Nezgar
-
Posts
3,628 -
Joined
-
Last visited
Content Type
Profiles
Forums
Blogs
Gallery
Events
Store
Posts posted by Nezgar
-
-
No, Atari's 1050 diagnostic utility is not compatible with the US Doubler ROM. you'll have to put the original ROM back in to run those tests. The original ROM will work with the US doubler RAM module still present however.
Other ideas include using an adapter with a 2764 eprom and a switch to allow convenient toggling between the two "modes" .
-
Yeah, the first 4 pins on the back (J11) are for the write-protection infrared LED (pins 1 power & 2 ground), and the opposing photodiode (pin 3 collector & 4 also ground).
Check out if anything is connected at the top of J11. A simple "hack" I use to force allow writing is by simply jumping/shorting pins 2 & 3 of J11, force always protected by leaving nothing connected (Great when archiving disks to prevent oopsies)
I devised a simple switch as well which ignores the photodiode, for a completely manual toggle between always write protect and always allow write:
In your drive, it looks like only pins 1 & 2 were connected to your switch, so it would would seem that your switch just "cuts" the power to the emitter LED, resulting a toggle between "Normal operation" (Write protect follows what the notch allows), and "always allow write" (forcing the LED off so the photodiode never sees infrared light).
-
1
-
1
-
-
This is a known issue with the original rev 2 happy 1050 firmware, not necessarily a problem with DUMP1050, but you might just need to add the workaround as others have done in the past. It is easily reproducible by writing to a single density disk (maybe medium density as well, can't remember) in ultraspeed mode with the write buffer disabled.
SpartaDOS X's INDUS.SYS, the Happy 1050 utility disk "enable ultraspeed emulation" menu option, and MYCOPYR sector copier are three examples that work around this bug by also issuing a happy command to enable write buffering along with the switch to ultraspeed mode as a workaround. Hias highspeed OS patch also implements this workaround if it detects a happy drive - see readme.txt here: https://github.com/HiassofT/highspeed-sio/blob/master/README.txt - and an excerpt from that text:
First, the code tries to send a $3F command (get speed byte) to the drive. All ultra-speed capable devices support this command and will return the pokey-divisor byte. If this command succeeds, the byte is stored in the table. Then it's checked if the byte is $0A. If this is the case, the drive is most certainly a Happy 1050 and a $48 command with DAUX=$0020 is sent. This enables fast writes and is needed by the Happy drives, otherwise they might write corrupted data to the disk when writing in highspeed mode (thanks to ijor for this info!).
@phaeron wrote about some specifics of the bug as well:
-
55 minutes ago, Draxxs said:
Ouch on the fake ram. I am about to be in the market for a couple pieces of those very soon I hope. I am still fighting the 1050 drive and want to put a doubler on it.
If you haven't already seen, in this thread is a PCB to install the 2x6810's fully solderless/reversable... (though Still likely need to solder to change the Mask ROM/EPROM config jumpers)
I have some bare PCB's as well as fully assembled including eprom if you're interested in that, or you can order the PCB directly yourself using links from that thread:
-
12 hours ago, Draxxs said:
Humm, I just got in 64's , 128's, 256's, and 512's they were not advertised as new but best I can tell they were all new old stock. Pins were in perfect shape. Cost me something like $3 each. Some came on the slow boat from china some from a vintage video game repair place.
The world sending electronics recycling to China has resulted in them selling it all back to us now.
I also buy all my EPROMs this way, and some are fake/rebranded (2764--27512 have a "chip ID" the programmer can read to help reveal the TRUE brand) -- and usually most work, maybe 1 out of 10 are bad... OK for the discount. I recently took a bet on a large batch of 6810 SRAMs, but all of them were bad or fake so I lost out on that one.
-
1 hour ago, Larry said:
I've not seen a wiring diagram for a 3-way to a 27512. Can you show or link one?
39 minutes ago, DjayBee said:Isn't a 3-way just an on1-off-on2 switch?
Then on1 pulls up one of the two highest address lines, on2 pulls up the other one and off pulls up neither of them.
Both address lines need a pull-down resistor to be low when not connected.
