Nezgar

Looks like the answer according to Mathy's readme file is YES Looks like his resulting OBJ is 11 bytes shorter too, as I think there's some extra bytes in the one I uploaded. His download includes the M65 source, so I may play with that.

"The fix" merely allowed the program to properly show an error when it ran out of RAM from loading too big of a file instead of just locking up, so the patch wouldn't have made any difference to whatever it did to break bbxfer...

I think i got it - with the help of the source in the ANALOG article, and Omnivore disassembly: *= $252c LDX #$02 ; read three bytes - changed to #$01 to only read 2 bytes STX MEMTOP L2530 LDA D1,X Then down in the 3 bytes "D1:" *= $25e1 D1 .byte $44, $31, $3a Changed it to D1 .byte $44, $3a, $00 but it only reads the 1st 2 of those 3 bytes, resulting in just D : - so now it should operate on a file in the same drive the program was run from, at least in SDX. Please test, as I didn't test on a real DOS I present v1.2 - (could save a few bytes here by assembling from source instead of this binary patch...) streamln12.xex

Hehe, I've thought the same thing over the years.... but it does work in SpartaDOS... as long as your file is on D1: Edit: Looks like the three bytes that spell "D1:" are at 0x0347-00349 and maybe easy to change it to a different drive # , ie D2:, but it has to be three bytes. D:<00> puts a heart after.... and same if I shift the preceding RTS forward a byte with a NOP... Time to look at the source code...

Good Q, not sure -- looking in a hex editor offset 908 decimal is A4 (LDY Zero Page)... the above patch replaces it with 160, or A0 (LDY Immediate) , which seems right according to above? There I patched it, I also patched it to say "V1.1" instead of "V1.0" ? Hopefully I did it right! streamln11.xex

The Streamliner by James Hague, ANALOG #56 page 33: https://archive.org/details/analog-computing-magazine-56/page/n33/mode/2up Attached is my copy... Edit: Also, another useful utility is Super Packer, also attached. lets you compress individual segments, maybe re-save "optimized" (remove FFFF between segments) without packing too - can't remember. streamln.xex superpck.xex

Basically completely different games.

Fire buttons are handled by the GTIA chip, not the PIA, so first thing is to check continuity from both joystick ports pin 6 through to the GTIA chip pins 8 and 9. If not good, there may be a cracked solder joint that could use some solder reflowing. If it is good, next would be to try replacing the GTIA chip...

Untick "flash" ? - that's not a flash chip... Might also want to select a different "similar" chip like ST 27C64. Match up voltage. You can confirm the chip is indeed blank beforehand with the blank check button. (If it's already blank when you received it, it doesn't need a UV bath) At least those are my 1st thoughts...

Having grown up with the 1010... a 410 seems A) massive B) 110V AC going straight into it, oh my!

I feel the pain - I get about a 70% success rate programming 21/25V chips on a version 1 TL866... It claims max 21V but in reality it maxes out about 18V, so its borderline. I often have to use the tricks you mentioned... Slow down, repeat programming, etc... Interesting if injecting an external voltage boost didn't work for you. I have been planning to try this using this doohickey I picked up from aliexpress that can step up any variable voltage from 3-25V : https://www.aliexpress.com/item/1005001417851839.html?spm=a2g0s.9042311.0.0.792f4c4dkzidmK I guess if your 2732's aren't working for you maybe you'd be interested in a trade for some blank and or programmed chips for your high voltage ones... although your postage costs to Canada may negate the worth. Here is an eBay seller of 12.75V 2732's that worked well for me - $13USD and free shipping (from China) for 10: https://www.ebay.com/itm/123248646835

Could the traces all be straight through on the PCB if one SIO socket was on the opposite side of the PCB? (ugly, but curious if my assumption is correct)

Well, I guess you could simulate different speeds including 2x (3.58Mhz) using Altirra:

OK, so you can still run XRAM from another DOS. Also SysInfo will show some extended RAM info: http://drac030.krap.pl/en-si-pliki.php

Ah, so it's not really a resistor, just a 0-ohm jumper?

