Circuit and firmware to support Seiko-Epson G1216B1N000 dot graphics display
by Dick Cappel
A serial interface and bias supply for the Seiko-Epson G1216N000 using an AT90S2313 because there just aren't enough applications examples for this display on the web.
I was looking for an LCD display that I could use to display waveforms on the workbench. The selection criteria for the display module itself was straight forward:
1. Dot graphic with sufficient resolution to display a simple waveform,
2. Available for fast delivery,
3. Have sufficient documentation available online to design with,
4. Be easy to interface to mechanically, electrically, and firmware wise, and
5. Have a low price.
The Seiko-Epson G1216B1N000 was selected, apparently meeting all criteria, but I was unable to get it to work until Jesper Hanse kindly sent me his driver for the driver for the display in his VAMPP4 TAMPP (Yet Another MP3 Player) and comparing his code with my code, I found some mistakes I had made.
Connected to the a waveform capture unit, displaying a 15 kHz square wave.I think the poor contrast has more to do with the digital camera's low light signal to noise ratio than the performance of the display.
The display is a pretty old transflective gray mode TN LCD design with ok, but not great contrast. Its contrast is better than the screen shot above. The display's saving grace is that appeared to meet all of the primary criteria. The code to drive it was complicated somewhat by the fact that the 8 bit interface appears as two 64x64 displays for the left and right hand sides, and since I was determined to control it with an AT90S2313, I was a pin short of being able to have separate select pins for the left and right displays, and so implemented a Right/Left control pin, which produced some artifacts in the display, and avoiding these made it necessary to carefully code around them.
More about the hardware:
The controller was implemented with an AT90S2313, but it could have been done with an AT90S1200 and a firmware UART simulation since code is pretty small and doesn't rely on the RAM or timers that an AT90S2313 provide. The hardware needs to drive the data and control lines of the Seiko-Epson module, supply an adjustable negative bias voltage and + 5 volts to the module, and provide inverting buffers for the RS-232 interface.
The inverting RS-232 takes its negative supply from the host. The negative supply could have come from the LCD bias circuit (see below) but the handy thing about this circuit was that its output swung negative when connected to my RS-422 computer and only swung down to a little above ground when it was driven by the CMOS I/O port in the host on the waveform capture circuit that I connected it to after debugging.
The negative bias supply was generated by making 5V peak-to-peak pulses and using them to drive a 5 stage half-wave voltage multiplier and regulating the output with an opamp. The half-wave doubler is terribly inefficient using diodes for switch such a low voltage as each doubling of the 5 volts lost nearly half of it in the diodes, but I have a lot of diodes and miniature capacitors and wanted to pass the power signal through a single 5 volt regulator to minimize the exposure of the internal circuitry to accidental hazardous voltages on the cable. A single chip switched capacitor power supply would have reduced the parts count quite a bit.
The reference for the adjustable negative bias supply is the regulated +5 volt supply. The PNP and NPN saturating switches can be 2N2907 and 2N2222. The 330 Ohm resistor from the base of the PNP to +5V is necessary because the NPN emitter follower in the NE555 output stage cannot pull the output high enough to assure the PNP will shut off when the output is high.
More about the firmware:
After reset, the screen is erased, a line is drawn from the upper-left corner to the middle of the bottom edge, then another from there to the upper-right corner, which looks similar to a large letter "V" spanning the display. This is a visual indication that the display and serial controller are up and running. Its default is ASCII input mode, in which commands and the locations of dots to be written are received in ASCII- real nice for debugging with a terminal. Communications are at 19200 baud, 1 stop bit, no parity. The display can be switched to binary mode input, and in this mode commands and x,y coordinates are received as binary values, which is not easy to test with a terminal, but takes only half the transmission time to write a dot on the screen.
The ASCII conversion takes place within the GetASCIIorBinary routine, which returns a binary value regardless of whether the display is in the binary or ASCII input mode. If in the ASCII input mode, it also ignores commas, allowing it to graph comma delimited spreadsheet files.
