Console hardware bugs - Any exist?

Discussion in 'General Gaming' started by Trenton_net, Feb 27, 2015.

  1. Trenton_net

    Trenton_net AKA SUPERCOM32

    Joined:
    Apr 13, 2007
    Messages:
    2,381
    Likes Received:
    59
    Hey gang,

    I know software issues/bugs are quite common, but I rarely hear about hardware issues. Does anyone know of any instances where modern consoles have issues? Such as, PCB design errors, or chip logic errors which make performing some calculations impossible, or rendering some features un-useable? Perhaps they exist in abundance but people just don't hear about them?

    I hear about software issues all the time, so if people are so sloppy in software, I would suspect hardware would have an equal amount of bugs/problems?
     
  2. Anthaemia.

    Anthaemia. The Original VF3 Fangirlâ„¢

    Joined:
    Jun 4, 2004
    Messages:
    1,654
    Likes Received:
    221
    For a console notoriously difficult to produce alpha blended transparency on, I find it rather ironic that the Saturn has a strange hardware "quirk" relating to how translucent layers are handled, which can sometimes result in the desired effect being repeated in a tiling pattern. For examples of this, check out the text boxes in Mystaria: The Realms Of Lore (AKA Riglord Saga/Blazing Heroes) and the bonus version of Dural with a fishbowl for a head in Virtua Fighter Kids...
     
  3. GodofHardcore

    GodofHardcore Paragon of the Forum *

    Joined:
    Mar 31, 2007
    Messages:
    11,821
    Likes Received:
    454
  4. lolzvid

    lolzvid Peppy Member

    Joined:
    Jan 27, 2012
    Messages:
    340
    Likes Received:
    11
    The Sega Genesis had a memory glitch that even some games used it, but the problem was solved on the third model.

    Initially, the Famicom had a PPU glitch that would freeze up games (saw that on the recent Famicom topic), too.
     
  5. abveost

    abveost Peppy Member

    Joined:
    Dec 15, 2014
    Messages:
    321
    Likes Received:
    56
    It's not a bug. It's a standard Microsoft feature available in some for or another in all their products.
     
  6. sonicsean89

    sonicsean89 Site Soldier

    Joined:
    Dec 19, 2012
    Messages:
    2,207
    Likes Received:
    157
    Yeah, the only game that it doesn't work with is Gargoyles (other than Virtua Racing and add-on games) (http://segaretro.org/Gargoyles).

    Does the Dreamcast's Mill-CD exploit count as a hardware bug? It was intended to use media CDs (a thing that never ever caught on in the West, despite repeated attempts), but hackers exploited it to boot backups/homebrews with ease.
     
  7. MottZilla

    MottZilla Champion of the Forum

    Joined:
    Feb 1, 2006
    Messages:
    5,073
    Likes Received:
    112
    I've seen this elsewhere on the internet and it's a misleading comment. It's not related to "memory", but instead related to the processor and Gargoyles. http://segaretro.org/Gargoyles

    The Test and Set opcode does not behave like it would on normal 68000 systems. I've read it's related to how the pcb is wired causing this issue. I also read something that suggest that the "Test" part works as intended however the "Set" part gets ignored for some reason. Fixing this on the Genesis model 3 breaks Gargoyles.

    Unless you meant something else by a "memory glitch".

    I've heard about the earliest Famicom's having some issue causing Nintendo to recall them. I never heard much beyond that.

    The DreamCast Mil-CD is a feature, not a bug. If there was a "bug" it was not locking it down properly.
     
  8. wilykat

    wilykat Site Supporter 2013

    Joined:
    Mar 25, 2012
    Messages:
    991
    Likes Received:
    45
    Intellivision 2 had a "bug" in EXEC that Mattel Electronics added to block Coleco games but accidentally broke a few of their own games. One games won't work and a few had sound issues.

    Commodore 64 were much more complicated. Originally CBM intended to have a burst speed with disk drive like 128 and 1571 drive has but a few boneheaded mistakes ruined that. One: the high speed wire intended to connect to the serial port was accidentally deleted when they were routing the PCB and adding mounting holes. Add to that the timing bug between 2 chips on C64 side resulted in badly timed interrupt that hobbled the disk drive performance, they were forced to bit-bang serial port at SLOWER rate than the disk drive on Vic-20. It's a miracle that C64 still became #1 computer when it has the worst disk drive performance of any home computer.

