different schemes

ZXNet echo conference «hardware.zx»

From Kirill Frolov To Alexander Araktcheew 18 March 2000

Press RESET immediately, Alexander! 14 Mar 00 16:51, Alexander Araktcheew wrote to Felix Knajzev: VV>>> a.turbation of the yellow Scorpio board AA> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ FK>> A big archive is gone. ;) AA> Come on, come on, I think a lot of people would like to see what is underlined and how AA> make it less bloody and without glitches, if this is of course possible :) There is nothing complicated there except the smallest detail - it will be achieved that the control signals did not diverge in time, and this is almost unrealistic. Here's a logic table: __ ___ ____ H0 RAS IORQ TURBO CLK 1 1 x x 1 0 0 1 1 1 0 1 x x 0 1 0 x x 0 0 0 0 x 0 0 0 x 0 0 You will make a circuit like this, which from the first four input arguments does CLK with a delay time of <~20ns and you will have a turbo. You only need a ROM Set it to 150ns. TURBO, when turbo mode is turned on, has level 1. Switch TURBO is possible when there are no memory accesses, the easiest way is programmatically through ports. Turbo coefficient 175%. If you wish, you can mix BORDER and get more. Overclocking the processor to 14Mhz without overclocking the memory is of little use sense. ____WAIT is always formed as (not(H1) and H0). Processor WAIT recognized only in turbo mode. Picture of the layout: CLKF - a bunch of processor in turbo mode. CLKS - a bunch of CPU in non-turbo mode. ┌───┐ ┌───┐ ┌───┐ ┌───┐ ┌───┐ ┌───┐ ┌───┐ ┌───┐ CLKF ──┘ └───┘ └───┘ └───┘ └───┘ └───┘ └───┘ └───┘ └── ┌───┐ ┌───┐ ┌───┐ ┌───┐ CLKS ──┘ └───────────┘ └───────────┘ └───────────┘ └────────── __ ┌────────┐ ┌───────┐ ┌───────┐ ┌───────┐ H0 ─┘ └───────┘ └───────┘ └───────┘ └────── cpu ┌───────────────┐ cpu H1 ─┘ video └───────────────┘ video └────────────── ___ ──────┐ ┌───────┐ ┌───────┐ ┌───────┐ ┌── RAS └───────┘ └───────┘ └───────┘ └───────┘ ___ ┌────────┐ ┌───A───┐ ┌───────┐ ┌─── A A └───────┘ └───────┘ A └────── ____─┐ ┌───W───┐ ┌───W── WIT └────────────────────────┘ W or WRBUF __ ───┐ ┌─────── WE └───────────┘ (put the ruler to the screen...) Instead of a buffer on IR23, you need to put IR22, WRBUF is taken straight.Instead of WRBUF you can also submit WAIT, or instead of WAIT you can submit WRBUF... It's just a matter of delays. The letters AA indicate latching the higher half of the address in pens, with the letters WW polling the WAIT signal processor. The operating logic was copied from KAY-256 and improved. In green scorpion or meter Almost everything is the same as I drew. In non-turbo mode the processor will not have no WAIT loops. The picture with the clocks is very ideal, if you take the oscills in your hands and look that in reality you won’t see anything like it. The main problem here so that the clock does not start too late - the address must be latched, and so that the clump didn’t end too early - if the brake levers were on, you wouldn’t have time to count. That's why IR22 is installed. Well, if the clock is delayed, then WE is in the trigger will not turn on - you can try from the D input to feed directly to WE, in which case it will done. Only in Kay the Sims don’t work, but Kay is generally one big mystery. Think about how to do all this on 2 microcircuits (not PLM!). One of them is a trigger, it clings to reading from 1FFD and 7FFD and turns on/off turbo mode. That leaves you with half of TM2 and choose the second chip yourself, but everything should work quickly. It is possible on a multiplexer 8->1 or 4->1. AA> PS: I almost completed the diagram of the yellow Scorpio in pseudographics. Remaining AA> only the arrangement of elements on the board and the catch :)) AA> Who needs the thread or did I just suffer for a week? 8) Masochist. After all, I threw myself into the echo jpeg - on the first pussy I came across with You can print it using the printer. Are you aware that there are errors in the diagram of the yellow scorpion? Everything is on the board not as depicted. By the way, the clock in my computer is now as written CLKS, and the white is generally at +5v connected It works... Made on one 155LA3. I didn’t play anything fancy with WE.

