MK808 Wifi Antenna fix! VIDEO to help you do it!

[AlbaJoe]

We are working on it hard! Should be soon.

Right now some people in Russia are testing my ug802 system with 2dark's new MK808 kernel. He needs direct feedback for the kernel in native language

Bob

Good to see the yanks and ruskies working together ..... it was worth it Gorbie = Perestroika

[crasdan]

i solved my problem with mk808. new speed

[dbp876]

Has anyone tried an ipad 2 antenna. You can get them on amazon pretty cheap. just curious as to how it will work. I am at work offshore so it os impossible for me to try right now.

[Abducted]

Im curious if anybody has opened up their wireless router yet to check the antennas? After all those are pretty much all made in China too.

[vudude]

Antenna Jack

I decided to add an antenna jack to my device. My goal was to be able to defeat the mod if I wished to.
I decided to use the spring loaded pins (wifi) as intended. I added 2 contact pads insulated with electrical tape from the original antenna. I did 5 tests of speed each the increase in download speed was 27.8%.

[crystawth]

I inadvertently ripped off the pcb trace for the antenna too lol.. Guess I'm stuck with Wired.

Not bad though, because with the cheapy adapter i have i get ~4MB/s which is more than enough for anything i need to do =)

Just so aggravating trying to fix this wifi issue just to ruin it!!!

[andrep182]

I inadvertently ripped off the pcb trace for the antenna too lol.. Guess I'm stuck with Wired.

Not bad though, because with the cheapy adapter i have i get ~4MB/s which is more than enough for anything i need to do =)

Just so aggravating trying to fix this wifi issue just to ruin it!!!

[ATTACH=CONFIG]2271[/ATTACH][ATTACH=CONFIG]2272[/ATTACH][ATTACH=CONFIG]2273[/ATTACH]

Hey buddy don't lose hope! I ripped it big time, but managed to trace back which pin on the chip the wireless node is. Then, I soldered the + side directly to that pin and whalaa! it works again =).

FYI, I have the timing power RK3066 stick so I can't speak for your stick configuration

[crystawth]

well luckily for me, this stick is hooked up to my living room TV which sits literally a foot above my network switch. So using a cable is easier lol.

[Tagon]

...many thanks, Bob!

My MK808 was fine when it was new. But recently it started having SERIOUS wifi issues!
[quick fix by removing shielding GND shortcut]

Many, many thanks, Bob! Your video was essential help for me get an acceptable WiFi level on my new MK808B. As sort of a very little "thank you", I document and post my findings.

Here is the report:

Outside, my stick looks identical to the one in the video, but inside it has at least two modifications.
One is a different (but equally stupid) type of screw-up:

As you can see in the picture, instead of the usual shielded cable, my MK808B has two little spring-loaded connector rods to contact "antenna" and "ground" from the circuit board to the WiFi antenna foil.

In principle, this is actually a very good idea - no shielded cable necessary, and nothing can break off when you open it.

Unfortunately, the antenna foil has a single, unbroken, copper "connect plate", which connects "antenna" and "ground" signals to the same piece of copper, producing a short-circuit.

For an initial "non-destructive" test, I put some duct tape over the GND connection point (where the red arrow points to). Here is how it looks now:

[Note: from my measurements, the 'Duck' tape I have is a good insulator. Sellotape would probably work as well, or just a tiny cut-out piece of strong insulating plastic foil. The insulating tape I have is very soft, and I feared the pointy connector tip of the rod would punch through it.]

This mini-modification improved WiFi reception a lot. Close to the router I now get "full five bars". In the living room (at the TV set) WiFi reception changed from "practically unusable" to "sort of ok".

It is still not brilliant, therefore I want to experiment a little and replace the antenna with something better (maybe just a piece of insulated wire with the correct length for a 2.4GHz antenna), but for the moment it is ok.

