Gazoo

Well, that won't work since we need to use the DSR that we're dumping. It's a bit more complicated, but it can be done. Select memory editor. Press F1, the cursor will move to the "CRU Base" in the lower right corner. Make the values there 1100 FF. Press F1 again. Make the top line read "c4000 Start 4000 Finish 5FFF". Make the 2nd line read "Dest cC000" Press F2, this will move the eprom contents to >C000. Press F1, the cursor will move to the "CRU Base" in the lower right corner. Make the values there 1100 00. (this deselects the TI FDC eprom so it can be used by the system) Press F1 again. Make the top line read "cC000 Start C000 Finish DFFF". Move the cursor to the page number. Press F8. The contents of the eproms will be saved to the file 'DSK1.TRIALDUMP'. Gazoo

Select memory editor. Press F1, the cursor will move to the "CRU Base" in the lower right corner. Make the values there 1100 FF. Press F1 again. Make the top line read "c4000 Start 4000 Finish 5FFF". Move the cursor to the page number. Press F8. The contents of the eproms will be saved to the file 'DSK1.TRIALDUMP'. Gazoo

No, that's something else. Can you dump the eproms? You can do it with the memory editor in the XB v2.7 Suite cart if you have no other way. Gazoo

I believe that there's nothing else that can be accomplished with this particular patch, although I may be proven wrong, I hope. I've tried changing the values, but it doesn't get any better, only worse. The 2nd patch looks at the data after the branch in the first patch and alters the delay based on the data. I've tried other data there, but to no avail. The original data was >0302,>0202,>0202 for 20ms and >0301,>0101,>0101 for 12ms. Here's the source I disassembled from the Guion Eproms, I had to alter it a little as the CLR R6 didn't fit in the area of the 2nd patch (2 bytes short at the end of the Eprom), so I had to move it to the end of the first patch and adjust the return accordingly. If anyone can figure out how to make the head step faster, have at it. Here's the source for it: * * First patch * AORG >45CC B @>5FC2 DATA >0301,>0101,>0101 DATA >0000,>0000,>0000 DATA >0000,>0000,>0000 CLR R6 * * Second patch * AORG >5FC2 MOVB @>5FF0,R6 SLA R6,1 SBZ >01 SBO >01 JOC AA LI R6,>7530 AB SRC R5,4 SRC R5,4 DEC R6 JNE AB AA MOVB R0,R0 JGT AC SOCB @>45D0,R0 MOVB @>004C(R9),R6 SRL R6,8 AC SZCB @>45D0(R6),R0 B @>45E2 * Gazoo

Yes.

Do you use the print buffer? If not, pull the Z80 chip, it's been known to cause problems and may clear up your problem. Gazoo

To decrease the head step time from 20ms to 12ms on the TI Floppy Controller, make the changes below. This will work with any of the 80 track eprom sets as well as the original 40 track eproms. Enjoy! Gazoo At offset >05CC in the U26 Eprom, you will find this code: >D1A0,>5FF0,>0A16,>1E01,>1D01,>1806,>0206,>7530,>0B45,>0B45,>0606,>16FC Replace it with this code: >0460,>5FC2,>0301,>0101,>0101,>0000,>0000,>0000,>0000,>0000,>0000,>04C6 At offset >0FC2 in the U27 Eprom you will find this code: >FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF >FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF,>FFFF Replace it with this code: >D1A0,>5FF0,>0A16,>1E01,>1D01,>1806,>0206,>7530,>0B45,>0B45,>0606,>16FC >D000,>1507,>F020,>45D0,>D1A9,>004C,>0986,>5026,>45D0,>0460,>45E2

Well, I disassembled the Guion Eprom dumps and managed to put the code for the 12ms head step time in place of the 20ms code in my 80/80/40 eprom set. All seems to work well so far. Wanna try it out and see what you think?

