warerat

4.38V is pretty low. Check a +5V somewhere else, like pin 16 on the 'LS138 above it. If it looks good, run a temporary jumper from pin 16 on the 'LS138 to pin 17 on the POKEY. If it fixes it the problem, you probably have some leaky electrolytics in the DC filter where the power connector is or your power adapter might be flaky.

Are you saying you can barely hear POKEY sounds, or not at all? Are you using the RF mod output or direct from the A/V port? When you say "turn it up", do you mean your volume on your display or the software controlled volume? Could be several things. 1) Check to make sure you have +5V on POKEY pin 17. There are individually filtered +5V supplies on the 130XE. 2) Bad op-amp at U1. Labeled LM358 or 1458. Look for bad joints around here. Mixed output comes from pin 1, POKEY input at pin 6, and GTIA input at pin 2. In a pinch, short pin 2 to 6 to see if any audio comes out. If you hear the POKEY, check connections at R1-R4, and C17. 3) Bad POKEY If you get no audio from the RF output, but you do at the monitor jack, then either your RF mod is bad or there's a bad connection. Good luck!

Hey, that's pretty neat. Sounds about the same on my 800 through the speaker, but a little scratchier. Never could figure out the words before the "Stay a while"...

They normally do, if you use Axlon/Byrd boards. These are just replacement plug in boards-- no motherboard modifications required. This mod turns the 800 into an XL/XE (XL OS, complete memory map, and PORTB switching so you lose ports 3 and 4). With the modified boards plugged in, it is functionally identical to a Peterson 576K 130XE (64K main + 512K banked). Only thing that is wierd is the System Reset key gets clobbered by some software because it is a NMI interrupt on the 800, rather than a true hardware reset. Which is why I have to use a 800 NMI handler in the XL OS. Permanent solution is to wire the reset key to the reset circuit in the 4050 buffer, but then if you went back to 800 mode you'd get a cold start everytime you pressed reset. Other than that, it runs all XL/XE software and I even have to hold down option if I switch to 320K mode to disable BASIC.

Hmm, no ROM upgrade could fix this. As Bryan explained it's a limitation of the controller. The FDC won't start the format it it doesn't see the index pulse. There are some fixes, but they involve more than just ROM change. Also, not every XF-551 can read and write to the flippy side. It depends on the mechanism. Some mechanism don't become ready if they don't see the index hole (disregarding the FDC issue). It seems that most units sold in US don't have this problem. I think that the FAQ mentions which brand mechanisms can access the flippy side and which not. It is possible that this could be fixed messing with the drive's jumper. Some have jumpers to select the condition to become ready. But this is only for being able to read or write to the flippy side. Formatting is still limited by the FDC. Well, I believe it did come with a ROM. heres the link: http://www.nleaudio.com/css/products/XF551enh.htm I've got some XF551's with Enhancers. It's a potted module with a bunch of wires soldered to the mainboard. The XF551 I have has a CSS ROM for 3.5" floppies and Ultraspeed.

No, but I'll borrow your CF card with the IDE converter. I'll trade you two 1050's and a XEGS for it.

