Muzi Works R1 Neo support by khudson · Pull Request #2007 · meshcore-dev/MeshCore

khudson · 2026-03-12T02:06:17Z

Resolves #1275.

Adds support for Muzi Works R1 Neo device.

Adds new RTC: Epson Seiko RX8130CE.

All features on the board currently work with the exception of software shutdown. Workaround: Hold the button for 8 seconds to put the unit into DFU mode with the USB cable disconnected. The device will shut down.

Support for R1 Neo hardware. New variant and baseboard class. * Known issues: - power management is not currently supported - power off via long button press is not implemented Add support for Epson Seiko RX8130CE I2C Real-time clock.

weebl2000

Thx for the implementation @khudson - I hope my review is useful ;)

variants/muziworks_r1_neo/R1NeoBoard.cpp

variants/muziworks_r1_neo/variant.h

weebl2000 · 2026-03-12T02:41:37Z

src/helpers/RTC_RX8130CE.cpp

+    buf[1] = bin2bcd(t->tm_sec) & 0x7F;
+    buf[2] = bin2bcd(t->tm_min) & 0x7F;
+    buf[3] = bin2bcd(t->tm_hour) & 0x3F;
+    buf[4] = bin2bcd(t->tm_wday) & 0x07;


The RX8130CE WEEK register uses a bitmap format (bit 0 = Sun, bit 1 = Mon, ..., bit 6 = Sat), not BCD. For example Wednesday (tm_wday=3) should be 0x08 (bit 3), but bin2bcd(3) gives 0x03 (bits 0+1 = Sun+Mon). The ArtronShop library has the same bug - opened a fix upstream: ArtronShop/ArtronShop_RX8130CE#1

Suggested change

buf[4] = bin2bcd(t->tm_wday) & 0x07;

buf[4] = (1 << t->tm_wday) & 0x7F;

weebl2000 · 2026-03-12T02:41:37Z

src/helpers/RTC_RX8130CE.cpp

+    t->tm_sec =  bcd2bin(buff[0] & 0x7F);
+    t->tm_min =  bcd2bin(buff[1] & 0x7F);
+    t->tm_hour = bcd2bin(buff[2] & 0x3F);
+    t->tm_wday = bcd2bin(buff[3] & 0x07);


Same bitmap issue as setTime() — the WEEK register is a bitmap, not BCD. This needs to find which bit is set and convert it back to an index. See ArtronShop/ArtronShop_RX8130CE#1

Suggested change

t->tm_wday = bcd2bin(buff[3] & 0x07);

// WEEK register is a bitmap (bit 0=Sun, bit 1=Mon, ..., bit 6=Sat)

uint8_t wday_bits = buff[3] & 0x7F;

t->tm_wday = 0;

while (wday_bits >>= 1) t->tm_wday++;

weebl2000 · 2026-03-12T02:41:37Z

src/helpers/AutoDiscoverRTCClock.cpp

+  if (i2c_probe(wire, RX8130CE_ADDRESS)) {
+    MESH_DEBUG_PRINTLN("RX8130CE: Found");
+    rtc_8130.begin(&wire);
+    rtc_8130_success = true;
+    MESH_DEBUG_PRINTLN("RX8130CE: Initialized");
+  }


Nit: begin() return value is ignored. Unlikely to fail since i2c_probe() already confirmed the device is present, but worth checking defensively to match the pattern of the other RTCs above.

