mike_z Posted November 11, 2021 Share Posted November 11, 2021 Recently I came across my old TI99. I thought I'd try it and see if it would work. Seemed to have a few connection problems and the fuse let go on the power supply, but after a couple hours of cutting and soldering, it came to life. Then I tried writing a short basic program with the intention of saving it to cassette and then seeing if I could load it in again. I wrote three lines of basic, connect the cassette and typed SAVE CS1 After responding to the prompts, the program was saved. I listened to the tape and you can hear a leader and then the data burst. There is no trailer, is that correct? So Next I tried to load the program and it would not work. I turned off the TI99 so that there was nothing in it. Typed OLD CS1 started the cassette at the correct prompt. The tape recorder made the leader and data sounds, but the TI99 did not respond. After a while the TI99 responded NO DATA. After playing around with it a while I found that the DB9 I/O cord was bad. Replaced the DB9 connector and tried again, still no soap. I tried changing the volume (the only adjustment on the recorder) but still no data. Next I down loaded Audacity and tried that instead of my old cassette. Once again I could record the BASIC program but could not load it back to the TI99. Again I played with the volume etal but no soap, any ideas.? Thanks Mike Quote Link to comment Share on other sites More sharing options...
Count9929A Posted November 11, 2021 Share Posted November 11, 2021 Usually, if the TI99 responds with NO DATA, one should lift up the recorder volume. I would recommend to clean the heads of the recorder Quote Link to comment Share on other sites More sharing options...
mike_z Posted November 11, 2021 Author Share Posted November 11, 2021 Thanks I tried that, but to no avail. Does anyone know what the recording format is? What is the leader character? How many of them should there be? What is the data format. Is the data just straight? How is the memory addresses sent? Is there any checking CRC? Is there a trailer? Thanks Mike Quote Link to comment Share on other sites More sharing options...
Casey Posted November 11, 2021 Share Posted November 11, 2021 From Thierry's page - this may answer some of your questions: https://www.unige.ch/medecine/nouspikel/ti99/cassette.htm Quote Link to comment Share on other sites More sharing options...
Fritz442 Posted November 11, 2021 Share Posted November 11, 2021 You may need a recorder with a tone control. Quote Link to comment Share on other sites More sharing options...
mike_z Posted November 11, 2021 Author Share Posted November 11, 2021 Thanks for the link. The cassette.htm is what I was looking for up the TI99 and see if I can capture the data at the TMS9901. Mike 2 Quote Link to comment Share on other sites More sharing options...
mike_z Posted November 11, 2021 Author Share Posted November 11, 2021 Well...... I believe that I see the rough format of a recording. I found what I think is a record. It starts with 8 bytes of 00 followed by one byte of FF followed by 64 bytes of data and then one byte for checksum. My recording is a short BASIC program. My question is what is the data format of the 64 bytes. Is it ASCII text? I would think that this maybe the case, how else could the program be listed? Is the data something else? Maybe tomorrow I can look a little closer and try and determine if the data is ASCII text. Thanks Mike 1 Quote Link to comment Share on other sites More sharing options...
+arcadeshopper Posted November 11, 2021 Share Posted November 11, 2021 you may need an inline amplifier, modern computer sound cards are too quiet, so are some tape players 1 Quote Link to comment Share on other sites More sharing options...
Ed in SoDak Posted November 11, 2021 Share Posted November 11, 2021 (edited) What sort of cassette machine? The old basic mono portable works best. A hi-fi deck might record fine, but only has line-out, not speaker level. Sound card output from computer via headphone jack is often weak as well. Look for about a volt AC at the output. If played through the cassette machine's built-in speaker instead of through the output jack into the TI, it would be quite loud, as it would normally be set at just about full volume for the TI to properly "hear" it. Little "walkman style" player/recorders usually have headphone out reduced so dummies won't blow out their ears. Unless you have an adapter, any stereo output will only connect one channel into a mono cable. You'd lose half the output level. As stated, "NO DATA" is just about always too low of playback level's volume. Assuming cable is good. Edited November 11, 2021 by Ed in SoDak 4 Quote Link to comment Share on other sites More sharing options...
mike_z Posted November 12, 2021 Author Share Posted November 12, 2021 I'm using an old mono Realtone cassette recorder. I tried a newer GE and the Audacity. I even amplified the signal some and still no soap. So I opened the unit up and looked at the signal going into the 9901 on pin 30. I get a nice square wave. I can not see the signal to well with the scope, but the leader looks like it might be OK. Maybe tomorrow I can drag out my logic analyzer and try can capture the wave form. Either the recording is not correct or there is something wrong with the 9901 or some other more inter parts. Thanks Mike Quote Link to comment Share on other sites More sharing options...