The way everything is wired, the 800 looks exactly like a Peterson 576K XE (without selectable BASIC, and the switch ). But all you have to do is swap wires for PB1 with the basic select (instead of an address), and you get a 320K XE machine with BASIC on the same hardware. ANTIC enhanced is always implied, and not selectable with PB5, as there is no EMMU in the 800, and it doesn't have separate CAS banks like the XE (this is a SRAM based design anyway to minimize ICs, so I don't use CAS/RAS) so ANTIC always sees the same memory as the CPU. I could implement this, but it will require some major changes because HALT isn't present on the memory board to signal an ANTIC cycle. I'd have to move to a bigger CPLD to do this because I'd need more I/O pins. But this design works as is, considering it is EMMUless and FREDDIEless. Just think of each GAL as fancy quasi-EMMUs/MMUs for the 800. There are a few things missing from this design, mainly some PBI signals: - No MPD to disable $D800-$DFFF for a PBI ROM - No CASINH to disable main memory access - EXTSEL is different on the 800 - CAS is generated locally on RAM cards and not present on any connector First two can be implemented rather easily if I had a bigger GAL (I'm using all the pins on it right now). There is an unused pin on the OS connector, one could theoretically run the $D1XX enable out from the OS board. Is PBI possible with this? Absolutely. Most of the signals are there. I know a 1064 memory board wouldn't work, but I don't think it's that practical to run PBI on the 800-- yet. As Mr. Atari's MyIDE stuff is made to run only on $D1XX or $D5XX, and slot 3 has $D6XX, I'm going to modify the OS board to send the $D1XX select in it's place to slot 3 so I can use all the internal MyIDE stuff. Would be nice to have a $D6XX version. I've got part of the PCB laid out for the memory board-- time is limited this week to work on it. Will probably save some cost to use SMDs instead of DIPs. If I do away with the last 16K board and replace it with SRAM, I can do away with the analog RAS delay circuitry on the OS board and have a simpler design, but no DRAM will be able to be used in the machine at all, at least while in XL/XE mode. Considering the 512K SRAM was $5.90, and given the small memory addressing capability of the Atari, I'm all for SRAM based designs. If you wanted 1088K you could just piggy back another 512K SRAM on top of this one with a 74LS00. It would cost you the selftest, but I like Mathy's idea of keeping compatibility with his 1MB DRAM design, so I might implement that later. I'll be in the San Antonio/New Braunfels area late next week if you want to see a demo of this on your hardware. All plug in, no soldering required. Seeing is believing.

But does it retain 100% compatibility with the 800? If I wanted to rig up a switch I could retain the 800 OS and the XL OS. But using ports 3 and 4 in XL mode is mutually exclusive. That *is* what makes the XL/XE what they are. Had Atari decided to use four joystick ports on the XL/XE and do the switching with a latch or another PIA, we wouldn't be having this discussion. The 800 is a different design than the XL/XE. I'd rather have SpartaDOS, large ramdisks, and the ability to run newer software than two joystick ports. If I want to run an 800 game I can switch cards, or run the translator disk. I can have the best of both worlds, it is just a card swap away now. All I've done is added the necessary logic using the existing 800 signals to make it look like an XL/XE memory map. But I still retain both cartridge slots. And it is always ANTIC enhanced (cannot be switched-- ANTIC will always see the memory the CPU does). It's just a hybrid now. So, if I want to use this "upgrade" to "downgrade" I could: 1) Unplug the connectors from ports 3 and 4 and plug in the original personality module. If you're a purist, you can plug the original 16K boards back in too. 2) Unplug the connectors and hack a 32-in-1 OS in there. I believe it connects to PHI2, RESET, and something else. Then you're two joystick ports away from an XL. 3) Run a translator disk 4) Patch the misbehaving 800 software to run

After weeks of working on and off on this and ironing out bugs, I'm finally able to declare victory. What I have now is a 100% XL/XE compatible 800, with 576K, including switchable selftest and built in BASIC. What's really nice is that the power consumption of the GALs and SRAMs are much lower than the two 16K RAM cards it replaces, so the whole machine runs a lot cooler. I'll probably end up replacing the remaining 16K card with a SRAM to drop consumption even more. Nothing like a 512K ramdisk on a nice keyboard. The top board is a highly modified OS board, it contains a slightly modified XL OS, the BASIC ROM, and RAM (effectively the missing 16K on a 48K machine). It receives PB0 from port 3 for the OS RAM/ROM select (blue wire on right), and has two additional inputs (two wires entering on left) for the BASIC and selftest ROM enables that come from the RAM board. The RAM board is on the bottom. It goes in the middle memory slot and replaces the last two 16K modules. It provides the regular 32K and an additional 512K in the normal $4000-$7FFF zone. The wires going out on the right go to ports 3 and 4 for the memory enable, bank selects, basic, and selftest. The two wires on the left connect to the OS board to provide the additional enables for switchable BASIC and selftest ROMs. Now that slot 3 is free, I'm going to make a custom MyIDE based card to have an internal CF disk. In theory, as the 800 has +12V on the slots and I've dropped most of the power hungry DRAMs out, I might even be able to run a low-power harddrive off it. Don't mess with the beige.

I have two XF551's with bad mechs. I could use a pair if you're willing to part with them. Amen to that-- you have to take care of whatever stocks are out there, I know I do!

Pin 18 of the MMU (CO61818) controls the BASIC select. On a stock 130XE, this is connected to PB1, pin 11 on the 6520 PIA if you want it controlled through software.

