Jump to content
Have You Played Atari Today?
2600|5200|7800|Lynx|Jaguar|Forums|Store  

ijor

Members
 View Profile  See their activity

  • Content Count

    2,621

  • Joined

  • Days Won

    2

Posts posted by ijor

  2. Posted

    On 5/22/2021 at 1:53 AM, Nezgar said:

    I used one 4K Happy 810 ROM dump marked "pre-v7" with CRC32 942EC3D5 as well as an 8K  v7 ROM CRC32 982D825D. Both allowed the menu to run.... once - until the drive has been virtually power-cycled - not just a Shift-F5 cold-start in Altirra, it needs a full close+rerun of Altirra. (3.90-test10)

     

    That sounds strange. Would you mind posting those roms?

  4. Posted · Edited by ijor

    23 hours ago, drpeter said:

    Fascinating reading (the thread that is -the schematic is all Greek to me!)

     

    Perhaps @ijor, @Bryan or @Rybags have some thoughts on the cause/mechanics of this phenomenon, as I'm just speculating wildly...

    I can't check the whole thread right now and give a detailed answer. But we talked about this before with some experiments that, IIRC, Bryan did years ago. As I explained in that thread with the reverse engineered schematics, GTIA has some weak synchronization paths. That means that synchronization logic is at the edge and might break in some special conditions that depend on the specific chip, the temperature AND the voltage. On that thread I also described the one that is perhaps the most significant path that might break sync.

     

    Exactly what happens here in this case I can't really comment without reading the thread and checking what the program exactly does, and I'm afraid I don't have the time right now.

    • Like 2
    • Thanks 1

  5. Posted

    22 minutes ago, Faicuai said:

    You mean traffic from A8 host to SIO peripheral, or peripheral to A8 host?

     

    What I meant is that push pull drivers can overcome the extra capacitance that those caps add. The usage of a push pull driver, as opposed to open drain, is up to the device that drives the signal. In the case of SIO IN (from peripheral to computer), the driver is at the peripheral side. That means that any SIO2XX device using push pull drivers, or a SIO hub with a push pull buffer, would solve the issue in this direction (from peripheral to computer)

     

    But the SIO OUT signal (from computer to peripheral) is handled directly by Pokey. Unless you replaces Pokey with something like PokeyMax, it would always be open drain. Using a buffer at the hub won't help too much because the problem is, in this case before the buffer, between the computer and the hub.

     

    So I am saying, if divisor 0 works with the hub, even with all the capacitors, then probably it has an additional stronger pullup at the SIO OUT signal (may be at SIO IN as well).

    • Like 1

  6. Posted

    6 minutes ago, Faicuai said:

    Well, in reality it is doable (for sure) but not needed. At least not with today's SIO-hub solutions:

    Somebody has schematics for that hub?

     

    I would guess that SIO input works, even with caps, because the hub has a push pull (not open drain) buffer. When using push pull drivers the caps are not a problem, even at the highest SIO synchronous mode bitrate (close to 600 KHz).

     

    But push pull drivers won't help for SIO output. May be the hub adds a stronger pull up to the SIO out signal? I think the Speedy docs recommend adding a strong pull-up at the drive, so may be Lotharek included it at the hub.

  7. Posted

    3 minutes ago, Mazzspeed said:

    Based on FJC's video, removing the caps definitely made a difference. I don't use PokeyMax and it definitely made a difference in my case.

    Removing the caps is needed for the higher bitrates, we all agree. But we are talking about some 600XL computers with PokeyMax that don't work even at standard SIO speed.

    • Like 1

  9. Posted

    14 hours ago, Mazzspeed said:

    Before cap removal the best I could get was divisor 5 stable.

    - After removing caps on RX/TX and handshake the best I could get was divisor 1 stable.

    - Removing all SIO caps got divisor 0 rock stable under RespecQt.

    I'm curious what's the reason for that. The other capacitors should not harm the transmission because the signals are not used at all or they toggle at a very low rate. Anybody investigated this or has any theoretical explanation?

