(Topic ID: 191079)

Stern reliability: S.A.M. vs SPIKE

By halflip87

6 years ago


Topic Heartbeat

Topic Stats

  • 368 posts
  • 94 Pinsiders participating
  • Latest reply 4 years ago by kvan99
  • Topic is favorited by 21 Pinsiders

You

Linked Games

No games have been linked to this topic.

    Topic poll

    “Which system is more reliable and repair friendly?”

    • S.A.M. 176 votes
      90%
    • Spike 20 votes
      10%

    (196 votes)

    Topic Gallery

    View topic image gallery

    Samiam (resized).jpg
    20170814_094849 (resized).jpg
    IMG_20170811_064200610 (resized).jpg
    1389487650-0 (resized).jpg
    SPIKE.jpg
    Wiley E Coyote Tricked.jpg
    download (resized).png
    1141cb3f415534bc4b87b6066071d20ffadc483e (resized).jpg
    pinball-stepper-repair-02 (resized).jpg
    IMG_8109 (resized).JPG
    Stern Wiring Diagram IMG_2198 (resized).jpg
    1268_1289005 (resized).png
    system_drawing_xNodes1 (resized).jpg

    You're currently viewing posts by Pinsider Joe_Blasi.
    Click here to go back to viewing the entire thread.

    #39 6 years ago
    Quoted from barakandl:

    Which is fine, just not what pin heads are used to.
    Say in 5 years from now, Stern comes out with a new system called SPUNK instead of SPIKE. How long does Stern keep making SPIKE node boards. How long did they sell WhiteStar boards?
    People can make a new replacement classic Bally MPU because of the schematics and servicing guides Bally provided. If no detailed info on Spike is available, the future repairability of the games 5-15 years from now is questionable.

    well they are on a newer cpu board. But what they can do is split up the game code / os / node board firmware into there own updates or do some thing so after they stop updating an game it will still work node boards that have newer firmware on them.

    #40 6 years ago
    Quoted from HighVoltage:

    This Spike architecture sounds flawed. In a distributed architecture like this, one of the points and benefits if you spread out the processing, is to reduce a single central point of failure. Ideally, if a node went down it should only result in partial failure - the relevant subsection of the game - but the whole should keep operating. That not being the case, suggests a poor design in the software, and what we actually have is a system which is worse than a monolithic design: it has multiple points of failure, all of which render the whole game inoperable. This is a defect in the overall operating system, but it is something that could potentially be addressed by a redesign / update.
    This sounds like what happens when you have a game designer develop the operating system / architecture.
    It's not surprising the first incarnation of Spike could not support video / LCD as was intended.

    Well the way networking / data bus flows though each board it can be like the old bus networking where one bad part can jam up the network same thing with e-net hub. Switches are better but the way stern is wiring them has data in and out of board so anything past it can fail if it goes down.

    Now if they had some kind of switch with each board running a cable to it then one may not take down an full game but that costs more.

    #98 6 years ago
    Quoted from cooked71:

    So this happened on my ASLE today: mid game, finished a song, animation for Super came up and the game froze. But flippers/lights/music worked, could still hit the ball, but nothing else was working, no scoring, no magnet no sounds. Until I hit the scoop and it just sat in there.
    Then another game, it went dark, and all but shut down, except flippers were working. Could hit the ball for about 2 minutes, with no lights/sound etc. then it eventually did a full reset.
    Never had anything like it happen on any other pre-Spike 2 game.

    I have seen spike 1 games lockup looked like some kind of software crash / reboot mid game.

    #120 6 years ago
    Quoted from Bendit:

    That's true indeed! With Ubuntu's "universal" drivers, chances are a mobo swap of the same generation would just work indeed.
    Good point.

    same gen should work unless they build there own kernel with stuff taken out.

    #122 6 years ago
    Quoted from jwilson:

    Stern is in the business of selling you a new machine. Making old ones expensive to service seems right in line with that.

    sounds like apple but unlike app stern sells parts to end users.

