Vectrex Programming TOC

Vector display

Just let me drop a few words on how vectrex positions the beam on the monitor. How the BIOS does it, and how one can use that information to tickle a bit more speed out of the whole system.

I won't really go into the hardware level, if you want a good describtion of a general way how a vector is drawn/positioned look at the 'internal.txt' by Keith Wilkins.

There are a couple of highly important things to know in order to understand the following. Look at the schematics, at the 'internal.txt' and at the reference sheet (or the documentation of Vectrex Frogger) when I lose you.

~ZERO (=VIA CA2)
~BLANK (=VIA CB2)
~RAMP (=VIA PB7)
VIA Timer 1 (One shot mode, PB7 enabled)
VIA Shift Register Mode 4 (Shift Out Under T2 Control)
VIA Interrupts
VIA Auxiliary Control Register
VIA Control Register
Integrators (look at 'internal.txt' good introduction!)

If the above mentioned don't tell you anything, you are probably better off learning a bit more an come back later :-)!

First I'll describe how the BIOS positions the beam, than how it draws vectors and after that... well, let's see.

The BIOS upon initialization sets some VIA registers, if the programmer does not change them, they will be relevant until you switch your vectrex off. The BIOS most of the time assumes some of the settings, if they should not be as expected your vectors probably won't be drawn correctly anymore.

BIOS initializes the data direction registers of VIA port A all for output and port B all for output except bit 5 and 6 (which are input). It than pokes $98 to VIA_aux_cntl.

There are mainly two aspects of interest in that setting, first the shift mode, which is shift mode 4. Which means, that shifting is enabled by poking something to the shift register, and that the bit shifted out from bit 7 is circulated back to bit 0. Shifting continues even after 8 shifts. Shifting is done at 1/2 the system clock speed. The bit that is shifted out is put into CB2 (~BLANK). ~BLANK is zero active, that means if the bit is zero, output is switched off. The other interesting bit is the one that enables PB7 output when timer1 is active. PB7 is the ~RAMP signal which is zero active, but becomes inverted. What it comes to is that the integrators are switched on, when the bit PB7 (= ~RAMP) is set. Integrators on means, that the vectorbeam is moving.

Perhaps you can allready at this point see where this is leading us. The beam will be positioned for a certain time, the time is stored to timer 1. While timer 1 is active the beam is moving. Depending on the value in the Shift register, we have this or that output. If there is %00000000 in the Shift register we have no output at all, since at each shift a zero will be stored to the ~BLANK line, which disables all output. If a %10101010 would be stored to the Shift register we would see something like a dotted line. Depending on the 'strength' of the vectors we would see short 'dots' or wider 'dots'(!). If a %11111111 was set we would see a solid line!

Following is some part of the BIOS Moveto_d function:

Moveto_d:       STA     <VIA_port_a     ;Store Y in D/A register
                CLR     <VIA_port_b     ;Enable mux
                PSHS    D               ;Save D-register on stack
LF318:          LDA     #$CE            ;Blank low, zero high?
                STA     <VIA_cntl
                CLR     <VIA_shift_reg  ;Clear shift regigster
                INC     <VIA_port_b     ;Disable mux
                STB     <VIA_port_a     ;Store X in D/A register
                CLR     <VIA_t1_cnt_hi  ;timer 1 count high
                PULS    D               ;Get back D-reg
                JSR     Abs_a_b
                STB     -1,S
                ORA     -1,S
                LDB     #$40
                CMPA    #$40
                BLS     LF345
                CMPA    #$64
                BLS     LF33B
                LDA     #$08
                BRA     LF33D

LF33B:          LDA     #$04            ;Wait for timer 1
LF33D:          BITB    <VIA_int_flags
                BEQ     LF33D
LF341:          DECA                    ;Delay a moment
                BNE     LF341
                RTS
:HERE
LF345:          BITB    <VIA_int_flags  ;Wait for timer 1
                BEQ     LF345
                RTS

In D is the Y, X position, the first thing the function does is send the Y strength to Port A of the VIA, which also is input to the DAC. The DAC output can be distributed VIA a MUX to different destinations since this is the Y coordinate we enable the MUX with MUX settings which send the DAC data to the Y integration data storage (the strength which will be added at each cycle to the Y integrators, once additions are allowed), the MUX settings which enable sending to Y is 00 (two bits of VIA Port B). Enabling the MUX itself is also a bit of VIA Port B, bit 0, MUX is enabled by clearing that bit. Here a small describtion of the bits of VIA port B:

VIA_port_b $D000:

Than the functions stores register D on stack, which also serves as a short delay befor disabling the MUX again. Right after that we load register A with $CE, which as a bitmask is:

This sets the ~BLANK to low, which disables output and ~ZERO to high, which disables zeroing. If zeroing is enabled, the vector beam can not be moved, since it is 'grounded', it remains in zero position. The settings thus allow the vector beam to be moved.

