fabrice montupet Posted May 8, 2018 Share Posted May 8, 2018 (edited) If we resume the 99/2 ROM sets dumped, we have the ones of Klaus and I, do we have another one? Edited May 8, 2018 by fabrice montupet Quote Link to comment Share on other sites More sharing options...
+FarmerPotato Posted May 9, 2018 Share Posted May 9, 2018 If we resume the 99/2 ROM sets dumped, we have the ones of Klaus and I, do we have another one? Hi 99/2 owners, I've been following this thread eagerly. I have a 99/2, though it stopped working some years ago. It looks to have 32K ROM after piggy-backing a 2564 (28 pin) with some Kynar wiring. I have a plan to dump the ROMS in-circuit if I can't power it up. 2 Quote Link to comment Share on other sites More sharing options...
fabrice montupet Posted May 9, 2018 Share Posted May 9, 2018 (edited) Thank you FarmerPotato to answering the call. Sure, A dump of your 99/2 ROM set would be interesting :-)Can you describe in details the trouble on your 99/2? Maybe we can revive it! Edited May 9, 2018 by fabrice montupet Quote Link to comment Share on other sites More sharing options...
+FarmerPotato Posted May 9, 2018 Share Posted May 9, 2018 Can you describe in details the trouble on your 99/2? Maybe we can revive it! It has not powered up. I do not see full voltage on the power brick, with no load, so that may be the issue (it is just a linear transformer in there?) I don't have a bench power supply. I purchased a LCR meter so I'm going to start checking components. Starting with the big capacitor. Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 9, 2018 Author Share Posted May 9, 2018 (edited) Fascinating ... Just to make sure, I deliberately ignored the information in the Ground Squirrel specs that the CPU clock be divided by 2 so we have 5.35 MHz. Then I tried to save a program. This is not a proof, but a very strong indication that the CPU is driven with 10.7 MHz. I mean, they dropped the 2 KiB RAM idea and took 4 KiB already in Fabrice's console, although there is an explicit statement about 2 KiB RAM. [Edit: The 10.7 version can be loaded by the 99/4A console! Unfortunately, it does not work with the 99/2 yet. The bad thing is that I simply copied that stuff from the 99/4A emulation, but I don't know why only SAVE is working and not OLD. Have to look further.] cass992_5.wav cass992_10.wav Edited May 9, 2018 by mizapf 1 Quote Link to comment Share on other sites More sharing options...
fabrice montupet Posted May 10, 2018 Share Posted May 10, 2018 Both of your files don't work on my 99/2. I retried with audio files programs generated by Js99er and they work fine. Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 10, 2018 Author Share Posted May 10, 2018 Mmm. Then I don't understand why I can load the 10.7 Mhz version in the TI emulation. I think we proved somewhat long ago that the MAME wav files can be loaded on the real TI, although the signal is a bit different; MAME produces a square signal, while the real iron produces these [trying to describe with my hands in front of my monitor] spikes. So can you at least say that the 10.7 MHz version sounds similar to what you saved from your 99/2? The program was this here: 10 CALL CLEAR 20 PRINT "HALLO" 30 END Quote Link to comment Share on other sites More sharing options...
fabrice montupet Posted May 10, 2018 Share Posted May 10, 2018 Yes, the signals are differents between the ones generated by Mame and the ones generated by Js99er. Js99er produces the signal of a real TI99 computer.Here is one of the audio files I used with success. You will be able to compare: 99_2-speedtest.wav 1 Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 10, 2018 Author Share Posted May 10, 2018 It loads in the emulated TI-99/4A, not in the 99/2. Sigh. I am using the cassette interface in the exact same way; I guess I will now have to learn how cassettes work in MAME. By the way, your file cannot be loaded directly in MAME because it uses floating point. It must be 16bit PCM. I transformed it in Audacity. Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 10, 2018 Author Share Posted May 10, 2018 (edited) Another speed test for you. CALL POKE(-3872,2,2,0,100,4,193,6,1,22,254,6,2,22,252,4,91) CALL MCHL(-3872) You can also increase the loop count (0,100) if you find it too fast to reliably take the time. Please make sure you do a CALL CLEAR before. Edited May 10, 2018 by mizapf Quote Link to comment Share on other sites More sharing options...
+FarmerPotato Posted May 10, 2018 Share Posted May 10, 2018 Here is a photo of mine. It shows 4x 2564s with date code 8327 for 32k and 2x 6116 RAMs for 4k . Thank you FarmerPotato to answering the call. Sure, A dump of your 99/2 ROM set would be interesting :-)Can you describe in details the trouble on your 99/2? Maybe we can revive it! 2 Quote Link to comment Share on other sites More sharing options...
fabrice montupet Posted May 10, 2018 Share Posted May 10, 2018 (edited) Thank you for the pictures! :-) Your 99/2 seems to be the same as the Klaus' 99/2. 1983 week 27, six months have past since the 99/2 was shown at the 1983 winter CES and still no finalized motherboard... It is a very very long period in computers market. If TI had finally sold this 99/2, this could be about one year after the computer was unveiled to the public. Too long, unsuccess assured. Edited May 10, 2018 by fabrice montupet Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 10, 2018 Author Share Posted May 10, 2018 (edited) The 99/8 docs have some entertaining qualities; you can spot signs of sarcasm here and there. I believe at this point, also in the 99/2 group, "the air was out", as we say. If they did not know already about the pending ti-exit, they may have heard rumors. Edited May 10, 2018 by mizapf Quote Link to comment Share on other sites More sharing options...
