Important changes to repositories hosted on mbed.com
Mbed hosted mercurial repositories are deprecated and are due to be permanently deleted in July 2026.
To keep a copy of this software download the repository Zip archive or clone locally using Mercurial.
It is also possible to export all your personal repositories from the account settings page.
Dependencies: mbed FastIO FastPWM USBDevice
TLC5940/TLC5940.h@54:fd77a6b2f76c, 2016-04-30 (annotated)
- Committer:
- mjr
- Date:
- Sat Apr 30 17:43:38 2016 +0000
- Revision:
- 54:fd77a6b2f76c
- Parent:
- 48:058ace2aed1d
- Child:
- 55:4db125cd11a0
TLC5940 with SPI DMA setup in interrupt handler (not quite working)
Who changed what in which revision?
| User | Revision | Line number | New contents of line |
|---|---|---|---|
| mjr | 26:cb71c4af2912 | 1 | // Pinscape Controller TLC5940 interface |
| mjr | 26:cb71c4af2912 | 2 | // |
| mjr | 26:cb71c4af2912 | 3 | // Based on Spencer Davis's mbed TLC5940 library. Adapted for the |
| mjr | 54:fd77a6b2f76c | 4 | // KL25Z and modified to use SPI with DMA to transmit data. The DMA |
| mjr | 54:fd77a6b2f76c | 5 | // scheme results in greatly reduced CPU load. This version is also |
| mjr | 54:fd77a6b2f76c | 6 | // simplified to remove dot correction and status input support, which |
| mjr | 54:fd77a6b2f76c | 7 | // the Pinscape Controller app doesn't use. |
| mjr | 26:cb71c4af2912 | 8 | |
| mjr | 26:cb71c4af2912 | 9 | |
| mjr | 26:cb71c4af2912 | 10 | #ifndef TLC5940_H |
| mjr | 26:cb71c4af2912 | 11 | #define TLC5940_H |
| mjr | 26:cb71c4af2912 | 12 | |
| mjr | 54:fd77a6b2f76c | 13 | #include "FastPWM.h" |
| mjr | 54:fd77a6b2f76c | 14 | |
| mjr | 38:091e511ce8a0 | 15 | // Data Transmission Mode. |
| mjr | 38:091e511ce8a0 | 16 | // |
| mjr | 38:091e511ce8a0 | 17 | // NOTE! This section contains a possible workaround to try if you're |
| mjr | 38:091e511ce8a0 | 18 | // having data signal stability problems with your TLC5940 chips. If |
| mjr | 40:cc0d9814522b | 19 | // things are working properly, you can ignore this part. |
| mjr | 33:d832bcab089e | 20 | // |
| mjr | 38:091e511ce8a0 | 21 | // The software has two options for sending data updates to the chips: |
| mjr | 38:091e511ce8a0 | 22 | // |
| mjr | 40:cc0d9814522b | 23 | // Mode 0: Send data *during* the grayscale cycle. This is the default, |
| mjr | 40:cc0d9814522b | 24 | // and it's the standard method the chips are designed for. In this mode, |
| mjr | 40:cc0d9814522b | 25 | // we start sending an update just after then blanking interval that starts |
| mjr | 40:cc0d9814522b | 26 | // a new grayscale cycle. The timing is arranged so that the update is |
| mjr | 40:cc0d9814522b | 27 | // completed well before the end of the grayscale cycle. At the next |
| mjr | 40:cc0d9814522b | 28 | // blanking interval, we latch the new data, so the new brightness levels |
| mjr | 40:cc0d9814522b | 29 | // will be shown starting on the next cycle. |
| mjr | 40:cc0d9814522b | 30 | |
| mjr | 38:091e511ce8a0 | 31 | // Mode 1: Send data *between* grayscale cycles. In this mode, we send |
| mjr | 38:091e511ce8a0 | 32 | // each complete update during a blanking period, then latch the update |
| mjr | 38:091e511ce8a0 | 33 | // and start the next grayscale cycle. This isn't the way the chips were |
| mjr | 38:091e511ce8a0 | 34 | // intended to be used, but it works. The disadvantage is that it requires |
| mjr | 40:cc0d9814522b | 35 | // the blanking interval to be extended long enough for the full data |
| mjr | 40:cc0d9814522b | 36 | // update (192 bits * the number of chips in the chain). Since the |
| mjr | 40:cc0d9814522b | 37 | // outputs are turned off throughout the blanking period, this reduces |
| mjr | 38:091e511ce8a0 | 38 | // the overall brightness/intensity of the outputs by reducing the duty |
| mjr | 38:091e511ce8a0 | 39 | // cycle. The TLC5940 chips can't achieve 100% duty cycle to begin with, |
| mjr | 40:cc0d9814522b | 40 | // since they require a brief minimum time in the blanking interval |
| mjr | 38:091e511ce8a0 | 41 | // between grayscale cycles; however, the minimum is so short that the |
| mjr | 38:091e511ce8a0 | 42 | // duty cycle is close to 100%. With the full data transmission stuffed |
| mjr | 38:091e511ce8a0 | 43 | // into the blanking interval, we reduce the duty cycle further below |
| mjr | 38:091e511ce8a0 | 44 | // 100%. With four chips in the chain, a 28 MHz data clock, and a |
| mjr | 38:091e511ce8a0 | 45 | // 500 kHz grayscale clock, the reduction is about 0.3%. |
| mjr | 33:d832bcab089e | 46 | // |
| mjr | 40:cc0d9814522b | 47 | // Mode 0 is the method documented in the manufacturer's data sheet. |
| mjr | 40:cc0d9814522b | 48 | // It works well empirically with the Pinscape expansion boards. |
| mjr | 40:cc0d9814522b | 49 | // |
| mjr | 38:091e511ce8a0 | 50 | // So what's the point of Mode 1? In early testing, with a breadboard |
| mjr | 38:091e511ce8a0 | 51 | // setup, I saw some problems with data signal stability, which manifested |
| mjr | 38:091e511ce8a0 | 52 | // as sporadic flickering in the outputs. Switching to Mode 1 improved |
