Dear Editor, By now I suppose you have heard from people who have tried the Byte-Back UM-64 mod (SWN 2/5). I now have Fred’s Memotext running beautifully in ram, but I had some problems with the connections in the article.
Instruction 6 pg. 12 says, “Install a jumper from pin 5 of U10 to ‘C’.” This connects pin 5 to pin 6. Pin 5 of U10 should be jumpered to “D” after the trace cut instruction (1). “D” is hooked to U9 pin 5 on the UM 64, “C” isn’t.
The drawing on page 11 shows “C” connected to pin 5 U10 by a UM 64 trace when it is actually connected to pin 6. The jumper should be shown from pin 5, U10 to “D” as mentioned above and not as it is shown.
These changes feed pure A15 to U9 and not the Gated A15 (GA15) and allows for an easy pick off point for the A15 input to the buffer.
James Holder, Huntsville, AL
I have pondered your questions for some time (as you know) and finally figured out what is wrong. The problem is on page 11. The “Backside of UM-64,” should be “Backside of M-64.” Jeff Moore did the UM 64 mod and Mark Fisher (CATS) and I did the M-64 mod and the subsequent drawing. However, there are several different boards out for the UM-64 and the new ones have this mod already done for you.
Dear Fred, Thanks for your continued work in supporting our computers. Keep up on those “One Chip Mods.” I have played around with the Cassette Connection Mods and offer some comments.
1. Removing the Cap on the Mic circuit really squares up the Save square wave.
2. The suggested 1800 ohm resistor for R12 on the Load side is rather touchy. 3400 Ohms gives better volume control range. The piggy-backing resistor does not seem to give as good a result as the single value (I’m not sure why).
3. Bumping the 6 Volts on the recorder to 7.5 Volts improves the recorder output more reliably than changing the value of the last resistor in the output line (change of waveform).
4. The 7.5 V. and 3400 Ohm resistor usage make the load conditioners less helpful, perhaps less needed. Uploading from the 1000 with these changes is better accomplished by setting the volume at the lowest possible loading value.
Marinus Heusveldt, Lantana, FL
Dear Basil, I saw your “Delete” machine code routine and I would like to include it as a subroutine in the 8 to 16K area and activate it by RAND USR 8462. I don’t understand the spaces around the LPRINT in the 1 REM (SWN 2/4).
Dan Sullivan, Seattle, WA
Dear Dan, First, the spaces around the LPRINT token are actually part of the ONE byte key word. Each keyword also includes spaces front and back, which makes for easy reading. They only appear in the display file and not in the memory location the holds the LPRINT token. Remember, this is one of Sinclair’s great memory saving tricks. Don’t be fooled by it.
The listing (Fig. 4.3 pg 16 SWN 2/4) can be used by changing the three JP NZ instructions to relative jumps (JR NZ). This makes the routine 3 bytes shorter, but adds a few microseconds to the execution time. A modified listing follows:
16514 221 <> LD IX,
16515 33 5 16509
16516 125 ?
16517 64 RND
16518 62 Y LD A,
16519 117 ? 117
16520 60 W INC A
16521 221 <> INC IX
16522 35 7
16523 221 <> CP(IX+0)
16524 190 ▒
16525 0
16526 32 4 JR NZ,
16527 249 RAND -7
16528 221 <> CP(IX+1)
16529 190 ▒
16530 1
16531 200 COS RET Z
16532 62 Y LD A,
16533 234 REM 234
16534 221 <> CP(IX+5)
16535 190 ▒
16536 5 ▒
16537 32 4 JR NZ,
16538 235 FOR -21
16539 62 Y LD A,
16540 18 > 18
16541 221 <> CP(IX+6)
16542 190 ▒
16543 6 ▒
16544 32 4 JR NZ,
16545 3 ▒ 3
16546 221<> PUSH IX
16547 229 FAST
16548 193 AT POP BC
16549 60 W INC A
16550 221 <> CP(IX+6)
16551 190 ▒
16552 6 ▒
16553 32 4 JR NZ,
16554 219 <= -37
16555 221 <> PUSHIX
16556 229 FAST
16557 225 LPRINT POP HL
16558 237 GOSUB SBC HL,BC
16559 66 PI
16560 35 7 INC HL
16561 35 7 INC HL
16562 35 7 INC HL
16563 197 VAL PUSH BC
16564 221 <> POP IX
16565 225 LPRINT
16566 221 <> LD(IX+4),H
16567 116 ?
16568 4 ▒
16569 221 <> LD(IX+3),L
16570 117 ?
16571 3 ▒
16572 201 TAN RETRevision of Figure 4-3 of Basil’s COMPENDIUM (SyncWare News, Vol. 2 Number 4)