fabrice montupet Posted May 10, 2018 Share Posted May 10, 2018 I am sad for the technical TI-99/x teams, for months to continue working on home computers projects that executing team no longer believed. Quote Link to comment Share on other sites More sharing options...
kl99 Posted May 11, 2018 Share Posted May 11, 2018 I am preparing the 99/2 setup now for the desired speed tests.Please try to stay available to answer upcoming questions. Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 11, 2018 Author Share Posted May 11, 2018 If you can wait for about 4 hours, I can prepare another test which blanks the screen before the loop. Quote Link to comment Share on other sites More sharing options...
kl99 Posted May 11, 2018 Share Posted May 11, 2018 (edited) those are the speed tests that i found in the thread. please check the timing from the video. now trying to save a program to cassette tape. Edited May 11, 2018 by kl99 Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 11, 2018 Author Share Posted May 11, 2018 I was referring to my two lines in posting #160. Quote Link to comment Share on other sites More sharing options...
kl99 Posted May 11, 2018 Share Posted May 11, 2018 Hi again! The cable situation was not easy. There is no audio out on the 99/2, at least not without a NTSC Tuner in your Television. This never was a problem since there is no sound output, and therefore a PAL Tuner can be used for the black/white video. Therefore I can not tell you, whether the 99/2 would output the tape recording sound on the television like it does on the 99/4(A) when recording/loading a program. So you don't hear what you record or read in. The 99/2 is not featuring a 9 pin port for the tape recorder, but two mono(?) jacks, one for input, one for output, and no motor control for the tape recorder. I was able to record a single line program to tape, did it twice. ti-99-2-basic-prg.zip The line is 10 CALL CLEAR (or line number 100, i am unsure now). I was not able to OLD that program back into memory. Maybe it was some cable problem, since I only have and used stereo jacks on both ends. After I recorded the tape content via line out onto my laptop, which is currently having some hardware issues with sound output (it is silent unless headphones attached(, so it stayed a bit uneasy. The only program I found was Sound recorder which only provided wma as save format. So please bare in mind and transform to the target format. I can do more speed tests tonight. Shouldn't the display be rather filled with characters beyond ascii code 127 to trigger the EOL and allow more cpu time. I am not sure whether ascii char 32 triggers this. Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 11, 2018 Author Share Posted May 11, 2018 (edited) Here is the new test; please run it on the TI-99/2. I verified that it runs on MAME, but only on the 32K version (there is a large enough space in on-chip RAM). I found out that the workspace is at F0BA, that is, also on-chip. This version blanks the screen with the line VIDENA, counts down 400*65536, turns on the screen again, and returns. Also, it inhibits interrupts in the loop and restores the mask before returning. > CALL POKE(-4056,3,0,0,0,2,12,224,14,30,0,2,2,0,400,4,193,6,1,22,254,6,2,22,252,29,0,3,0,0,4,4,91) > CALL MCHL(-4056) The source code of this machine program: AORG >F028 LIMI 0 LI R12,>E00E SBZ 0 LI R2,400 CLR R1 L DEC R1 JNE L DEC R2 JNE L SBO 0 LIMI 4 RT Edited May 11, 2018 by mizapf Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 11, 2018 Author Share Posted May 11, 2018 This is a similar test for the 99/2 24K (Fabrice). I had to split the program in two parts. > CALL POKE(-8192,2,12,224,14,30,0,2,2,1,144,4,193,4,96,240,224) > CALL POKE(-3872,6,1,22,254,6,2,22,252,29,0,4,91) > CALL MCHL(-8192) The emulated machine returns with a RUNTIME ERROR, but this is not a problem, since it suffices that it returns somehow. Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 11, 2018 Author Share Posted May 11, 2018 Shouldn't the display be rather filled with characters beyond ascii code 127 to trigger the EOL and allow more cpu time. I am not sure whether ascii char 32 triggers this. You assume that CALL CLEAR fills the screen with spaces. In fact, it fills it with FF. Quote Link to comment Share on other sites More sharing options...
kl99 Posted May 11, 2018 Share Posted May 11, 2018 You assume that CALL CLEAR fills the screen with spaces. In fact, it fills it with FF. It does that on the 99/4 at least. I can check with GCHAR tomorrow. Will run your tests tomorrow. Quote Link to comment Share on other sites More sharing options...
+mizapf Posted May 11, 2018 Author Share Posted May 11, 2018 You can watch it in MAME's debugger, and indeed, it does make sense to implement CALL CLEAR that way for the 99/2. I'm also pretty curious whether you will be successful with tape loading/saving. Quote Link to comment Share on other sites More sharing options...
fabrice montupet Posted May 11, 2018 Share Posted May 11, 2018 This is a similar test for the 99/2 24K (Fabrice). I had to split the program in two parts. > CALL POKE(-8192,2,12,224,14,30,0,2,2,1,144,4,193,4,96,240,224) > CALL POKE(-3872,6,1,22,254,6,2,22,252,29,0,4,91) > CALL MCHL(-8192) The emulated machine returns with a RUNTIME ERROR, but this is not a problem, since it suffices that it returns somehow. When I execute the code, the display stays, the cursor disappears during the execution and reappears after 1'07". No crash, no error. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.