| mjr | 38:091e511ce8a0 | 53 | // the signal stability considerably. I'm therefore leaving this code |
| mjr | 38:091e511ce8a0 | 54 | // available as an option in case anyone runs into similar signal problems |
| mjr | 38:091e511ce8a0 | 55 | // and wants to try the alternative mode as a workaround. |
| mjr | 38:091e511ce8a0 | 56 | // |
| mjr | 38:091e511ce8a0 | 57 | #define DATA_UPDATE_INSIDE_BLANKING 0 |
| mjr | 33:d832bcab089e | 58 | |
| mjr | 26:cb71c4af2912 | 59 | #include "mbed.h" |
| mjr | 30:6e9902f06f48 | 60 | #include "SimpleDMA.h" |
| mjr | 47:df7a88cd249c | 61 | #include "DMAChannels.h" |
| mjr | 26:cb71c4af2912 | 62 | |
| mjr | 54:fd77a6b2f76c | 63 | |
| mjr | 26:cb71c4af2912 | 64 | /** |
| mjr | 26:cb71c4af2912 | 65 | * SPI speed used by the mbed to communicate with the TLC5940 |
| mjr | 26:cb71c4af2912 | 66 | * The TLC5940 supports up to 30Mhz. It's best to keep this as |
| mjr | 33:d832bcab089e | 67 | * high as possible, since a higher SPI speed yields a faster |
| mjr | 33:d832bcab089e | 68 | * grayscale data update. However, I've seen some slight |
| mjr | 33:d832bcab089e | 69 | * instability in the signal in my breadboard setup using the |
| mjr | 33:d832bcab089e | 70 | * full 30MHz, so I've reduced this slightly, which seems to |
| mjr | 33:d832bcab089e | 71 | * yield a solid signal. The limit will vary according to how |
| mjr | 33:d832bcab089e | 72 | * clean the signal path is to the chips; you can probably crank |
| mjr | 33:d832bcab089e | 73 | * this up to full speed if you have a well-designed PCB, good |
| mjr | 33:d832bcab089e | 74 | * decoupling capacitors near the 5940 VCC/GND pins, and short |
| mjr | 33:d832bcab089e | 75 | * wires between the KL25Z and the PCB. A short, clean path to |
| mjr | 33:d832bcab089e | 76 | * KL25Z ground seems especially important. |
| mjr | 26:cb71c4af2912 | 77 | * |
| mjr | 26:cb71c4af2912 | 78 | * The SPI clock must be fast enough that the data transmission |
| mjr | 26:cb71c4af2912 | 79 | * time for a full update is comfortably less than the blanking |
| mjr | 26:cb71c4af2912 | 80 | * cycle time. The grayscale refresh requires 192 bits per TLC5940 |
| mjr | 26:cb71c4af2912 | 81 | * in the daisy chain, and each bit takes one SPI clock to send. |
| mjr | 26:cb71c4af2912 | 82 | * Our reference setup in the Pinscape controller allows for up to |
| mjr | 26:cb71c4af2912 | 83 | * 4 TLC5940s, so a full refresh cycle on a fully populated system |
| mjr | 26:cb71c4af2912 | 84 | * would be 768 SPI clocks. The blanking cycle is 4096 GSCLK cycles. |
| mjr | 26:cb71c4af2912 | 85 | * |
| mjr | 26:cb71c4af2912 | 86 | * t(blank) = 4096 * 1/GSCLK_SPEED |
| mjr | 26:cb71c4af2912 | 87 | * t(refresh) = 768 * 1/SPI_SPEED |
| mjr | 26:cb71c4af2912 | 88 | * Therefore: SPI_SPEED must be > 768/4096 * GSCLK_SPEED |
| mjr | 26:cb71c4af2912 | 89 | * |
| mjr | 26:cb71c4af2912 | 90 | * Since the SPI speed can be so high, and since we want to keep |
| mjr | 26:cb71c4af2912 | 91 | * the GSCLK speed relatively low, the constraint above simply |
| mjr | 26:cb71c4af2912 | 92 | * isn't a factor. E.g., at SPI=30MHz and GSCLK=500kHz, |
| mjr | 26:cb71c4af2912 | 93 | * t(blank) is 8192us and t(refresh) is 25us. |
| mjr | 26:cb71c4af2912 | 94 | */ |
| mjr | 38:091e511ce8a0 | 95 | #define SPI_SPEED 28000000 |
| mjr | 26:cb71c4af2912 | 96 | |
| mjr | 26:cb71c4af2912 | 97 | /** |
| mjr | 26:cb71c4af2912 | 98 | * The rate at which the GSCLK pin is pulsed. This also controls |
| mjr | 26:cb71c4af2912 | 99 | * how often the reset function is called. The reset function call |
| mjr | 38:091e511ce8a0 | 100 | * interval is (1/GSCLK_SPEED) * 4096. The maximum reliable rate is |
| mjr | 26:cb71c4af2912 | 101 | * around 32Mhz. It's best to keep this rate as low as possible: |
| mjr | 26:cb71c4af2912 | 102 | * the higher the rate, the higher the refresh() call frequency, |
| mjr | 40:cc0d9814522b | 103 | * so the higher the CPU load. Higher frequencies also make it more |
| mjr | 40:cc0d9814522b | 104 | * challenging to wire the chips for clean signal transmission, so |
| mjr | 40:cc0d9814522b | 105 | * minimizing the clock speed will help with signal stability. |
| mjr | 26:cb71c4af2912 | 106 | * |
| mjr | 40:cc0d9814522b | 107 | * The lower bound depends on the application. For driving lights, |
| mjr | 40:cc0d9814522b | 108 | * the limiting factor is flicker: the lower the rate, the more |
| mjr | 40:cc0d9814522b | 109 | * noticeable the flicker. Incandescents tend to look flicker-free |
| mjr | 40:cc0d9814522b | 110 | * at about 50 Hz (205 kHz grayscale clock). LEDs need slightly |
| mjr | 40:cc0d9814522b | 111 | * faster rates. |
| mjr | 26:cb71c4af2912 | 112 | */ |
| mjr | 40:cc0d9814522b | 113 | #define GSCLK_SPEED 350000 |
| mjr | 26:cb71c4af2912 | 114 | |
| mjr | 26:cb71c4af2912 | 115 | class TLC5940 |
| mjr | 26:cb71c4af2912 | 116 | { |
| mjr | 26:cb71c4af2912 | 117 | public: |
| mjr | 54:fd77a6b2f76c | 118 | uint64_t spi_total_time;//$$$ |
| mjr | 54:fd77a6b2f76c | 119 | uint32_t spi_runs;//$$$ |
| mjr | 54:fd77a6b2f76c | 120 | |