My point, vague as it might be, is that the dual antic upgrade isn't really adding anything new to the system, but just using the same technology that the 8-bit was shipped with in a different way. But in this is one problem: how many new ANTIC chips you can buy today? You must destroy one Atari to rebuild other, so...? Sorry, for me better way are using modern components - more Ataris will be live... 1030555[/snapback] Not true! BOTH Best Electronics AND B&C carry Antic chips that aren't from destroying other Atari's, they are REAL REPLACEMENT parts bought from Atari. I'm sure they aren't the only ones to carry them either, but if they are, they both seem to have plenty. Obviously you need to pick yourself up the Best Rev. 10 catalog, the greatest electronic (and Atari) catalog in the world! 1030626[/snapback] I've always wondered exactly how many GTIAs/ANTICs/POKEYs are left. 100s? 1000s? 10000s? Think how many have ended up in landfills. Someone has a VHDL model of FREDDIE, and the equations for the MMU and EMMU are out there. I just worry about these three core chips. I believe there's a VHDL model of POKEY, see the Asteroids game on www.fpgaarcade, the code looks about 90% complete (missing keyboard, serial port). It would be neat to make a quad POKEY and have a single chip solution.

If you built your interface with a 75189 chip try changing it out for a 14c89, or try placing an in-line resistor on the Atari data-out line. The input impedance on the 75189 is much lower than a 1489 and can cause this problem. Steve 1030011[/snapback] I'm using a MAX232. The SIO cable to the small PCB is about three feet long. The DB9 cable I used is about six feet to the PC. SIO 2x mode works fine.

As I understand the Atari documentation, there is no clear definition for $d600/$d700, but it can be understood from various drafts that these pages are NOT to be used by PBI devices, EXCEPT for just one: the 1090 expansion box. At the other hand, the definition for $d100 - a large part of it, $d100-$d1af - is completely clear: the I/O registers of PBI devices should be decoded there. There is even some support on the external bus for that - namely the D1xx signal, which facilitates register addressing at the device's side of affairs. There is no such signal for $d6/$d7. So a floppy/harddisk/printer interface has no point in using anything except $d100. This has been suggested to him few days ago. Alternative locations proposed are: $d020-$d03f, $d700-$d71f or some other location on page $d6 (maybe the upper half of it or something). The first one would be the best IMHO. ANTIC does. 1029817[/snapback] I like the idea of mapping it in the GTIA range, keeps the video registers together. The 800 has the $D6XX select on the third memory slot.

Sounds exactly like my problem! I will shorten the length to see if that improves noise immunity. I figured at first maybe it was something in the software I could tweak, but nothing seemed to help it. This is on a 1.7 GHz P4.

Technically, yes. In use, no. My 800 to XL mod uses a 32K SRAM for the RAM under the OS, but I ground A14 to use only 14K of the chip (remember $D000-$D7FF can never be mapped out). One could technically wire a switch to it and have two 14K RAM OS's. But no compatible way to control it from software, unless someone decided "from here on out address $XXXX will control the OS banks). I know $D1XX is used by PBI, I'd like to see a comprehensive list of devices that map $D5XX, $D6XX, and $D7XX so any new devices won't bump into each other and that way we can design around everyone else's memory maps. For example, I'd like to see that new video board with the 512K mapped into the GTIA space assuming all the registers can fit.

I haven't got mine to work reliably in APE 3x mode. It will sputter and I usually have to end up hitting the BREAK key to have it finish loading something. Then I'll get error 139/140 trying to do further reads. Haven't figured if it's my interface (I have a MAX232) or the software settings, but changing the timing didn't seem to help.