blackbox Posted November 13, 2021 Share Posted November 13, 2021 (edited) BITD there were many cassette players that simply would not work with the TI - to the extent that in the UK TI published a list of brands then on sale in the UK that would work. In the USA the official TI list of working recorders was: Superscope C2L00LP; Panasonic RQ2309A; Sears 2165; Sears 21686; JC Penney 6568. The TI Recorder was a GE model. Older cassette machines may be in need of demagnetising, cleaning and possibly realigning! As already mentioned here, modern devices have too low an output signal voltage to satisfy the console. Adding an extra level of amplification may work. It was recommended that for most recorders the tone setting be HIGH. One big problem was not using a mains supply for a cassette recorder- amazing how many tapes failed for that simple reason. The voltage drop in the console was too great. The signal the TI needs is neither square wave nor sine wave- rather peculiar with some quite sharp bits. Cassette recorders were designed for sine waves, and if one insists on a pure sine wave output it probably won't work. This is probably why cheaper machines seemed to work better. There were several articles in the TI magazines of the time comparing different models, and providing lists of ones that would work and those that didn't.There is a great article at http://www.unige.ch/medecine/nouspikel/ti99/cassette.htm Edited November 13, 2021 by blackbox 2 Quote Link to comment Share on other sites More sharing options...
mike_z Posted November 13, 2021 Author Share Posted November 13, 2021 I'm using a Realtone recorder, this is the same recorder that I used many years ago. I spent a lot of time cleaning it and it will record and play back well. My GE recorder does not output enough signal so I have abandoned using that one. I'm not sure I understand what "mains supply" is. Are you suggesting that I sub a power supply for the recorders batteries? The square wave signal I mentioned is at the 9901 chip on pin 30. This is the condition signal from the recorder. Using a scope I can see the frequency shift signals you mention on the output of the recorder. My logic analyzer captured a recorder signal from this pin 30 on the 9901 chip. This should be the digital signal the TI captures. BUT the leader, which the documentation says is 768 bytes of 00h is not that, but some thing else. It does look like it repeats I have to take some time to see what the bytes are and to see if I can find the DATA MARK. I saved a simple BASIC program 10 AAAAAAAA 20 BBBBBBBB 30 CCCCCCCC I hope to be able to find the 8 A's, 8 B's and 8 C's, I figure that they should be ASCII bytes. I don't know whether the LSB or the MSB are transmitted first or if there are any other bits included with the bytes. The documentation seems to indicate that there are no extra bits. Mike Quote Link to comment Share on other sites More sharing options...