    3 weeks later
    #157 6 years ago
    Quoted from merccat:

    I don't see JJP's architecture as comparable:
    - Off the shelf CPU board with a linux base OS vs proprietary CPU board with proprietary base OS
    - Traditional single power driver board and power supply board with hole-through compnents vs node board modules with surface mount components.
    In this way JJP's architecture is much more similar to P2K but actually simplified because there is no PRISM boards to worry about.
    The one similarity is the lighting sytem which is a data driven modular design. I actually don't really see any other way to efficiently and cost effeively run RGB lighting if you want to have it, and I think this is one area where it is worth it.
    However still not that similar because while an issue here will impact lights it won't take down the entire game. Also they have since addressed the issues caused by a serial chain by moving to a parallel system with a hub so if a light board goes out only the one light board is impacted. If nodes were architectured more in this manner I actually would be a little less concerned.

    well JJP still has there own usb dongle and io board. The pc is just an PC. Now an SDD maybe over kill vs say an SD card and the full OS does give some added overhead + issues with putting the games on line if os updates can mess-up the system and they don't update the os out side of game code.

    Stern has pc and IO all in one board.

    Now can the stern boards run in hub/switch setup with some kind of switch vs being Daisy chained? Higher cost for that.

    Stern has talked about putting games on line for some time but with there slow pace of updates that is an bad idea.

    #159 6 years ago
    Quoted from Bendit:

    JJP's CPU "boards" are just desktop computer motherboards, the generic kind. As long as it's the x64 architecture of motherboards, the generic Linux drivers should just work. There's no dedicated video card in the JJP metal case, which means that they use generic video drivers that are build into the Linux install, using the onboard video device of the desktop motherboard. That's sustainable and universal.
    I've had Linux installs run on VERY old motherboards. I would also guess that finding NOS motherboards in the future might not be that hard. There are way more desktop motherboards in the world today than Spike CPU boards. And that's a fact that will remain for decades.
    I think it's funny how people here defend the longevity of proprietary Spike boards versus generic x64 desktop motherboards. You kidding me? x64 motherboards have been in production for what? a decade already?

    Once the generic x64 motherboards are gone? They're all over the planet! And you know what? The capacitors on desktop motherboards can be replaced too! They are not the size of flea testicles.
    Limitations and a mixed blessing. Because modern motherboards already has bluetooth hardware built in, it was a natural thing to do for JJP to comply with it and use it in their games. Granted, older x64 motherboards does not support bluetooth.
    Now, if Stern wants to implement Bluetooth in their games, they'll need another version of the Spike board (version 4 now?), and add all the components required for Bluetooth communication...

    Most desktop boards with Bluetooth do it over usb so even with out it on the board they should work with any usb Bluetooth device that has drives.

    Stern can add drives for usb Bluetooth as well.

    #160 6 years ago
    Quoted from Bendit:

    What are you talking about? The USB extension cable?

    The USB plugs into the JJP IO board and they use some kind of an usb security dongle.

    #162 6 years ago

    Now will JJP / stern and others if right to repair laws pass will they give out code / info needed for people bypass / repair any of there hardware?

    #163 6 years ago
    Quoted from Bendit:

    No way. All the motherboards that I have seen support bluetooth without additional USB devices. BlueTooth has nothing to do with USB. In fact, Universal Serial Bus has been around 10+ years before the first version of Bluetooth.
    Drives? You mean drivers?
    But yes, you could add bluetooth support via a USB device. My point was that you don't have to, if the motherboards has it onboard already.

    The motherboards that have it put it on the usb bus. and yes drivers that was an typo

    #166 6 years ago
    Quoted from smokedog:

    What happens when the base Linux OS causes a bug that doesnt get fixed by JJP? Doesn't WoZ and Hobbit suffer a bug when on for extended periods of time? I think I've read that a few operators are having to run older software in order for them to be used on route.

    well they added an auto reboot if on for more then 24 hours but how can you mess up linux that bad to need that?

    #168 6 years ago
    Quoted from Bendit:

    Has there been a real-world documented case of this?
    Linux is known for its generic drivers that work on pretty much anything.

    an update can break an app or they can get stuck on a old linux distros.

    Will all of there games still work say 5-10 years from now on an new distros as it's likely that old distro (there image with an old kernel / drives) may not work to well with new x86-64 hardware.