From Felix Knajzev To Alexander Araktcheew 19 March 2000

Приветствую тебя, о досточтимый(ая) Alexander! [Вторник 14 Март 2000] в [16:51:25] Alexander Araktcheew начертал(а) письмецо к Felix Knajzev. Обсуждали они тему "схемы разные". Продолжим начатое... VV>>> а.турбирование желтой платы "Скорпиона" AA> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ FK>> Ба-а-а-льшой архив ушел мылом. ;) AA> Hу-ка, ну-ка, думаю много народу хотело бы увидеть AA> подчеркнутое и как его сделать меньшей кровью и без глюков, ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ AA> если это конечно возможно :) Щас, разбежался. :) Вот древняя _каловая_ схема (ююк). Более продвинутый способ, удачно опробованный на практике, - ТУРБО а-ля КАЙ-1024. Полную схему и FAQ для этого варианта пока лень рисовать/писать. ;) Получи эту мешанину из ASCII для изготовления файла "ZX&TURBO.$Z": section 1 of 1 of file ZX&TURBO.$Z < Lara Croft 4.50 UUE > begin 644 ZX&TURBO.$Z M6E@F5%520D]:25`,#``-Y7]:6"M4=7)B;U=X=(L6%_8+L:OQAP,!1<7_"QQT MP!ZH`N(`)Z#X`7J`5@K(`X)2@!60!10!=D`0P/XX`6]`'&`'!`%V0%H**`*2 M:59ZF'%^89P0:4_;C7TK7?L1YX=X05Y1L<#+>I`"APR,03R("QPBR%#^S&'T MO,DOC!EY/&@5_(`SP`I(`KP`-U`$R`%!@![`_P`SP`K4`6Y`_#L3G)^=G5OP M0QG*4%+^H&%ZV,^8-$$`_B@2EH_!SF'R0)GB"3!7WG9`US?2;5E^R"397.:#=X'=;V)P#7`"M`M M"QHA@,#8!_5AT/@A&B2XELG?^8:EYN*T0OZ8`5L@`LSDJ.$5#LA@X;`SPX: M&),8*,,SD&#)``'X.8R!IS%#1^Q*`1@V<'=$V-&D3(Q^[O;<+)<#AA#Y,&&M3;M,,A#P)1)DA) M+#R-D_B=9(H,"!@[B`5Q'(]`'AX8K79'9)!PZT3AYRA<&-`'IHT*`1-`)-H`DTZ(4AIX-Z MR,&*MSNM;O+.!W30(.&X3JA=LECZ%].5_ICR0+&+1;]/W#3)ID9F<' MO('PA&%(OXIX)S`JIP8-0!]A$VA@HGQR*"^!P$2V,=NL<&K0<$XU&?+K!)SR MR::MI)S*']..I'C3DH[2H'=)6B@1Z5IHC+Q"4+Y9%$>PPW(P$1Y0G]/LBT< M_D(!2?DS*8_%L.L+)UGB/I+R-##"N^0;!22N%(&:K5PP08PMJD61=`H+8??F^'5C1ID41!<+5-^(]"47OQ7&C/ZH^Y%< M017>(!;4)[E!