    DC technically has a bug. It was supposed to have copy protection but it only checked for legit GD-ROM and not check CD-ROM so it was very easy to play ripped games burned on CD-R. Had it been done properly, the security would have checked regardless of the disk format. If there's no security, assume the disc is audio CD only

    Commodore PET has a killer bug... literally. The first series of PET was slow and creative people found a way to speed things up by poking a certain value in certain memory. Later PET models changed it a bit, and poking the value in same memory location was actually dangerous issue that can kill the CRT on PET. It's probably the first case of killer poke.
     
  9. sonicsean89

    sonicsean89 Site Soldier

    Joined:
    Dec 19, 2012
    Messages:
    2,207
    Likes Received:
    157
    If we can talk about computers, the Apple III would overhead to much that the chips would detach from the board, and users were told to try lifting it up (mind you, the thing had its monitor included, and probably weighed far over 20 lbs) and dropping it to try and get the chips to reseat.

    Apparently Steve Jobs forced them to not put a fan or air vents in the machine, because it needed to run quietly.
     
  10. BM-Viper

    BM-Viper <B>Site Supporter 2013</B><BR><B>Site Supporter 20

    Joined:
    Sep 7, 2008
    Messages:
    383
    Likes Received:
    0
    The Atari Jaguar also had many hardware bugs, most notably in how it worked with controllers. They're well documented in the official documentation, even from the initial releases.
     
  11. TriMesh

    TriMesh Site Supporter 2013-2017

    Joined:
    Jul 3, 2008
    Messages:
    2,373
    Likes Received:
    785
    I'm not sure if this counts, because they caught it before release, but the original PlayStation CPU had a cache controller bug that could make the machine crash - this is why the early SCPH-1000s show signs of soldering around the CPU - they pulled the original CPU off the board and replaced it with the "A" stepping version that had the bug fixed.

    The PSX also had another bug that did get out the door - which was that the first generation VRAM based GPU ignored the texture flip bits for the textured rectangle command.
     
  12. Riki

    Riki Peppy Member

    Joined:
    Sep 2, 2013
    Messages:
    313
    Likes Received:
    5
  13. proarturs

    proarturs The force is with me

    Joined:
    Dec 23, 2013
    Messages:
    820
    Likes Received:
    49
    The 72 pin connector for the NES was badly designed in my opinion.
    Although the first Xbox 360 models were also badly designed, so I guess those are not bugs, just bad design.
     
    Last edited: Mar 1, 2015
  14. Druidic teacher

    Druidic teacher Officer at Arms

    Joined:
    Jun 6, 2006
    Messages:
    3,643
    Likes Received:
    129
    x
     
    Last edited: Jun 22, 2017
  15. sonicsean89

    sonicsean89 Site Soldier

    Joined:
    Dec 19, 2012
    Messages:
    2,207
    Likes Received:
    157
    I'm not sure if it counts as a bug, and it didn't really effect the system during its lifespan, but the 3DO had zero copy protection. CD burners weren't widespread, and downloading a full CD would take hours or maybe a day on dial-up, so it didn't kill the system.
     
  16. -=FamilyGuy=-

    -=FamilyGuy=- Site Supporter 2049

    Joined:
    Mar 3, 2007
    Messages:
    3,097
    Likes Received:
    1,046
    Mil-CD is a feature. The Mil-CD exploit is a design flaw. No bug per se there.

    EDIT: Seems I wasn't the only one to point it out... I would add to Druid's comment saying that they probably though people wouldn't figure out the gd-rom soft-reset command. Security through obscurity is never a viable option.
     