Exactly. The 4xOS board has 2 pull-up resistors (circuit designers choice/preference for pull-up or down I guess, doesn't really matter), so the middle "not connected" switch position results in a "high" status for the 2 address lines. (resulting in the "fourth" bank on a 27512) Up or down would select bank 2 or 3, with bank 1 inaccessible. The 3 terminals of a 3-way switch would be wired so that the center pin goes to ground (one of either the two inner-facing pins of the 2 double-pin jumper blocks) and the outer terminals wired to the outer pins of SW1/SW2 in that board as.
For a basic switcher using the retro innovations 2364 adapter it's the same idea: 3-way switch terminals go to 2 address lines and middle pin to ground. (the 4 through holes in the corner) -- this PCB has spots for 3 pullup resistors, but for 1 3-way switch you would technically only need 2...
-
1
-
-
5 hours ago, Larry said:
I also liked the DIY 27512 4-way, but that typically requires 2 switches/holes.
That mod can easily select between three OS's on a 27512 with a 3way switch, just 1 of the OS's would go unused/inaccessible. Same goes for the 2364 adapter PCB to switch between three BASIC ROM's with a 27C256 - leaving one unused.
-
Just now, reifsnyderb said:
Yeah. I want to use it for the BASIC socket as well. ?
Sorry I re-read what you typed as soon as I hit post and realized you meant YOUR board supported both. That's great! Reminds me of the AtariMax 32-in-1 board for the OS which is fancy because it allows switchless software selection of OS... APE Warp+ OS Upgrade Module (atarimax.com) What switching method are you planning to use to chose between the OS's you can fit on a SST39SF010 (I believe up to 8x16KB=128KB?)?
For multiple BASIC ROM's, I've been using this PCB - again using older form factor chips which may be harder to obtain as you say: 2364 Adapter - RETRO Innovations (go4retro.com)
-
5 minutes ago, reifsnyderb said:
The board is configured so that it could be used in either a 24 pin or 28 pin socket, depending upon how the jumpers are configured and how many pins are added to the bottom.
I think this "feature" of the 600XL applies to the BASIC ROM socket, not the OS ROM socket, which is always 28 pins, 27C128 compatible.
-
Why limit yourself to one? With this simple PCB, you can program an eprom with 2 or 4 OS's and with a switch you can activate whichever OS you like. Original for compatibility, OmniView 80 (even better choose Omnivew XL or XE) for 80 columns, and 2 more (ie an OS patched with HSIO routines, and maybe BOSS X2 V2 for other 800 game compatibilitiy. (There are dozens of alternatives....)
Here is an example open source PCB to do this solderlessly: DIY 4xOS Switch. I have a few onhand if you're looking for one already assembled...
But yes, if your 600XL still only has 16KB, upgrading to at least 64KB is the first order of business...
PS: OmniView and OmniMon to do not preserve PBI routines.
-
On 7/16/2022 at 8:13 PM, vinniet said:
Do you have extra pcb board assembled? I bent pin and it broke off.
Sorry for delayed response - PM sent!
-
3 hours ago, Peter Rabitt said:
I do not know the type of mod but it does format in sd, and dd, single side.
As @BillC mentioned, there is only one very rare mod that enabled double-density on the 810 - I'm very curious if you could confirm the board that is in your drive, and maybe post some pictures to help. I am particularly interested to compare the firmware ROM against that dumped from the only other one known to exist currently.
If you are not immediately willing to open the RF shield, there are some utility disks specific to the NCT Turbo 810 here: https://atariage.com/forums/topic/291937-neanderthal-computer-things-810-turbo/
Pictures of the PCB here: https://archive.org/details/Neanderthal_810Turbo/mode/2up
Manual here: https://archive.org/details/810Turbo
-
On 6/5/2022 at 3:34 PM, Nermal said:
Removed the EPROM and installed the (Happy / Lazer clone) board.
Power on and I have a Happy 1050
Run the option 2) diagnostic. The result of the RAM test will show which version of the 1050 happy ROM you have: 'PASS' for v1 and 'Pass' for v2, or possibly a failure for any other modified version...
I believe "laser" upgrades I've seen previously here on the AtariAge forums had the v1 ROM. V1 does not support the UltraSpeed protocol on power-up, and must be enabled with the utility disk. It also probably doesn't exhibit a bug where sectors are currupted with unbuffered single density disk writes.