... is when you spend a few hours replacing that old AT 300 watt power supply in your Geneve tower with a new 650 watt ATX power supply and everything starts working like it should again, other than that flakey 3.5" floppy drive which was easily replaced with a good one. I had a good day Today. Gazoo

Has anyone ever messed with the TI FDC head step times? It seems that they could be speeded up considerably from the original, especially if one is using 3.5" drives. Gazoo

That's the general idea, they have PC cases about that size. This one looks pretty good for what I have in mind: http://www.newegg.com/Product/Product.aspx?Item=N82E16811133240&nm_mc=KNC-GoogleAdwords-PC&cm_mmc=KNC-GoogleAdwords-PC-_-pla-_-Computer+Cases-_-N82E16811133240&gclid=CjwKEAjwq-OsBRDd95aryprR9wQSJACQnU3GQz95TdTyxN9mTdZXWvVOdBDT6Xl7F8j1tP-KQKkk-RoCWBHw_wcB&gclsrc=aw.ds Gazoo

I'm thinking of taking an HFDC card and an HXC device, finding a small case with power supply that they will fit into and use Stuart's console-to-card cable to build a small storage device. Paired with a console that has internal 32k and speech would make for a nice small system with this approach. Gazoo

Me and my Dad did that!

It's preferable to just fix the filename in the XBasic loader, rather than have 2 different programs with the same filename. I shouldn't need to post this, but here it is in case anyone doesn't know how to edit a line in Extended Basic. XB27SPEECH.dsk 060715XB27.dsk Gazoo

Check the file you're loading in line 110 of the LOAD program. Is it 'GROMLOAD', or the program previously used - 'GROMCFG'.

I just tried this on my real TI, and it seems there is too long of a pause between 'Instruments' and 'Home'. You could just tell me I'm nuts, there's nothing wrong with the way it is, and we'll move on. Also, how many times should the speech play before it is replaced by a beep? I'd suggest 2 or 3 times before it got annoying. What's cool is that even after loading BOOT in the Gram space, there's still over 2k left there (the Gram space) to store stuff that won't normally be touched, so you could store all sorts of data and variables there. I should be able to put a byte there and have it count to 2 and then change the Speech to a beep after two Speeches. I would think that Rich would be interested in integrating that 15k of Gram in the Ubergrom cart into RXB, there's plenty of space to store things you need that won't be touched unlike the Ram space on the TI. Gazoo EDIT: I just played with this a little, and it seems the speech is in the wrong place. It should be after the main menu on the Games menu after the spacebar is pressed. The text for what is spoken is even on that screen! The main menu (blue) should beep all the time. Thoughts? Ideas?

The latest version is 4 messages up, message #508.

