Hi there,
anyone out there getting the ESP32 working with this nice project?
Any advice how to realise?

Br
JIP0815

I think the leftover command is wrong... I'm not sure how it works though.

Look into the code here:

	uint16_t leftover = buffersize & 0x1f;
	if( leftover ) leftover = 32 - leftover;
	for( place = 0; place < leftover; place++ )
	{
		*(bufferpl++) = 0;
		*(bufferpl++) = 0;
	}

Hey Charles-- love all your stuff!

Been using this project in one of my own for some months now. In the beginning it was working fine, but recently I started getting some flickering. I couldn't find any hardware issues so I plugged into the board and lo and behold it was crashing randomly. Reflashed the firmware and it got stuck in a boot loop. Figured it was a hardware malfunction and bought a replacement.

On the new one, I flashed it and found it to be crashing with a "MEM CHECK FAIL" error-- but only when I sent it UDP packets above a certain size. SSID/wifi functionality was working fine. I reflashed again and it too is now stuck in a boot loop.

These things were just sitting on a breadboard, in a setup which I had working great with 300+ lights for months. Do you have any advice for me to track this down? What should I look for to fix/what could be breaking it like this?

You probably mean #ifdef USE_2812_INTERRUPTS?

Your repo didn't compile for me (conflicting copies of strcat so I probably have a different sdk setup).
So I tried to replace ws2812.c (from your other repo) in my current project with ws2812_i2s.c.
It compiles when I replace the #if with an #ifdef. (also copied ws2812_i2s.h and some other headers of course).

In your README.MD you are saying this uses the RX1 pin of the ESP8266 to shift out I2S bits.
Am i correct that this means that with any development boards where the USB serial bridge is hardwired (like. NodeMCU) I2S usage is impossible?

Hey cnlohr,

thanks for sharing your work!
I'm trying to get it to work on a ordinary esp-01 breakout board. What I did so far:

• adjusting the paths in the Makefile to my system
• installing this Toolchain https://github.com/esp8266/esp8266-wiki/wiki/Toolchain (so far i only programmed the esp via arduino)
• run make and make burn in your repo

what works: flashing
what doesn't work: I get this via serial when i power up the esp

Fatal exception 0(IllegalInstructionCause):
epc1=0x40251c3f, epc2=0x00000000, epc3=0x00000000, excvaddr=0x00000000, depc=0x00000000

 ets Jan  8 2013,rst cause:2, boot mode:(3,6)

load 0x40100000, len 30088, room 16 
tail 8
chksum 0x77
load 0x3ffe8000, len 876, room 0 
tail 12
chksum 0x5b
ho 0 tail 12 room 4
load 0x3ffe8370, len 2504, room 12 
tail 12
chksum 0x44
csum 0x44
don't use rtc mem data
Fatal exception �d

are you asuming certain other bin files flashed to other areas of the memory? which?

thanks again!

Hi!

I think the values for WS_I2S_BCK and WS_I2S_DIV are incorrect.
Currently -1 is subtracted from both before setting the values.
Instead they should be used as is.

The calculation given in the code for the frequency is:
// 1600000000L/(div*bestbck), I assume there's one '0' too many :)
Which I would assume means:
160000000L / (WS_I2S_DIV * WS_I2S_BCK) = bit frequency in Hz

For an example I'll use WS2812_FOUR_SAMPLE, so:
#define WS_I2S_BCK 17
#define WS_I2S_DIV 4
And for the buffer sizes:
#define WS_BLOCKSIZE 4000
And 32 byte for the 'zero' buffer.
That's 4032 bytes for all the DMA buffers.

By triggering an interrupt on EOF on one of the buffers, I can quite accurately get the Hz of going through all the 4032 buffer bytes.
According to the caluculation and dividers this should be:
(160000000 / (4032 * 8) ) / (4 * 17) = 4960.317 / (4 * 17) = 72.94584500466853 Hz
But the value I measure is close to 103.3 Hz..
Now If I use the '-1' values instead: 3 and 16 we get:
4960.317 / (3 * 16) = 103.3399470899471 Hz
Which is pretty much identical to the measured value :)

So my suggestion is to not subtract -1 of the two values, and then use:
#define WS_I2S_BCK 16
#define WS_I2S_DIV 3
Which are the actual values used.
And correct the comment with the formula to:
// 160000000L / ( WS_I2S_DIV * WS_I2S_BCK )

Cheers :)

Hello, Nice work!

After taking off power i noticed that some leds have a flicker (often a little flash to red).
Have you ever had this kind of problem?

Thank you!

I just tried it out myself and it seems that sometimes when I send ws2812 network data, the ESP crashes and reboots. This seems to happen frequently, every 2-3 seconds! I'll look into it more tonight, but, it looks to be recent changes. Anything @con-f-use could have done?

Hi!

Quick question/remark, as issues cause I didn't see any contact details :)

Since this way of driving the WS2812's uses pin: RX1/(pin 25)/U0RXD/i2SO_DATA (lots of names for one pin :) ).
Which is also the RX pin when flashing the ESP8266.
Am I in any danger of breaking my ESP by flashing this firmware?