As Bob recommends in the video, I also checked the Bluetooth antenna, here is what I found:

From what I can see, it looks as if both Antenna and Ground of the antenna cable connect to the same solder pad (hard to see, because the solder blobs are so big).
Hm... I don't know if this is actually part of an omitted "shielding cage", or a real signal-carrying part.

From what I know, normally this would be another "botch-up" - the shield of the cable should be connected to ground (GND), the inner wire to the antenna signal.

But in this case it is actually beneficial, because the antenna itself looks like this:


[For the non-electronics-savvy reader: The problem with this antenna - the yellow foil bit to the right - is that between the "folded antenna" part (left) and the "ground plane" part (right) the conducting copper foil - which is glued to the plastic casing - seems to have no gap. But without gap, you short-circuit the "antenna" and the "ground plane" signals. Here it probably only works because a second mistake - soldering cable wire and shielding to the same signal - prevents a short-circuit.]

As Bluetooth works well enough, at the moment I see no reason to solder or cut off the "GND" connection. Yet. :-) (The potential of a short-circuit will not let me be at ease until I have made sure it is none.)

Again, many thanks, Bob, for the video (and for the whole forum in general)!
BR-Tagon

[Finless]

Great write up! But so you know both the antenna's are for wifi. There is no BT antenna. Most BT devices don't have an antenna because BT is not expect to be long range.

So if you mod the other patch antenna wifi should improve even more.

Bob

[Tagon]

...the story continues...

Great write up! But so you know both the antenna's are for wifi. There is no BT antenna. Most BT devices don't have an antenna because BT is not expect to be long range.
So if you mod the other patch antenna wifi should improve even more.

Yessir! - thanks again, Bob! Also meanwhile I discovered the other umpteen posts from people who seem to have a similar model to mine. [Now I feel a little bit sheepish... baaaaa! ]

Meanwhile, I looked at the RK903 specs - its antenna output is marked as for WiFi and Bluetooth. Makes sense - both protocols work in the same ISM 2.4GHz band.

For my device, I continued modification:

1. (a) On the antenna foil connected via shielded cable, I cut a gap between 'groundplane" and 'antenna' parts to remove the short-circuit - normally this should not be necessary, because (I also looked up some PCB antenna designs) this is a modified "inverted F" antenna design, and having the connection should not matter for HF. But it only 'does not matter', if the antenna is actually 'matched' to the TX/RX frequency, and I doubt that anybody in the factory did any tuning. This is how it looks now:
   
2. To the antenna foil bit that's connected via spring rods (where I insulated 'ground' with duct tape), I soldered an insulated "loose wire" extension (to the foil, not to the rod), now a flexible red wire sticks out of my MK808B stick. Best would be a multiple-quarter-wavelength wire (so with reduction factor for insulation and everything else it should be multiples of about two centimeters), but without any HF measuring equipment all I could do is to check signal strength via an Android app (WiFi Analyzer), and successively cut the wire by a few millimeters in the hope to catch the reception (and transmission) maximum.

There is not much visible WiFi gain (in the best stick position, 'WiFi Analyzer' display went from -71db to -67db), but data transfer rate and stability definitely seem to have improved a lot: before, HD videos did not play at all, now they play decently well. Android itself reports link speeds up to 65MBit/s.
So this is still not really good, but then WLAN reception in my living room in this corner is bad (my Android mobile has no better reception).

Again, thanks for this forum, it is a great platform for the exchange of ideas, and also for reference, and also many thanks for your work on the Finless ROMs!
(Still have to try one... but first we need to watch that one other dance video.... )

[pe1dnn]

For my device, I continued modification:

1. (a) On the antenna foil connected via shielded cable, I cut a gap between 'groundplane" and 'antenna' parts to remove the short-circuit - normally this should not be necessary, because (I also looked up some PCB antenna designs) this is a modified "inverted F" antenna design, and having the connection should not matter for HF. But it only 'does not matter', if the

Yes, I've seen this before. But what appears to be a short is not a short at all at 2.4 GHz. Basically an antenna is a 1/4 or 3/4 wavelength. Think of a Sinus curve. It starts at 0, is max at 1/4, then 0, then another extreem at 3/4 and so forth. That's also how the impedance is distributed. Calculation is a bit more difficult, the wavelength depends on the frequency and the speed of travel, in copper the speed is it lower than lightspeed (in general about 0.98 in copper if I remember correctly) and folding will have an effect too.