The Epson documentation refers to the long axis as the Y axis and the coordinate 0,0 as being the upper-left corner of the screen (the latter is conventional for graphics systems). The driver converts this such that the coordinates 0,0 corresponds to the lower-left corner and 127,63 corresponds to the upper-right corner, which conventional for graphing.
The bits in a received byte (whether received as two characters in ASCII mode or a single byte in byte mode) are used according to the following rules:
Bit 7, if set, indicates that the byte is a command to be interpreted as one of the instructions listed below.
Bit 6 is the most significant bit of the Y address or 0 if it is an X address.
Bit 5 through 0 are the lower 5 bits of the X and Y addresses.
Four commands are supported by the serial controller:
$80 Erase display and reset
$81 Switch to ASCII input mode
$82 Switch to Binary input mode
$83 Restart the controller
Because of the way the Seiko-Epson module's interface works, dots are written in 7 row by 1 column segments (like 1 dot wide characters), and the current firmware doesn't read the display's internal ram to add a dot to it, rather it just writes a new segment of 1x7 dots, so this display is limited to simple graphs.
Link to Site: http://cappels.org/dproj/sed/sed.htm
by Dick Cappel
A serial interface and bias supply for the Seiko-Epson G1216N000 using an AT90S2313 because there just aren't enough applications examples for this display on the web.
I was looking for an LCD display that I could use to display waveforms on the workbench. The selection criteria for the display module itself was straight forward:
1. Dot graphic with sufficient resolution to display a simple waveform,
2. Available for fast delivery,
3. Have sufficient documentation available online to design with,
4. Be easy to interface to mechanically, electrically, and firmware wise, and
5. Have a low price.
The Seiko-Epson G1216B1N000 was selected, apparently meeting all criteria, but I was unable to get it to work until Jesper Hanse kindly sent me his driver for the driver for the display in his VAMPP4 TAMPP (Yet Another MP3 Player) and comparing his code with my code, I found some mistakes I had made.
Connected to the a waveform capture unit, displaying a 15 kHz square wave.I think the poor contrast has more to do with the digital camera's low light signal to noise ratio than the performance of the display.
The display is a pretty old transflective gray mode TN LCD design with ok, but not great contrast. Its contrast is better than the screen shot above. The display's saving grace is that appeared to meet all of the primary criteria. The code to drive it was complicated somewhat by the fact that the 8 bit interface appears as two 64x64 displays for the left and right hand sides, and since I was determined to control it with an AT90S2313, I was a pin short of being able to have separate select pins for the left and right displays, and so implemented a Right/Left control pin, which produced some artifacts in the display, and avoiding these made it necessary to carefully code around them.
More about the hardware:
The controller was implemented with an AT90S2313, but it could have been done with an AT90S1200 and a firmware UART simulation since code is pretty small and doesn't rely on the RAM or timers that an AT90S2313 provide. The hardware needs to drive the data and control lines of the Seiko-Epson module, supply an adjustable negative bias voltage and + 5 volts to the module, and provide inverting buffers for the RS-232 interface.
The inverting RS-232 takes its negative supply from the host. The negative supply could have come from the LCD bias circuit (see below) but the handy thing about this circuit was that its output swung negative when connected to my RS-422 computer and only swung down to a little above ground when it was driven by the CMOS I/O port in the host on the waveform capture circuit that I connected it to after debugging.
The negative bias supply was generated by making 5V peak-to-peak pulses and using them to drive a 5 stage half-wave voltage multiplier and regulating the output with an opamp. The half-wave doubler is terribly inefficient using diodes for switch such a low voltage as each doubling of the 5 volts lost nearly half of it in the diodes, but I have a lot of diodes and miniature capacitors and wanted to pass the power signal through a single 5 volt regulator to minimize the exposure of the internal circuitry to accidental hazardous voltages on the cable. A single chip switched capacitor power supply would have reduced the parts count quite a bit.
The reference for the adjustable negative bias supply is the regulated +5 volt supply. The PNP and NPN saturating switches can be 2N2907 and 2N2222. The 330 Ohm resistor from the base of the PNP to +5V is necessary because the NPN emitter follower in the NE555 output stage cannot pull the output high enough to assure the PNP will shut off when the output is high.