-
At one point I toggled the partition name option and it caused all the partitions to become unreadable, since partition names cause each partition to incur a +1 sector offset. Maybe try toggling that?
-
2 minutes ago, Calab said:
- should I be formatting using a normal skew or the ultra skew?
Format in normal "skew" (the word "interleave" would be more correct) when the disk is going to be used in non-ultraspeed mode - ie with a stock 1050, or with an OS or other software that only operates in 19.2K mode. The modified firmware formats an interleave that's about 10% faster than other drives. In the case of single density, both are technically a 9:1 interleave, but optimizing the interleave to account for the space taken by the gap between the first and last sector of the formatted track.
When you choose UltraSpeed skew, SpartaDOS uploads an alternate sector interleave to the drive and then issues a custom format command that is optimized for the reduced time that it takes to transmit a sector between the drive and the computer so that the next sector arrives at the read head sooner than a regular formatted disk. A regular single density disk is about 9:1 (the next sector is available after skipping 8 sectors), the "optimized" standard interleave is about 8:1, and a US Doubler ultraspeed single density disk is 5:1.
An Indus GT syncromesh interleave is 4:1, and is too tight to be read at full speed on a US doubler, causing the drive to basically only be able to read 1 sector per rotation. Similar when reading a disk formatted with the "UltraSpeed" 5:1 interleave with a drive operating at regular speed - it will be a lot slower, reading only 1 sector per rotation.
4 minutes ago, Calab said:- should "Optimize" be enabled?
I believe all this does is allow use of the last sector of a disk by the file system. In the tradition of Atari DOS, the last sector of a disk was not usable by the file management system, and SpartaDOS I guess followed it. The post ICD/FTe releases of SpartaDOS from the SpartaDOS X Upgrade Project introduced this option in the formatter.
22 minutes ago, Calab said:- does this allow me to increase the SIO speed used to communicate with the drive?
SpartaDOS will try to engage the "UltraSpeed" protocol with all drives by default. You can tell the difference by the audibly higher-pitched and shorter sound of the sector transfer beeps. Since the US doubler has no track buffer, this will only help with the overall throughput with disks that have been physically formatted with the "UltraSpeed" interleave aka skew. It's a cool hack with minimal additional hardware which made it popular since it was cheap, but this is also where the more expensive enhancements like the Happy, Duplicator, Indus GT with Super-Synchromesh shined because their track buffer allowed maximum speed no matter the sector interleave of the disk...
-
2
-
-
Well, the 800 version is a prototype.
-
@Calab I made a patched US Doubler ROM that speeds up the stepper delay to timing similar to a Speedy 1050, and also replaced the format sector interleave tables for Enhanced Density (taken from Turbo 1050) and Double Density (taken from Percom AT-88). Note the original and this modified ROM are optimized for drives with a Tandon (made in Singapore) mechanism.
https://atariage.com/forums/topic/158768-atari-1050-roms/page/3/?tab=comments#comment-4594602
The other 1050 variant, made in Hong Kong, has a World Storage Technologies mech, and uses an alternate phase encoding that will result in a 1/4 track misalignment if the ROM is mismatched. Here is a link to a patched US Doubler ROM for these drives. It provides a faster stepping rate equivalent to the original Atari ROM for these drives, as well as the proper phase encoding. I haven't yet made a patch of this so it will format faster ED+DD interleaves.
https://atariage.com/forums/topic/156462-1050-roms/page/2/?tab=comments#comment-4060813
If you're interested in an assembled or unassembled 6810-doubler PCB I have a couple assembled and numerous unassembled here, or you can order your own from a fab. I really like this idea because it makes the upgrade completely reversible rather than permanently soldering two 6810 chips together. Details on that project here:
-
2
-
-
Can you take a better picture of the board from the top so the chip numbers are readable? Or just indicate what the leftmost chip there is...
My guess is that PCB is either a clone happy or duplicator board, and since U8 has double-stacked 6810 chips, it may have also had a switch to toggle to run as a US Doubler compatible mode...
I can see the EPROM is a 2764, which would be the size to house only a Happy or Duplicator Rom, not also the US Doubler ROM. Can you check if there is another ROM present in the U10 socket?
