did a great job did not think about it.


    A personal example I've had a few.

    One time I left something under the 12V gel - this one discharged and the equipment was to be thrown away (secured ROM).
    More than 10 years passed before I got the device back on the Allegro (artisanal work).

    An already rotten path or moisture / corrosion (extremely common) can be the reasons for this.
    These are the realities of reality from the hobbies of a young technician to cases of damage in factories.
  • #18
    deus.ex.machina
    Level 32  
    carrot wrote:
    Interesting article Link


    It looks like an urban legend and the mention of the Oder as a DSP-capable computer (radar signal) proves a very vivid imagination.

    I dare say that the F-117 also uses processors from the mid 80's or early 90's. New technologies are used with extreme caution in military technology. I know the radar data processing system (15 radar stations) - all the functionality is implemented on the DSP TI (320C30 + 320C40) and there is a lot of it.
  • #19
    CMS
    Administrator of HydePark
    carrot wrote:
    Interesting article Link


    Interesting article indeed.
    I wonder how these pigs / fudges acted on people, since the radiation could from several meters and in a short time (train travel) have such a significant impact on the operation of the computer, which I suspect was in a thick metal housing.

    Best regards.
    CMS
  • #20
    deus.ex.machina
    Level 32  
    The article is unbelievable - the author did not even want to illustrate it with a photo of the SM1800 machine, only the photo of the DWK-2 set was used ... there is nothing to write about the rest of the "sensation" ...
  • #21
    Hektar Zahler
    Level 34  
    Before the speaker, read how the Win95.CIH mechanism works. It cleverly describes what you will do with such an EPROM. I had the opportunity to change such a BIOS. The method described may be questionable, but don't deny it when there are other reasons for concern.

    You also have to separate deletion from destroying content are two completely different things.
  • #22
    HD-VIDEO
    Level 40  
    Hektar Zahler wrote:
    describes what to do with such an EPROM and


    CIH will do nothing to EPROM
  • #23
    deus.ex.machina
    Level 32  
    Hektar Zahler wrote:
    Before the speaker, read how the Win95.CIH mechanism works. It cleverly describes what you will do with such an EPROM. I had the opportunity to change such a BIOS. The method described may be questionable, but don't deny it when there are other reasons for concern.

    You also have to separate deletion from destroying content are two completely different things.


    Do not confuse FLASH with EPROM - FLASH is designed in such a way as to program it in the system and the EPROM requires an external programmer (the need to use higher voltages, for example) although I could imagine a system with a built-in EPROM programmer, however, in such a situation the problem is usually handled differently by installing NVRAM be simulating the ROM in RAM and loading the correct code with a small bootloader.
  • #24
    TechEkspert
    Editor
    As for the material Link , maybe the disruptions in the system's operation resulted from ionizing dust which covered the wagons, or maybe the reason was still another, e.g. an electric locomotive pulling this train, emitted radio interference of a specific spectrum or power?
  • #25
    Hektar Zahler
    Level 34  
    @ deus.ex.machina I know the theory, but it is exceptionally capable of undermining. Old things getting along with each other in the vast maze of circuits. It is about their multitude, because the chances of chemical damage increase - corroded circuits, especially connectors, put the contents at the mercy of the highest. Let's add power surges and the problem is ready. The vast majority of them will be eternal as it was meant to be, but not always. I will tell you something from practice, chemistry likes to stick to electricity.
  • #26
    deus.ex.machina
    Level 32  
    Hektar Zahler wrote:
    @ deus.ex.machina I know the theory, but it is exceptionally capable of undermining. Old things getting along with each other in the vast maze of circuits. It is about their multitude, because the chances of chemical damage increase - corroded circuits, especially connectors, put the contents at the mercy of the highest. Let's add power surges and the problem is ready. The vast majority of them will be eternal as it was meant to be, but not always. I will tell you something from practice, chemistry likes to stick to electricity.