More about the firmware:
After reset, the screen is erased, a line is drawn from the upper-left corner to the middle of the bottom edge, then another from there to the upper-right corner, which looks similar to a large letter "V" spanning the display. This is a visual indication that the display and serial controller are up and running. Its default is ASCII input mode, in which commands and the locations of dots to be written are received in ASCII- real nice for debugging with a terminal. Communications are at 19200 baud, 1 stop bit, no parity. The display can be switched to binary mode input, and in this mode commands and x,y coordinates are received as binary values, which is not easy to test with a terminal, but takes only half the transmission time to write a dot on the screen.
The ASCII conversion takes place within the GetASCIIorBinary routine, which returns a binary value regardless of whether the display is in the binary or ASCII input mode. If in the ASCII input mode, it also ignores commas, allowing it to graph comma delimited spreadsheet files.
The Epson documentation refers to the long axis as the Y axis and the coordinate 0,0 as being the upper-left corner of the screen (the latter is conventional for graphics systems). The driver converts this such that the coordinates 0,0 corresponds to the lower-left corner and 127,63 corresponds to the upper-right corner, which conventional for graphing.
The bits in a received byte (whether received as two characters in ASCII mode or a single byte in byte mode) are used according to the following rules:
Bit 7, if set, indicates that the byte is a command to be interpreted as one of the instructions listed below.
Bit 6 is the most significant bit of the Y address or 0 if it is an X address.
Bit 5 through 0 are the lower 5 bits of the X and Y addresses.
Four commands are supported by the serial controller:
$80 Erase display and reset
$81 Switch to ASCII input mode
$82 Switch to Binary input mode
$83 Restart the controller
Because of the way the Seiko-Epson module's interface works, dots are written in 7 row by 1 column segments (like 1 dot wide characters), and the current firmware doesn't read the display's internal ram to add a dot to it, rather it just writes a new segment of 1x7 dots, so this display is limited to simple graphs.
Link to Site: http://cappels.org/dproj/sed/sed.htm
| CODE |
;HOME;********************************* ;2313 serial to Sekio Epson 1216 display panel interface program ;********************************* .include "2313def.inc" ;Driver for Seiko 1216 LCD display. 128 X 64, with separate drivers for left and right ;halves of the screen. See the specification and app. notes for Seiko G1216B1N000. ; Here is how the AT90S2313 is connected to the display: ;Port B is connected to the 8 bit data input/ouput on the display. ;Port D, bits 0 and 1 are not available as the corresponding pins are used for the UART. ; these bits should be written as zero. ;Port D, bit 2 is connected to the R/W input on the display. This is ; more of a data direction and mode pin. ; Actual writes are clocked by the Enable pin. ;Port D, bit 3 is conntected to the D/I (Data/Instruction) pin on the display. ; Low signifies insturction on the data port. ;Port D, bit 4 is connected to the enable input on the display. Instructions and ; data are clocked into the display on the negative edge. ;Port D, bit 5 is to the RESET input on the display. Low resets the display. ;Port D, bit 6 is connected to CS1 on the dispaly and to the input of an inverter, ; the output of which, drives CS2. It is therefore a right side/left side ; select signal. Low