I'm worried about this because my serial interface might make it's TX high, at the same time the ESP boots with this firmware, which would set the pin as out and also make it high to set the leds to their initial state?

I'm not experienced enough to know if this is an issue or not, but I'm sure I remember never hooking up two outputs together ;)
Obviously once it's properly flashed I can remove the serial interface, and use OTA for further updates, but I'm worried about the short period inbetween :) Especially since I can't remember if the ESP reboots directly after flashing XD

Any suggestions/workarounds? :)
Thanks! :)

I dont understand what I need to change to fix this ? I am using the toolchain installed from this:
https://github.com/esp8266/esp8266-wiki/wiki/Toolchain

jon@esp8266dev:~/esp8266-examples/esp8266ws2812i2s$ make
/esp-open-sdk/xtensa-lx106-elf/bin/xtensa-lx106-elf-gcc -mlongcalls -I/home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/include -Imyclib -Iinclude -Iuser -Os -I/home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/include/ -Icommon -DICACHE_FLASH driver/uart.c common/mystuff.c common/flash_rewriter.c common/http.c common/commonservices.c common/http_custom.c common/mdns.c common/mfs.c user/custom_commands.c user/ws2812_i2s.c user/user_main.c -flto -nostdlib -Wl,--relax -Wl,--gc-sections -L/home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib -L/esp-open-sdk/xtensa-lx106-elf/lib/gcc/xtensa-lx106-elf/4.8.2/libgcc.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/liblwip.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/libssl.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/libupgrade.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/libnet80211.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/liblwip.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/libwpa.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/libnet80211.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/libphy.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/libcrypto.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/libmain.a /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib/libpp.a ~/esp-open-sdk/xtensa-lx106-elf/lib/gcc/xtensa-lx106-elf/4.8.2/libgcc.a -T /home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/ld/eagle.app.v6.ld -B/home/jon/esp-open-sdk/esp_iot_sdk_v1.5.2/lib -o image.elf
FW 0x00000.bin
/home/jon/esp-open-sdk/esptool/esptool.py -eo image.elf -bo 0x00000.bin -bs .text -bs .data -bs .rodata -bc -ec
usage: esptool [-h] [--port PORT] [--baud BAUD]
{load_ram,dump_mem,read_mem,write_mem,write_flash,run,image_info,make_image,elf2image,read_mac,chip_id,flash_id,read_flash,verify_flash,erase_flash}
...
esptool: error: argument operation: invalid choice: 'image.elf' (choose from 'load_ram', 'dump_mem', 'read_mem', 'write_mem', 'write_flash', 'run', 'image_info', 'make_image', 'elf2image', 'read_mac', 'chip_id', 'flash_id', 'read_flash', 'verify_flash', 'erase_flash')
Makefile:72: recipe for target '0x00000.bin' failed
make: *** [0x00000.bin] Error 2

if I let wsend run and watch the usb serial port on 115200, I see the device reboot every time after it connects to the local wifi network and receives a udp packet from wsend.
on the webui however, it looks to do what it's supposed to without rebooting.

ets Jan  8 2013,rst cause:2, boot mode:(3,7)

load 0x40100000, len 32492, room 16
tail 12
chksum 0xf0
ho 0 tail 12 room 4
load 0x3ffe8000, len 884, room 12
tail 8
chksum 0x10
load 0x3ffe8380, len 2716, room 0
tail 12
chksum 0x39
csum 0x39
��r�l��H����2XX Web-GUI
Version: v0.5-14-g1d7d7-dev - Build sø. 11. sep. 00:13:34 +0200 2016 with -DICACHE_FLASH -DWEB_PORT=80 -DCOM_PORT=7777 -DBACKEND_PORT=7878  -DSLOWTICK_MS=5Loading Settings: af / 0 / 69 / 69
 Settings Loaded: ESP_881DBF / Default
Opmode: 1
Station mode: "MYSSID":"MYPASS" (bssid_set:0)
 Boot Ok.
IGMP Joining: 6b25a8c0 fb0000e0
 STAT: 5
IP: 192.168.37.107
NM: 255.255.255.0
GW: 192.168.37.1
WCFG: /MYSSID/MYPASS/
IGMP Joining: 6b25a8c0 fb0000e0
MFS Not found at regular address (ffffffff).
MFS Found at: 00010000

esp8266ws2812i2s seems like exactly what i'm looking for, a way to have the esp8266 be a dumb front-end to a bunch of WS2812 LED's. Unfortunately I can't seem get the any output from it at all.

I've successfully got it running on the esp8266 and can access the website. When I run the top command, the website's LED display does cycle.

I've attempted to use both the THREE & FOUR sample types and have played with moving the WS_I2S_BCK up and down to no avail.

The fact that the LED's don't even flicker suggests to me that something's fundamentally mis-wired somewhere, but I am connecting the data-in line to the RX pin. The LED's do work with the neopixel library on the same device, (but on different PIN). I've had problems running neopixel in conjunction with network activity though, so I thought your library would be ideal.

Issue #2 mentions some debugging output. Can you provide some hints on how to access that? Since the serial RX port is busy with I2S, I don't think I can access the debug output that way.

