The outer packaging is red:
No network cable is included in the accessories.
Power output specification is 12V/2A
The casing is not difficult to see.
The gold “WI-FI6 Turbo” logo is printed on the four antennas, but not on the X18G.
The back of the four support blind gigabit network port, there is a physical press to build “Mesh”, two routers press the Mesh button can be automatically grouped, no need to manually set. The instruction manual says that it supports up to 8 units.
After the bottom two screw threads were removed, the shell was opened.
Saw the larger familiar radiator:
The lower two antenna circuit has two amplifiers, yes, the man did not lie to me, there is an external amplifier chip, the size is not small, the power is not low look. Unfortunately, look at the antenna feed line interface solder joints are 2.4G.
Carefully turn over the back of the motherboard and see that there is a flat aluminum plate to help cool the back of the motherboard. The Mercury X18G does not have this aluminum cooling plate on the back of the motherboard.
After going down to the heatsink on the front of the motherboard and seeing a shield, the Mercury X18G does not have this shield. But the other two chips and the heatsink between no heat-conductive silicone pad, that is, the heat of the chip is not intended to be dissipated from the front of the motherboard.
Before lifting the shield to collect the thermal conductive silicone on top, something that can also be used later.
The shield is taken off and there is no silicone pad between the chip and the shield.
The chip model inside the shield is MT7621DAT, an integrated memory MT7621AT, dual-core 880MHz, we are all familiar with the Mercury X188G memory will not be larger than the X18G 128MB?
The TTL is connected and the readout shows that the memory is still 128MB, as shown in the following message.
MT7621 stage1 code done
CPU=500000000 HZ BUS=166666666 HZ
U-Boot 1.1.3 (Jul 24 2020 – 16:09:27)
Board: Ralink APSoC DRAM: 128 MB
relocate_code Pointer at: 87fa0000
set ALL LAN Partition
Config XHCI 40M PLL
flash manufacture id: 1c, device id 70 17
Warning: un-recognized chip ID, please update bootloader!
*** Warning – bad CRC, using default environment
Press reset button to enter recovery mode.
Autobooting in 300 ms
verifying uboot partition…
verifying kernel and romfs partition…
copy the last 0x300000 bytes of the partition to ram …
To boot, bootcmd = go 0xbfc30000.
## jump to 0xBFC30000, PHYS_FLASH_1 0xBFC00000
The other two chips are MT7975DN and MT7905DAN, wireless chips that support WI-FI 6. Support 2x2mimo of 2.4G and 2x2mimo of 5G, because 5G does not support 160MHz bandwidth, so the highest rate of 5G 1201Mbps, so the wireless specifications of this is AX1800.
2.4G and 5G are led from the MT7675DN chip, and 2.4G and 5G have an external independent amplifier chip.
As shown above, the red box is the 5G amplifier chip, model SKY85755-11, which integrates PA, LNA and transceiver switching.
The parameters of SKY85755-11 are as follows.
Integrated 802.11ax , 5 GHz PA, LNA with bypass, and T/R switch
Fully-matched input and output
Integrated logarithmic power detector and directional coupler
Transmit gain: 32 dB
Receive gain: 14 dB
MCS11, +18 dBm, –43 dB EVM
MCS9, +20.5 dBm, –35 dB EVM
MCS7, +22.5 dBm, –30 dB EVM
Small QFN (16-pin, 3 x 3 mm) package (MSL3, 260 °C per JEDEC J-STD-020)
Looking at the chip area, we know that the amount of gallium nitride is not much and the power will not be very high. SKY85755-11 is a little stronger than the QPF4550 used in Redmi AX5 on 802.11ax.
SKY85755-11: MCS11, +18 dBm, -43 dB EVM
QPF4550: +16dBm MCS11 HE80 -43dB Dynamic EVM
Dismantled here I was also curious than no external amplifier Mercury X18G how much stronger, after I simply tested the C point and D point, later said the speed measurement results.
The blue picture above is the 2.4G amplifier chip, the model is QPF4216, look at the size of the chip, gallium nitride dosage is more, the power is estimated not low.
The parameters of the QPF4216 are as follows.
POUT = +22 dBm MCS11 HE40 -43 dB Dynamic EVM
POUT = +24 dBm MCS8/9 VHT40 -35 dB Dynamic EVM
POUT = +25 dBm MCS7 HT20/40 -30 dB Dynamic EVM
POUT = +26 dBm MCS0 HT20 Spectral Mask Compliance
Optimized for +5 V Operation
32 dB Tx Gain
1.8 dB Noise Figure
15.5 dB Rx Gain & 7 dB Bypass Loss
15 dB 5 GHz Rejection on Rx Path
Integrated Quasi-Log Power Detector
From the above data, we can see that it belongs to medium-high power.
The last thing to look at is the flash memory chip, model QH64A-104HIP, with a capacity of 8MB. much larger than the Mercury X18G’s flash memory capacity.
To make a summary of all chip models.
TTL pin definition of Mercury X188G.
When I put the machine back together, I put a piece of heat-conductive silicone pad on the CPU, as follows.
There are two electrolytic capacitors on the right side of the picture above, and only one for the Mercury X18G.
In order to facilitate a comparison with Mercury X18G, I list the speed measurement results with some parameters in the following table. 2.4G I will not put up, because the difference is not obvious and meaningless. It may also be related to the time period of the test, I test 2.4G will usually be in the morning or zero morning. The broadband is 300Mbps, using Xiaomi 4C phone running speedtest app speed test, three consecutive speed test, are written down, only take the downlink data, because the uplink is about the same.
CPU, wireless chip model memory capacity a ghost-like will not say. (The firmware version of X188G is 1.0.1, X18G version is 1.0.3)
From the two positions of the speed measurement, Mercury X188G relative to the X18G without external amplifier to improve is very obvious, the speed measurement in the D point than the Redmi AX5 is better.
In summary, the Mercury X18G only cost 30 yuan more, well worth it, in exchange for the 5G to improve a lot, the power supply specifications have also improved, because there is an external amplifier, making the machine’s chance also increased, so the 12V/2A power supply is necessary, after a few hours of operation, touching the power supply does not feel hot.