for left side, high for right side. ;I/O Pin assignment .equ Enable =$04 ;Display control codes .equ instructionleft = $30 .equ instructionright = $70 .equ writedataleft = $38 ;.equ writedataright = $78 ;.equ readstatusleft = $34 ;.equ readstatusright = $74 ;Displays Instructions .equ displayon = $3F ;.equ displayoff = $3E .equ XtoZero = $B8 .equ YtoZero = $40 .equ Slinetozero = $C0 ;Flag bit assignments for reigster YH ;YH bit 0 if = 1 then in bindary mode for serial data, if 0, ASCII-HEX mode ;YH bit 1 if = 1 then adds dot without erasing 8 bits in column (read-modify-write; slow) ;UART clock parameters .equ clock = 4000000 ;clock frequency .equ baudrate = 19200 ;choose a baudrate .equ baudconstant = (clock/(16*baudrate))-1 ;Register Assignments .def temp = r16 .def xtarget = r17 .def ytarget = r18 .def dispdata = r19 .def temp2 = r20 .def delayreg = r23 .def delay2 = r24 ;.def inchar = r25 ;.def outchar = r26 .org $00 rjmp reset ;Reset handle reset: rjmp init ;Start init. init: ldi temp,ramend;low(ramend) out spl,temp ;Set the stack pointer ldi temp,baudconstant out ubrr,temp ;Load UART bard rate sbi ucr,txen;enable UART xmit sbi ucr,rxen;Enable UART receiver ldi YH,$00 ;Clear all flags rcall waitlongtime ;Allow dispay to come on. ldi temp,$FF ;Issue reset. rcall odd ldi temp,$00 rcall opd ldi temp,instructionleft;Address left half of display rcall opd ldi temp,$FF rcall odb ldi temp,displayon ;Turn left display half on. rcall opb rcall strobeE ldi temp,SlinetoZero;Set left start line to zero. rcall opb rcall shortdelay rcall strobeE ldi temp,instructionright;Address right half of display. rcall opd ldi temp,$FF rcall odb ldi temp,displayon ;Turn right half on. rcall opb rcall strobeE ldi temp,SlinetoZero;Set rightstart lien to zero. rcall opb rcall shortdelay rcall strobeE rcall erasedisplay ;Fill display with zeros (clear screen). ldi xtarget,$00 ;Draw A "V" on the screen. uses each row and each column. ldi ytarget,$00 moredots: rcall WriteDotAtXY inc xtarget inc ytarget cpi xtarget,63 brne moredots moredotsT: rcall WriteDotAtXY dec xtarget inc ytarget cpi xtarget,$FF brne moredotsT defaultloop: ;*****Default loop**** ;Until a command is received, which is distinguished by the MSb (bit 7) ;being set hight, receive x value, then y value, then write a dot there, ;notify when the dot write is complete, and get another. When a command ;is received, to interpret it. ;Message format: ;bit 7: 1 if command. ;bit 6: msb of y address, or zero if x address ;bits 5 hrough 0 remaining bits of address ;Host should send x, then y, then wait for completion notification. ;IN THIS ROUTINE, THE AXES ARE CORRECTED. X IS THE HORIZONAL AXIS (VALUE 0..127) ;AND Y IS THE VERTICAL AXIS (VALUE 0..63). ;SENSE IS CORRECTED ALSO. O,O IS LEFT,BOTTOM rcall GetASCIIorBinary;Get char and put into xtarget. mov temp2,temp andi temp2,$80 brne ItsaCommand andi temp,$7F mov ytarget,temp rcall GetASCIIorBinary;Get char and put into ytarget. mov temp2,temp andi temp2,$80 brne ItsaCommand andi temp,$3F ldi temp2,$3F sub temp2,temp mov xtarget,temp2 ; mov xtarget,temp rcall WriteDotAtXY ;Write dot. ldi temp,$2A ;Notify write is complete. rcall emitchar rjmp defaultloop GetASCIIorBinary: ;Get byte value from host to temp. If YH bit 0 is set, gets binary ;from host. If YH bit 0 is set, then gets ASCII-HEX from host. rcall recvchar sbrc YH,0 ret ;Binary mode - return after first byte cpi temp,$30 brmi GetASCIIorBinary;If comma, ignore and use next char instead mov temp2,temp;Get second char and make binary. No error checking. rcall recvchar ;If comma, out of sync - start receiving byte over cpi temp,$30 brmi