Suggested change

if (i2c_probe(wire, RX8130CE_ADDRESS)) {

MESH_DEBUG_PRINTLN("RX8130CE: Found");

rtc_8130.begin(&wire);

rtc_8130_success = true;

MESH_DEBUG_PRINTLN("RX8130CE: Initialized");

}

if (i2c_probe(wire, RX8130CE_ADDRESS)) {

MESH_DEBUG_PRINTLN("RX8130CE: Found");

rtc_8130_success = rtc_8130.begin(&wire);

if (rtc_8130_success) {

MESH_DEBUG_PRINTLN("RX8130CE: Initialized");

}

}

@khudson probably good to do this check

weebl2000 · 2026-03-12T02:41:37Z

src/helpers/RTC_RX8130CE.cpp

+    if (!i2c_dev->begin()) {
+       return false;
+    }
+


Per the RX8130CE datasheet init flowchart (fig. 42, page 55), after a power loss the VLF flag (bit 1 of Flag Register 0x1D) indicates clock data may be unreliable. A dummy read followed by a VLF check should happen before the register configuration. See also ArtronShop/ArtronShop_RX8130CE#1

Suggested change

if (!i2c_dev->begin()) {

return false;

}

if (!i2c_dev->begin()) {

return false;

}

// Dummy read per init flowchart (page 55)

read_register(0x1D);

// Check VLF flag (bit 1 of Flag Register 0x1D) — indicates data loss

uint8_t flags = read_register(0x1D);

if (flags & 0x02) {

// VLF set: clock data unreliable after power loss, clear flag

write_register(0x1D, flags & ~0x02);

}

weebl2000 · 2026-03-12T02:41:37Z

variants/muziworks_r1_neo/variant.h

+#define PIN_VBAT_READ (31) // P0.31 (39) ADC_VBAT
+#define PIN_BAT_CHG   (34) // P1.02 (26) BAT_CHG_STATUS
+
+#define ADC_MULTIPLIER (3 * 1.73 * 1.187 * 1000)


The Meshtastic variant uses an ADC multiplier of 1.667, which was calibrated by simon-muzi via a 15-hour discharge test (Meshtastic commit f4e260e0f). Accounting for nRF52's 3.6V default AREF that works out to 3.6 * 1.667 * 1000 = 6001.2. The current formula gives ~6159 which is about 2.6% off. Could you verify with a multimeter to confirm which value is closer?

Suggested change

#define ADC_MULTIPLIER (3 * 1.73 * 1.187 * 1000)

#define ADC_MULTIPLIER (3.6 * 1.667 * 1000)

variants/muziworks_r1_neo/target.cpp

variants/muziworks_r1_neo/target.h

variants/muziworks_r1_neo/platformio.ini

weebl2000 · 2026-03-12T03:11:30Z

src/helpers/RTC_RX8130CE.cpp

+
+
+bool RTC_RX8130CE::stop(bool stop) {
+    write_register(0x1E, stop ? 0x040 : 0x00);


Suggested change

write_register(0x1E, stop ? 0x040 : 0x00);

write_register(0x1E, stop ? 0x40 : 0x00);

weebl2000 · 2026-03-12T03:11:30Z

src/helpers/RTC_RX8130CE.cpp

+    for (int i = 1; i <= len + 1; i++) {
+       buf[i] = value[i - 1];
+    }


Suggested change

for (int i = 1; i <= len + 1; i++) {

buf[i] = value[i - 1];

}

for (int i = 1; i <= len; i++) {

buf[i] = value[i - 1];

}

src/helpers/RTC_RX8130CE.cpp

weebl2000 · 2026-03-12T03:52:08Z

variants/muziworks_r1_neo/variant.h

+#define PIN_QSPI_SCK  (3) // P0.03 (29) BUZZER
+#define PIN_QSPI_CS  (26) // P0.26 (34) USER_BUTTON
+#define PIN_QSPI_IO0 (30) // P0.30 (33) MCU_SIGNAL
+#define PIN_QSPI_IO1 (29) // P0.29 (32) SOFT_SHUTDOWN
+#define PIN_QSPI_IO2 (28) // P0.28 (31) BLU_LED_RAK
+#define PIN_QSPI_IO3  (2) // P0.02 (30) GPS_PPS


MeshCore convention is PIN_GPS_TX = GPS module's TX output = MCU receives, so PIN_SERIAL1_RX should map to PIN_GPS_TX (and vice versa). Every other nRF52 variant does it that way.

