Elektroda.com
Elektroda.com
X
Elektroda.com

Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits

manekinen 224494 342
This content has been translated flag-pl » flag-en View the original version here.
  • Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits
    Atmega fusebit doctor, as the name says, is a device for repairing mismatched fusebits in microcontrollers from the AVR family. The biggest problems are setting the wrong clock source (CKSEL fusebits), disabling SPI programming (SPIEN fusebit), or setting the reset pin to I / O mode (RSTDISBL fusebit). This simple device will repair the microcontroller by giving it factory settings in a fraction of a second.

    While in the first case you can use a clock generator or an RC / quartz generator, in the second and third case, it is not possible to bring the microcontroller back to life using the SPI serial programmer. Few people decide to build a parallel programmer, because it is inconvenient to use, and because it is cheaper to buy a new microcontroller than to play some repairs.

    The presented device uses the possibility of parallel (and serial programming from version 2.03) high voltage. The device is extremely simple and cheap to build, you only need ATmega8 as a doctor, two LEDs, a jumper, a stabilizer, transistors. Signatures of 96 AVR microcontrollers and their factory fusebits are saved in the memory, just insert the "released" uC into the socket and press the button START and the system will perform the required operation and our patient will be brought back to life. The entire procedure takes a fraction of a second.

    There are three sockets on the main board for processors that are pin compatible with atmega8, atmega16, and attiny2313, i.e. the most popular ones. In addition, the board has a female goldpin connector with all necessary signals led out, for connecting adapters:
    " # 1 HVPP adapter "As an HVPP extension for 20pin Attiny26 and 40pin Atmega8515 compatible processors
    " HVSP adapter "For HVSP 8pin and 14 pin attiny processors, which cannot be programmed in parallel due to too few pins.
    It is possible to make your own additional adapters for other types of processors in DIP or SMD casings. However, you do not need to make an adapter to repair one chip, you can do it with a contact plate by connecting the signals to the appropriate pins. As? Check the datasheet of your AVR, go to "memory programming" and then "parallel programming" - the names of the signals and pins as on the tray. All pins are signed under the DIP40 socket, and in the attachment there is also a project of an "empty" adapter.

    Single-sided plate, 55mm x 92mm. Several jumpers should be soldered on the top part, or the board can also be double-sided. 12V stabilized power supply. Resistors R7 to R23 can have values from 100 ohms to 1K, I suggest rather 330ohm. It should be remembered that the three pins of the bits from the data line are also used by the ISP programmer to update the program - the device will not work properly if, for example, we solder to them with the programmer. The board also has a connector described as RS232, it is a UART output, by connecting to it (38000bps), we will learn everything about the course of the repair operation - an example screenshot below. Of course, the terminal is not necessary, we will find out everything from the diodes themselves :)
    Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits

    ATTENTION! When installing the DIP40 socket, remove the metal connectors from pin 29 to 37 from it! The paths passing in these places cannot be electrically connected to the pins of the inserted processor, and they run there to simplify the board itself. I have marked which pins in the picture below:
    Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits

    Diode markings :
    glows green - fusebits fixed and verified, system fixed. If the lockbit protection is set, it only checks whether the fuses correspond to the factory ones, and if so, it will also light this diode.
    shines red - problem with reading the signature, no system or no signature in the database.
    blinking green - signature read, fusebits do not match the factory ones, but lockbits are set and you need to allow memory wiping to fix them (read on).
    flashing red - signature read, lockbits disabled, but new fusebits cannot be saved for some reason.

    Jumper ALLOW ERASE allows you to erase the entire memory in the case of set Lockbits (without deleting them, it is not possible to change the Fusebits). After connecting the system and pressing the START button, the program initiates the parallel programming mode, or if an HVSP adapter is used - initiates the serial programming mode. The patient responds with a high status on the RDY / BSY pin. The first thing the doctor does is wipe all memory if allowed by the user. After that, it reads the signature of the connected microcontroller and checks if it is able to handle it. Then the lockbits are checked, and if they do not block access, the doctor reads the fusebits and compares them with the factory ones stored in the database. If they differ, it saves the factory settings, taking into account whether the given patient model has extended fusebits or not. There are also processors with only one fusebit byte and that is also included. At the end, the program verifies their correctness and lights the appropriate LED. Information via rs232 is sent on a regular basis.

    The program was written on the basis of the description of parallel (and serial) programming contained in most AVR microcontroller catalog notices, (memory programming - parallel / serial programming).