`]WB-K*K/S`:VD3*;TX&5]!TXN_Z27?0@(FL_BV`K[Z8_33 M'2=@DI'*Z)QCX,8_?!@$`Q^(!'P9N=F!.*(OS!FJ1`^#`$70T6N.T$CNIW3@.;<`L MD&98F?B$N:W(233(0/MXMQ>SFY(B,`<-I@N4+2`.E?S&29T,A+B?(!5D*$I8 M%:$Q$L2X`$X$3P*R,W:,0T&-LITI[86,'4J'>2H$'5K&JQX(F5A@>F'3II MD.&0@W$[G4J<@W20#$R"1P1:K7!,!P&#%.)MGF!LXF MLJ:+37+#!/`)WQO4S:[MR33`"[`#K``SP`LH?IP`)R`($H!08`7!UUEYC1_ M-L?/%V@_/TI`.U&/>`.%J`%V`%V%.SM,_6FFA5&7Z@HV@)T`.*`"]`#Q!^ M0/;C!%@!>D`44$R%%W)3`.R$#==8@!Y@!3AQ?J`.L`)U@!Y@!;`!=D`0$(02 MKS7@"T0R0)5.,Z^%HV.7@29`F]I?X`=$&O_`"?`#A1%Q8$Z(`[P`^2` M(E#;50/D`#O`*L@#R@BE!=?8`7D`59`&&`'%&R='`%N"H_P`[0`ZR`[`?J M`+,44`1D`7%`$N`&!`&90RI>G&"8@"X@`W0`^P`OPX1.W/SL1":=X';`%Q MH`[0O];RG;H!>8`6X'I^U44N3CT.<`,U8`;$`1XUMV``JT28)8"@@`WP`J(M'9` M+G6`"]`+P58`7:`'6`Y]`%1@%8*L&UKM3I$0I^F=#O)X];_P20G3IY!`.)_P,NE3H0#[@"L4`B2B!!M2".Q+X--BCC@#C`L<@#XH`8+AN0 MB^Q`3?(#K(!P/R`)T`*<@")0D/8!28!% MNYB_]-D`[0=B`46`8Q1KZ36@",@"@@#]!^BE``UUA_B/-C5?`TZ'E20K3M&` M0"PZ0+#ORF<#8@$R,N"+P,"HCK`#>`#0JV/FY-%]E@#E(`X4`?J`*OTY1@= M%X82(#00#M2T;T`6$,/E^=]Y!UCNB^"I]S??-,'$U(`BP`Y*`.,`+U,I^ M?YQ`+(Y")TX4$^`VRE<"K`#K(`8?`UP`_@`+#Z@3H@"8@#W(`B(,:@++X$ M:)H^^0%Y@!7@FN8!=D`1D`3$`5&`'6`%^'F%(5[@O?XH)D^`]>,$F(S>0KUG MW/G'*2H.*`*R`-W_-FX'.`%:H.3Y@3J`C2;U`ZP`6;&K9Z_O'>`&6+YN`](` M"Z>W1SB+CL@"'`#`@;G.`'%#Y`#8J8[(`YP`[P`"['[`QROD!"7"0<2ON/ MR`($R3Q`(ML_X`VH`O&`-:"R]TF`QN>M;AV@!P0!7H#&'AVH`V2E_P,X`6^0 M4K^H$"TG]`1!$.<`-%0!X@,/R`),"B[`-LOC2",^@3X'(`1G>S>9U&N` M%N`&:`%R@$3SR/1BG0`AE#]:`:"='"S7X&7&)@%V0!Q@!>2!BM$[P`FP,&-* M`7%`&%!`.H8).3M=T`2H`7*@"#`+RIF<.,5XDWO-A':+*C=U6PAJ#F7 M%,B1^0IN`@3!#]T')`%Z`!N0,H<#7'JCR#N"UH@/R`$'OI@LELP;O`ML'Z`&! M2$F@#@@"4@EGB*X<[@C;J1X0!K@!@3R9N/(K`%V0!`H0&P'2'Y^`_*`.""4 MV0&AO;U$'>![N0>4<+[BA`38AMY@9=C:_$-:8?(:8`>PM/:*.R`(2"!#`CR% M+*=)D8W2N6O&`S*#*;DK.)C_K,N_ZY?"[MS-#[SO-S,O.CI-G=AHT!%3S[" MI^2?WL8%B[ZOMQ M+X!'9/P_/NK;'>Y'>