+Lee Stewart Posted November 13, 2021 Share Posted November 13, 2021 39 minutes ago, mike_z said: I hope to be able to find the 8 A's, 8 B's and 8 C's, I figure that they should be ASCII bytes. I don't know whether the LSB or the MSB are transmitted first or if there are any other bits included with the bytes. The documentation seems to indicate that there are no extra bits. I don’t know whether you are familiar with TMS9900 Assembly Language Code (ALC). If you are, the snippet in the following spoiler of the ALC source by TI programmers of the cassette routines from the console ROM may offer you some insight: Spoiler * * THIS IS A SET OF ROUTINES DEFINED FOR AUDIO CASSETTE * COMMUNICATION. THEY ARE ENTERED THROUGH A GRAPHICS * LANGUAGE PROGRAM, WHICH GIVES INFORMATION LIKE THE NO. * OF BLOCKS (=64 BYTES) TO BE WRITTEN, OR THE NUMBER OF * FREE BLOCKS IN CASE OF READ MODE; THE VDP START ADR. * AND THE BAUD RATE. * THE ROUTINES TAKE CARE OF THE NECESSARY ENCODING/ * DECODING AND THE ERROR CHECKING. * * THE BIPHASE FORMAT, USED IN THIS SET OF ROUTINES, HAS * THE FOLLOWING REPRESENTATIONS FOR THE BINARY DIGITS: * * --------------- --------- ---- * | | | | | * --- --- --- ------- * <-----"0"-----> <---- "1"---------> * * REPRESENTATIONS OF THE BITS MAY BE CHANGED IN PHASE BY * 180 DEGREES, DEPENDING UPON THE VALUE OF THE BIT STREAM * AFTER THE PREVIOUS BIT. * * PRINCIPLES OF OPERATION * * WRITING * THE ACTUAL IMPLEMENTATION OF THE BIPHASE RECORDING SCHEME * IS RELATIVELY SIMPLE. THE VALUE FOR THE DATA RATE, AS INDICATED * BY THE GRAPHICS LANGUAGE PROGRAM, IS USED AS A TIMER VALUE FOR * THE INTERNAL TMS9985 TIMER/COUNTER. IT IS USED AS A TIMER VALUE * FOR HALF A BIT CELL. * * ------------ ------ ---- * | | | | | * --- --- --- ------ * |<----><----> <-----><-----> * DRATE DRATE * * EACH BIT CELL THUS CONSISTS OF TWO TIMER INTERVALS. THE TIMER * INTERRUPT AT THE BEGINING OF EACH BIT CELL CAUSES THE OUTPUT LINE * TO CHANGE VALUE. THE NEXT TIMER INTERRUPT, IN THE MIDDLE OF THE * BIT CELL, ONLY CHANGES THE VALUE OF THE OUTPUT LINE IF THE BIT TO * BE OUTPUT EQUALS A BINARY "1". * * READING * ON READING BACK, THE BASIC TIMER INTERVAL TIME IS SET TO 1.5 * TIMES THE DRATE OF THE WRITE SECTION. THE TIMER IS SYYNCHRONIZED * ON THE FLUX CHANGE AT THE BEGINING OF THE BIT CELL. AFTER THE TIMER * HAS GIVEN AN INTERRUPT, THE CURRENT INPUT LINE VALUE IS COMPARED TO * THE VALUE AT THE BEGINING OF THE BIT CELL. IF THIS VALUE HAS * CHANGED, THE BIT VALUE IS ASSUMED TO BE "1" IF NOT, IT WILL BE * A "0" * TO PROVIDE A TIME-OUT MECHANISM THE TIMER AUTOMATICALLY * RESTARTS ITSELF WITH THE SAME RATE. IF THE TIMER TIMES OUT BEFORE * THE NEXT FLUX CHANGE, AN ILLEGAL BIT LENGTH IS ASSUMED, AND AN * ERROR RETURN CODE IS PRODUCED. * ********************************************************************* * * CASSETTE WRITE ROUTINE * * WRITES N BLOCKS OF 64 BYTES TO THE AUDIO CASSETTE. * * THE OUTPUT FORMAT USED IS: * - ZERO LEADER CONSISTING OF LDCNT ZEROES * - SYNC BYTE (8 "1" BITS) * - NUMBER OF BLOCKS TO FOLLOW (8 BITS) * - CHECKSUM (8 BITS) * - 2*N BLOCKS, CONSISTING OF: * - 8 BYTES OF ZERO * - 1 BYTE OF ONES * - 64 BYTES OF INFORMATION * - CHECKSUM (8 BITS) * - EACH BLOCK IS REPEATED TWICE. THE LEADING ZEROES AND * ONES ARE USED FOR TIMING AND TO R * - TRAILER OF EIGHT "1" BITS * * REGISTER DEFINITION * * R0 TEMPORARY * R1 PHASE FLAG REGISTER * R2 COUNTER/BYTE COUNTER * R3 DATA RATE COUNTER * R4 BYTE * R5 BLOCK COUNT * R6 LOOP COUNT * R7 CHECKSUM * R8 SET/RESET OUTPUT INSTRUCTION * R12 CRU BASE REGISTER = TIMER CRU * ********************************************************************* * THE GRAPHICS LANGUAGE PROGRAM PROVIDES THE FOLLOWING * INFORMATION: * (R1) DATA LENGTH IN BYTES * (R1) +2 VDP RAM START ADDRESS * * THE DATA SHOULD BE READ BACK AT THE SAME BAUD RATE AT WHICH * IT WAS WRITTEN. * HX0010 DATA >0010 * = BR TABLE * * WRITE ROUTINE * WRITE CLR R0 NO OFFSET FOR VIDEO ADR. LI R2,>0300 LOAD LEADER COUNT FIRST LI R8,>1E19 LOAD SBZ INSTN LI R3,>0023 SET DATA RATE AT 1500 BAUD BL @INIT INIT. FURTHER LI R0,BOUT SAVE BYTES HX0300 LIMI 1 LABEL IS NASTY! WRITE0 CLR R4 BL *R0 WRITE BYTE OF LEADER DEC R2 COMPLETE LEADER JNE WRITE0 AND RTN IF NOT COMPLETED SETO R4 BYTE =>FFFF (SYNC BYTE) BL *R0 MOV R5,R4 WRITE BLOCKCOUNT TO TAPE SWPB R4 BL *R0 AND OUTPUT AGAIN MOV R5,R4 TWICE (ONE CHECKSUM) SWPB R4 BL *R0 CLR R9 USED TO SEE IF 2ND RECD. WAS WRITTEN REWRI LI R2,8 WRITE 8 BYTES OF ZERO LEAD CLR R4 ZERO BL *R0 WRITE THEM DEC R2 MORE? JNE LEAD YES SETO R4 WRITE ALL ONES BYTE BL *R0 WRITE IT MOVB @R10LSB,*R15 RESTORE VDP ADR. LI R2,64 LOAD COUNT WITH ONE BLOCK MOVB R10,*R15 CLR R7 CLEAR CHECKSUM WRBYT CLR R4 MOVB @VRDOFF(R15),R4 VDP DATA TO R4 MSB A R4,R7 ADD BYTE TO CHEKCSUM BL *R0 OUTPUT BYTE TO TAPE DEC R2 DEC. BYTE COUNTER JNE WRBYT WRITE ALL 64 BYTES MOV R7,R4 OUTPUT CHECKSUM BL *R0 INV R9 2ND TIME RECORD WRITTEN? JNE REWRI REWRITE AI R10,64 BUMP ADR. TO NEXT RECD DEC R5 DONE ALL BLOCKS? JNE REWRI NO... NOT YET, CONTINUE *==LES==* SPIN WAIT FOR COMPLETION OF LAST CYCLE WSPIN JMP WSPIN WAIT FOR COMPLETION OF LAST CYCLE *==LES==* B @STTIM STOP TIMER AND EXIT * * INITIALIZATION ROUTINE * INIT MOV *R1+,R5 COPY NO. OF BLOCKS IN BLKCNT AI R5,63 COMPUTE IN BLOCKS OF 64 BYTES SRL R5,6 DIVIDE BY 2**6=64 SOC *R1,R0 COPY ADR. + OFFSET IN R0 MOV R0,R10 SAVE ADR FOR WRITING DUPLICATES MOVB @R0LSB,*R15 COPY IN VIDEO ADR CLR R1 INIT THE ERROR FLAG CLR R12 CRU BASE MOVB R0,*R15 COPIED... SOC @HX0020,R14 SET TIMER INTPT FLAG SBZ 2 TURN OFF VDP INTPT SBZ 12 TURN OFF ??? INTPUT LDCR R3,15 LOAD TIMER SBZ 0 ENABLE INTS SBZ 1 TURN OFF EXTERNAL INT. SBO 3 ENABLE CLOCK INTPT RT AND RTN * * BYTE OUTPUT ROUTINE * BOUT LI R6,8 LOAD LOOP COUNTER INV R4 INVERT FOR ENABLING "0" BIT JUMP *==LES==* BOUT0 SPIN IDLE AROUND UNTIL LAST CYCLE OF BOUT0 JMP BOUT0 IDLE AROUND UNTIL LAST CYCLE OF *==LES==* X R8 ALWAYS FLIP AT START OF BIT XOR @HX0300,R8 HXSPIN EQU $ *==LES==* D3F0 SPIN PREVIOUS BYTE+1ST OF THIS BYTE D3F0 JMP D3F0 PREVIOUS BYTE+1ST OF THIS BYTE *==LES==* MOV R4,R4 TEST HIGHEST BIT FOR NEW FLIP CHECK JLT BOUT1 BIT SET MEANS NO FLIP (SEE INV) X R8 FLIP OUTPUT LINE XOR @HX0300,R8 REVERSE INSTN BOUT1 SLA R4,1 GET NEXT BIT READY FOR OUTPUT DEC R6 HAS BEEN COMPLETED... THEN LOOP JNE BOUT0 UNTIL BYTE COMPLETED RT RETURN IF BYTE COMPLETED * * TIMER INTERRUPT ROUTINE * TIMER SBZ 0 MAKE SURE WE'RE IN INTPT MODE SBO 3 CLEAR DECREMENTER INTPT MOV R1,R1 TEST FLAG REGISTER JLT TIMER1 FLAG SET -> ERROR EXIT LWPI WKSC BACK TO REGULAR WORKSPACE C *R14,@HXSPIN DOING A SPIN? JNE TIMER1 NO, DON'T INC ADR. INCT R14 BYPASS SPIN TIMER2 RTWP AND RETURN TIMER1 LWPI