| mjr | 26:cb71c4af2912 | 121 | /** |
| mjr | 26:cb71c4af2912 | 122 | * Set up the TLC5940 |
| mjr | 54:fd77a6b2f76c | 123 | * |
| mjr | 26:cb71c4af2912 | 124 | * @param SCLK - The SCK pin of the SPI bus |
| mjr | 26:cb71c4af2912 | 125 | * @param MOSI - The MOSI pin of the SPI bus |
| mjr | 26:cb71c4af2912 | 126 | * @param GSCLK - The GSCLK pin of the TLC5940(s) |
| mjr | 26:cb71c4af2912 | 127 | * @param BLANK - The BLANK pin of the TLC5940(s) |
| mjr | 26:cb71c4af2912 | 128 | * @param XLAT - The XLAT pin of the TLC5940(s) |
| mjr | 26:cb71c4af2912 | 129 | * @param nchips - The number of TLC5940s (if you are daisy chaining) |
| mjr | 26:cb71c4af2912 | 130 | */ |
| mjr | 26:cb71c4af2912 | 131 | TLC5940(PinName SCLK, PinName MOSI, PinName GSCLK, PinName BLANK, PinName XLAT, int nchips) |
| mjr | 47:df7a88cd249c | 132 | : sdma(DMAch_TLC5940), |
| mjr | 47:df7a88cd249c | 133 | spi(MOSI, NC, SCLK), |
| mjr | 26:cb71c4af2912 | 134 | gsclk(GSCLK), |
| mjr | 26:cb71c4af2912 | 135 | blank(BLANK), |
| mjr | 26:cb71c4af2912 | 136 | xlat(XLAT), |
| mjr | 33:d832bcab089e | 137 | nchips(nchips) |
| mjr | 26:cb71c4af2912 | 138 | { |
| mjr | 54:fd77a6b2f76c | 139 | spi_total_time = 0; spi_runs = 0; // $$$ |
| mjr | 54:fd77a6b2f76c | 140 | |
| mjr | 40:cc0d9814522b | 141 | // start up initially disabled |
| mjr | 40:cc0d9814522b | 142 | enabled = false; |
| mjr | 40:cc0d9814522b | 143 | |
| mjr | 33:d832bcab089e | 144 | // set XLAT to initially off |
| mjr | 30:6e9902f06f48 | 145 | xlat = 0; |
| mjr | 33:d832bcab089e | 146 | |
| mjr | 33:d832bcab089e | 147 | // Assert BLANK while starting up, to keep the outputs turned off until |
| mjr | 33:d832bcab089e | 148 | // everything is stable. This helps prevent spurious flashes during startup. |
| mjr | 33:d832bcab089e | 149 | // (That's not particularly important for lights, but it matters more for |
| mjr | 33:d832bcab089e | 150 | // tactile devices. It's a bit alarming to fire a replay knocker on every |
| mjr | 33:d832bcab089e | 151 | // power-on, for example.) |
| mjr | 30:6e9902f06f48 | 152 | blank = 1; |
| mjr | 30:6e9902f06f48 | 153 | |
| mjr | 26:cb71c4af2912 | 154 | // Configure SPI format and speed. Note that KL25Z ONLY supports 8-bit |
| mjr | 26:cb71c4af2912 | 155 | // mode. The TLC5940 nominally requires 12-bit data blocks for the |
| mjr | 26:cb71c4af2912 | 156 | // grayscale levels, but SPI is ultimately just a bit-level serial format, |
| mjr | 26:cb71c4af2912 | 157 | // so we can reformat the 12-bit blocks into 8-bit bytes to fit the |
| mjr | 26:cb71c4af2912 | 158 | // KL25Z's limits. This should work equally well on other microcontrollers |
| mjr | 38:091e511ce8a0 | 159 | // that are more flexible. The TLC5940 requires polarity/phase format 0. |
| mjr | 26:cb71c4af2912 | 160 | spi.format(8, 0); |
| mjr | 26:cb71c4af2912 | 161 | spi.frequency(SPI_SPEED); |
| mjr | 33:d832bcab089e | 162 | |
| mjr | 33:d832bcab089e | 163 | // Send out a full data set to the chips, to clear out any random |
| mjr | 33:d832bcab089e | 164 | // startup data from the registers. Include some extra bits - there |
| mjr | 33:d832bcab089e | 165 | // are some cases (such as after sending dot correct commands) where |
| mjr | 33:d832bcab089e | 166 | // an extra bit per chip is required, and the initial state is |
| mjr | 54:fd77a6b2f76c | 167 | // unpredictable, so send extra bits to make sure we cover all bases. |
| mjr | 54:fd77a6b2f76c | 168 | // This does no harm; extra bits just fall off the end of the daisy |
| mjr | 54:fd77a6b2f76c | 169 | // chain, and since we want all registers initially set to 0, we can |
| mjr | 33:d832bcab089e | 170 | // send arbitrarily many extra 0's. |
| mjr | 33:d832bcab089e | 171 | for (int i = 0 ; i < nchips*25 ; ++i) |
| mjr | 54:fd77a6b2f76c | 172 | spi.write(0x00); |
| mjr | 33:d832bcab089e | 173 | |
| mjr | 33:d832bcab089e | 174 | // do an initial XLAT to latch all of these "0" values into the |
| mjr | 33:d832bcab089e | 175 | // grayscale registers |
| mjr | 33:d832bcab089e | 176 | xlat = 1; |
| mjr | 33:d832bcab089e | 177 | xlat = 0; |
| mjr | 29:582472d0bc57 | 178 | |
| mjr | 39:b3815a1c3802 | 179 | // Allocate our DMA buffers. The transfer on each cycle is 192 bits per |
| mjr | 40:cc0d9814522b | 180 | // chip = 24 bytes per chip. Allocate two buffers, so that we have a |
| mjr | 40:cc0d9814522b | 181 | // stable buffer that we can send to the chips, and a separate working |
| mjr | 40:cc0d9814522b | 182 | // copy that we can asynchronously update. |
| mjr | 40:cc0d9814522b | 183 | dmalen = nchips*24; |
| mjr | 48:058ace2aed1d | 184 | livebuf = new uint8_t[dmalen*2]; |
| mjr | 54:fd77a6b2f76c | 185 | memset(livebuf, 0x00, dmalen*2); |
| mjr | 40:cc0d9814522b | 186 | |
| mjr | 40:cc0d9814522b | 187 | // start with buffer 0 live, with no new data pending |
| mjr | 48:058ace2aed1d | 188 | workbuf = livebuf + dmalen; |
| mjr | 40:cc0d9814522b | 189 | dirty = false; |
| mjr | 40:cc0d9814522b | 190 | |
| mjr | 30:6e9902f06f48 | 191 | // Set up the Simple DMA interface object. We use the DMA controller to |