Any idea's on other things to try?

../user/pattern.c: In function ‘hex_pattern’:
../user/pattern.c:52:37: error: ‘NULL’ undeclared (first use in this function)

if stdlib.h is included in pattern.c , it compiles fine.

Hi Charles. I compiled the project with no changes and flashed it to my board. I used the 'top' application to send packets to the module but none of the LEDs flashed. I used a logic analyzer to see the signal coming out of the U0RXD_U/I2S0_DATA pin. This is before sending running 'top':

Zoomed in pulse:

After running the 'top':

Any idea what the problem is?

Would this library work with ws2813, if not what do you recommend to be done to make it do so..

I'm trying to add APA104 support with another timings (T0H: 0.35µs, T0L: 1.36µs, T1H: 1.36µs, T1L: 0.35µs).

Can you explain your I²S frequency setup formula?

12000000 / (div * bestbck * 2)

WS2812B timings is
T0H: 0.4µs, T0L: 0.85µs, T1H: 0.8µs, T1L: 0.45µs
1,25µs per led / 3 sample = 0,416 µs per sample

Default settings is

define WS2812_THREE_SAMPLE

define WS_I2S_BCK 22

define WS_I2S_DIV 4

Then..
12000000 Hz — system frequency
2 — I²S frequency (half of system frequency)
4 — I²S frequency divider
22 — secondary I²S frequency divider

12000000 / (4 * 22 * 2) = 68181,81 Hz I²S bit speed
68181,81 / 1000000 = 0,0681 µs per bit

Aw, but there is 0,416 µs per sample.. Something goes wrong.

And why you need bitpattern arrays?

When connecting a led strip to the required pin the leds light up in the expected pattern, but with a lot of noise.
Up to 10 leds are flickering random colors and the whole strip flickers off and on.

Thanks for making this I have been playing with it but I am having issues with sending data over UDP this is not an issue with your code but i am unsure of where to post it :)

If I look at the data going out via wsend it looks like 00ff00 but if I echo data to the UDP port
echo "00ff00ff00000000ff" > /dev/udp/192.168.4.1/7777

I get a different set of colors that affect more than the expected first 3 led's being red, green and blue?

any input on what I am doing wrong will be appreciated.

Dear Charles

I've managed to flash the bin files provided to my esp-12e and successfully blink some leds with the provided top program.

The only thing I can't make work is the web interface. Whenever I'm connected to the esp and browse to 192.168.4.1 I get

File not found.

Do I need to flash some additional files to the esp in order for the web interface to work properly?

Thanks a lot for the great work and your videos on youtube!
Cheers,
gabe0815

i installed latest version and compiled with SDK 1.4.0, but i am unable to get mDNS to advertise.
this is my output:

esp8266 ws2812 driver
Settings Loaded: ESP_9E5F33 / Default
Opmode: 1
Station mode: "SSID":"pass" (bssid_set:0)
IGMP Joining: b60aa8c0 fb0000e0
STAT: 5
IP: 192.168.10.182
NM: 255.255.255.0
GW: 192.168.10.1
WCFG: /SSID/pass/
IGMP Joining: b60aa8c0 

So i see that JoinGropMDNS() was called but i cant find advertisement with commad avahi-browse -at from another linux station.

This version doesn't seem to fall back to AP mode, when connecting to another network fails. ColorChord firmware seems to do that.

The generic buffer is currently unaligned. This can cause serious issues with reading from the MFS.

Hi!

I'm (slowly) trying to clean up/improve(?) your ws2812 i2s code:
https://github.com/cranphin/esp8266-dev/blob/master/esp8266-neo/dev/driver/ws2812_i2s.c

One thing I'd like to do is actually enable/use DMA interrupts.
This is certainly possible, someone has done this rather elegantly (I think) for the Arduino core:
https://github.com/Makuna/NeoPixelBus/blob/master/src/internal/NeoEsp8266DmaMethod.h#L282
But he has the benefit of the yield() method the Arduino core provides:
https://github.com/Makuna/NeoPixelBus/blob/master/src/internal/NeoEsp8266DmaMethod.h#L241
Which he uses to 'block' the update call until the DMA buffer has space, I don't think it's easy to do something like that in esp-open-sdk/Espressif NON-OS.

My current idea is to store the received/update bytes (just the RGB bytes) in a buffer, possibly a ring buffer size 2, with each part the length of the led string size. And then a far smaller DMA buffer (ring buffer as well I think, with at least 2 parts for converted RGB bytes), which is fed in small parts from the previous buffer in the interrupt, possibly in combination with a 'zero' buffer once it's done.
So the update method writes to the first buffer (overwriting the previous value if we get data faster then we can send it to the leds), And the interrupt moves it in small parts to the two DMA buffers, marking the first buffer done when it's finished.

Any thoughts on this? I'm really not used to working with interrupts and DMA, usually I work in Java :)
You seem pretty awesome at it though ;)

Need to switch to dynamic memory for much more. The footprint of this whole collection of stuff, like MDNS, etc, just takes up too much memory by default.