Also, the PIN_QSPI_* block and EXTERNAL_FLASH_* defines should be removed — the R1 Neo has no QSPI flash, and those defines alias active GPIOs (buzzer, user button, soft shutdown, LED, GPS PPS). If the BSP ever initializes the QSPI peripheral it would reconfigure those pins.

Suggested change

#define PIN_QSPI_SCK (3) // P0.03 (29) BUZZER

#define PIN_QSPI_CS (26) // P0.26 (34) USER_BUTTON

#define PIN_QSPI_IO0 (30) // P0.30 (33) MCU_SIGNAL

#define PIN_QSPI_IO1 (29) // P0.29 (32) SOFT_SHUTDOWN

#define PIN_QSPI_IO2 (28) // P0.28 (31) BLU_LED_RAK

#define PIN_QSPI_IO3 (2) // P0.02 (30) GPS_PPS

#define PIN_SERIAL1_RX (PIN_GPS_TX) // MCU receives <- GPS transmits

#define PIN_SERIAL1_TX (PIN_GPS_RX) // MCU transmits -> GPS receives

variants/muziworks_r1_neo/platformio.ini

brycv · 2026-03-24T17:24:58Z

Where are we at with this? Waiting for more feedback? Thanks again! Can't wait!

mem101296 · 2026-03-31T11:40:50Z

Hello. Wondering what needs to be done to merge this? I'm excited to get the r1 neo on meshcore.

khudson · 2026-04-01T01:23:24Z

Just waiting for it to be merged at this point. Hopefully @ripplebiz or someone can give it a once over and push it through.

…

On Mar 31, 2026, at 4:41 AM, mem101296 ***@***.***> wrote: mem101296 left a comment (meshcore-dev/MeshCore#2007) Hello. Wondering what needs to be done to merge this? I'm excited to get the r1 neo on meshcore. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>

Co-authored-by: Wessel <wessel@weebl.me>

liamcottle · 2026-04-01T11:17:44Z

src/helpers/AutoDiscoverRTCClock.cpp

+  if (i2c_probe(wire, RX8130CE_ADDRESS)) {
+    MESH_DEBUG_PRINTLN("RX8130CE: Found");
+    rtc_8130.begin(&wire);
+    rtc_8130_success = true;
+    MESH_DEBUG_PRINTLN("RX8130CE: Initialized");
+  }


@khudson probably good to do this check

liamcottle · 2026-04-01T11:22:32Z

variants/muziworks_r1_neo/R1NeoBoard.h

  #define BATTERY_SAMPLES 8

  uint16_t getBattMilliVolts() override {
+    MESH_DEBUG_PRINTLN("R1Neo: Sampling battery");


probably don't need this debug line

liamcottle · 2026-04-01T11:34:59Z

Just waiting for it to be merged at this point. Hopefully @ripplebiz or someone can give it a once over and push it through.

Hey @khudson, thanks for the PR. Sorry for being so slow at reviewing it. Have been working on lots of other things at once. Originally when I looked at it, I saw a bunch of review comments from @weebl2000 that had gone unanswered.

Was waiting to see any movement on those. Most of the changes are variant specific, except for the RTC clock stuff, so if there's any issues it should only affect the new variant.

I added some minor comments, that would be good to resolve, but will merge now anyway so things can move forward :)

Note: I also merged in a couple of minor changes related to the uint8_t vs int8_t for tx power.