    Last edited: Mar 1, 2015
  17. sp193

    sp193 Site Soldier

    Joined:
    Mar 29, 2012
    Messages:
    2,235
    Likes Received:
    1,077
    PlayStation 2 hardware bugs and design flaws:

    Some EE bugs that may be applicable to the DTL-T10000(H) as SCE documented, but the SCPH-10000, SCPH-15000 and SCPH-18000 (GH-003) share the same CXD9542 EE:
    1. Improper VIF0/VIF1-Related INTC Interrupts by the i Bit of VIFcode.
    2. Problem with the VU1 Microinstruction XGKICK.
    3. Data value hardware breakpoint inaccurate (Unable to stop accurately, program may malfunction if break'ed at a SYNC instruction, EE may BREAK at different address).
    4. Transfers via the MFIFO drain channel may become unstable if the ring buffer address is changed during a DMA transfer.
    5. DMA-tag mismatch may occur when VIFn_ERR.ME0=0
    6. When DMA channels 1 and 9 are simultaneously transferring data, the transfers may stall.
    7. Divide instructions of COP1 and COP2 may result in an incorrect result, under certain conditions.
    8. DMA transfers unable to function properly under certain conditions.
    9. Short loop bug (branches in loops that are shorter than 5 steps may be incorrect).
    10. Rarely, accesses to RAM beyond 32MB will not be accessed properly. Will not happen on retail or Debugstation sets.
    11. MFIFO drain channel may stall if DMA packet tags end with refe, ref or refs. An empty packet may be incorrectly detected.

    GS bugs for the DTL-T10000(H), and possibly for the early PlayStation 2 sets:
    1. Dept test may be incorrectly performed when ZTE is ON/OFF switched.
    2. When the interrupt handler is handling SIGNAL, FINISH, HSync or VSync, further interrupt assertions may fail to be detected.
    3. When textures are not transferred with the required widths, part of the data may not be written correctly.

    i.Link controller:
    1. Apparently no support for hardware byte-swapping, hence why the third DMAC is undocumented and unused.
    2. Not sure if it's a requirement by the IEEE1394 spec, but the local node cannot be set to anything above S100 speed (transfers will fail) if it's the root node. Doesn't seem to happen with my SCPH-39006, but happens with a lot of older consoles.

    USB OHCI controller:
    1. When a packet of 64-bytes is sent out, it may be incorrectly split by the hardware. SCE never mentioned how widespread this issue is though.

    IOP:
    1. Too slow, causing the peripherals connected to the expansion interface to lose performance. :(
    2. Its DMA channel 8 (DEV9) doesn't seem to have support for chain DMA transfers, at least the SCE DEV9 and DMACMAN drivers don't support that.

    I still hate how slow this thing is. Even if you gave it a good 100Mbit Ethernet connection, it will only give you about a quarter of its achievable speed. I learned that SCE got only about 28Mbit with their EE-side network protocol stacks, so it's most likely a hardware limitation.

    SPEED (Expansion device):
    1. EMAC3 registers sometimes cannot be read properly unless read twice.

    Stuff that is not documented, but observed with the SCPH-10000 and SCPH-15000:
    1. CD/DVD drive eject mechanism is weak, so SCE adjusts the tray of refurbished (and later?) sets to eject and retract faster to prevent it from seizing up.
    2. Inadequate cooling, leading to thermal shock. The fans of refurbished (and later?) units sound like they're spinning faster.
    3. Known issues, which SCE solved via software updates: the Protokernel lacks support for the HDD unit, ExecPS2 not designed properly, OSD config functions lack support for (worldwide) regional data, OSD can freeze if it finds a corrupted save and is unable to pass arguments to the update.
    4. Known issue, which SCE solved in newer SDKs: the RPC server will call mkdir() after calling remove() because the programmer forgot to put a break statement. But this bug exists in the boot ROM of all consoles, regardless of the model (for backward-compatibility?).

    Well known stuff throughout its lifespan, up to the SCPH-75000:
    3. MECHACON bugs (lasers burning out etc), whether deliberate or not.

    IOP DMAC2 (for the PSX only):
    1. Apparently there was some issue with SIF DMA transfers, which SCE was working around at SDK v3.0.
     
    Last edited: Mar 1, 2015
  18. gladders

    gladders Robust Member

    Joined:
    Nov 1, 2013
    Messages:
    256
    Likes Received:
    15
    The Atari Jaguar suffered from some severe CPU bugs, according to Wikipedia:

     
  19. Riki

    Riki Peppy Member

    Joined:
    Sep 2, 2013
    Messages:
    313
    Likes Received:
    5
    I've seen this happening even on 39 and 50 models.
     
  20. PrOfUnD Darkness

    PrOfUnD Darkness Familiar Face

    Joined:
    Mar 13, 2004
    Messages:
    1,122
    Likes Received:
    48

    Talking about this command, is there any official reason for it exist?
     

Share This Page