| mjr | 30:6e9902f06f48 | 192 | // send grayscale data updates to the TLC5940 chips. This lets the CPU |
| mjr | 30:6e9902f06f48 | 193 | // keep running other tasks while we send gs updates, and importantly |
| mjr | 30:6e9902f06f48 | 194 | // allows our blanking interrupt handler return almost immediately. |
| mjr | 30:6e9902f06f48 | 195 | // The DMA transfer is from our internal DMA buffer to SPI0, which is |
| mjr | 30:6e9902f06f48 | 196 | // the SPI controller physically connected to the TLC5940s. |
| mjr | 54:fd77a6b2f76c | 197 | SPI0->C2 &= ~SPI_C2_TXDMAE_MASK; |
| mjr | 54:fd77a6b2f76c | 198 | sdma.attach(this, &TLC5940::dmaDone); |
| mjr | 48:058ace2aed1d | 199 | sdma.destination(&SPI0->D, false, 8); |
| mjr | 30:6e9902f06f48 | 200 | sdma.trigger(Trigger_SPI0_TX); |
| mjr | 30:6e9902f06f48 | 201 | |
| mjr | 30:6e9902f06f48 | 202 | // Configure the GSCLK output's frequency |
| mjr | 26:cb71c4af2912 | 203 | gsclk.period(1.0/GSCLK_SPEED); |
| mjr | 33:d832bcab089e | 204 | |
| mjr | 33:d832bcab089e | 205 | // mark that we need an initial update |
| mjr | 54:fd77a6b2f76c | 206 | forceUpdate = true; |
| mjr | 33:d832bcab089e | 207 | needXlat = false; |
| mjr | 40:cc0d9814522b | 208 | } |
| mjr | 40:cc0d9814522b | 209 | |
| mjr | 40:cc0d9814522b | 210 | // Global enable/disble. When disabled, we assert the blanking signal |
| mjr | 40:cc0d9814522b | 211 | // continuously to keep all outputs turned off. This can be used during |
| mjr | 40:cc0d9814522b | 212 | // startup and sleep mode to prevent spurious output signals from |
| mjr | 40:cc0d9814522b | 213 | // uninitialized grayscale registers. The chips have random values in |
| mjr | 40:cc0d9814522b | 214 | // their internal registers when power is first applied, so we have to |
| mjr | 40:cc0d9814522b | 215 | // explicitly send the initial zero levels after power cycling the chips. |
| mjr | 40:cc0d9814522b | 216 | // The chips might not have power even when the KL25Z is running, because |
| mjr | 40:cc0d9814522b | 217 | // they might be powered from a separate power supply from the KL25Z |
| mjr | 40:cc0d9814522b | 218 | // (the Pinscape Expansion Boards work this way). Global blanking helps |
| mjr | 40:cc0d9814522b | 219 | // us start up more cleanly by suppressing all outputs until we can be |
| mjr | 40:cc0d9814522b | 220 | // reasonably sure that the various chip registers are initialized. |
| mjr | 40:cc0d9814522b | 221 | void enable(bool f) |
| mjr | 40:cc0d9814522b | 222 | { |
| mjr | 40:cc0d9814522b | 223 | // note the new setting |
| mjr | 40:cc0d9814522b | 224 | enabled = f; |
| mjr | 40:cc0d9814522b | 225 | |
| mjr | 40:cc0d9814522b | 226 | // if disabled, apply blanking immediately |
| mjr | 40:cc0d9814522b | 227 | if (!f) |
| mjr | 40:cc0d9814522b | 228 | { |
| mjr | 40:cc0d9814522b | 229 | gsclk.write(0); |
| mjr | 40:cc0d9814522b | 230 | blank = 1; |
| mjr | 40:cc0d9814522b | 231 | } |
| mjr | 40:cc0d9814522b | 232 | |
| mjr | 40:cc0d9814522b | 233 | // do a full update with the new setting |
| mjr | 54:fd77a6b2f76c | 234 | forceUpdate = true; |
| mjr | 40:cc0d9814522b | 235 | } |
| mjr | 29:582472d0bc57 | 236 | |
| mjr | 30:6e9902f06f48 | 237 | // Start the clock running |
| mjr | 29:582472d0bc57 | 238 | void start() |
| mjr | 29:582472d0bc57 | 239 | { |
| mjr | 26:cb71c4af2912 | 240 | // Set up the first call to the reset function, which asserts BLANK to |
| mjr | 26:cb71c4af2912 | 241 | // end the PWM cycle and handles new grayscale data output and latching. |
| mjr | 26:cb71c4af2912 | 242 | // The original version of this library uses a timer to call reset |
| mjr | 26:cb71c4af2912 | 243 | // periodically, but that approach is somewhat problematic because the |
| mjr | 26:cb71c4af2912 | 244 | // reset function itself takes a small amount of time to run, so the |
| mjr | 26:cb71c4af2912 | 245 | // *actual* cycle is slightly longer than what we get from counting |
| mjr | 26:cb71c4af2912 | 246 | // GS clocks. Running reset on a timer therefore causes the calls to |
| mjr | 26:cb71c4af2912 | 247 | // slip out of phase with the actual full cycles, which causes |
| mjr | 26:cb71c4af2912 | 248 | // premature blanking that shows up as visible flicker. To get the |
| mjr | 54:fd77a6b2f76c | 249 | // reset cycle to line up more precisely with a full PWM cycle, it |
| mjr | 54:fd77a6b2f76c | 250 | // works better to set up a new timer at the end of each cycle. That |
| mjr | 54:fd77a6b2f76c | 251 | // organically accounts for the time spent in the interrupt handler. |
| mjr | 54:fd77a6b2f76c | 252 | // This doesn't result in perfectly uniform timing, since interrupt |
| mjr | 54:fd77a6b2f76c | 253 | // latency varies slightly on each interrupt, but it does guarantee |
| mjr | 54:fd77a6b2f76c | 254 | // that the blanking will never be premature - all variation will go |
| mjr | 54:fd77a6b2f76c | 255 | // into the tail end of the cycle after the 4096 GS clocks. That |
| mjr | 54:fd77a6b2f76c | 256 | // might cause some brightness variation, but it won't cause flicker, |
| mjr | 54:fd77a6b2f76c | 257 | // and in practice any brightness variation from this seems to be too |