As far as troubleshooting goes, if you have another 1050 I'd probably start by returning it to a stock configuration by swapping in an original ROM (U10) and 6507 CPU (U9) and possibly single 6810 SRAM (U8) to rule out the upgrade...
-
1
-
-
Says on the label of your OS chip "APE Warp+ OS 4-in-1" - so I guess it's an earlier version of the current "32-in-1" product.
https://www.atarimax.com/warpos/documentation/
So my guess are some combination of:
800XL OS800 OS
1200XL OS
Warp+ OS
-
2 hours ago, Beeblebrox said:
This effect seems like it could be done pretty easily with pre-rendered vertically short animation frames for the cog-wheels in GR.9 mode, and using simple DLI tricks to swap frames and repeat down the screen. The overlay text would presumably have to be done with PMG.
-
1
-
-
Official ICD source code is not "out there". However, there is a disassembly of unknown origin and limited usefulness due to lack of comments here:
https://atariage.com/forums/topic/115434-new-indus-gt-firmware/?tab=comments#comment-1396039
I believe @phaeron did a disassembly as well, and his comments are usually stellar, which I can't find the post about at the moment.
Keep in mind that the stock 1050 ROM works with the USD doubled-6810 SRAM installed, so it's just a matter of swapping the USD vs Stock ROM. you can put it on a single 2764 and install a toggle switch between the two ROMsand the drive itself can be left in the "Mask ROM" configiration too with an adapter like this:
-
All versions of SpartaDOS merely print the directory entries in the order they are on disk. You can solve this with a utility called SORTDIR which sorts and rewrites the directory as per your specified preferences. I think it came in the 3.2 construction set, maybe it's in the SDX one too.
-
My first thought is this is the result of a custom/replaced OS...
Checking some of my XL/XE OS's...
- OmniMon XL - BOOT ERROR
- OmniView XE - LOAD ERROR
- XL BOSS V1 - LOAD ERROR
- XL BOSS V2 - BOOT ERROR
- OSN XL - BOOT ERROR
- SuperMon '86 - BOOT ERROR
Funny XL BOSS switched back to "BOOT ERROR" in V2...
-
1
-
On 4/18/2022 at 5:06 AM, 1050 said:
It's only 192K, you don't get to count the original 64K because they are just sitting there using up power and doing nothing at all.
There is literally only 256KB total in the system. 64KB + 192KB. Look at the motherboard in the original post - 8x256kbit chips. The chips on the additional PCB only contains logic chips.
@Steverd - there is a slightly updated version of that memory tester available here:
http://atari.sk/extended-ram-test-0-22-0-xram0220-xex/
For reference, here is what 0.22 looks like from a system of mine with an ICD Rambo 256K 800XL. I'm curious if the newer version presents differently that this on your system. I'm curious about the difference of missing row at $Ax vs $8x.
I also buy all my EPROMs this way, and some are fake/rebranded (2764--27512 have a "chip ID" the programmer can read to help reveal the TRUE brand) -- and usually most work, maybe 1 out of 10 are bad... OK for the discount. I recently took a bet on a large batch of 6810 SRAMs, but all of them were bad or fake so I lost out on that one.
Like us on Facebook
Follow us on Twitter
AtariAge on Instagram
Watch us on YouTube
Chat on Discord
Atari 810 happy board??
in Atari 8-Bit Computers
Posted · Edited by Nezgar
It's the data separator -- an add-on recommended by WDC's disk controller specifications that Atari chose to ignore in early production units leading to unreliability on disk reads especially on the inner tracks. Initially Percom, then later Atari rectified this by using this data separator board plugged into the WDC chip's socket to perform the function, and plugging the WDC chip into it.
Here's some excerpted details from Percom & WDC.
Radio Shack used the same chip in the TRS-80, and also made the same mistake. Percom's board worked in both the TRS-80 and the 810 -- before Atari produced their own offering.
There are four things in an 810 that has all of the "grass valley" upgrades:
My own experience showed that trying to add a data separator board to a pre-grass valley 810 without the upgraded power board will barely function as it can't produce enough power to spin the drive... Clearly showing me their need for the upgraded power board... Plus the RPM control on the original power/analog board is EXTREMELY touchy, also contributing to unreliability of pre-grass valley drives.