    The project started in 2008, but abandoned due to lack of time, now redone.

    Fusebits: Internal 8MHz clock, and EESAVE bit on.

    As it is 2in1 (HVPP and HVSP), the 8kB Atmega8 memory turned out to be insufficient and not all the bells and whistles came in ...
    1.Not all processors display names after rs232, but the most popular ones. However, this has no effect on the operation of the system.
    2. Some text sent by rs232 has been stored in EEPROM. Even if you are not going to use this output, you MUST program the eeprom with EEP.BIN or EEP.HEX! Using a chip with an unprogrammed or badly programmed eeprom will do more harm than good!



    If the green LED is on, you can be 100% sure that the fusebits have been reset. If the chip still does not correspond to the normal ISP programmer, it means that it has a damaged hardware SPI, it has other damage, or you have uploaded the eeprom. If the red LED is flashing, the only thing you can do is check what the doctor is sending via rs232 - then I can help you fix the problem.

    Number of processors supported: 96
    Number of supported sockets: 53

    The rest are SMD cases, there are no adapters for them yet.

    Full list of supported processors:
    (tested marked green)

    1kB:
    AT90s1200, Attiny11, Attiny12, Attiny13 , Attiny15
    2kB:
    Attiny 2313 , Attiny26, Attiny261, Attiny28, AT90s2333, Attiny22, Attiny25 , AT90s2323, AT90s2343
    4kB:
    Atmega48, Atmega48P, Attiny461, Attiny43U, Attiny4313, Attiny48, AT90s4433, AT90s4414, AT90s4434, Attiny45
    8kB:
    Atmega8515, Atmega8535 , Atmega8 , Atmega88 , Atmega88P, AT90pwm1, AT90pwm2, AT90pwm2B, AT90pwm3, AT90pwm3B, AT90pwm81, AT90usb82, Attiny861, Attiny88, Attiny85
    16kB:
    Atmega16 , Atmega16U4, Atmega16M1, Atmega161, Atmega162, Atmega163, Atmega164, Atmega164P, Atmega165, Atmega168 , Atmega168P, Atmega169, AT90pwm216, AT90pwm316, AT90usb162
    32kB:
    Atmega32 , Atmega32U4, Atmega32M1, Atmega324, Atmega324P, Atmega325, Atmega3250, Atmega325P, Atmega3250P, Atmega328, Atmega328P, Atmega329, Atmega3290, AT90can32
    64kB:
    Atmega64, Atmega64M1, Atmega649, Atmega6490, Atmega640, Atmega644, Atmega644P, Atmega645, Atmega6450, AT90usb646, AT90usb647, AT90can64
    128kB:
    Atmega103, Atmega128, Atmega1280, Atmega1281, Atmega1284, Atmega1284P, AT90usb1286, AT90usb1287, AT90can128
    256kB:
    Atmega2560, Atmega2561

    Gallery:
    Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits

    Eagle 3D render:
    Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits

    Prototype - Atmega8 reads Atmegi32 signature and fuses with damaged SPI, and displays info on LCD:
    Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits

    Adapters:
    Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits

    Compatibility with stands - what we put where:
    Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits

    In the attachment all files:
    Design of the board and adapters (eagle 5.4.0) version 2.0d: BRD, SCH, and printable PDF versions
    Object code (bascom 1.11.9.0) version 2.03: HEX and BIN
    Eeprom image version 2.03: HEX and BIN

    A brief explanation for those not in the topic:
    HVPP = high voltage parallel programming = high voltage parallel programming.
    HVSP = high voltage serial programming = high voltage serial programming.
    So programming methods that allow you to get to the processor despite the disabled reset or isp, i.e. our dead man. Serial programming is used by processors with a small number of pins, because this type requires only a few lines. Parallel programming is used by processors that have at least 20 pins, i.e. everything above tiny2313 inclusive.

    The above description and attachment contain all updates,while the full description and all attachments with different hardware and software versions - for the inquisitive - are available on the project website:
    http://diy.elektroda.eu/atmega-fusebit-doctor-hvpp/
    There are also a lot of comments there, I recommend reading if you have any problems, but here I am also happy to answer if I have time. The system was made by a lot of people, so it is tested and works properly.




