Xiegu G90 - backlight for Front panel keypad

There are not many features missing from the Xiegu G90... At least, when is about hardware, in the firmware, there is another story.

When I was replacing the STM32F103 in the front panel, I was asking myself if I can add backlight to the front panel keypad. At that moment I was focused on that repair job and I put the ideea on other shelf.

These days I will get involved into a big project and I will have to let down this kind of PMR (Pimp My Radio)  projects so I thought this is the right time to gave it a chance.

I have around some SMD LEDs and some ideea so, I get to the bench and opened, once more, the front panel assembly.

I did some tests about how the light is diffused in the rubber keypad and I investigated the places where I can put some LEDs.

The space between the PCB and the keypad is none so we need some alteration in the keypad rubber to accomodate the LEDs and the wires.

First try was with the tip of the soldering iron but, to my surprise, nothing had happened.

Then, I tried to use a 3.5mm drill to remove the rubber in an "organised" manner... nothing! The rubber act like rubber and get back into the hole.

Short story: I couldn't found the proper way to make 3mm holes with low risk of destruction.

The problem is that I must embedd not only LEDs but also the wires that goes to them. About the wiring, I had in plan to use 26 Gauge wrapping wire (around 0.5mm) so a proper way to make them too has to be found.

Here is the optimal position for a bunch of wide angle SMD LEDs to diffuse the light into the keypad.

NOTE 1: the botom line of keypad is already illuminated by the backlight from the LCD. Not very bright but enough to be seen into the dark.

NOTE 2: the colour of the LEDs has to be different than the two LEDs for Function and CW decoder because the light will be conducted in those too and will cause them to appear lit.

Custom ATMEGA328p board

 There is a time when an Arduino is too big and has useless features.

An ATMEGA328p on a breadboard, using internal clock at 8 MHz with a common Arduino Bootloader has to be made.

ATMEGA328 datasheet.

Some references here, notes to myself but also usefull (maybe) for others.

Pin map:

The Reset circuit should be a resistor from RST pin to Vcc and a 0.1uF capacitor from RST to whatever....

Xiegu G-90 Audio Gain for quiet nights

 I use to listen to CW to train the reception but not always I am not so willing to use the headphones. Sometime, late at night when I am working at some projects the sound can be heard in the bedrooms and I already had some complaints about this.

On Xiegu G90, the minimum audio level is still too loud for this and some faint signals need to keep the RF Gain at higher settings so something about the audio has to be done.

The audio volume is digitally defined into the audio codec in the Main unit and even I sent the issue to Xiegu, nothing has been done yet (FW 1.77).

Therefore, I felt I must do it myself...

The audio amplifier is made around a wellknown LM386 in SOIC package, not the best for thermal dissipation, if you ask me...

LM386 is a very versatile audio power amplifier designed to be used in portable applications. 

It feature a wide supply voltage range, between 4V and 12V (some versions from 5V to 18V), and can deliver 125 mW on 8 Ohm speaker with 0.2% distorsion.

Because it is designed for minimum external parts, it has an internal fixed value resistor (1.35 kOhm) for an amplification factor of 20.

Internal schematic of LM386:

This internal resistor can be bypassed with an external network to achieve higher gain (up to 200 or 46db) and Xiegu put there a 10uF capacitor and a 100 Ohm in series.

Sidenote: I found a nice online gain calculator for LM386.

The values used by Xiegu in G90:

According to the formula, the Gain is 124.5.

After tweaking with some Rext values, I replaced the original 100 Ohm with a 300 Ohm for a Gain of 76.

Now, the audio is pretty nice, the first step of volume settings is just where it should be for a quiet night.

Audio circuit, PCB layout in Xiegu G90:

Yeah, in the picture is the original resistor. I didn't take a picture after the mod :-( 

I played a lot there with various values for resistor and capacitor then i remind to take a picture. Well... please forgive me :-)

Trick to protect the RJ45 connector tab

 Many transceivers have RJ45 connectors on the microphones and many of them broke at transport or even in the storage box when searching for stuff.

I do have a well equipped laboratory with all kind of stuff, including the proper tools to repair a broken RJ45 but... I do have a problem with colours! Yeah, I barely can distinguish between red and green, at least the red and the green used in wiring. And in resistor code. Traffic light pose no problem, I can see that green and that red so, no worry with me on the streets at the same time with you, HI.

So, I thought myself it's better to prevent than to cure (this came with the age, probably) so I searched for a quick, cheap and reliable solution to protect that little shit tab.

An ideea popped out from the void and here I am, making protective caps from some piece of heatshrink tube.

Find a suitable HS tube, put the connector inside it; the HS tube has to overpass the tab with around 3mm not more!

Then gently apply heat observing the molding. When it wrap the connector stop heating and cut the useless part outside the connector.

After that, gently move the HS tube to extract the connector.

And this is all... 

I suggest to use a dummy RJ45 or some kind of ethernet cable as a jig. 

Do not heat the HS tube with the real RJ45 inside, you may mess it up and cry a lot.

HiRes pictures from the XIEGU G90 transceiver

 Here are some pictures of the PCBs in the Xiegu G90 transceiver.

The Main unit.

        Control board.

RF Front End Bandpass Filters 

As far as I can identify, there are 10 bandpass filters switched with PIN diodes. Far away from the schematic!

 

 

 

PA RF Board

Reverse voltage protection detail

NTC (Negative Temperature Coefficient resistor) problem

There is a serious gap between the NTC which compensate the bias for the temperature increase and the PA RF transistor case. This is one potential cause for premature failure in the final stage especially in TX intensive modes like FT8.

The solution is to gently bent the pins of the NTC to make physicall contact between them and apply a small blob of thermal silicone GLUE, not usual paste.

This will prevent a future gap due to vibrations or shocks and will increase the mutual temperature transfer thus an adequate bias voltage. 

Before:

After:

Schematic (work in progress)

-I removed the schematic as I found error in the AIC3204 area and I must redraw it. I will find a way to have it updated.