Quick way with SpartaDOS X is MEM /X Then run v0.22.0 of XRAM: https://atari.sk/extended-ram-test-0-22-0-xram0220-xex/

Some advantages that come to mind: Cartridges load instantly.. faster than loading from disk as you suggest. It will always be just as fast on every computer, vs SIO which could be as slow as 19.2Kbps depending on the users configuration. With only two conventional single sided disk drives, even at double density you only get 360KB of storage, not 1024KB. It would take 5 double-density disk sides. There is a disk version of AtariBlast that requires 1MB of RAM, but the EXE will take a LOOOng time to load over SIO if you don't have a really good ultraspeed SIO divisor working with a SIO2PC-USB or a cartridge based storage like SIDE. AtariBlast is kinda cool that it runs on a 400/600XL with only 16KB of RAM, but it has full instantaneous access to the 1MB of ROM data via the cartridge The bad apple demo streams the 1MB of compressed video data from the cartridge. Adding additional latency waiting for unpredictable disk I/O would make it basically not work well in most situations... A RAM-load version for machines with 1MB of RAM would be possible but was not released by R0ger since there is relatively few machines with this much RAM. The 1MB cartridge works on ALL machines.

If the left column's 8 chips for the base 64K is still 4164's then it's not required. If the 2nd column's 8 chips for the extended banks has 41256's, but it's only setup for 64K extended so far and not 256 (for 320 total) then it will be required for those...

That's a mouthful of a sentence there. Generally, if you wanted to run most disk based games, and cracked cartridge games from disk on the computers, you needed at least 48K RAM. Then without a cartridge present, there is RAM in the cartridge address space that the programs can be loaded into. Since the machine code generally makes use of JMPs to specific memory addresses, without recompiling the whole darn thing it was easiest for the crackers to patch as little as possible, and letting most of the code run in the same memory address range as the 5200 (From 16-48K in the 64K memory map) For more details of how that was done, you can listen to ANTIC Interview 90 - Glenn the 5200 Man where he talks about his work. Most games released by Atari themselves for both platforms were actually completely different releases. The 5200 "ports" to the computers were generally from the cracker/pirate community. One notable exception, Centipede released on cassette from Atari Europe, was the 5200 version found to actually have Glenn the 5200 man's handler included - CTRL-G was a a dead giveaway it had Glenn's handler See this thread: https://atariage.com/forums/topic/318326-my-centipede-story/ I'm not sure of specific cartridges originally released with 32KB ROM... there are many that have been repackaged by enthusiasts in the years following.

I'd say AtariBlast! - 1 Megabyte. (aka 8 Megabit) This one is also cool that it also runs on a machine with only 16KB RAM. (400 & 600XL) http://www.atarimania.com/game-atari-400-800-xl-xe-atariblast_s25209.html Also Bad Apple demo by @R0ger is 1MB / 8Mb: https://www.pouet.net/prod.php?which=70220 The computer can only directly address 16K in the cartridge, and common banking methods either switch the full 16K, or 8K of the 16, while 8K stays static.

The only difference is the ROM was shrunk, I/O addresses were shifted in the memory map (Still in the same range as the ROM though.... I don't see any advantage, maybe less electronic components) , and the directly addressable cartridge ROM area was increased to 32KB compared to 16KB in the computers. Most 5200 games were directly ported to the 8-bit computers with a simple remapping of the I/O locations, and a VBI translation routine to convert the 8-way digital inputs the computers use, to the analog potentiometer based 360-degree signal that the games were written for. Only 1 or 2 games could not be played well with the more limited 8-way joystick input, and maybe those controller keypad buttons too, which could be translated to input from the keybaord.

What does the Atari version look like? The commodore one I see in this ebay listing sure looks a lot like an XC12! https://www.ebay.ca/itm/203299443257

Yeah the original Atari shells have the dumb clips, yours appear to be based on the design by Sikor and Mq which is a big improvement.

Trigger input from the joystick ports to go the GTIA, and the rest of the inputs lead to the 6520 PIA, so the 6520 chip might be the thing to try swapping out. Unfortunately that involves desoldering since there's no sockets in XEGS. You can swap one from any other Atari 8-bit computer. Also test teh "Bad" one in that other computer. If you don't have another computer, 6520's are a commodity part easily found today.

That would be hot glue that has come loose. At least for the "ribbed" style XEGS cartridge, there are tabs on the left+right of one of the halfs, that will break if pried. You may have success using a vice grip, example in this post: https://atariage.com/forums/topic/284679-repos-for-sale-what-do-you-think/?do=findComment&comment=4150380 See what the inside of the cart looks like here: https://atariage.com/forums/topic/289505-xe-super-cart-128-midi-maze-offering/?do=findComment&comment=4244634