WiFi & NTP Setup via Captive Portal

When the unit is unconfigured, factory-reset, or otherwise unable to associate with a WiFi network, it broadcasts its own access point named Config-XXXXXX. This is the only way to onboard the older firmware revision (MAS8710BN- hostname units), since once those units join a WiFi network they firewall TCP/80 from the LAN side.

The newer-firmware revision (ESP- hostname units) leaves the same web UI listening on TCP/80 from the LAN side after WiFi join, so for those, see http-api.md — but the captive-portal form is identical.

Captive-portal credentials

Field	Value
SSID	Config-XXXXXX (per-device suffix)
Password	33669999
Gateway / config URL	http://192.168.4.1/
Phone DHCP lease	192.168.4.X

The XXXXXX suffix is not a MAC tail — it's an opaque per-device identifier set at manufacture. Treat it as a unit serial.

Procedure (vendor)

1. Power on the clock; wait for the blue LED on the back to flash.
2. Turn off mobile data on the phone. If the phone keeps mobile data on, it may abandon the captive AP (which has no upstream).
3. Connect the phone to Config-XXXXXX using password 33669999. If prompted "this WiFi has no internet, use anyway?" — choose Use and disable auto-switch.
4. Browse to http://192.168.4.1. Three buttons:
   • Configure WiFi — set home SSID + password (case-sensitive).
   • NTP Server — set the NTP host the clock syncs against.
   • System Setup — DST toggle and integer UTC offset.
5. After SAVE the clock joins the home WiFi; blue LED goes solid when connected.

The home page shows lastSync: YYYY.MM.DD HH:MM:SS — useful for confirming sync after changes.

Procedure (Linux desktop with WiFi NIC)

If you have a USB or built-in WiFi adapter on a workstation, you can drive the captive portal with nmcli + curl instead of using a phone. Useful for batch reconfigurations and for units that are otherwise inaccessible from the LAN.

# Confirm the device is broadcasting its captive AP
nmcli -t -f IN-USE,SSID,SIGNAL,SECURITY device wifi list --rescan yes | grep -i 'config-'

# Connect (the clock acts as a DHCP server, you'll get 192.168.4.X)
nmcli dev wifi connect "Config-XXXXXX" password "33669999" ifname <wifi-iface>

# Read current state
curl -s http://192.168.4.1/ | grep -oE 'lastSync:[^<]+'

# Set NTP server
curl -s "http://192.168.4.1/ntpSave?ntpServer=<your-ntp-server>"

# Set DST + timezone (integer hour offset, -12..+12)
curl -s "http://192.168.4.1/setupSave?summerTime=1&timeZone=2"   # UTC+2 with DST on
curl -s "http://192.168.4.1/setupSave?summerTime=0&timeZone=0"   # UTC, no DST

# Done
nmcli dev disconnect <wifi-iface>

Captive AP intermittency

Empirically the captive AP is not always broadcasting even on a unit that's already on the LAN. Observations on a unit running the older firmware:

• Sometimes visible from a nearby scanner at -60 dBm; sometimes not visible at all.
• Re-appearance does not require power-cycling — wait, re-scan, retry.
• After a successful setupSave, the unit appears to drop the AP for a short period before re-broadcasting.

If the captive AP is genuinely gone and the unit is on the LAN, on older-firmware units there is no way back into configuration without a factory reset (long-press SET ≥ 5 s — see osd-buttons.md).

Dual-band caveat

If the home WiFi uses a single SSID for both 2.4 GHz and 5 GHz, the clock will fail to associate (it is 2.4 GHz only). Vendor recommends splitting the SSIDs and pointing the clock at the 2.4 GHz band specifically.