Jump to content

flashjazzcat

Members
  • Posts

    19,279
  • Joined

  • Days Won

    30

flashjazzcat last won the day on October 14 2022

flashjazzcat had the most liked content!

About flashjazzcat

  • Birthday 12/19/1972

Contact / Social Media

Profile Information

  • Custom Status
    Solidly unarguable
  • Gender
    Male
  • Location
    United Kingdom
  • Interests
    Playing jazz guitar, music, reading, writing, PCs and anything to do with computers, movies

Recent Profile Visitors

74,695 profile views

flashjazzcat's Achievements

Quadrunner

Quadrunner (9/9)

14.8k

Reputation

  1. The repository of plugins is in the firmware download. As for OS ROMs and 8K ROMs: there are quite a few places hosting these, such as: http://ftp.pigwa.net/ You're describing something along the lines of the existing XEL-CF3 which Mytek designed for the 1088XEL. It's little more than an IDE connector with some buffers on the address bus, and the U1MB is able to drive the attached media as a HDD. Regarding SIDE3, however, although things like the on-cart SDX are redundant in the presence of U1MB, not much else is, given the capabilities of the hardware: The SIDE3 RTC is exploited by the loader's integral FAT DOS for the purposes of time and date stamping of files in FAT partitions (admittedly the U1MB RTC could be used when present, but in order to keep the FAT DOS as small as possible, it isn't). The SIDE3 loader heavily exploits DMA (SRAM to SRAM, SD to SRAM) as part of its basic operation, as well as the large amount of SRAM and ROM on the cartridge SIDE3 uses SD cards, which itself implies a resonable amount of hardware and software complexity The SIDE3 loader stores its configuration in the RTC chip's battery-backed NVRAM So, a 'simplified' or cost-reduced SIDE3 could perhaps give up the on-board SDX (which is neither here nor there, since it occupies a tiny amount of the cartridge's 8MB of ROM), but not much else if full functionality (including cartridge emulation) is to be retained. My pleasure - thanks for the kind comments!
  2. The partition table is read from the APT on the card right at the start of the OS boot process, well before carts are initialised. The SIDE loader itself runs as a cart which disallows a disk boot, but the PBI BIOS still initialises beforehand and the partition table in RAM is present and correct by the time the UI comes up.
  3. Thanks. That's clear now, although I'm slightly reluctant to implement something which won't be required during normal operation (it might be preferable to address the problem of the partition table in RAM being lost for reasons unknown when the AVG selector is active). It's kind of odd, actutally, since the partition map is stored in the U1MB 'IORAM' at $D1xx/$D5xx, so why or how it is lost in these circumstances is a complete mystery to me.
  4. Please spell out what your complaint is because I struggle to understand what you're trying to accomplish.
  5. Good stuff. I'm chomping at the bit for some quality coding time, although I do feel we've yet to fully exploit existing hardware from a software point of view (or completely nail stability on the hardware, for that matter), and I think time spent on R&D will pay off big time even if this means a longer wait for stuff to actually become available. That's just my view, however.
  6. I expect the answer will be that the firmware isn't written yet, and that this is my fault. I did, however, explain to Lotharek the last time the matter was brought up that I have just moved house, am backed up with work, and spent the last three months of 2023 with one of my fingers in a splint. I also pointed out that I'd spent several months in 2022 doing a complete re-write of the SIDE3 Loader which appeared to have gone completely unnoticed but which nevertheless counts in large part towards the firmware which will ultimately run on the devices currently in development. I also note that upgrades which don't depend on my input at all (Simple Stereo 4, VBXE 3, etc) appear no closer to release than Incognito Lite, Incognito 3 and External U1MB.
  7. Most certainly, especially since the same trick is employed on VBXE ANTIC adapters. Although one precision socket inserted into another is commonly an extremely tight fit, the problem here is that there's a PCB between one socket and the one it's being plugged into. Therefore, the pins don't go all the way in, and the upgrade isn't necessarily a snug fit in the precision socket on the motherboard. PokeyMAX is a different matter again: it uses gold header pins, and although these do go 'all the way in' in a precision socket, they're not that snug and all it takes is a bit of deformation (for example, plugging an AKI into the same socket beforehand) to make it even looser. The original machine sockets do grip things pretty well, but the problem there is that the device plugged into it will tend to spring right out of the socket (almost ejected by the tension on the socket contacts) if knocked or not installed just so. In any case, all of this results in a situation where poor connectivity is often a real issue which can imply problems in other areas which don't actually exist.
