The SA868 and similar modules from NiceRF have a quite long delay between pressing PTT and the transmitted audio actually coming "on". Some online sources say that this delay is about 700ms which is much to long for many applications. I wanted to build a mumble gateway with these modules (interface between radio and VoIP). There, a long delay at the start of TX results in missing the first words of a transmission. The transceiver IC used in these modules is the RDA1846S which is also used in many Chinese radios. These other radios don't have this delay so I thought there must be a component in the SA868 module that controls the delay time. For sure, I found it! Strictly speaking, the capacitor for the analog reference voltage is the bad guy. This capacitor gets charged every time when the IC switches to TX mode and seems to filter the analog reference for the mic preamp. You only get loud audio output when this cap is fully charged. That's why the audio slowly fades in. In the SA868 modules a 10µF cap is used. The datasheet of the RDA1856S even calls for a 47µF cap. This makes no sense to me because the delay would be even bigger. I changed the cap to 1µF which should reduce the delay by a factor of 10 (70ms). Sure enough, this solved the problem! No words get lost now. See the picture below to find the cap on the PCBA of the radio module.

In addition I changed the mic coupling capacitor to 100nF (left to the cap discussed above). I don't know if it affects the delay but it could be.

Size of reference voltage cap: 0603 (inches), 0402 (inches) also fits
Size of the coupling cap: 0402 (inches)

Changing the output power does not work

While further testing the radio module I found out that the H/L pin (pin number 7) has no effect on the output power. After investigating the schematic of the SA868 I saw that there is a switching transistor inside the module that can pull H/L to low (switch to low power). The base of this transistor is connected to the modules microcontroller and the collector connects to the H/L pin but there is no known command to change the output power by software. Measuring the output of the microcontroller pin revealed that the output power is always switched to "low power" (microcontroller output high). To make the external H/L pin functional, one could remove the switching transistor or permanently connect its base resistor to GND (transistor switched off). Please see the picture below.

Extra tips

• In addition, I would recommend to adding a small amount of heatsink compound to the top of the metal stud near the final stage transistor. This creates a thermal coupling to the shield above.
• Somehow I already managed to destroy one final stage transistor, so always make sure that you have a matched antenna connected or a dummy load.
• Measured "High" output power: 2.5W
• Measured "Low" output power: 1.8W
• Without modification output power is always 1.8W