WKSC MOV @WKSE+R6+R6,R14 GET EXIT ADDRESS JMP TIMER2 AND EXIT * * READ REGISTER DEFINITIONS * * R0 TEMPORARY * R1 ERROR EXIT/INPUT LEVEL FLAG * R2 COUNTER/BYTE COUNTER * R3 DATA RATE COUNTER * R4 BYTE * R5 BLOCK COUNT * R6 BIT LENGTH ERROR EXIT ADR * R7 CHECKSUM * R8 RETRY COUNT/BYTE COUNT * R12 CRU BASE REGISTER = TIMER CRU * * * CASSETTE READ ROUTINES * HX2100 BYTE >21,0 * * DEVIATION OF UP TO -25 TO +50 PERCENT OF THE * NOMINAL BAUD RATE IS PERMITTED * VERIFY SOC @HX0010,R14 SET VERIFY BIT CLR R0 SET ADR BIT TO READ FROM VDP JMP READA * = BR TABLE READ SZC @HX0010,R14 RESET VERIFY BIT LI R0,>4000 SET ADR. BIT TO WRITE TO VDP READA LI R3,>002B SET BAUDRATE IN HX002B EQU $-2 BL @INIT AND INIT MOV R10,R7 SAVE VDP ADR. CLR R0 INDICATE FIRST RECORD HX20_ MOVB @HX20_+1,@ERRCD ASSUME TIME OUT ERROR *==LES==* HX20 to HX20_ READ12 LI R8,30000 LOAD RETRY COUNT FOR 10 SECS LIMI 1 ALLOW TIMER INTPTS LI R6,READ0 LOAD TIME-OUT RETRY ADR READ00 LI R3,>2B IF FALSE START READ0 ANDI R1,>FF CLEAR FLAG TO PREVENT ERROR HANG UP HX00FF EQU $-2 DEC R8 COUNT FOR ONE RETRY JEQ STTIM RETRY ERROR LI R2,48 COUNT AT LEAST 48 ZEROES MOV R0,R0 AM I IN THE MIDDLE? JNE READ1 YES LOOK FOR 48 ZEROES A R2,R2 NO LOOK FOR MORE READ1 BL @RDBIT GET A BIT ** TIME OUT GIVES RETRY JMP READ15 JMP READ0 "1" RETURN READ15 DEC R2 DETECTED A "0" BIT, COUNT IT JNE READ1 AND CONTINUE TILL DONE * --- CALCULATE ACTUAL BIT RATE BASED ON 8 ZERO BIT TIMES --- LI R9,>7FFF BIG COUNT FOR TIMER LI R8,8 NUMBER OF BITS TO AVERAGE LDCR R9,15 SET TIMER SBZ 0 TURN ON 9901 INTPTS SBO 3 ENABLE CLOCK INTPT CRLOOP BL @FLUXC FIND BIT *==LES==* LOOP to CRLOOP JMP ENDL CHANGE-DEC NO. OF BITS JMP CRLOOP NO CHANGE, KEEP LOOKING *==LES==* LOOP to CRLOOP ENDL DEC R8 LOOK FOR MORE ??? JNE CRLOOP YES *==LES==* LOOP to CRLOOP SBO 0 SET 9901 TO CLOCK STCR R3,15 READ TIMER S R3,R9 HOW MUCH TIME USED? MOV R9,R3 SAVE CELL TIME SLA R9,2 CELL TIME * 4 A R9,R3 MAKE NEW 3/4 BIT TIME SRL R3,6 5/8 CELL TIME/2 ORI R3,>0001 SET 9901 TO CLOCK LI R10,READ5 RTN ADR FOR SUBSUBROUTINE CI R3,>001F MAKE SURE FREQ. NOT TOO SMALL JLT READ00 TOO SMALL, LOOK FOR REAL ZEROS B @RDBIT2 LOAD NEW BIT TIME * READ5 BL @RDBIT AGAIN READ BIT INPUT JMP READ5 "0" IS STILL OK * ENTRY NOT KNOWN D4B6 LI R2,7 SHOULD BE "ALL ONE" BYTE READ3 BL @RDBIT NEXT BIT PHASE JMP READ0 FALSE ALARM-NOT A REAL LEADER DEC R2 ALL SYNC BYTE DONE? JNE READ3 NO...KEEP ACCEPTING "1" BITS LI R6,ERROR1 GET SET FOR ERROR EXIT * AT THIS POINT WE ARE SYNCHRONIZED ON THE DATA STREAM * COMING IN FROM THE CASSETTE, BOTH ON BIT AND R4 LEVEL * ALL DATA COMING IN NOW SHOULD BE CORRECT. IF NOT WE * TAKE THE ERROR EXIT MOV R0,R0 DO I READ BLOCK COUNTS? JNE DUMP3 I ALREADY READ THEM MOVB @HX2100,@ERRCD ASSUME NOT ENOUGH BLOCKS MOV R7,R0 CLR R7 INIT CHECK SUM BL @RDBYTE CALL READ BYTE ROUTINE C R5,R4 ENOUGH SPACE AVAILABLE?? JL STTIM ** ERROR ** NOT ENOUGH SPACE MOV R4,R5 READ NO. OF BLOCKS ON TAPE INC R5 AND INC FOR EASE OF PROGRAMMING NEG R7 PREPARE FOR BLOCK COUNT CHECK BL @RDBYTE READ CHECK SUM JNE STTIM BAD BLOCK COUNT DATA JMP DUMP4 LOOP AROUND TO FIRST READ RDLOOP ANDI R7,>FF USE ONLY LOWER BYTE OF CHKSUM NEG R7 NEGATTE FOR ZERO RESULT BL @RDBYTE READ RECORDED