Thanks for contributing!

khudson · 2026-04-01T12:33:47Z

Thanks - I had mainly just gotten frustrated at the pushy, AI-generated code review and my inability to test these commits before merge. I had asked weebl2000 to submit PRs to my branch directly, without conflicts, so I could test them locally and that hadn't been done. A lot of the review points to code paths that aren't even in use, so it's not particularly important to address them now before the code gets merged.Thanks - I look forward to seeing this live in the next release!On Apr 1, 2026, at 4:35 AM, Liam Cottle ***@***.***> wrote:liamcottle left a comment (meshcore-dev/MeshCore#2007) Just waiting for it to be merged at this point. Hopefully @ripplebiz or someone can give it a once over and push it through. Hey @khudson, thanks for the PR. Sorry for being so slow at reviewing it. Have been working on lots of other things at once. Originally when I looked at it, I saw a bunch of review comments from @weebl2000 that had gone unanswered. Was waiting to see any movement on those. Most of the changes are variant specific, except for the RTC clock stuff, so if there's any issues it should only affect the new variant. I added some minor comments, that would be good to resolve, but will merge now anyway so things can move forward :) Thanks for contributing! —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: ***@***.***>

khudson added 2 commits March 8, 2026 11:46

Support for muzi works R1 Neo device.

5d1f513

Support for R1 Neo hardware. New variant and baseboard class. * Known issues: - power management is not currently supported - power off via long button press is not implemented Add support for Epson Seiko RX8130CE I2C Real-time clock.

R1 Neo: Turn off debugging, rtc cleanup

a56419d

khudson mentioned this pull request Mar 12, 2026

Add firmware for Muzi Works R1 NEO device #1275

Open

weebl2000 reviewed Mar 12, 2026

View reviewed changes

variants/muziworks_r1_neo/platformio.ini Show resolved Hide resolved

weebl2000 reviewed Mar 12, 2026

View reviewed changes

variants/muziworks_r1_neo/platformio.ini Show resolved Hide resolved

liamcottle and others added 2 commits April 2, 2026 00:23

Apply suggestion from @weebl2000

8fc9e17

Co-authored-by: Wessel <wessel@weebl.me>

Apply suggestion from @weebl2000

fcad0dc

Co-authored-by: Wessel <wessel@weebl.me>

liamcottle approved these changes Apr 1, 2026

View reviewed changes

liamcottle merged commit b934daa into meshcore-dev:dev Apr 1, 2026
9 checks passed

	buf[4] = bin2bcd(t->tm_wday) & 0x07;
	buf[4] = (1 << t->tm_wday) & 0x7F;

-    t->tm_wday = bcd2bin(buff[3] & 0x07);
+    // WEEK register is a bitmap (bit 0=Sun, bit 1=Mon, ..., bit 6=Sat)
+    uint8_t wday_bits = buff[3] & 0x7F;
+    t->tm_wday = 0;
+    while (wday_bits >>= 1) t->tm_wday++;

-    if (!i2c_dev->begin()) {
-       return false;
-    }
+    if (!i2c_dev->begin()) {
+       return false;
+    }
+    // Dummy read per init flowchart (page 55)
+    read_register(0x1D);
+    // Check VLF flag (bit 1 of Flag Register 0x1D) — indicates data loss
+    uint8_t flags = read_register(0x1D);
+    if (flags & 0x02) {
+       // VLF set: clock data unreliable after power loss, clear flag
+       write_register(0x1D, flags & ~0x02);
+    }

	#define ADC_MULTIPLIER (3 * 1.73 * 1.187 * 1000)
	#define ADC_MULTIPLIER (3.6 * 1.667 * 1000)



		bool RTC_RX8130CE::stop(bool stop) {
		write_register(0x1E, stop ? 0x040 : 0x00);

	write_register(0x1E, stop ? 0x040 : 0x00);
	write_register(0x1E, stop ? 0x40 : 0x00);

Conversation

khudson commented Mar 12, 2026

Uh oh!

weebl2000 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

weebl2000 Mar 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

brycv commented Mar 24, 2026

Uh oh!

mem101296 commented Mar 31, 2026

Uh oh!

khudson commented Apr 1, 2026 via email

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

liamcottle commented Apr 1, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

khudson commented Apr 1, 2026 via email

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

weebl2000 Mar 12, 2026 •

edited

Loading

liamcottle commented Apr 1, 2026 •

edited

Loading