| mjr | 54:fd77a6b2f76c | 258 | // small to be visible. |
| mjr | 54:fd77a6b2f76c | 259 | armReset(); |
| mjr | 26:cb71c4af2912 | 260 | } |
| mjr | 26:cb71c4af2912 | 261 | |
| mjr | 39:b3815a1c3802 | 262 | /* |
| mjr | 39:b3815a1c3802 | 263 | * Set an output |
| mjr | 26:cb71c4af2912 | 264 | */ |
| mjr | 26:cb71c4af2912 | 265 | void set(int idx, unsigned short data) |
| mjr | 26:cb71c4af2912 | 266 | { |
| mjr | 39:b3815a1c3802 | 267 | // validate the index |
| mjr | 39:b3815a1c3802 | 268 | if (idx >= 0 && idx < nchips*16) |
| mjr | 39:b3815a1c3802 | 269 | { |
| mjr | 40:cc0d9814522b | 270 | // If the buffer isn't dirty, it means that the previous working buffer |
| mjr | 40:cc0d9814522b | 271 | // was swapped into the live buffer on the last blanking interval. This |
| mjr | 40:cc0d9814522b | 272 | // means that the working buffer hasn't been updated to the live data yet, |
| mjr | 40:cc0d9814522b | 273 | // so we need to copy it now. |
| mjr | 54:fd77a6b2f76c | 274 | // |
| mjr | 54:fd77a6b2f76c | 275 | // If 'dirty' is false, it can't change to true asynchronously - it can |
| mjr | 54:fd77a6b2f76c | 276 | // only transition from false to true in application (non-ISR) context. |
| mjr | 54:fd77a6b2f76c | 277 | // If it's true, though, the interrupt handler can change it to false |
| mjr | 54:fd77a6b2f76c | 278 | // asynchronously, and can also swap the 'live' and 'work' buffer pointers. |
| mjr | 54:fd77a6b2f76c | 279 | // This means we must do the whole update atomically if 'dirty' is true. |
| mjr | 54:fd77a6b2f76c | 280 | __disable_irq(); |
| mjr | 40:cc0d9814522b | 281 | if (!dirty) |
| mjr | 40:cc0d9814522b | 282 | { |
| mjr | 54:fd77a6b2f76c | 283 | // Buffer is clean, so the interrupt handler won't touch 'dirty' |
| mjr | 54:fd77a6b2f76c | 284 | // or the live/work buffer pointers. This means we can do the |
| mjr | 54:fd77a6b2f76c | 285 | // rest of our work with interrupts on. |
| mjr | 54:fd77a6b2f76c | 286 | __enable_irq(); |
| mjr | 54:fd77a6b2f76c | 287 | |
| mjr | 54:fd77a6b2f76c | 288 | // get the current live data into our work buffer |
| mjr | 40:cc0d9814522b | 289 | memcpy(workbuf, livebuf, dmalen); |
| mjr | 40:cc0d9814522b | 290 | } |
| mjr | 40:cc0d9814522b | 291 | |
| mjr | 54:fd77a6b2f76c | 292 | // Figure the DMA buffer location of the output we're changing. The DMA |
| mjr | 54:fd77a6b2f76c | 293 | // buffer has the packed bit format that we send across the wire, with 12 |
| mjr | 54:fd77a6b2f76c | 294 | // bits per output, arranged from last output to first output (N = number |
| mjr | 54:fd77a6b2f76c | 295 | // of outputs = nchips*16): |
| mjr | 39:b3815a1c3802 | 296 | // |
| mjr | 39:b3815a1c3802 | 297 | // byte 0 = high 8 bits of output N-1 |
| mjr | 39:b3815a1c3802 | 298 | // 1 = low 4 bits of output N-1 | high 4 bits of output N-2 |
| mjr | 39:b3815a1c3802 | 299 | // 2 = low 8 bits of N-2 |
| mjr | 39:b3815a1c3802 | 300 | // 3 = high 8 bits of N-3 |
| mjr | 39:b3815a1c3802 | 301 | // 4 = low 4 bits of N-3 | high 4 bits of N-2 |
| mjr | 39:b3815a1c3802 | 302 | // 5 = low 8bits of N-4 |
| mjr | 39:b3815a1c3802 | 303 | // ... |
| mjr | 39:b3815a1c3802 | 304 | // 24*nchips-3 = high 8 bits of output 1 |
| mjr | 39:b3815a1c3802 | 305 | // 24*nchips-2 = low 4 bits of output 1 | high 4 bits of output 0 |
| mjr | 39:b3815a1c3802 | 306 | // 24*nchips-1 = low 8 bits of output 0 |
| mjr | 39:b3815a1c3802 | 307 | // |
| mjr | 39:b3815a1c3802 | 308 | // So this update will affect two bytes. If the output number if even, we're |
| mjr | 39:b3815a1c3802 | 309 | // in the high 4 + low 8 pair; if odd, we're in the high 8 + low 4 pair. |
| mjr | 39:b3815a1c3802 | 310 | int di = nchips*24 - 3 - (3*(idx/2)); |
| mjr | 39:b3815a1c3802 | 311 | if (idx & 1) |
| mjr | 39:b3815a1c3802 | 312 | { |
| mjr | 39:b3815a1c3802 | 313 | // ODD = high 8 | low 4 |
| mjr | 40:cc0d9814522b | 314 | workbuf[di] = uint8_t((data >> 4) & 0xff); |
| mjr | 40:cc0d9814522b | 315 | workbuf[di+1] &= 0x0F; |
| mjr | 40:cc0d9814522b | 316 | workbuf[di+1] |= uint8_t((data << 4) & 0xf0); |
| mjr | 39:b3815a1c3802 | 317 | } |
| mjr | 39:b3815a1c3802 | 318 | else |
| mjr | 39:b3815a1c3802 | 319 | { |
| mjr | 39:b3815a1c3802 | 320 | // EVEN = high 4 | low 8 |
| mjr | 40:cc0d9814522b | 321 | workbuf[di+1] &= 0xF0; |
| mjr | 40:cc0d9814522b | 322 | workbuf[di+1] |= uint8_t((data >> 8) & 0x0f); |
| mjr | 40:cc0d9814522b | 323 | workbuf[di+2] = uint8_t(data & 0xff); |
| mjr | 39:b3815a1c3802 | 324 | } |
| mjr | 39:b3815a1c3802 | 325 | |
| mjr | 54:fd77a6b2f76c | 326 | // if we weren't dirty before, we are now |
| mjr | 54:fd77a6b2f76c | 327 | if (!dirty) |
| mjr | 54:fd77a6b2f76c | 328 | { |
| mjr | 54:fd77a6b2f76c | 329 | // we need an update |
| mjr | 54:fd77a6b2f76c | 330 | dirty = true; |
| mjr | 54:fd77a6b2f76c | 331 | } |
| mjr | 54:fd77a6b2f76c | 332 | else |
| mjr | 54:fd77a6b2f76c | 333 | { |
| mjr | 54:fd77a6b2f76c | 334 | // The buffer was already dirty, so we had to write the buffer with |
| mjr | 54:fd77a6b2f76c | 335 | // interrupts off. We're done, so we can re-enable interrupts now. |