  10. Posted

    On 5/15/2021 at 2:37 PM, AirKon said:

    Hello,
    I think I found the problem.
    The output (SIO5) is too low ( 2 volts ) and very noisy.
    I replaced the 4.7 k resistor by a 1.5 k one to reduce the noise and an AOP (x2) to raise the signal.
    Now it works. Could it come from the POKEY which would start to get tired ? 

     

    Pokey has nothing to do with this. Pokey outputs are open drain and then it doesn't control the signal at the high state.

     

    You normally get low voltage when the bitrate is too high and also depends on the capacitor. Are you sure you removed the correct capacitor? Which frequency/bitrate you were using at that scope capture?

     

    I'm not sure it is a good idea to replace the pullup with such a strong one. You might damage Pokey. At least you are certainly going out of specs for the max current.

    @tf_hh recently posted an issue with 600XL because they lack in line resistors. Not sure this is related to your problem though.

    • Like 1

  11. Posted

    I am certain don't remember seeing this: 1982 Happy Computing Slow it Down

    Not quite sure about the others. There are a couple that might be missing, or they might be included in other Happy docs that are floating around. Hard to tell until seeing the actual document.

     

    May be @Nezgar can tell if some documents are missing?

  12. Posted

    On 5/14/2021 at 2:05 AM, cx2k said:

    There were some other Happy Docs with this.

    - 1982 Happy 810 Enhancement Installation

    - 1982 Happy 810 Enhancement Information Package

    - 1982 Happy Computing Slow it Down

    - 1982 Happy Backup Program Instructions

    - 1986 Happy 810 & 1050 Enhancement Manual

    - 1983 Spartan Software The Chip

     

    Let me know if these are of interest as well.

    Some of these are common and already available. But a couple of them I don't remember seeing them. Yes, please, scan them if you can.

     

    Thanks a lot for uploading the Customizer manual!

    • Like 1

  13. Posted · Edited by ijor

    12 minutes ago, carlsson said:

    I'm still curious what the Greaseweazle community may come up with regarding support and software. It is by far the cheapest but also most DIY of all solutions, and I understand if it never will have the same specs/capacity that a $100++ solution has.

    I don't know where you got that idea. The Greaseweazle is even more powerful than both the Kryoflux and the SCP. The lower cost has nothing to do with lower specs. The rock bottom cost is because it leverages modern MCU hardware that has almost everything that is needed at minimal cost; and because you can build it yourself. The Kryoflux and the SCP are based on much older hardware and they are both sold for profit.

    • Like 1

  14. Posted

    24 minutes ago, bfollowell said:

    Yeah, and the Kryoflux is route is expensive. The Kryoflux iteslf is expensive.

     

    The Greaseweazle is a similar device, actually using more modern hardware. There are a few different options available to purchase. But it is completely open source. The smallest version, if you build it yourself, costs close to nothing.

    • Thanks 1

  15. Posted

    I have personally used a SC1124 monitor, as many other electronic equipment, with a step down converter (that provides 110V 50Hz) for years without any problem whatsoever. There were some electronic devices that were sensitive to the power frequency. I remember table clocks that used the power frequency to actually measure the time. Probably this design is not used anymore. But other than, there shouldn't be any problem mixing the power frequency, at least for consumer devices.

     

    It is difficult to know what the problem exactly is in this case. But I would bet is not the power frequency.

     

  16. Posted · Edited by ijor

    On 5/10/2021 at 1:36 PM, Alfred said:

    but after that, so what. If somebody really wants to know the history, go ask Gustafson, he's on LinkedIn.

     

    Are you serious? Asking the original author is no substitute at all for the original work. He probably doesn't remember everything after 40 years. He might not have the source code anymore. And even if it has, he might be reluctant to release anything since he might not be sure he has the rights to do it.

     

    Quote

    Sure Mike own the rights, but apparently he's not too hung up on granting permissions, or he'd have been all over DLT about them publishing their version. He's had 40 years to do something with any of the stuff he purchased from ICD, and actually I don't think he himself has a copy of anything anymore, although there's a miniscule chance I'm wrong on that point.

     

    Are we talking about just one source only, SpartaDos 3.5? What happened to the source for older versions, like SpartaDos 1.1, are they lost? And what about sources for all the other ICD software and hardware? Do you have any idea if Gustafson still has any source material?