    Well, once again - at 45 years of age, I say that no CIH will reprogram / damage EPROM in the style of 27128 - 27512 (because this is what BIOS used to be like). It's not raining because angels pee (I've heard such interpretations of this phenomenon) ...
  • #27
    DJ_Opornik
    Level 20  
    Jacekser wrote:
    At the moment, it would probably be more useful to gain experience with the durability of writing common FLASH memories which are an extension of the EEPROM memory. Of course, they do not have a window and the need for UV exposure to erase, but I wonder how resistant they are to penetrating environmental radiation. spontaneous (?) deletion of this type of memory ........


    I will heat the cutlet because I am curious about the reasons for the self-deletion of FLASH.
    In my life, I have encountered 93C46 memory erasing many times in a computer from a certain machine (excavator). It stores the settings related to the controller configuration (equipment and available control functions in the machine). There were two variants of damage:
    1. Self-resetting for reasons unknown to me - in these cases, the customer reported machine malfunction - delivery of a driver and programming of the full FF memory.
    2. Deletion at the stage of disassembly / desoldering (maybe some "Zeus") - a few cases that happened to me during the transplant (reworking a functional controller from the donor machine for the recipient's needs). When, after inserting the programmer, the reading gave FFs, I knew that it fell out when disassembling
  • #28
    Jacekser
    Level 21  
    DJ_Opornik wrote:
    ... I have encountered 93C46 memory erasing many times in my life ...

    The same thing happened in CRT monitors several years ago - and I think I remember on this type of memory. Reason-? I suspected punctures in the lamp. It was troublesome to get both the charge and the service mode. I even looked for some way to replace them with a PROM with a serial bus but I did not find and gave up the service of such a damaged CRT.
    I got the application about the discharge in the lamp erasing the memory after such an incident - the lightning hit the high block next to which I live (it went through the cable TV); on the second day I had a lot of reports on the website about out of tune TVs; during repairs, either the serial EEPROM or just re-program the programs.
  • #29
    Hektar Zahler
    Level 34  
    Are you talking about deletion, or maybe a damage that cannot be saved again? Are you confident in a programmer that can read the memory with the correct or non-standard settings? I am asking because you should not identify the programmer with the device, because most often (though not necessarily) these are completely different systems and parameters. For example, very often with one programmer I easily read the memory that is not working in the device, and with the other programmer I do not read it completely or not at all. As for the causes of memory, microwires are simply not secured against interference, and this is common for this very reason, so it's no wonder. Strange, however, was the susceptibility to I2C memory distortions, with which I encountered several times, for example the Blaupunt RCR-27 radio and a damaged (I do not remember exactly) tuner, which, despite blocking the bus, somehow functioned there. As a result, there was a self-edition of the eeprom, whose part of the map was responsible for the options on this radio that did not exist, and existed in the twin model. And this is how the rear speaker adjustment of the radio, which only has front speakers, has come. After finding the cause and fixing the fault, the situation did not happen again.

    Added after 45 [minutes]:

    Coming back to the deletion of eprom UV in the system in which it is located (eg Win.CIH) - for doubters that it is impossible, and yet it is.

    Suppose we have programmed:

    1.01 (hex) - 001 (bin)
    2.05 (hex) - 101 (bin)

    Can I re-save eprom UV for example 03 (hex)?

    Ad.1 001 (bin) -> 011 (bin) = 001 (bin)
    - I can not

    Ad.2 101 (bin) -> 011 (bin) = 001 (bin)
    - YES I can, because I will get 01 (hex)

    Can I delete the eprom UV Win.CIH?
    - NO or NOT necessarily

    Can I destroy the UV Win.CIH eprom?
    - YES, because the virus generates random strings to do this.

    I saw it personally, because I lost a motherboard with a Pentium 166MMX processor this way and I have this board as a souvenir until today.
  • #30
    DJ_Opornik
    Level 20  
    I am talking about erasing - the memory full of FFs was then able to be correctly saved. There were also a few cases of destruction, but I am excluding these as unrelated to the topic.