CHKSUM JEQ DUMP **NO ERROR** CHKSUM CORRECT ERROR1 MOV R5,R5 ERROR ON 2ND RECORD JLT STTIM YES, FLAG ERROR MOVB @R0LSB,*R15 RESTORE VDP ADR. NEG R5 NO, TRY 2ND READING MOVB R0,*R15 JMP READ12 READ AGAIN DUMP MOV R5,R5 HAS 2ND RECORD BEEN READ? JLT DUMP2 YES, SO DON'T READ AGAIN LI R2,73 SKIP NEXT RECORD DUMP1 LI R6,DUMPX BL @RDBYTE WAIT FOR A BYTE DUMPX DEC R2 MORE TO READ? JNE DUMP1 READ SOME MORE DUMP2 AI R0,64 BUMP RECORD ADR. IN VDP MOVB @R0LSB,*R15 OUTPUT ADR. TO VDP ABS R5 SET FLAG FOR 1ST RECORD MOVB R0,*R15 DUMP4 CLR R7 CLEAR CHKSUM AFTER DUMPING DEC R5 REDUCE NO. OF BLOCKS TO READ JNE READ12 READY...MORE BLOCKS TO GO JMP DONE NO MORE TO READ DUMP3 LI R2,64 START OUT FOR 128 BYTES CLR R7 ZERO CHKSUM FOR NEXT RECORD RDL0 BL @RDBYTE READ ONE BYTE SWPB R4 SWAP TO UPPER BYTE COC @HX0010,R14 IS IT A VERIFY JNE STO NO, SO STORE IT SB @VRDOFF(R15),R4 IS IT RIGHT? JEQ STO1 YES, CONTINUE CI R5,1 LAST RECORD? JEQ STO1 YYES, IGNORE BAD COMPARISONS JMP ERROR1 NO, TRY READING 2ND RECORD STO MOVB R4,@VWDOFF(R15) WRITE DATA TO VDP STO1 DEC R2 SEE IF DONE JNE RDL0 NO.. NOT YET JMP RDLOOP COMPLETE READ LOOP DONE MOVB @HX002B,@ERRCD CLEAR ERROR CODE * ------ ALSO ERROR EXIT -------- STTIM SZC @HX0010,R14 FREE UP BIT SZC @HX0020,R14 TURN OFF TIMER INTPT FLAG SBZ 3 TURN OFF TIMER INTPT MASK BIT SBO 12 ENABLE INTERRUPT SBO 1 ENABLE EXT INTPT SBO 2 ENABLE VDP INTPT B @NEXT * * BIT INPUT ROUTINE * * READ ONE BIT FROM THE INPUT STREAM. RETURN TO CALLER+2 * IF BIT READ IS "1" * THE VALUE OF THE BIT CELL IS COMPUTED BY DETERMINING * THE INPUT LINE VALUE AT 3/4 OF THE BIT CELL LENGTH. IF THE * INPUT LINE LEVEL HAS CHANGED DURING THAT PERIOD, THE BIT * READ = "1"; IF NOT, THE BIT READ = "0" * THE NEXT FLUX CHANGE SHOULD COME WITHIN 3/4 OF A BIT * CELL, IN ORDER TO ACCEPT THE BIT * RDBIT MOV R11,R10 SAVE RTN ADR. IN LEVEL 2 *==LES==* SPIN WAIT FOR END OF CYCLE RSPIN JMP RSPIN WAIT FOR END OF CYCLE *==LES==* BL @FLUXC SEE IF WE HAD A FLUX CHANGE INCT R10 IF SO... UPDATE RTN ADR ORI R1,>FF00 SET FLAG FOR ERROR EXIT RDBIT2 CZC @HX00FF,R1 WHICH WAY IS CHANGED FLUX? JEQ F2 INPUT IS A ZERO F1 TB 27 WAIT FOR A CHANGE JNE RDBIT1 CHANGED JMP F1 NO CHANGE F2 TB 27 WAIT FOR A CHANGE JNE F2 NO CHANGE, WAIT SOME MORE RDBIT1 LDCR R3,15 AN ERROR EXIT SBZ 0 ENABLE INTPTS SBO 3 ENABLE CLOCK INTPT ANDI R1,>FF RESET ERROR FLAG TO PREVENT CHANGE XOR @HX00FF,R1 RECORD FLUX CHANGE B *R10 TIMER... THEN EXIT ON LEVEL 2 * ----- BYTE READ ROUTINE ------ * THE R4 READ ROUTINE IS SIMPLE AND STRAIGHT FORWARD. * ONLY REMEMBER THAT WE ARE DEALING WITH THREE LEVELS * OF SUBROUTINES HERE, SO WE MUST SAVE THE RTN ADR IN R9 * RDBYTE LI R8,8 LOAD BYTE COUNTER CLR R4 CLEAR RESULT MOV R11,R9 SAVE RTN ADR. IN LEVEL 3 RDBYT0 SLA R4,1 RESULT ONE BIT TO THE LEFT BL @RDBIT READ ONE BIT JMP RDBYT1 AVOID COUNTING ZEROS INC R4 IF IT'S A "1", INC. BYTE RDBYT1 DEC R8 CONTINUE?? JNE RDBYT0 YEP... A R4,R7 UPDATE CHKSUM *** CONDITION CODES B *R9 RTN FROM BYTE READING (LEVEL 3) * FLUX CHANGE CONTROL ROUTINE * RETURNS TO (R11) IF FLUX HAS CHANGED, (R11) +2 IF FLUX * HASN'T CHANGED SINCE LAST OUTPUT VALUE. * FLUXC TB 27 CONTROL VALUE OF INPUT LINE JEQ FLUXC1 INPUT LINE SET CZC @HX00FF,R1 WAS INPUT LINE HIGH LAST TIME? JEQ FLUXC2 INPUT LINE WAS LOW FLUXC0 XOR @HX00FF,R1 FLIP FLAG(INPUT LINE CHANGED) RT & RETURN FLUXC1 CZC @HX00FF,R1 INPUT LINE HIGH AGAIN? JEQ FLUXC0 INPUT LINE WAS LOW -> CHANGE FLUXC2 INCT R11 UPDATE RTN ADR, NO CHANGE RT ...lee 1 Quote Link to comment Share on other sites More sharing options...