GetASCIIorBinary sbrc temp2,6;convert high byte subi temp2,$f7;if bit 6 is set, add $09 andi temp2,$0F sbrc temp,6;convert low byte subi temp,$f7;if bit 6 is set, add $09 andi temp,$0F lsl temp2 ;combine them lsl temp2 lsl temp2 lsl temp2 or temp,temp2 ret ;convert ascii in inbyteh,inbytel to byte in inbytel ItsaCommand: cpi temp,$80 ;Command $80 erase display and reset breq ClearAndReset cpi temp,$81 ;Command $81 switch to ASCII mode breq SwitchToAscii cpi temp,$82 ;Command $82 switch to binary mode breq SwitchToBinary cpi temp,$83 ;Command $83 restart controller brne drs1 rjmp init drs1: ; cpi $84 ; breq ????? ldi temp,$21 ;Else, Nack command and reset rcall emitchar ;Exclaimation mark is printable Nack rjmp defaultloop SwitchToAscii: andi YH,$FE rjmp defaultloop SwitchtoBinary: ori YH,$01 rjmp defaultloop ClearAndReset: ;Clear display and reset default interperetation loop (can be used to sycn) rcall erasedisplay rjmp AckAndReset AckAndReset: ldi temp,$2A ;Send asterisk, which is printable ack rcall emitchar ;and go back to start of default loop. rjmp defaultloop erasedisplay: ldi dispdata,$00 ldi xtarget,$00 anotherx: ldi ytarget,$00 anothery: rcall WriteByteAtXY inc ytarget cpi ytarget,128 brne anothery inc xtarget cpi xtarget,8 brne anotherx ret WriteDotAtXY: ;Write a dot at X Y location ;(0=<X=<63, 0=<Y=<127. ;Enter with X in xtarget and ;Y in ytarget (nothing to do with Y but pass it along. push xtarget ;Save xtarget on stack for later push ytarget andi xtarget,$3F andi ytarget,$7F mov temp,xtarget ;Make display data from 3 lsb of xtarget andi temp,$07 rcall bitpositiontoormask mov dispdata,temp ;Make X page address from xtarget. Put in xtarget. lsr xtarget lsr xtarget lsr xtarget andi xtarget,$07 rcall WriteByteAtXY pop ytarget pop xtarget ret WriteByteAtXY: ;Write 8 bits at X and Y locaton per data sheet sbi portd,2 ;Put R/W on display in Read to ldi temp,instructionleft;prevent ghost right write in column 1. mov temp2,ytarget andi temp2,$40 breq notrights ldi temp,instructionright notrights: rcall opd mov temp,xtarget ldi temp2,XtoZero add temp,temp2 rcall opb rcall strobeE mov temp,ytarget andi temp,$3F ldi temp2,YtoZero add temp,temp2 rcall opb rcall strobeE notright: sbi portd,3 ;Make this a data write. rcall shortdelay mov temp,dispdata rcall opb rcall strobeE ldi temp,$00 ;Prevent glitching starting line register rcall opb ret odd: out ddrd,temp ret opd: out portd,temp rcall shortdelay ret odb: out ddrb,temp ret opb: out portb,temp ; rcall shortdelay ret strobeE: ; rcall shortdelay cbi portd,4 rcall shortdelay sbi portd,4 ret bitpositiontoormask:;Convert bit position (0..7) to OR mask value ;Enter with bit position in temp, Exit with ;mask in temp ldi temp2,$07 cp temp2,temp brmi error ldi temp2,$01;Put 0000 0001 pattern into temp for shifting rotatesomemore: cpi temp,$00 breq finishedshifting lsl temp2 dec temp rjmp rotatesomemore finishedshifting: mov temp,temp2 ret error: ;Report error - for now, just return. ldi temp,$21 ;Else, Nack command and reset rcall emitchar ;Exclaimation mark is printable Nack ret emitchar: ;Send byte in temp to terminal sbis usr,udre;wait until the register is cleared rjmp emitchar out udr,temp;send the byte ret ;go back recvchar: ;Receive a byte from the terminal and put into temp sbis usr,rxc ;Wait for byte to be received rjmp recvchar in temp,udr;Read byte ret shortdelay: ldi delayreg,$04 delaymore: dec delayreg brne delaymore ret longdelay: ldi delayreg,$00 ldi delay2,$00 delaymore2: dec delayreg brne delaymore2 dec delay2 brne delaymore2 ret waitlongtime: rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay rcall longdelay ret |