Jump to content


+AtariAge Subscriber
  • Content Count

  • Joined

  • Last visited

Community Reputation

233 Excellent

About J-Data

  • Rank
    Chopper Commander
  • Birthday 06/07/1966

Contact / Social Media

Profile Information

  • Custom Status
    No Expert
  • Gender
  • Location
    Melbourne Florida
  • Interests

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Looks what's going to be reviewed by @jonecool on Wagner's tech Talk! Can't wait to hear his thoughts.
  2. Another easy thing to do is just copy the TIPI disk directory from \\tipi\tipi to a backup folder on your PC. You can then dump it right back into any new image. Does not bring all the configs with it, but way faster than creating and burning an ISO. Stuart's browser mouse (9.3) and TI Artist mouse both work for me with TIPI 2.4.
  3. I'd be happy to help test, although my issues have settled down quite a bit. "CALL TIPI" issue do still occur but at a much less frequent rate. (The SNEK high score issue is still 100% of the time.) All my testing was with external power using an official RPi power supply.
  4. The work you've done is awesome and I'm grateful for everything you've contributed for this community. The TIPI brings me joy every day. I repeated my test a sidecar TIPI and RPi zero with similar results, so I'm 80% sure its just the lower performance of the zero bring out a corner case that may not show up under normal testing. For now, 1.61 works great for me, and I'm stick with it. Let me know if there is anything I can do to help.
  5. I repeated the same test with another console, a different TIPI/32K and a different RPi zero and got the same results. (no IP at times, CRUBASE >2000, and generally flaky behavior) I repeated a third time with a TIPI/32K but with RPi 3b (connected to wifi) and DID NOT see the Ethernet or CRUBASE issues. (I did still see the SNEK high score issue though). I suspect this is a RPi Zero W specific problem, like @jedimatt42 suggested earlier. I could do a test with the original Sidecar TIPI and a RPi Zero If anyone thinks that datapoint would be useful. I should mention, the longer I test, the more stable 2.4 seems to get. I seem to remember a similar behavior when upgrading to 1.59 or so, that it took some time for the upgrade to "settle in" on the RPi zero. I know it doesn't make much sense, but It does feel similar. If anyone needs it, I posted a back-up ISO of 1.61 (and 1.43) if you want back out of 2.4, and don't have an image to go to: http://www.lilyandgrace.com/ti/TIPI-32K/ISO/ Obviously the .zip file is much smaller and will be faster to download.
  6. Just to add another data point, I'm seeing some similar issues with 2.4 on my TIPI/32K with RPi Zero W. I burnt a fresh SD card with 2.3 form the new ISO, and then upgraded to 2.4. so its a fairly clean image. (I haven't attempted upgrading from 1.6x to 2.x, since I've had issue with upgrades, so I avoid them unless there's something I really need, and I do keep back up copies of all the old images as ISO files.) So my issues are somewhat different. Sometimes I get an IP address, sometimes I don't. When I get an IP address, the CRU address is >1000, when I don't its >2000, which should literally be impossible with the new rev 3 of the CPLD (Ignores CRU MSB). The answer changes from one "CALL TIPI" to the next, and the first boot on power up seems to be the most unreliable. Also, the myti99.com apps (Chatti, Chess, SNEK) are all slightly broken with 2.4, either forgetting my logins, not launching reliably, or having new odd bugs that make them unusable. I've tried going back to my old V1.61 image and these issues seem to go away. I think you've already explained what might be causing the app issues a few posts ago. Anyway, I'll going back to my 1.61 image for now.
  7. I just wanted to see if it was possible, and to get a better understand the NanoPEB operation and design. To me, the disk capability of the TIPI is so much better than the NanoPEB, its not even comparable. The CF disks were great for its time, but getting a bit long in the tooth. My only reason to add RS232 was to play the my old terminal emulator 2 cart and 300 baud modem. Unfortunately, the phone number for CompuServe doesn't appear to be good anymore. Still have many old TE2 sessions saved on tape. Also wanted to see if a similar approach could be added to the TIPI/32K. NanoPEB did it with a TMS9902 (Still available), an EPROM and and very small CPLD, (XC9536) and we already have 2 of three with room to spare, so I have no doubt it could be done. Just not sure there's a need.
  8. I wasn't sure it was possible, but apparently you can run a TIPI/32K and a NanoPEB at the same time. Just can't figure out how to connect the SAMS card yet. (Disable NanoPEP 32K ram, and add a header to edge card adapter to connect the SAMS and the NanoPEP?) Needs more integration.
  9. Five more TIPI/32K boards built and fully testing (and have many more partially populated boards if ever needed). Will be in the mail to @arcadeshopper in the morning.
  10. I probably should know this, but does CIB (Complete In Box) mean everything that originally came in the box but used, or unused and still in the original box (i.e. New Old Stock)?
  11. Cool. I need to add the shutdown mod. What GPIO pin did you end up using for the shutdown input? Between the TIPI interface and the CPLD JEDEC interface, the ones commonly used are already taken (GPIO 17 & 21). Can you post the python script you used?
  12. Love it! Is that label, clear or metallic? I haven't been able to get the original lettering off myself yet. Acetone usually works but not in this case.
  13. Or you could put the TIPI activity LED there. Its in about the right spot.
  14. I'm really sorry. I posted the wrong version of the updated file, which was an earlier test build that still had a bug. The correct one, that had been fully tested is now attached. (I was in a rush to get back to work after lunch). The bad version does disable the RAM with the 4th jumper set, but will always have a CRU address of >1100 with ANY jumper set, including the 4th (used "&&" rather than "&" on the CRUBASE mask operation to strip out the 4th bit in Verilog). This corrected version give address >1000 and ram disabled with only the fourth jumper set (the goal), and from 1100 to 1700 with first 3 jumpers sets regardless of the fourth jumper. Sorry for the confusion. Hopefully this resolves the issue you're seeing (both TIPI and NanoPEB with CRU address of >1100 might explain the behavior.) tipi-32K_top-v3.jed
  15. Updated CPLD file for the TIPI/32K is attached. In this version now ignore the 4th CRUBASE jumper for CRU addressing but still uses it for RAM disable. This will allow a CRU addresses of >1000 to >1700 with or without the RAM disabled. I tested with a SAMS and work, but don't have a NanoPEB. You do need to cycle power between jumper changes or the CRU address will not update. RAM disable is immediate. Instructions for loading the CPLD using the Pi are located here: https://github.com/jgparker/tipi/wiki/CPLD-Programming but you will have to manually load the attached JEDEC file into the pi and update the file name. <<Deleted file>>
  • Create New...