mike_z Posted November 13, 2021 Author Share Posted November 13, 2021 Lee.... thanks a million for the information. No, I'm not familiar with ALC. But I'm an old DEC and CPM guy. I know other assembly languages, so I believe I can understand this. There is also Google if I get stumped. I may have found a reason for my trouble. It may be recorder after all. Looking at the timing on the logic analyzer, the time of the zeros seems to vary. My sample rate on the Logic analyzer is 5 uSec, I would expect some difference but not more that 5-10 uSec. I'm seeing 50 uSec or more. This could screw up the bits. I'm going to check the belt in the recorder, it could have stretched and thus wobble the audio. The belt was changed a few years ago, but the recorder is not used a lot and could have kinked at the motor. I've seen this before. So, I suppose I should just replace the belt. If the sensing is 1.5 times the data rate, there should be plus or minus 362 uSec that the signal could vary before the bit could be read wrong. Did I read this correctly, the error would not be cumulative, because the timing starts at the flux change interrupt each time.? So, I'll see if I can improve the recorder and then check how that affects the data stream. Thanks again. Mike 1 Quote Link to comment Share on other sites More sharing options...
Ed in SoDak Posted November 13, 2021 Share Posted November 13, 2021 (edited) "Mains" voltage is a term in the UK for normal AC house wiring, 120v AC for us across the pond. He means that batteries are often not able to both run the motor and the amplifer at high volume. Output from Audacity should not show any time lag, compared to a slow-running cassette player. On the cassette side of things, even if speed is not correct nor head alignment, so long as they're not too far off, a recording made on that machine is inherently correct when played back on the same machine. It would only be when using a tape made on a different machine or trying this tape in another player, then the errors would show up. As you suspect, there may well be bad parts in the console. That seems to be a rare thing, much more often it's some external glitch or error in the load process. I made this load dongle from an old stereo cassette VU meter. That gives me a stable reference for correct level and I can watch the pointer's wiggle to help detect bad oxide or wrinkles in the tape itself, dropouts, etc. Somewhere I have a crude schematic of it which I'll add when I find it. Edited November 13, 2021 by Ed in SoDak Quote Link to comment Share on other sites More sharing options...
Ed in SoDak Posted November 13, 2021 Share Posted November 13, 2021 (edited) Found the meter pic with schematic added. Here's a couple related threads: Edited November 13, 2021 by Ed in SoDak 1 Quote Link to comment Share on other sites More sharing options...