    _______________________________
    NOTE UPDATE # 11 - HERE -
    -145 supported chips
    -sds for Atmega88, Atmega88P, Atmega168, Atmega168P, Atmega328, Atmega328P.
    - two-way communication via UART, you can e.g. send your own papers
    -dooo much lots of fixed bugs, also on the board

    A shortened list of changes and improvements can be found in the changelog, but I recommend reading the whole topic (only 5 pages).
    FAQ - frequently asked questions and their answers.
    _______________________________

    Cool? Ranking DIY
    Can you write similar article? Send message to me and you will get SD card 64GB.
    About Author
    manekinen
    Level 29  
    Offline 
    manekinen wrote 1629 posts with rating 2315, helped 75 times. Live in city Kętrzyn. Been with us since 2006 year.
  • #2
    TormentedSoul
    Level 10  
    Cool! I just recently dumped 2 ATmegi8 with these fusebits. ;)
  • #3
    profesorek_96
    Level 16  
    Nice design, very useful.
    I have a question about pcbs that are very nicely made?
    Did you make the layer with the inscriptions and the tile itself with an iron or a laminator?
  • #4
    Duch__
    Level 31  
    Or maybe the author would convert the main program into some other processor than Atmega8. As you know, it's hard to get at a reasonable price, and there are many other procks in DIP28 housings, with the same pinology.
  • #5
    Jdsoul
    Level 23  
    Can this toy be used to unlock the processor in the TSOP housing - are there any adapter designs ???
  • #6
    krzysiek_p
    Level 14  
    Duch__ wrote:
    Or maybe the author would convert the main program into some other processor than Atmega8. As you know, it's hard to get at a reasonable price, and there are many other procks in DIP28 housings, with the same pinology.

    I also support the quoted one - especially since I haven't used m8 for a long time and the only one I have now is permanently soldered in the logic analyzer, so I won't use it. But I have 48 (which is out because of flash), 88 and 168. It would be nice to put together something like this, especially since my m16 is blocked. And there is also the possibility of using the uC reset pin jao gpio.
  • #7
    piotrva
    VIP Meritorious for electroda.pl
    Jdsoul, you can make an adapter, but the cost would be enormous (CLIP-type sockets or special top-mounted contact connectors cost a lot, and getting them for ordinary electronics is often close to a miracle). I personally saw such a stand live only once, the price is only ~ PLN 130 if I remember correctly. You can always connect the wires to the PCB where you have the uP soldered in the TSOP, as long as other elements do not interfere with the work. or you can use something like this: http://diy.elektroda.eu/przystawka-smd-do-programatora/#more-634
    As for the design itself, it's very useful. And here is my request, because I don't have atmega8 but other uP, so could you please include the source code of the software? possibly a compilation for atmega168 (I would use such a uP for these purposes, I have a pair of meg2560 blocked :D )
  • #8
    omicronNs
    Level 21  
    Well, wouldn't it be better to make a parallel programmer? You will do more with it, because by the way you can program the processor, as you need the reset pin as and / o there is no problem with changing the program when uK is already programmed, and after all, the parallel programmer looks simpler in construction because you do not need a microcontroller. The only disadvantage of the parallel programmer is that you have to play with the computer with each unlocking, and the LPT port can cause some problems.
  • #9
    piotrva
    VIP Meritorious for electroda.pl
    well, the most interesting solution, although I don't know if it would work, would be the ISP-> HVPP / HVSP converter
    I personally would invest in prog. parallel as long as it wasn't on LPT or if I found virtual LPT on USP
  • #10
    Gostek
    Level 17  
    Hmm, a used STK500 will probably be cheaper ....
  • #11
    leonow32

    Level 30  
    How do you do the inscriptions on the tiles? This is a rosin coated toner with a solvent? What?
  • #13
    pwlu10
    Level 12  
    "cheap"? After all, Atmega8 now costs even PLN 15
  • #14
    GruX
    Level 18  
    Gostek wrote:
    Hmm, a used STK500 will probably be cheaper ....


    Yes, and you will definitely unlock fusebits;]
  • #15
    piotrva
    VIP Meritorious for electroda.pl
    stk500 is an ISP, not an HV
    EDIT: Incorrect information
  • #17
    ZbeeGin
    Level 39  
    GruX wrote:
    Yes, and you will definitely unlock fusebits;]

    piotrva wrote:
    stk500 is an ISP, not an HV

    Everyone should realize that the STK500 is prototype board which, in addition to the startup possibilities contains too parallel programmer . You can find a copy of it on the web. Therefore, please refresh your STK500 news before listing any more nonsense .

    manekinen wrote:
    As it is 2in1 (HVPP and HVSP), the 8kB Atmega8 memory turned out to be insufficient and not all the bells and whistles came in ...