Next we clear the Shift register, that means, when the integration starts, the vector moves, there will be no output, since at every shift out a zero will be moved to ~BLANK. Integration will start as soon as ~RAMP is zero (which as said before will be negated befor reaching the integration circuit).

After that we disable the MUX, remember, the enable bit for that was bit 0 of VIA port B. It gets disabled when the bit is set. This means that the DAC value now only gets to the X integration storage (which allways gets the DAC value, the MUX setting doesn't change that).

Than we store the X strength to VIA port A and thus to the DAC register. The next step is to enable the timer by writing to the hi byte. The high byte is cleared, since we integrate at most to $ff time, which is the highest scale value. The scale value was set befor even calling the move function into the low byte of the timer.

The next lines are IMHO not neccessary, all this computing of some extra cycles to wait. And actually I don't understand it. Since the additional waiting is done AFTER the timer reaches zero, that means the ~RAMP is disabled anyway, so no change to any vector position will occur.

Anyway, let us assume that we reach the point marked ':HERE'. Register B is loaded with $40, which is the bitmask for timer 1 interrupt testing. The loop's only purpose is to wait for the time to expire. After the time has gone by the ~RAMP will be again not active, no integration is done and position of the beam will not change anymore (presumable), since ~RAMP is disabled, as soon as the timer reaches zero. Why on earth the BIOS programmers wait for some more cycles to go by - I don't understand. Apart from that - you see that once you know where to put the correct timers, and what registers are used for what... it is sort of understandable how a vector is positioned.

The Vector drawing routines are essentially the same, apart from the Shift register, which is not cleared, but set to $ff = %11111111, which when shifted results in ~BLANK to be non active, thus the beam is illuminated.

Here is a simple BIOS vector plot routine:

Draw_Line_d     STA     <VIA_port_a     ;Send Y to A/D
                CLR     <VIA_port_b     ;Enable mux
                LEAX    2,X             ;Point to next coordinate pair
                NOP                     ;Wait a moment
                INC     <VIA_port_b     ;Disable mux
                STB     <VIA_port_a     ;Send X to A/D
                LDD     #$FF00          ;Shift reg=$FF (solid line), T1H=0
LF3ED:          STA     <VIA_shift_reg  ;Put pattern in shift register
                STB     <VIA_t1_cnt_hi  ;Set T1H (scale factor?)
                LDD     #$0040          ;B-reg = T1 interrupt bit
LF3F4:          BITB    <VIA_int_flags  ;Wait for T1 to time out
                BEQ     LF3F4
                NOP                     ;Wait a moment more
                STA     <VIA_shift_reg  ;Clear shift register (blank output)
                LDA     $C823           ;Decrement line count
                DECA
                BPL     Draw_VL_a       ;Go back for more points
                JMP     Check0Ref       ;Reset zero reference if necessary

That routine is part of some other routines, because of that some things are done, which are not directly related to vector drawing. The corresponding source lines do not interest us here. As you can see the routine starts of exactly as the Moveto routine, by setting the DAC to the Y strength. The MUX becomes enabled, we wait a moment, than we disable the MUX again. This time the routine assumes, that the ~ZERO flag is allready set correctly, so we don't bother with it. Then we poke the X strength to the DAC. When integration starts we have the correct strengths set to the integrators. Now we prepare the shift register, we poke a $ff to it, which means we will be plotting a solid line. Notice, that we are not integrating yet. But the light is allready switched on (so to say), since shifting starts the moment we set a new shifting value. This is one of the reasons, why all lines we draw have a bright spot at the beginning. The light is always switched on a few cycles to early! After that we clear the hi value of timer one, this enabled the ~RAMP, integration, now the vector beam is moving with the light switched on, that means we are actually drawing a vector now. This drawing will last till timer one expires. The movement will cease the moment the timer reaches zero. Unfortunately the light is not switched off automatically. We therefore wait in a loop, till we get a message, that timer one has expired ($40 (bit 6) in the VIA_int_flags register). After that we switch the light off by clearing the shift register. Done, the complete line is drawn now. As you can see there is no fiddling with any strange additional wait loops, while drawing lines, but otherwise it is exactly the same thing as positioning the vectors (apart from $ff in the Shift register instead of $00). This sort of confirms my suspicion, that these strange loops in the positioning functions are sort of useless. In my experiments I have not seen any differences, if I left them out.

Next page Last Page

Vectrex Programming TOC