Ed in SoDak Posted November 14, 2021 Share Posted November 14, 2021 Console schematic showing cassette circuitry. TI99_3.TIF 1 Quote Link to comment Share on other sites More sharing options...
mike_z Posted November 29, 2021 Author Share Posted November 29, 2021 Well.... today I received a new drive belt for my cassette. It made no difference. It will record, I can hear it, but will not read a recording. So, I dragged out my GOOD reel to reel tape recorder and I get the same results. There must be something wrong inside the 99. Mike Quote Link to comment Share on other sites More sharing options...
+Schmitzi Posted November 29, 2021 Share Posted November 29, 2021 ... or the cable, maybe ? Quote Link to comment Share on other sites More sharing options...
GDMike Posted November 29, 2021 Share Posted November 29, 2021 2 hours ago, Schmitzi said: ... or the cable, maybe ? Or caps in the ti, slave a 9901 on top of the old one. Quote Link to comment Share on other sites More sharing options...
Fredrik Öhrström Posted November 30, 2021 Share Posted November 30, 2021 You can also try the generated tape sounds from this page: http://nivelleringslikaren.eu/ti994a_basic 1 Quote Link to comment Share on other sites More sharing options...
mike_z Posted November 30, 2021 Author Share Posted November 30, 2021 Well... here's what I did. Opened the 99 up and looked at 9901 pin 30 (think that is what it was) and connected my logic analyzer. I saw square waves that looked like data. I figured that maybe if I saved a simple program, just a bunch of A's, I might be able to see the A's in the data stream. Have not done that yet. I also pourchased a used 9901 on eBay and installed that, again no soap. Who knows if it was OK. I'm starting to run out of things to try, other than trying to see the A's in the data stream. Mike 1 Quote Link to comment Share on other sites More sharing options...
Tursi Posted November 30, 2021 Share Posted November 30, 2021 If it's telling you NO DATA FOUND and not DATA ERROR, then it's not hearing the leading tone or can't lock in on it - there's no data in that leading tone, it's just used to calibrate the read loop. If you see a square wave at pin 30, check the voltage of it - maybe it's too low? Also check that it's regular with even spacing between the highs and lows - if your tape deck is filtering the output it could be corrupting the data too much for the circuit to parse it. The old advice was tone control all the way up, volume most of the way up. Nothing with an equalizer or similar filters - basically keep the edges of the input sharp as possible. It's analog with a hacky input circuit... so it's really going to come down to a case by case basis. The official TI recorder was just a rebranded brick-shaped unit... someone found and posted the original service manual for it, and as a kid we had the exact same recorder with a Sears logo on it. I also remember getting a friend's walkman to work. Cheaper and dumber the better, I think! 2 Quote Link to comment Share on other sites More sharing options...
mike_z Posted November 30, 2021 Author Share Posted November 30, 2021 The response is 'NO DATA FOUND'. I know that there is a good leader tone. I can hear it on each recorder that I've used. My logic analyzer doesn't help with voltage levels, just timing. I can use a scope, but it is hard to sync on the changing signal. My GOOD reel to reel has plenty of volume. It can blow you out of the room. In fact with the volume up all the way, I have to turn off the speakers. Earlier in this thread someone gave a reference to the data stream characters. I was trying to see them, but the time of the bits was varying. That is why I thought that maybe the drive belt on the cassette was a problem. I suppose that I should try and find the characters using the reel to reel recorder. DO you know what bit pattern is used for the leader? I could easily see that, because it is repeating and last rather long. Mike Quote Link to comment Share on other sites More sharing options...
+arcadeshopper Posted November 30, 2021 Share Posted November 30, 2021 7 minutes ago, mike_z said: The response is 'NO DATA FOUND'. I know that there is a good leader tone. I can hear it on each recorder that I've used. My logic analyzer doesn't help with voltage levels, just timing. I can use a scope, but it is hard to sync on the changing signal. My GOOD reel to reel has plenty of volume. It can blow you out of the room. In fact with the volume up all the way, I have to turn off the speakers. Earlier in this thread someone gave a reference to the data stream characters. I was trying to see them, but the time of the bits was varying. That is why I thought that maybe the drive belt on the cassette was a problem. I suppose that I should try and find the characters using the reel to reel recorder. DO you know what bit pattern is used for the leader? I could easily see that, because it is repeating and last rather long. Mike you should hear the tone THROUGH the computer sound.. do you? 2 Quote Link to comment Share on other sites More sharing options...
Like us on Facebook
Follow us on Twitter
AtariAge on Instagram
Watch us on YouTube
Chat on Discord
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.