    As you wrote software in BASCOM Basic, unfortunately everything could not fit. I propose to rewrite the software in C dialect (AVR GCC). As practice shows, the resulting program will shorten by a few percent, freeing up some memory for use.

    pwlu10 wrote:
    "cheap"? After all, Atmega8 now costs even PLN 15

    That is why it is worth removing the blinders and looking around for other systems from the AVR family, not only ATMega8 / Tiny2313 / Mega16. It becomes boring to watch designs done in one hoof.
  • #18
    piotrva
    VIP Meritorious for electroda.pl
    ZbeeGin, right, I meant AVR PROG USB partially compatible with stk500. AVR PROG USB only has an ISP.
    I found atmega32 for the price of 8, also 8 megi does not pay off at all.
    I still keep asking the author for the source codes as long as he wants to share them
  • #19
    MasMas
    Level 16  
    A very nice device. I myself recently "hit" the m8 with a bad clock fuse, but fortunately for me the external RC saved me. A very ingenious device.
    I have a question: In the description it writes that if it is the fault of the lockbits, you need to give a jumper that will allow you to completely erase the memory ... And now I wonder. After all, there is a setting in lockbits that prevents reprogramming or reading the program memory. So if the author (theoretically any person who knows electronics) can clean these lockbits and get to the processor, what is this security for? What does it protect against when they can be turned off so easily?
    And this is such that when you clear these lockbits, you also clear the entire program memory and therefore the program cannot be read?
  • #20
    piotrva
    VIP Meritorious for electroda.pl
    lockbits prevent uP reading and programming
    they are designed to protect the software against unauthorized copying (e.g. you buy a device, copy electronics, download the program and upload it to a new device).
    Lockbits can only be removed via Chip Erease, i.e. erasing the Flash memory and Eeprom of the chip, exactly as you say at the end.
  • #21
    PO.
    Level 20  
    That you wanted to :) ... I have two sockets soldered on the spider - dip28 and dip40 and I change the necessary lines with jumpers and then a simple load for three bytes and one unlocks the other (which depends on the needs). The rest of the games of course in the ISP after that.
  • #22
    Duch__
    Level 31  
    PO. wrote:
    That you wanted to :) ... I have two sockets soldered on the spider - dip28 and dip40 and I change the necessary lines with jumpers and then a simple load for three bytes and one unlocks the other (which depends on the needs). The rest of the games of course in the ISP after that.


    Could a friend expand his thought more widely?
  • #23
    PO.
    Level 20  
    Duch__ wrote:
    PO. wrote:
    That you wanted to :) ... I have two sockets soldered on the spider - dip28 and dip40 and I change the necessary lines with jumpers and then a simple load for three bytes and one unlocks the other (which depends on the needs). The rest of the games of course in the ISP after that.


    Could a friend expand his thought more widely?


    Nothing to brag about, but on request ...
    Prock descriptions are in datasheets, most of the lines are on the same pins so you don't even need redundant lines, just one or two, the rest depends on which processor is sending.
    ISP derived for each separately. Jumpers from resets. I do not remember the reset control at the moment, which I unlocked, you have to re-solder if necessary or add jumpers. I made diodes for information purposes for power supply, resets and signaling. Power from the computer and the isp programmer, loaded into the running program, the program waits for a while and then sends what you need to the blocked one.
    Just unlocking fusebits is three bytes to send + control. Since lockbits cannot be unlocked anyway, the erase chip can be done after the ISP.

    Used for procków in these housings only, so obviously less functionality than the presented project.

    Thought:
    Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits
  • #24
    biglolo94
    Level 16  
    Great design.
    I am about to save my ATMEGE168, but I have a problem and I would ask the author to provide the exact fusebit settings he used in the heart of the whole system, i.e. ATMEGA8.