    #173 6 years ago
    Quoted from Bendit:

    So there you have it guys. Awesome discussion. I suspect the issues listed above are from the JJP software that runs on top of Linux. So in 10 years, if nothing changes in the JJP software, you'd still have that sound popping. If reverting to software 5.x in WOZ fixes the problem, it's software-related, not hardware.
    This thread seems to be about hardware sustainability. But since we're talking about JJP hardware, I guess we're off topic...

    Does 5.x use an older base Linux os vs the newer builds?

    #178 6 years ago
    Quoted from Gexchange:

    I can only speak for what Ive seen in the Industry. Megatouch uses Linux as well and only works with X boards. I believe Raw thrills does as well and same issue (Raw thrills does use Video Card Drivers) Its been a HUGE issue in the business. Going back to 486 games that can be repaired anymore due to Bad Mother boards. Pb2000 Being one of these.
    JJ

    486
    The dts system in movie theaters used 486 systems with dos on a rom card.

    #182 6 years ago
    Quoted from Bendit:

    The 486 was a 32bit CPU. x64 is backwards compatible with 32bit systems. I bet you could run that DTS software engine today and it would work.

    well back then they used and scsi card (ISA) and the dts decoder card (ISA) as well with MS? DOS on a ROM DOS card (ISA).

    #184 6 years ago
    Quoted from pinball_keefer:

    Spoken like someone without much Linux experience.

    I can guarantee you that is not the case.

    Servers that aren't pushing out a ton of AV I'd imagine.
    Most of the issues we've found can be traced back to generally-Intel-provided drivers.

    I'm not worried about it. We're already x64 anyway.

    Well have seen issues with trying to run older Linux distros on new hardware. Is your software easy to move so you don't end up being like Windows XP where you are stuck on it as the software does not work that well on new windows builds even if an game that had it's last update 4-5 years ago and you are moving to a new Linux distro.

    Now days new hardware really does not have XP drivers any more.

    Even moving to say windows 7 just about all software will work with 10 but some new hardware does not have drives for 7.

    the one thing with your setup that can been seen as issue (From an IT stand point) is that you software is shipped wrapped with the OS.

    But say you have an app with as an example CentOS 6.X and it works with 6.X latest but it's only shipped as an image with say centos 6.2 and you have new hardware that needs at least 6.5 to work.

    #185 6 years ago
    Quoted from Bendit:

    You are 100% correct. In this case, email and data-related servers, no AV.
    I have a billion questions for you. Intel-provided drivers for onboard audio and video drivers, correct?
    You guys use Intel-specific drivers? Not the generic-device ones?
    You guys use OpenGL libraries for audio and video? Or did you code your own low-level engines?
    How about the Bluetooth stuff, generic Linux drivers?
    (I have a thread on the google group called "Hobbit - Sound Question (sound glitches)", I am extremely curious about this minor issue which I would guess is drivers-related)

    also are you locked in to intel??? right now AMD is kicking there ass and vendor lock in is bad even more so with X86-64 based software.

    #193 6 years ago
    Quoted from merccat:

    Precisely! Software engineers of today should be completely aware of mis-steps of the past and there are well established practices today to minimize such risk. Binding yourself to specific technologies is now a known anti-pattern and the value of abstractions have been realized by most. From Keith's comments JJP obviously gets this and has positioned their software architecture properly.

    More software QA is also needed in a lot of stuff and the QA team needs the power to do just about anything as well being able to manually set up different scenarios and not be tied down to a deadline of just passing over bugs to meet it.

    It seems the like the easier it is to update something. You get ship now patch later.

    3 weeks later
    #252 6 years ago
    Quoted from DDDwingmaster:

    If the spike communication protocol is known, then even in 20 years you would be able to design a new node board. The actual hardware used on them is not that important, as long as they understand the network protocol to talk to it.
    Lucky for us the mission pinball framework project has already partly reverse engineered the spike networking protocol. The hope is that they are able to fully decode the protoco. Even better would it be if Stern release the protocol specification.
    So the wait is for the full protocol spec and someone designing a third party node board.