XB27SPEECH.dsk Same disk as before, but it will verify now with an F6. Gazoo

XB27SPEECH.dsk Modification for XB v2.7 Suite cartridge This should make the TI speak whenever the XB v2.7 Suite menu screen is presented. Thanks for the code, Tursi. I've posted the changes I made to the source below to get it to work in the cart. No guarantees, as I'm really tired. And you won't get a good verification screen if you try that, I haven't updated the checksum for this version. Gazoo * Simple code to speak TEXAS INSTRUMENTS HOME COMPUTER from the built-in * vocabulary. Based on the examples in the E/A manual, which I found * presented somewhat confusingly. REF GPLLNK DEF START AORG >C000 START B @START1 TEXAS EQU >6696 HOME EQU >363E COMP EQU >2034 READIT EQU >8330 * where to store read routine H8000 DATA >8000 * speaking flag H10 BYTE >10 * READ DATA H50 BYTE >50 * SPEAK HAA BYTE >AA * ROM header byte EVEN START1 * save return address MOV R11,R9 * copy in the scratchpad code LI R0,READIT LI R1,CODE LI R2,6 * 6 words CPLP MOV *R1+,*R0+ DEC R2 JNE CPLP * Check if there IS a synth CLR R0 * going to read >0000 BL @SETADR MOVB @H10,@>9400 * READ DATA command * need to delay 12 uS, about 36 cycles - each NOP is 10+4 (8-bit) NOP NOP NOP * about 42 cycles BL @READIT * read one byte into R0 CB R0,@HAA * is it >AA? JEQ SPEAKY * no synth, so just return B @GOBACK * we have a synth, so make some noise SPEAKY * Say TEXAS INSTRUMENTS HOME COMPUTER. Note we will block until COMPUTER starts. LI R0,TEXAS BL @SAYIT LI R0,HOME BL @SAYIT LI R0,COMP BL @SAYIT * back to caller B @GOBACK * * Scratchpad code for reading data * CODE MOVB @>9000,R0 * E/A doesn't give exact time needed here, but they use 3 NOPS (30 cycles in scr NOP NOP NOP B *R11 * * set speech data address from R0 * Uses R1,R2 * SETADR LI R2,4 * 4 nibbles to load SETALP SRC R0,4 * Start with LS nibble MOV R0,R1 SRC R1,4 * move desired nibble to right place ANDI R1,>0F00 * mask it ORI R1,>4000 * OR with ADDRESS command nibble MOVB R1,@>9400 * send the command DEC R2 JNE SETALP * do all four nibbles LI R1,>4000 MOVB R1,@>9400 * write the final nibble * need to delay 42 uSec, about 126 cycles - SRC is 12+4+2C (8-bit) SRC R1,15 * 46 SRC R1,15 * 46 SRC R1,12 * 40 (A little extra to be safe) B *R11 * * Speech the vocabulary word stored in R0 * SAYIT MOV R11,R10 * save return address * set the address BL @SETADR * Send speak command MOVB @H50,@>9400 * Actual delay not specified again, assume 12? * need to delay 12 uS, about 36 cycles - each NOP is 10+4 (8-bit) NOP NOP NOP * about 42 cycles * wait for speech to end WAIT BL @READIT * with no address just loaded, it retu COC @H8000,R0 JEQ WAIT B *R10 GOBACK BLWP @GPLLNK DATA >7854 END

You can only use one of your Qumetrak drives as DSK3. then. You may want to change to the 40/40/80 eprom set so you can use your drives as DSK1. and DSK2. Gazoo

Yes it's possible. Yes it would drive people crazy. Before we get extravagant with options, post the assembly code that speaks what you want it to, I'll see if I can squeeze it in a cart. Gazoo

No, but by using an ubergrom cart, one could put a GPL header in it that autostarted, and then branched to the ROM code with 6 byes of code. So if your ROM cart started code execution at >6022, You could have your GPL code as this >BF00,>6022,>0FF0. The GPL code would look like this at address g>6000. >AA01,>0000,>6010,>0000,>0000,>0000,>0000,>0000 >0000,>6016,>0000,>BF00,>6022,>0FF0 Gazoo EDIT: You may want to add: >3100,>018F,>DD00,>6000 before the branch to lock in bank 1 of the ROM, so you'd then have this GPL code: >AA01,>0000,>6010,>0000,>0000,>0000,>0000,>0000 >0000,>6016,>0000,>3100,>018F,>DD00,>6000,>BF00 >6022,>0FF0

Has anyone ever tried making a cable with two 44 pin card edge connectors, one on each end, and then plugging one into the side of the TI console and the other into the card edge connector on the Speech synthesizer? I assume it might either work right off the bat, or some small modification to the speech synthesizer might be necessary to get it to work. It would be an easy way to install a speech synth into a console, especially if that console already had 32k installed. One would then have a stand alone game console that would work with the large gamecarts. Gazoo

In that case just use a Horizon Ramdisk with Rambo installed. The switch is already there to shut it off, too. You can have whatever menu program you write autostart. You'll have to make a few changes in the Forths to the banking schemes of Rambo instead of cart banking. But the authors of both the Forths you mentioned are here and might actually help with the code. Gazoo

Extended Basic v2.7 Suite autostarts.