     

    Yeah, everything matters, and matters a lot, for historical purposes, even unreleased versions, like SpartaDos 1.0 if there was such a thing, would be great.

    • Like 1

  17. Posted

    9 minutes ago, Alfred said:

    Anyway, what's the point of releasing the source ? There is all kinds of source code on Carsten's wiki, and nobody has done anything with it in ten years. Jesus, there isn't even an 80 column editor for VBXE and that thing has been out since what, 2006 ? So the idea that anybody is going to study the Sparta code and do something useful with it is plainly wishful thinking, just like Mike.

    Seems you don't understand the importance of original source code. We don't need the source code for technical purposes. It is not too difficult to disassemble and, obviously, it has been reverse engineered already. But the original source code, with the original comments, may be with commented out code, has an enormous historical value.

     

    Original source code (or original schematics for that matter) is simply invaluable and, sorry if I am being harsh, but I think that people that has original source code (not just Sparta Dos) and it keep it to himself, it not doing the community any favor at all

    • Like 4

  18. Posted

    19 hours ago, cx2k said:

    LOL. Sorry, no idea what the hell I actually uploaded. I edited my post and uploaded the correct ATR. Try it now @ijor

     

    Wow !!! Thank you so much. For quite some time I believed this program was  a myth! LOL

     

    1 hour ago, phaeron said:

    I checked the code that is producing this message, and what is happening is that the program is uploading code to the drive under the 'Q' command to read back $17FE-17FF, then checks it via an obfuscated algorithm that XORs those two bytes into a buffer and checks the sum of the result. The check will only pass if the two bytes at that address range are $90 $68. ...

    V7 ROM does have an obviously unused LDY #1 instruction in those two bytes

     

    I think we talked about that word already that seems to be some kind of serial number. And the Customizer, or at least this version, seems to be tailored to a specific H810 serial #.

     

  20. Posted

    On 5/7/2021 at 10:55 PM, carlsson said:

    Assuming there were custom machines used for disk duplication rather than regular floppy drives connected to some computer, wouldn't those duplicate both sides at once so fitting everything on one disk side or on two sides might not make much of a difference?

     

    I don't think there were such drives at that time.

     

    14 hours ago, Rybags said:

    The thing I'm interested in though is whether you could have coexistent boot sectors for both ST and Amiga. 

     

    Yes, you can have both boot sectors in the same track. That's how most of the dual format commercial games work.

    The Amiga and the ST both use the same MFM encoding. But the low level sector format (sync, header, checksum) is different. So each platform would read its own sectors without conflict

     

    • Like 1

  23. Posted

    19 hours ago, Wildstar said:

    MOS Technology officially has a 6502C which is clockable to 4 Mhz (without overclocking but officially rated for). The Sally is custom and may even been based on the MOS 6502C core ...

     

    We talked about this already years ago, in this very same thread I believe. There is no such a thing as a 6502C core, or 6502B core for that matter. There are no speed grade specific cores or masks. All 6502 NMOS parts are essentially the same with very minor mask modifications. The speed grades are obtained just by binning (using a term that nowadays is very popular, but I'm not sure it was already used back then).

     

    Anyway, I doubt very much that Sally was specified for  4MHz, or even 3 MHz. That would have reduced the yield and increased the cost unnecessarily.

  24. Posted · Edited by ijor

    On 5/3/2021 at 5:21 AM, Wildstar said:

    The 6502C is NMOS and was manufactured by Commodore / MOS.

    Just for the record because this doesn't alter your main point. But this is not accurate. Sally was not, or at least not only, manufactured by MOS. There are Sally chips made by Syntertek, Rockwell and others. I don't recall seeing one actually made by MOS/CSG. I think Atari intentionally avoided ordering chips from CSG, but somebody else might confirm if there was no MOS Sally at all.

     

    I don't recall if it is know exactly which company designed (designing is not the same as manufacturing) Sally, although it is a very minor and almost trivial modification to the original 6502 NMOS design.

    • Like 1
AtariAge
Forums
Store
General
Follow AtariAge

Copyright ©1998-2021 AtariAge

×
×
  • Create New...