    Now could they wire it up to work with some kind of a switch vs an hub / chained networking? Needs more cable maybe an switching / routering board.

    1 month later
    #338 6 years ago
    Quoted from spazzman90:

    At this point, I have a Star Wars where the screen randomly turns off (game still plays, must be power cycled), Game of Thrones that will randomly reboot, and a Ghostbusters where the top half playfield switches stop responding (again, reboot and all will be fine). Tough thing is all these issues only crop up once or twice a week. Machines get pretty good play. Stern is willing to send stuff, but they have to idea as to the true solution to any of these issues.

    would a fresh SD card / full clean install do any thing??

    Is something crashing at the OS level?

    #339 6 years ago
    Quoted from CaptainNeo:

    i shouldn't have to stock up on $300 node boards...and not just one type, at least 3 different types. So say you need 3 of each on hand for the distant future. That's 9 boards minimum. And that's if they don't change it again through another release. So I have to have $2700 in boards sitting around, instead of piles of .01 cent diodes, $2 chips, and .05 transistors? Screw that and screw stern for fawking over the guys that like to service their games instead of making them disposable!

    talking about do they have some kind of U-boot or what even is the hardware boot loader is recovery system from bad flash if there is any flash on the base board??

    I once had to use the serial port on some network hardware to reload the boot loader and fix it's config.

    1 month later
    #342 6 years ago
    Quoted from StylesBitchly:

    What bothers me is that if I cannot diagnose and repair my own pins, new pinball machines are going to become the domain of people wealthier than me just like everything else that floats, flies or fucks.

    sounds like cars you need to use the dealer for reading codes / tools / part install codes to fix it.

    2 months later
    #353 6 years ago
    Quoted from Phat_Jay:

    I see a factory warranty being offered in the future at point of sale for a price......just like autos. “These aren’t your dads pinball machine, they require professional service” they will tell us. They did it to cars, now they’re doing it to pinball. Next they will tell us it’s not our machine, we only “licensed” it’s use when we “paid” for it.

    more like how apple is now You have use our professional service / buy our parts at high prices to fix your hardware.

    #354 6 years ago
    Quoted from Phat_Jay:

    I see a factory warranty being offered in the future at point of sale for a price......just like autos. “These aren’t your dads pinball machine, they require professional service” they will tell us. They did it to cars, now they’re doing it to pinball. Next they will tell us it’s not our machine, we only “licensed” it’s use when we “paid” for it.

    more like how apple is now You have use our professional service / buy our parts at high prices to fix your hardware.

    Quoted from jackofdiamonds:I am Service dept manager for a major Broadway audio production facility.Very few component level repairs being done.Its mostly board swaps.We have several bench techs who do hands on repairs but its expensive and time consuming.Our access to schematics is based on whether or not we are authorized.Stern should either provide authorized "service centers"(this could be an op or restorer)or they should provide schematics.Basically some sort of "support" for their products.

    they should make node boards more universal and down the road maybe even an backwards compatible dip on newer ones.

    What is the plan if then they use the same broads but say now they are loading them with say node software V7 but your old game is stuck on OS V3.5 that does not work right with node software builds after V6.5 and that os build only came node software V5.9.

    #355 6 years ago
    Quoted from jar155:

    Star Wars had a code update that was causing a node board to fail, yes. It has been addressed though.

    I don't like idea of how they rap game code and os code in to one update. The node board software is part of the OS and most nodes can be swapped game to game.

    You're currently viewing posts by Pinsider Joe_Blasi.
    Click here to go back to viewing the entire thread.

    Reply

    Wanna join the discussion? Please sign in to reply to this topic.

    Hey there! Welcome to Pinside!

    Donate to Pinside

    Great to see you're enjoying Pinside! Did you know Pinside is able to run without any 3rd-party banners or ads, thanks to the support from our visitors? Please consider a donation to Pinside and get anext to your username to show for it! Or better yet, subscribe to Pinside+!


    This page was printed from https://pinside.com/pinball/forum/topic/reliability-sam-vs-spike?tu=Joe_Blasi and we tried optimising it for printing. Some page elements may have been deliberately hidden.

    Scan the QR code on the left to jump to the URL this document was printed from.