    Below is my fusebit setting.
  • #25
    raptor37
    Level 12  
    Hello,
    i made this layout. I have 4 atmegi8 blocked. When I put in the first one, after pressing the button, no diode lights up, when I put in the second one, the red diode is on (as if this atmega was not there), after inserting the next two atmegs, the green diode lights up. However, the atmeg is still not programmable. I tried to see what it throws on the RS232, however you can only see the bushes. I tried to set the speed as on the project page 38000, but it was not possible to set this speed. So I set 38400, there are bushes still ...
    I checked the board several times, I found a few short circuits, but it did not help.
    I also programmed the "doctor" circuit several times.
    I set the grounds so that the system would run on the internal 8MHz generator CKSEL = 0100. I marked the EESAVE bit and unchecked it in no case.
    Does anyone have any idea what to do with this arrangement? What could be the reason?
  • #26
    Gostek
    Level 17  
    GruX wrote:
    Gostek wrote:
    Hmm, a used STK500 will probably be cheaper ....


    Yes, and you will definitely unlock fusebits;]


    :) And the STK500 or 600 used to be it? These are kits that have all the possibilities of programming ... (i.e. almost, because PDI, for example, is needed with AVRISP or JTAGICE)
  • #27
    manekinen
    Level 29  
    Hello!
    Duch__ wrote:
    Or maybe the author would convert the main program into some other processor than Atmega8. As you know, it's hard to get at a reasonable price, and there are many other procks in DIP28 housings, with the same pinology

    krzysiek_p wrote:
    I also support the quoted one - especially since I haven't used m8 for a long time and the only one I have now is permanently soldered in the logic analyzer, so I won't use it. But I have 48 (which is out because of flash), 88 and 168. It would be nice to put together something like this, especially since my m16 is blocked. And there is also the possibility of using the uC reset pin jao gpio.

    piotrva wrote:
    And here is my request, because I don't have atmega8 but other uP, so could you please include the source code of the software? possibly a compilation for atmega168 (I would use such a uP for these purposes, I have a pair of meg2560 blocked

    ZbeeGin wrote:
    It becomes boring to watch designs done in one hoof.


    So why atmega8? When I started the project, these systems cost PLN 3.50 / pcs. Unfortunately they went up, it's hard, I will not be changing boards now to fit a different processor. Of course, you can use Atmega88, 168, or another compatible with the appropriate amount of memory. I do not provide the source codes, on request I can compile the program for a different processor, according to the provided scheme. I already compiled on M88 and M32, check out the comments on my website. Unfortunately, the authors of these comments did not indicate whether the programs were working properly (and I did not test them in practice). So go ahead and write the program, no problem recompiling.

    However, when it comes to some insight into the sources, I made available the sources of the previous "attiny fusebit doctor" system that supports only tiny HVSP Procki - you can download here https://www.elektroda.pl/rtvforum/viewtopic.php?p=8055689#8055689 Cała the initiation is the same, the principle is the same, the only difference is serial and not parallel transmission.

    Jdsoul wrote:
    Can this toy be used to unlock the processor in the TSOP housing - are there any adapter designs ???

    Of course, there are no contraindications. But as I wrote in the description, there are no adapters for SMD and probably won't be. You will have to solder directly to the chip or create your own adapter using the "empty" adapter design included in the package.

    piotrva wrote:

    It is only up to the ISP where there are three lines. This solution works, it is enough to press the chip well to the board. However, here where we have more than 20 lines, it will be a big problem to tighten the system so that they all have a good connection :(

    piotrva wrote:
    well, the most interesting solution, although I don't know if it would work, would be the ISP-> HVPP / HVSP converter

    I thought about it, about such a translation adapter for the ISP programmer - creating it will not be more difficult than creating such an independent unblocker, let alone relying on it. It is enough to receive commands and addresses from the ISP, recognize what and send it in parallel. I am not sure if the order is the same, maybe there are some slight differences.

    leonow32 wrote:
    How do you do the inscriptions on the tiles? This is a rosin coated toner with a solvent? What?

    A colleague below gave the link, thank you, everything is nicely explained there :)
    http://diy.elektroda.eu/plytki-drukowane-od-a-do-z/

    ZbeeGin wrote:
    As you wrote software in BASCOM Basic, unfortunately everything could not fit. I propose to rewrite the software in C dialect (AVR GCC). As practice shows, the resulting program will shorten by a few percent, freeing up some memory for use.
    These messages by RS should be treated rather as such an addition, and all these prock names take up a huge amount of flash. I optimized the program as much as I could, I also asked here on the electrode and, without any results, I could not reduce the program any more. Here https://www.elektroda.pl/rtvforum/topic1684304.html is the topic. You can see a fragment showing how the program "searches" for the appropriate probe after the next signature bytes, unfortunately I could not solve it otherwise, if someone has an idea, he can share it ;)

    MasMas wrote:
    In the description, he writes that if it is the fault of the lockbits, you need to give a jumper that will allow you to completely erase the memory ... And now I wonder. After all, there is a setting in lockbits that prevents reprogramming or reading the program memory. So if the author (theoretically any person who knows electronics) can clean these lockbits and get to the processor, what is this security for? What does it protect against when they can be turned off so easily?
    And this is such that when you clear these lockbits, you also clear the entire program memory and therefore the program cannot be read?