From Kirill Frolov To Alexander Araktcheew 24 March 2000

Press RESET immediately, Alexander! 22 Mar 00 23:04, Alexander Araktcheew wrote to Felix Knajzev: FK>> Full diagram and FAQ for this option too lazy to draw/write. AA> Come on, don't be lazy ;) AA> If it’s really bad, draw it even clumsily, we’ll understand :) AA> And then Scorpio people suffer, they want to have HD disks :) ┌─┬────┐ ┌──┬────────┐ /R7FFD>─────oS│T Q├────────┐ GND-───┬────┤X0│ MX │ /DOS >─────┤D│ │ │ ├────┤X1│ │ /RFSH >─────/C│ │ │ +5v+────────┤X2│ X ├───┐ /R1FFD>─────oR│ /Qo──┐ │ ├────┤X3│ │ │ └─┴────┘ │ └──────────────oCX│ │ │ 1533TM2 │ │ ├──┤ │ ├───> CLOCK │ ├────┤Y0│ │ │ CPU │ ┌────────────│────┤Y1│ │ │ │ │ +5v+───│────┤Y2│ Y ├───┘ │ │ └────┤Y3│ │ └─────┬──────────────oCY│ │ │ │ ├──┤ │ │ │ H0 >───────┤S0│ │ │ │ /RAS >───────┤S1│1533KP12│ ┌─┬────┐ │ │ └──┴────────┘+5v+──┬────oS│T Q├─────┘ │ /IORQ>──│────┤D│ │ │ TI >──│────/C│ │ │ └────oR│ /Qo─── │ └─┴────┘ │ 1533TM2 │ ┌─────────────┘ ┌─────────┐ │ ┌───┐ │ │ └──┤1 │ │ │ /MREQ >────┤ │ ┌─────V──────┐ │ D/M >────┤ ├──────────┤ 500 ├──┴──> /WAIT WRBUF >────┤ │ └────────────┘ └───┘ "Volume" 1533L??? It should work... But I could have forgotten something and didn’t draw it. :-( The setup must begin by connecting the CX input of the multiplexer to GND, CY is at +5v, WAIT is also at +5V. Everything should work. If it doesn't work, This means the memory does not have time to write. Treatment is long and difficult, easier and don't even try. But there shouldn't be any problems here. Another possible situation is when a very slow memory does not have time to read, here you can only put more other guns. When everything works without a turbo, you can connect CX and CY where necessary, WAIT is also in place and see what happens. The problem is that the memory is slow or, on the contrary, the percentage slows down here can no longer be, if without The turbos worked. But we need to start not with this scheme, but with completely different things.The exit from TR-DOS should not be determined by the WAIT signal as in the yellow scorpion done, but after reading from RAM (without a turbo, WAIT will always be one). Where /CSRAM is done on LP5 and the capacitor is generally the biggest glitch yellow board, it must be done according to normal logic. You may need a capacitor increase the RAS slightly - this will be visible during the circuit test even when turbo disabled. Well, and the Z80H processor, of course. I forgot the most important thing - yet it is necessary to replace IR23 with IR22, WRBUF is taken inverse for it. As for the operation of this circuit without a turbo, everything is ok. But for the rest not I answer. KP12, IR22 are better than 1533 series, trigger TM2 which forms RAS and TI in scorps they put 531 episodes. And mine costs 155... But as I wrote, reading and writing to memory does not depend on frequency if works in one mode, it should also work in the other. And if it doesn’t work in other, then you need to dig in the wrong place, most likely there is a problem with WAIT. Pictures I already drew it - this circuit implements that logic. Performance: IN RAM IN ROM WITHOUT TURBO 100% (93%) 100% (100%) IN TURBO MODE ~175% 200% In parentheses for the non-turbo yellow scorpion. Additionally, you can use the border to access the processor's RAM -- The performance can be increased a little more.