| mjr | 54:fd77a6b2f76c | 336 | __enable_irq(); |
| mjr | 54:fd77a6b2f76c | 337 | } |
| mjr | 39:b3815a1c3802 | 338 | } |
| mjr | 26:cb71c4af2912 | 339 | } |
| mjr | 40:cc0d9814522b | 340 | |
| mjr | 40:cc0d9814522b | 341 | // Update the outputs. We automatically update the outputs on the grayscale timer |
| mjr | 40:cc0d9814522b | 342 | // when we have pending changes, so it's not necessary to call this explicitly after |
| mjr | 40:cc0d9814522b | 343 | // making a change via set(). This can be called to force an update when the chips |
| mjr | 40:cc0d9814522b | 344 | // might be out of sync with our internal state, such as after power-on. |
| mjr | 40:cc0d9814522b | 345 | void update(bool force = false) |
| mjr | 40:cc0d9814522b | 346 | { |
| mjr | 40:cc0d9814522b | 347 | if (force) |
| mjr | 54:fd77a6b2f76c | 348 | forceUpdate = true; |
| mjr | 40:cc0d9814522b | 349 | } |
| mjr | 26:cb71c4af2912 | 350 | |
| mjr | 26:cb71c4af2912 | 351 | private: |
| mjr | 26:cb71c4af2912 | 352 | // current level for each output |
| mjr | 26:cb71c4af2912 | 353 | unsigned short *gs; |
| mjr | 26:cb71c4af2912 | 354 | |
| mjr | 30:6e9902f06f48 | 355 | // Simple DMA interface object |
| mjr | 30:6e9902f06f48 | 356 | SimpleDMA sdma; |
| mjr | 30:6e9902f06f48 | 357 | |
| mjr | 40:cc0d9814522b | 358 | // DMA transfer buffers - double buffer. Each time we have data to transmit to the |
| mjr | 40:cc0d9814522b | 359 | // TLC5940 chips, we format the data into the working half of this buffer exactly as |
| mjr | 40:cc0d9814522b | 360 | // it will go across the wire, then hand the buffer to the DMA controller to move |
| mjr | 40:cc0d9814522b | 361 | // through the SPI port. This memory block is actually two buffers, one live and |
| mjr | 40:cc0d9814522b | 362 | // one pending. When we're ready to send updates to the chips, we swap the working |
| mjr | 40:cc0d9814522b | 363 | // buffer into the live buffer so that we can send the latest updates. We keep a |
| mjr | 40:cc0d9814522b | 364 | // separate working copy so that our live copy is stable, so that we don't alter |
| mjr | 40:cc0d9814522b | 365 | // any data in the midst of an asynchronous DMA transmission to the chips. |
| mjr | 48:058ace2aed1d | 366 | uint8_t *volatile livebuf; |
| mjr | 48:058ace2aed1d | 367 | uint8_t *volatile workbuf; |
| mjr | 40:cc0d9814522b | 368 | |
| mjr | 40:cc0d9814522b | 369 | // length of each DMA buffer, in bytes - 12 bits = 1.5 bytes per output, 16 outputs |
| mjr | 40:cc0d9814522b | 370 | // per chip -> 24 bytes per chip |
| mjr | 40:cc0d9814522b | 371 | uint16_t dmalen; |
| mjr | 40:cc0d9814522b | 372 | |
| mjr | 40:cc0d9814522b | 373 | // Dirty: true means that the non-live buffer has new pending data. False means |
| mjr | 40:cc0d9814522b | 374 | // that the non-live buffer is empty. |
| mjr | 54:fd77a6b2f76c | 375 | volatile bool dirty; |
| mjr | 54:fd77a6b2f76c | 376 | |
| mjr | 54:fd77a6b2f76c | 377 | // Force an update: true means that we'll send our GS data to the chips even if |
| mjr | 54:fd77a6b2f76c | 378 | // the buffer isn't dirty. |
| mjr | 54:fd77a6b2f76c | 379 | volatile bool forceUpdate; |
| mjr | 40:cc0d9814522b | 380 | |
| mjr | 40:cc0d9814522b | 381 | // Enabled: this enables or disables all outputs. When this is true, we assert the |
| mjr | 40:cc0d9814522b | 382 | // BLANK signal continuously. |
| mjr | 40:cc0d9814522b | 383 | bool enabled; |
| mjr | 30:6e9902f06f48 | 384 | |
| mjr | 26:cb71c4af2912 | 385 | // SPI port - only MOSI and SCK are used |
| mjr | 26:cb71c4af2912 | 386 | SPI spi; |
| mjr | 26:cb71c4af2912 | 387 | |
| mjr | 26:cb71c4af2912 | 388 | // use a PWM out for the grayscale clock - this provides a stable |
| mjr | 26:cb71c4af2912 | 389 | // square wave signal without consuming CPU |
| mjr | 54:fd77a6b2f76c | 390 | FastPWM gsclk; |
| mjr | 26:cb71c4af2912 | 391 | |
| mjr | 26:cb71c4af2912 | 392 | // Digital out pins used for the TLC5940 |
| mjr | 26:cb71c4af2912 | 393 | DigitalOut blank; |
| mjr | 26:cb71c4af2912 | 394 | DigitalOut xlat; |
| mjr | 26:cb71c4af2912 | 395 | |
| mjr | 26:cb71c4af2912 | 396 | // number of daisy-chained TLC5940s we're controlling |
| mjr | 26:cb71c4af2912 | 397 | int nchips; |
| mjr | 26:cb71c4af2912 | 398 | |
| mjr | 26:cb71c4af2912 | 399 | // Timeout to end each PWM cycle. This is a one-shot timer that we reset |
| mjr | 26:cb71c4af2912 | 400 | // on each cycle. |
| mjr | 38:091e511ce8a0 | 401 | Timeout resetTimer; |
| mjr | 26:cb71c4af2912 | 402 | |
| mjr | 33:d832bcab089e | 403 | // Do we need an XLAT signal on the next blanking interval? |
| mjr | 33:d832bcab089e | 404 | volatile bool needXlat; |
| mjr | 54:fd77a6b2f76c | 405 | |
| mjr | 40:cc0d9814522b | 406 | // Reset the grayscale cycle and send the next data update |
| mjr | 26:cb71c4af2912 | 407 | void reset() |
| mjr | 26:cb71c4af2912 | 408 | { |
| mjr | 30:6e9902f06f48 | 409 | // start the blanking cycle |
| mjr | 30:6e9902f06f48 | 410 | startBlank(); |
| mjr | 33:d832bcab089e | 411 | |
| mjr | 54:fd77a6b2f76c | 412 | #if !DATA_UPDATE_INSIDE_BLANKING |