Can the 1772 handle that? What's the clock rate you'd have to use for HD? 16 MHz? Please note that I didn’t mean that would be easy. Just that it would easier, or less difficult if you want, than using a 5.25 HD drive. Unless that all you want is to use a modern 3.5 HD drive in DD mode. That wouldn’t be a problem, except for changing the drive ID because modern drives are jumper-less. Yes, clock should be 16 Mhz. Some, probably most, 1772 can support that much overclocking. And there is a version that Atari used in later ST models, nicknamed the Ajax, that is guaranteed to work at 16 Mhz. You would still have some complications. A custom ROM, the higher clock interface, etc. And this is assuming the drive’s CPU can cope with the faster transfer rate at all (which I don’t know). But with a 1.2MB drive, you’ll have all these complications plus some more. 1028515[/snapback] Wonder if this guy is really doing 1.44MB with a WD2797 and a 2 MHz clock for a Z80: http://home.swipnet.se/~w-68269/wd_fdc/wd_fdc.htm

I don't think that's the case; the Hyper-XF ROMs (linked to elsewhere on these boards) are both 8K, as was the dump that I made of Atari's original ROM (which I then modified with Bob Woolley's code for 720K support). 1028522[/snapback] I just verified on my XF551 that A12 is indeed strapped to ground, unless yours has been modified somehow. The files are 8K, but there is only 4K of code (look at the top half from $1000-$1FFF).

It would be quite complicated. Even for SD/DD disks, it won't work just like that. If you want to use an HD drive, then it would be much easier to use a 3.5 one. 1028484[/snapback] Can the 1772 handle that? What's the clock rate you'd have to use for HD? 16 MHz?

I've got a few 2764s and a programmer and could probably burn you an EPROM for a nominal cost (I'll even test it in my XF551 before I ship it). PM me if you're interested. 1024145[/snapback] Cool! I'll get to you soon on this! 1024149[/snapback] I believe the XF551 has A12 hardwired low so you can only use 4K. I have a CSS ROM in my 3.5" XF551 that does ultraspeed and fixes bugs on a 2732.

While he's at it, I think he's going to figure out how to power it with molasses (from the 7800 forum): 1028116[/snapback] The cancer is spreading, or should I say: The Cancer Is Spreading. Beware Of Young Trolls Who Don't Get Out Of The House.

Should be a group of jumpers on the board near the floppy connector. Even a newer model 1.2MB 5.25" I have from the PC days has them-- they're labeled S0-S3 on my drive, and the jumper is on S0.

I have! I have a semi-working prototype right now (Almost Rice has seen it). It makes the 800 into a 576K XL machine (gives 32K to main memory, and full 512K to ramdisk), only uses four ICs. That frees up slot 3, and slot 3 already has the $D6XX select for an internal MyIDE. The RAM board requires a few extra wires to the ports 3 and 4 for the bank selects and extra RAM enable. Now it's just a matter of cleaning up my connections-- I have some intermittent failure somewhere on my protoboard that causes it to work sometimes and not others, but the design is complete. Booting SpartaDOS on the 800 is surreal, running RD.COM and getting a ramdisk is insane. 1025904[/snapback] Sounds cool, but I'd need 64k; 32k restricts you further it seems to me making the other upgrades less effective( or pointless in the case of the XL OS), might as well keep the 48k ram and just do an Axlon compatible memory upgrade for ramdisk. What does it matter if you have a ramdisk and MyIDE if you can only run 32k programs (which are fewer and older)? Am I misunderstnding you somehow? Isn't the whole point of having the XL OS so that you can have a full 64k of main memory and compatibility with all the 64k software?!? 1026127[/snapback] It is a 32K RAM board with extra XL banked memory in slot 2, leaving slot 3 free. This is used in conjunction with the 16K in slot 1-- there's your 48K. The additional 16K RAM is on the OS board, so it *is* a full 64K 800XL with extra memory. The point of this whole upgrade is to make an 800 run the newer programs, not take away any functionality. So: OS Board 16K RAM/ROM + Slot 1 16K RAM + Slot 2 32K RAM & extra = 64K main memory + extra RAM with an XL OS in an 800. With both boards in place my 800 is a 576K 130XE, sans switchable BASIC, self-test, and HELP key.

I have! I have a semi-working prototype right now (Almost Rice has seen it). It makes the 800 into a 576K XL machine (gives 32K to main memory, and full 512K to ramdisk), only uses four ICs. That frees up slot 3, and slot 3 already has the $D6XX select for an internal MyIDE. The RAM board requires a few extra wires to the ports 3 and 4 for the bank selects and extra RAM enable. Now it's just a matter of cleaning up my connections-- I have some intermittent failure somewhere on my protoboard that causes it to work sometimes and not others, but the design is complete. Booting SpartaDOS on the 800 is surreal, running RD.COM and getting a ramdisk is insane.