croissantking

croissantking

Well-known member From London, UK
Joined
Last seen
Viewing thread Combating SE / SE30 voltage drop
Messages
2,718
Reaction score
186

Profile posts Latest activity Postings About

  • croissantking
    IMG_4689.jpeg

    Successfully moved a 466Mhz 7400 CPU onto a G3 CPU card - harvested from a Digital Audio card that I obtained for peanuts. The cache chips are rated at 250MHz, so I guess this thing will easily do 500MHz.

    The old 750L will go onto an iMac trayloader CPU card.
    • Like
    Reactions: cheesestraws, obsolete, Powerbase and 1 other person
    alectrona6400
    alectrona6400
    i wonder... whats your process to BGA soldering? i always struggle with it
    croissantking
    croissantking
    croissantking
    croissantking
    -Preheat donor PCB, use hot air to remove chip
    -Remove solder balls. Clean pads with braid/IPA
    -Line up stencil, tape to chip. Mask unused holes
    -Smear flux into holes
    -fill holes w/ solder balls
    -Heat chip w/ hot air @300C until balls melt to pads
    -Remove stencil + clean chip
    -Coat new PCB pads w/ flux
    -Line up chip
    -Preheat PCB + use hot air @300C to heat chip, should drop once soldered into place
    croissantking
    Got some OPTi chips for future Personality card mods. The plastic wrap on these Chinese imports always smells of citrus.

    IMG_4559.JPG
    • Like
    Reactions: obsolete
    View previous comments…
    croissantking
    croissantking
    @zigzagjoe I thought remarking usually is to disguise a lower speed grade as a faster one, or one with fewer features as having more - not relevant for this USB controller?
    If the chips work, and identify as 82c861, then would it matter if they were remarked?
    zigzagjoe
    zigzagjoe
    I'm given to understand it is more often cosmetics. Customers want new looking part not scratched up salvaged parts rather than malicious intent.

    For these I concur if it functions correctly then it's not really a huge concern, just good to know.
    croissantking
    croissantking
    croissantking
    This shutdown dialog under System 7.1 is particular to the 840AV. Normally, nothing happens when you press the power key under 7.1. It’s a customisation built into this machine’s system enabler.

    IMG_4397.jpeg
    Renegade
    Renegade
    it seems to me that many enablers implement this functionality under 7.1
    croissantking
    croissantking
    It is an ugly dialog box for sure! Probably made by an engineer not a designer. Still, interesting to see the progression towards the 7.5+ shutdown box we all know and love
    croissantking
    I’m hoping to save this 040 chip, soldered to one of @zigzagjoe’s QFP2PGA boards. I ground two corners with a Dremel to be able to attach repair wires to the missing pins, which were lost to corrosion. Might give you some ideas for yours @jmacz

    IMG_4343.jpeg
    • Like
    Reactions: joshc and cheesestraws
    jmacz
    jmacz
    Nice, haven’t played with the chips yet but will try to save that one with the broken pin.
    Joopmac
    Joopmac
    cool, what about this adapter? is this an accelerator or just adapter? looks sweet.. is there a topic?
    croissantking
    croissantking
    @Joopmac its an adapter that lets you use a quad flat pack 040 in a PGA socket, designed by @zigzagjoe. It’s freely available on his GitHub :)
    croissantking
    My SE/30s are very ill. Stock PSU issues, mostly. They’re recapped, but their +5V rails are too low. I might try tweaking the voltage pot - but +12V is already quite high.
    View previous comments…
    croissantking
    croissantking
    As I recall, anecdotally, they got worse after recapping. The big cap is a slightly different capacitance than the original, I hope that didn’t mess things up.
    robin-fo
    robin-fo
    I didn‘t change the big cap..
    croissantking
    croissantking
    Have you played with the pot yet? I’ve been putting it off as it’s a fiddle to do, but I need to try.
    croissantking
    Anyone know anything about this mystery USB Port '0' on the Yosemite board? (Next to the firewire module)

    IMG_3852.JPG
    View previous comments…
    Daniël
    Daniël
    The Yosemite schematic says J18 should be a test point for the 14MHz clock going to the IMI SG500 clock generator, so the silkscreen for a supposed USB port seems unrelated. There is mention of a removed internal USB port during the prototype phase in the schematic changelog, so it may just be orphaned silkscreening for that.
    croissantking
    croissantking
    That’s interesting!
    C
    cheesestraws
    yeah, agree with @obsolete and @Daniël - that's definitely an RF connector footprint of some description, probably is an SMA but I'm not familiar enough with 'em to be able to tell them by sight so I'll trust @obsolete on this one :-D. No idea what the usb0 silkscreen is doing there.
    croissantking
    Hit 94.7MHz bus speed on the Beige today, it’s stable but the clock generator can’t seem to handle more. I have some ideas to push it further.