  8. Yes: what this does to the fancier sockets I don't know, but any socket (such as those supplied with AKI and VBXE) that's been subjected to these solder-thickened pins is useless for anything else afterwards.
  9. You could, but my point is that the precision socket will be permenently slackened from the moment AKI is inserted into it. Running the machine with PokeyMAX or POKEY itself in that socket subsequently might prove problematic. The original machine socket often does a better job of gripping the legs, ironically.
  10. Not sure: I only have the older revision board here.
  11. Don't forget that IDE Plus required extra discreet connections on the XL in order to properly control cartridges.
  12. Here's something I noticed and had to combat when working on an 800XL with AKI USB and PokeyMAX recently (disregarding the unusual manufacturing defect on this particular PokeyMAX, which I'd fixed by this point). AKI USB has extra solder on the (quite short) socket legs protruding below the board, and the legs are protected by a socket during shipping. The extra solder is presumably there to ensure a tight fit in the socket. That AKI board, when pushed into a socket (whether the one supplied with it or another precision/turned socket), will tend to make said socket a bit slack when socket legs NOT coated in extra solder are subsequently inserted. Since the machine I was working on had problems with both PokeyMAX and AKI at various junctures, both upgrades were variously removed, and I found that the (newly fitted) POKEY socket on the motherboard (once AKI had been pushed into it a few times) became totally incapable of making decent contact with the slimmer pins on the PokeyMAX board if PokeyMAX was installed without AKI underneath it. The only solution that worked in this case (since I didn't want to replace the POKEY socket a second time, and it would have been instantly stretched again once AKI was insertedain) was to add extra solder on the legs of the PokeyMAX, using a lot of flux and desoldering braid to get rid of excessive amounts. Having encountered symptoms such as keyboard malfunction and SIO failure without the extra solder in place, after thickening the legs up, I was able to install AKI and PokeyMAX, or PokeyMAX on its own and have everything work reliably while troubleshooting other issues. That may have nothing to do with the problem at hand here, but it's something to watch out for since I've had to do this with several machines where the PokeyMAX was sitting a bit loose in its socket.
  13. I've had a close look at the photos of the underside of the motherboard, and the OS ROM and MMU sockets are ripe with potential issues (a couple of socket pins look barely soldered to the vias), so smarten them up and check them over properly, or get someone else to do it for you before we start looking elsewhere for issues. Every trace need inspecting and checking for stable continuity. When the soldering looks like that, it's first on the list of things to eliminate.
  14. Is the online configurator described in the update procedure not working?
  15. Yes - I understand now Michael. And I certainly didn't interpret any of your comments as derogatory. Even in the context it was 'snipped' out of, however, the implication seemed thus: if something isn't closed-source, it's open source; if something doesn't have a remote chance of ever reaching fruition, it presumably has a realistic chance of reaching fruition in a timely manner, and finally, if it isn't 'one guy working on it', there are presumably several developers sharing the workload. So this implied to me the opinion that if the project was open-source and worked on by several coders instead of one, it would have a realistic chance of being completed (and I assume within the fourteen year time frame already elapsed). It doesn't seem unnatural that a discussion regarding the pros and cons of open-sourcing broke out as a result of someone remarking that the fact the project is closed-source and maintained by one person with limited time is something of a logistical drawback (which could be reasonably argued), although my view is that the 'opposite' situation (open-source, several programmers, lots of time) doesn't necessarily incite progress, let alone guarantee it. If that's reading too much into the comment (which on reflection, it clearly is), my bad as I say. It happens to the best of us Michael, and I don't think on reflection that I worded my earlier posts very well either.
×
×
  • Create New...