| mjr | 48:058ace2aed1d | 413 | // We're configured to send new GS data during the GS cycle, |
| mjr | 48:058ace2aed1d | 414 | // not during the blanking interval, so end the blanking |
| mjr | 48:058ace2aed1d | 415 | // interval now, before we start sending the new data. Ending |
| mjr | 48:058ace2aed1d | 416 | // the blanking interval starts the new GS cycle. |
| mjr | 33:d832bcab089e | 417 | // |
| mjr | 48:058ace2aed1d | 418 | // (For the other configuration, we send GS data during the |
| mjr | 48:058ace2aed1d | 419 | // blanking interval, so in that case we DON'T end the blanking |
| mjr | 48:058ace2aed1d | 420 | // interval yet - we defer that until the end-of-DMA interrupt |
| mjr | 48:058ace2aed1d | 421 | // handler, which fires after the GS data send has completed.) |
| mjr | 33:d832bcab089e | 422 | endBlank(); |
| mjr | 48:058ace2aed1d | 423 | #endif |
| mjr | 48:058ace2aed1d | 424 | |
| mjr | 54:fd77a6b2f76c | 425 | // if we have pending grayscale data, update the DMA data |
| mjr | 54:fd77a6b2f76c | 426 | bool sendGS = true; // $$$ |
| mjr | 54:fd77a6b2f76c | 427 | if (dirty) |
| mjr | 54:fd77a6b2f76c | 428 | { |
| mjr | 54:fd77a6b2f76c | 429 | // The working buffer has changes since our last update. Swap |
| mjr | 54:fd77a6b2f76c | 430 | // the live and working buffers so that we send the latest updates. |
| mjr | 54:fd77a6b2f76c | 431 | uint8_t *tmp = livebuf; |
| mjr | 54:fd77a6b2f76c | 432 | livebuf = workbuf; |
| mjr | 54:fd77a6b2f76c | 433 | workbuf = tmp; |
| mjr | 54:fd77a6b2f76c | 434 | |
| mjr | 54:fd77a6b2f76c | 435 | // the working buffer is no longer dirty |
| mjr | 54:fd77a6b2f76c | 436 | dirty = false; |
| mjr | 54:fd77a6b2f76c | 437 | sendGS = true; |
| mjr | 54:fd77a6b2f76c | 438 | } |
| mjr | 54:fd77a6b2f76c | 439 | else if (forceUpdate) |
| mjr | 54:fd77a6b2f76c | 440 | { |
| mjr | 54:fd77a6b2f76c | 441 | // send the GS data and consume the forced update flag |
| mjr | 54:fd77a6b2f76c | 442 | sendGS = true; |
| mjr | 54:fd77a6b2f76c | 443 | forceUpdate = false; |
| mjr | 54:fd77a6b2f76c | 444 | } |
| mjr | 54:fd77a6b2f76c | 445 | |
| mjr | 54:fd77a6b2f76c | 446 | // Set the new DMA source to the live buffer. Note that we start |
| mjr | 54:fd77a6b2f76c | 447 | // the DMA transfer with the *second* byte - the first byte must |
| mjr | 54:fd77a6b2f76c | 448 | // be sent by the CPU rather than the DMA module, as outlined in |
| mjr | 54:fd77a6b2f76c | 449 | // the KL25Z hardware reference manual. |
| mjr | 54:fd77a6b2f76c | 450 | |
| mjr | 54:fd77a6b2f76c | 451 | // Start the new DMA transfer. |
| mjr | 54:fd77a6b2f76c | 452 | // |
| mjr | 54:fd77a6b2f76c | 453 | // The hardware reference manual says that the CPU has to send |
| mjr | 54:fd77a6b2f76c | 454 | // the first byte of a DMA transfer explicitly. This is required |
| mjr | 54:fd77a6b2f76c | 455 | // to avoid a hardware deadlock condition that happens due to |
| mjr | 54:fd77a6b2f76c | 456 | // a timing interaction between the SPI and DMA controllers. |
| mjr | 54:fd77a6b2f76c | 457 | // The correct sequence per the manual is: |
| mjr | 54:fd77a6b2f76c | 458 | // |
| mjr | 54:fd77a6b2f76c | 459 | // - reset the SPI controller |
| mjr | 54:fd77a6b2f76c | 460 | // - set up the DMA registers, starting at the 2nd byte to send |
| mjr | 54:fd77a6b2f76c | 461 | // - read the SPI status register (SPI0->S), wait for SPTEF to be set |
| mjr | 54:fd77a6b2f76c | 462 | // - write the first byte to the SPI data register (SPI0->D) |
| mjr | 54:fd77a6b2f76c | 463 | // - enable TXDMAE in the SPI control register (SPI0->C2) |
| mjr | 54:fd77a6b2f76c | 464 | // |
| mjr | 54:fd77a6b2f76c | 465 | if (sendGS) |
| mjr | 48:058ace2aed1d | 466 | { |
| mjr | 54:fd77a6b2f76c | 467 | #if 1 // $$$ |
| mjr | 54:fd77a6b2f76c | 468 | Timer t; t.start(); //$$$ |
| mjr | 54:fd77a6b2f76c | 469 | uint8_t *p = livebuf; |
| mjr | 54:fd77a6b2f76c | 470 | for (int i = dmalen ; i != 0 ; --i) { |
| mjr | 54:fd77a6b2f76c | 471 | while (!(SPI0->S & SPI_S_SPTEF_MASK)) ; |
| mjr | 54:fd77a6b2f76c | 472 | SPI0->D = *p++; |
| mjr | 54:fd77a6b2f76c | 473 | } |
| mjr | 54:fd77a6b2f76c | 474 | needXlat = true; |
| mjr | 54:fd77a6b2f76c | 475 | |
| mjr | 54:fd77a6b2f76c | 476 | spi_total_time += t.read_us(); |
| mjr | 54:fd77a6b2f76c | 477 | spi_runs += 1; |
| mjr | 54:fd77a6b2f76c | 478 | #else |
| mjr | 54:fd77a6b2f76c | 479 | // disable DMA on SPI0 |
| mjr | 54:fd77a6b2f76c | 480 | SPI0->C2 &= ~SPI_C2_TXDMAE_MASK; |
| mjr | 54:fd77a6b2f76c | 481 | |
| mjr | 54:fd77a6b2f76c | 482 | // reset SPI0 |
| mjr | 54:fd77a6b2f76c | 483 | SPI0->C1 &= ~SPI_C1_SPE_MASK; |
| mjr | 54:fd77a6b2f76c | 484 | |
| mjr | 54:fd77a6b2f76c | 485 | // set up a transfer from the second byte of the buffer |
| mjr | 54:fd77a6b2f76c | 486 | sdma.source(livebuf + 1, true, 8); |
| mjr | 54:fd77a6b2f76c | 487 | sdma.start(dmalen - 1, false); |
| mjr | 54:fd77a6b2f76c | 488 | |
| mjr | 54:fd77a6b2f76c | 489 | // enable SPI0 |
| mjr | 54:fd77a6b2f76c | 490 | SPI0->C1 |= SPI_C1_SPE_MASK; |
| mjr | 54:fd77a6b2f76c | 491 | |
| mjr | 54:fd77a6b2f76c | 492 | // wait for the TX buffer to clear, then write the first byte manually |