    IMG_3848.png
    croissantking
    I want a PM6100 now. :D
    croissantking
    IMG_3584.jpeg

    Beige G3 running a 87.3MHz bus. :)
    Swapped out the reference crystal.
    ASP still thinks it’s 83.3.
    • Like
    Reactions: Snial
    View previous comments…
    croissantking
    croissantking
    @Snial I wonder how it’s doing it’s memory performance calculation.
    • Like
    Reactions: Snial
    Powerbase
    Powerbase
    Theoretical maximum is one thing, i don't think I've ever seen even 200+ in Gauge Pro. I'm unsure what it's using to calculate either.
    stynx
    stynx
    "moving memory" = moving from one memory address to another = 1/2 the speed. If you add latency (333) you can subtract at least 50% again. Then there is wait-states from the controller, CPU and software induced (read-write cycle ... maybe with crc-check). The potential 266MiB/s memory (moving) throughput will more likely be maximum of 133MiB/s. Typical is about 1MiB/s per MHz so 87MiB/s would be considered normal.
    • Like
    Reactions: Snial
    croissantking
    Me complaining about my Beige G3 23 years ago lol

    IMG_3504.jpeg

    I really wanted to like Mac OS X when it came out, but I couldn’t. It took buying my first MacBook Pro in 2007 to properly feel comfortable with it.
    croissantking
    My PowerBook 165 keeps on finding new ways to piss me off. Now, it cold boots without sound or trackball/keyboard input. After doing a hard shutdown, and powering on again, it works normally. What in the hell… I’ve already replaced so many parts in this machine.
    • Sad
    Reactions: Snial
    View previous comments…
    croissantking
    croissantking
    Yes @Snial I’m pretty determined to get it fully working.
    My 180 is better behaved.
    Both are maxed out with rebuilt batteries.
    I like having one of each type of greyscale screen (passive and active) with their relative pros and cons.
    • Like
    Reactions: Snial
    Snial
    Snial
    The PB1xx series were special, because they revolutionised laptops. I had a PB100 (which I loved) until it got stolen when I accidentally left it in the foyer of the Kilburn building at Man Uni in 1998. I borrowed a PB150 for a couple of weeks a bit later (clunky, but cheap & had IDE drives).
    RetroTechTom
    RetroTechTom
    My 165 likes to reboot randomly when it feels like it
    croissantking
    IMG_3167.jpeg
    Good progress today fixing a 24X
    • Like
    Reactions: RetroTechTom, SophieRose, Snial and 5 others
    View previous comments…
    croissantking
    croissantking
    I found that one or more of the buffer chips was bad, but replaced them with wrong parts (too fast; overdriving inputs on other chips). Card worked for a while, but would crash upon resolution or bit depth change and sometimes displayed graphical corruption. After a while, the VCLK signal became unstable and I would lose sync, and now won't sync at all. Looking at VCLK output it wildly fluctuates between 0 and 8MHz.
    finkmac
    finkmac
    how'd you identify that the buffers were bad?
    croissantking
    croissantking
    @finkmac I just sort of guessed they might be given where they sit (between the Nubus interface and one of the ASICs) and it was easy to swap them.
    croissantking
    I disassembled my QuickTake 100 all the way, it has a tonne of electrolytic caps which I didn’t know about. Some are very well hidden. Anyone interested in a write-up?

    IMG_2830.jpeg
    croissantking
    IMG_2413.jpeg
    I wanted to test out my PowerBook 100 series modems, so I got two of them talk to each other directly, with all the associated sounds of nostalgia. Only at 2400bps though 😏
    • Like
    Reactions: RetroTechTom, Durosity, Snial and 3 others
    View previous comments…
    Snial
    Snial
    I had a friend whose Dragon 32 was connected to Prestel via a 1200/75 modem around 1984. In some ways it mirrors lots of Broadband performance. Prestel was based around the very weird GEC 4080, 16-bit minicomputer. It's mostly weird, because the Kernel is implemented in microcode! https://en.wikipedia.org/wiki/GEC_4000_series
    jmacz
    jmacz
    I've still got two 14.4 modems and a 28.8 connecting via a PBX. I probably have another modem around here somewhere -- actually like you I have an internal one in a PowerBook. I think my PBX can handle 6-8 lines. I should finish setting up my Hermes setup and then try having two other Macs dialed in at the same time. I don't remember if Hermes had any multi player door games. :)
    • Like
    Reactions: Snial
    Snial
    Snial
    At one level, the 68KMLA site is a BBS with hierarchical menus and articles. So, given the recent articles I've been reading about using FPGAs or fast MCUs to enable even lowly 8051s to handle elliptic-curve crypto I figure it's possible to design a PDS/NuBus/PCMCIA peripheral to offload the encryption and get BBS-ified, 68KMLA interaction :) !
  • Loading…
  • Loading…
  • Loading…
Top