    Completely erase memory, that is, delete everything. After that, you will, of course, unlock the security and select the memory, but what if it is already empty :) This protection is to protect the batch against reading and copying, and not the processor itself against reprogramming and use :)

    PO wrote:
    I have two sockets soldered on the spider - dip28 and dip40 and I change the necessary lines with jumpers and then a simple load for three bytes and one unlocks the other (which depends on the needs). The rest of the games of course in the ISP after that.

    Oh, that's what I started with :) There are also a lot of such dedicated chips (along with batches) to unlock only one processor model. Such a circuit sends only bytes with fusebits, simple as a wire and at the same time fulfills its role. My system is extensive, it differs in that it reads the signature first so that it knows what it is repairing, and after the repair, it also verifies the fuses. Placing almost 100 signatures and tea leaves and handling it, however, will take some memory. And such a simple system that everyone can do is 15 minutes with the nose in the "parallel programming" section of the catalog note, I recommend you to try it :)

    biglolo94 wrote:
    Great design.
    I am about to save my ATMEGE168, but I have a problem and I would ask the author to provide the exact fusebit settings he used in the heart of the whole system, i.e. ATMEGA8.

    Below is my fusebit setting.

    : for an attachment in the form of a 2MB BMP man! Anyway, set wrong, turn on the High G bit (i.e. eesave) so that the program does not delete the eeprom when reprogramming.

    raptor37 wrote:
    Hello,
    i made this layout. I have 4 atmegi8 blocked. When I put in the first one, after pressing the button, no diode lights up, when I put in the second one, the red diode is on (as if this atmega was not there), after inserting the next two atmegs, the green diode lights up. However, the atmeg is still not programmable. I tried to see what it throws on the RS232, however, you can only see kszaki. I tried to set the speed as on the project page 38000, but it was not possible to set this speed. So I set 38400, there are still kszaki ...
    I checked the board several times, I found a few short circuits, but it did not help.
    I also programmed the "doctor" circuit several times.
    I set the grounds so that the system would run on the internal 8MHz generator CKSEL = 0100. I marked the EESAVE bit and unchecked it in no case.
    Does anyone have any idea what to do with this arrangement? What could be the reason?

    The first and second circuits are probably dead, 3 and 4 are good and the doctor's circuit is working properly, if it lights up the green diode, it means that it reads the signature well - the problem is on your side. Not programmed or badly programmed eeprom. FIRST you set up the EESAVE fusebit and then program flash and eeprom with the appropriate file. As for the speed of 38400, this is correct, please indicate where I wrote 38000 - maybe a typo
  • #28
    anelrob
    Level 11  
    Hello.
    I made the layout according to your design and it works great.

    I only have a question whether the counter that counts the repaired microcontrollers gives any limitations or is it just such additional information.

    Atmega Fusebit Doctor (HVPP + HVSP) - fix fusebits
  • #29
    szymanczuk
    Level 2  
    I quickly built a spider to test and unblock a few AVRs by the way, it helped me with ATMEGA644p ATMEGA16, ATMEGA8, and tiny2313, so a project worth recommending.
  • #30
    Duch__
    Level 31  
    manekinen wrote:

    So why atmega8? When I started the project, these systems cost PLN 3.50 / pcs. Unfortunately they went up, it's hard, I will not be changing boards now to fit a different processor. Of course, you can use Atmega88, 168, or another compatible with the appropriate amount of memory. I do not provide source codes, I can compile the program for another processor upon request.


    Well, I would like a 168 atmega. I have nothing to do, and I can recover the blocked processors.

    manekinen wrote:

    As for the speed of 38400, this is correct, please indicate where I wrote 38000 - maybe a typo


    Quote on your part: "On the board, there is also a connector described as RS232, it is a USART output, by connecting to it (38000bps), we will learn everything about the course of the repair operation"