| mjr | 54:fd77a6b2f76c | 493 | while (!(SPI0->S & SPI_S_SPTEF_MASK)) ; |
| mjr | 54:fd77a6b2f76c | 494 | SPI0->D = livebuf[0]; |
| mjr | 54:fd77a6b2f76c | 495 | |
| mjr | 54:fd77a6b2f76c | 496 | // enable DMA to carry out the rest of the transfer |
| mjr | 48:058ace2aed1d | 497 | SPI0->C2 |= SPI_C2_TXDMAE_MASK; |
| mjr | 54:fd77a6b2f76c | 498 | |
| mjr | 54:fd77a6b2f76c | 499 | // we'll need a translate on the next blanking cycle |
| mjr | 54:fd77a6b2f76c | 500 | needXlat = true; |
| mjr | 54:fd77a6b2f76c | 501 | #endif |
| mjr | 48:058ace2aed1d | 502 | } |
| mjr | 54:fd77a6b2f76c | 503 | |
| mjr | 54:fd77a6b2f76c | 504 | #if !DATA_UPDATE_INSIDE_BLANKING |
| mjr | 54:fd77a6b2f76c | 505 | // arm the reset handler |
| mjr | 54:fd77a6b2f76c | 506 | armReset(); |
| mjr | 54:fd77a6b2f76c | 507 | #endif |
| mjr | 54:fd77a6b2f76c | 508 | } |
| mjr | 54:fd77a6b2f76c | 509 | |
| mjr | 54:fd77a6b2f76c | 510 | // arm the reset handler - this fires at the end of each GS cycle |
| mjr | 54:fd77a6b2f76c | 511 | void armReset() |
| mjr | 54:fd77a6b2f76c | 512 | { |
| mjr | 54:fd77a6b2f76c | 513 | resetTimer.attach(this, &TLC5940::reset, (1.0/GSCLK_SPEED)*4096.0); |
| mjr | 30:6e9902f06f48 | 514 | } |
| mjr | 30:6e9902f06f48 | 515 | |
| mjr | 30:6e9902f06f48 | 516 | void startBlank() |
| mjr | 30:6e9902f06f48 | 517 | { |
| mjr | 54:fd77a6b2f76c | 518 | //static int i=0; i=(i+1)%200; extern void diagLED(int,int,int); diagLED(i<100,i>=100,0);//$$$ |
| mjr | 54:fd77a6b2f76c | 519 | |
| mjr | 30:6e9902f06f48 | 520 | // turn off the grayscale clock, and assert BLANK to end the grayscale cycle |
| mjr | 30:6e9902f06f48 | 521 | gsclk.write(0); |
| mjr | 54:fd77a6b2f76c | 522 | blank = (enabled ? 1 : 0); // for the slight delay (20ns) required after GSCLK goes low |
| mjr | 30:6e9902f06f48 | 523 | blank = 1; |
| mjr | 30:6e9902f06f48 | 524 | } |
| mjr | 26:cb71c4af2912 | 525 | |
| mjr | 33:d832bcab089e | 526 | void endBlank() |
| mjr | 30:6e9902f06f48 | 527 | { |
| mjr | 54:fd77a6b2f76c | 528 | //static int i=0; i=(i+1)%200; extern void diagLED(int,int,int); diagLED(-1,i<100,-1);//$$$ |
| mjr | 54:fd77a6b2f76c | 529 | |
| mjr | 33:d832bcab089e | 530 | // if we've sent new grayscale data since the last blanking |
| mjr | 33:d832bcab089e | 531 | // interval, latch it by asserting XLAT |
| mjr | 33:d832bcab089e | 532 | if (needXlat) |
| mjr | 30:6e9902f06f48 | 533 | { |
| mjr | 26:cb71c4af2912 | 534 | // latch the new data while we're still blanked |
| mjr | 26:cb71c4af2912 | 535 | xlat = 1; |
| mjr | 26:cb71c4af2912 | 536 | xlat = 0; |
| mjr | 33:d832bcab089e | 537 | needXlat = false; |
| mjr | 26:cb71c4af2912 | 538 | } |
| mjr | 26:cb71c4af2912 | 539 | |
| mjr | 40:cc0d9814522b | 540 | // End the blanking interval and restart the grayscale clock. Note |
| mjr | 40:cc0d9814522b | 541 | // that we keep the blanking on if the chips are globally disabled. |
| mjr | 54:fd77a6b2f76c | 542 | if (enabled) |
| mjr | 54:fd77a6b2f76c | 543 | { |
| mjr | 54:fd77a6b2f76c | 544 | blank = 0; |
| mjr | 54:fd77a6b2f76c | 545 | gsclk.write(.5); |
| mjr | 54:fd77a6b2f76c | 546 | } |
| mjr | 26:cb71c4af2912 | 547 | } |
| mjr | 26:cb71c4af2912 | 548 | |
| mjr | 30:6e9902f06f48 | 549 | // Interrupt handler for DMA completion. The DMA controller calls this |
| mjr | 30:6e9902f06f48 | 550 | // when it finishes with the transfer request we set up above. When the |
| mjr | 30:6e9902f06f48 | 551 | // transfer is done, we simply end the blanking cycle and start a new |
| mjr | 30:6e9902f06f48 | 552 | // grayscale cycle. |
| mjr | 30:6e9902f06f48 | 553 | void dmaDone() |
| mjr | 30:6e9902f06f48 | 554 | { |
| mjr | 54:fd77a6b2f76c | 555 | //static int i=0; i=(i+1)%200; extern void diagLED(int,int,int); diagLED(i<100,-1,-1);//$$$ |
| mjr | 54:fd77a6b2f76c | 556 | |
| mjr | 48:058ace2aed1d | 557 | // disable DMA triggering in the SPI controller until we set |
| mjr | 48:058ace2aed1d | 558 | // up the next transfer |
| mjr | 48:058ace2aed1d | 559 | SPI0->C2 &= ~SPI_C2_TXDMAE_MASK; |
| mjr | 54:fd77a6b2f76c | 560 | SPI0->C1 &= ~SPI_C1_SPE_MASK; |
| mjr | 48:058ace2aed1d | 561 | |
| mjr | 33:d832bcab089e | 562 | // mark that we need to assert XLAT to latch the new |
| mjr | 33:d832bcab089e | 563 | // grayscale data during the next blanking interval |
| mjr | 54:fd77a6b2f76c | 564 | needXlat = true; |
| mjr | 33:d832bcab089e | 565 | |
| mjr | 33:d832bcab089e | 566 | #if DATA_UPDATE_INSIDE_BLANKING |
| mjr | 33:d832bcab089e | 567 | // we're doing the gs update within the blanking cycle, so end |
| mjr | 33:d832bcab089e | 568 | // the blanking cycle now that the transfer has completed |
| mjr | 33:d832bcab089e | 569 | endBlank(); |
| mjr | 54:fd77a6b2f76c | 570 | |
| mjr | 54:fd77a6b2f76c | 571 | // set up the next blanking interrupt |
| mjr | 54:fd77a6b2f76c | 572 | armReset(); |
| mjr | 33:d832bcab089e | 573 | #endif |
| mjr | 30:6e9902f06f48 | 574 | } |
| mjr | 30:6e9902f06f48 | 575 | |
| mjr | 26:cb71c4af2912 | 576 | }; |
| mjr | 26:cb71c4af2912 | 577 | |
| mjr | 26:cb71c4af2912 | 578 | #endif |