The impedance (resitance to HF) at the base is 0 and at the end it is infinite. The transmitter should see about 75 Ohm load, which means a small distance from the antenna base. At 75 Ohms the transmitter's energy is fully absorbed by the antenna, with a higher or lower impedance the energy is partially (or completely) reflected back into the transmitter. The energy absorbed by the antenna can only go one way, in the air!

When the antenna is tuned correct for transmission it's also tuned for reception (on the same frequency). This means that the energy catched by the antenna is transferred optimal to the receiver's front-end and not reflected back either.

I suspect the F patch antenna is reasonably tuned. However the antenna with the spring connections (if you have it) look susipicious because they are very close to each other and with a broad area of copper giving multiple paths for the current to flow. Maybe it is a 5 GHz antenna, but I do not trust its design for 2.4 GHz. So the only modification I did was isolate the ground pin with ductape.

Furthermore I only checked for shorts in the cable or at the motherboard connection. I disconnected the center wire from the patch antenna and measured with an ohm meter. I did not have a short and connected the center wire again.

The biggest improvent I had was not hardware but software. Using the MK808B with the stock 4.2 ROM, Finless 2.1 and the Hybrid ROM all gave a low speed reading and problems with streaming. However the Finless 1.7 ROM works perfectly and I get much higher speeds. And also with the 1.7 ROM I see access points which I do not see at all with the other ROM's.

So this proofs the antenna's actually work fine and the signal is not the problem. There is enough signal but the driver in ROM has to play nice with the Wifi chip and have enough priority to handle the chip in time. But since the driver source is kept secret there is no way to analyse this further or to get some statistics out of it.

So before trying a lot of weird modifications first try the Finless 1.7 ROM (or 1.7c for the MK808); if it works with that ROM than modification of the antenna is just trying to fix something which is not the root cause and is therefore futile.

[nc123]

Try downgrading to Finless 1.7(c) before you try an antenna mod!

Yes, I've seen this before. But what appears to be a short is not a short at all at 2.4 GHz. Basically an antenna is a 1/4 or 3/4 wavelength. Think of a Sinus curve. It starts at 0, is max at 1/4, then 0, then another extreem at 3/4 and so forth. That's also how the impedance is distributed. Calculation is a bit more difficult, the wavelength depends on the frequency and the speed of travel, in copper the speed is it lower than lightspeed (in general about 0.98 in copper if I remember correctly) and folding will have an effect too.

The impedance (resitance to HF) at the base is 0 and at the end it is infinite. The transmitter should see about 75 Ohm load, which means a small distance from the antenna base. At 75 Ohms the transmitter's energy is fully absorbed by the antenna, with a higher or lower impedance the energy is partially (or completely) reflected back into the transmitter. The energy absorbed by the antenna can only go one way, in the air!

When the antenna is tuned correct for transmission it's also tuned for reception (on the same frequency). This means that the energy catched by the antenna is transferred optimal to the receiver's front-end and not reflected back either.

I suspect the F patch antenna is reasonably tuned. However the antenna with the spring connections (if you have it) look susipicious because they are very close to each other and with a broad area of copper giving multiple paths for the current to flow. Maybe it is a 5 GHz antenna, but I do not trust its design for 2.4 GHz. So the only modification I did was isolate the ground pin with ductape.

Furthermore I only checked for shorts in the cable or at the motherboard connection. I disconnected the center wire from the patch antenna and measured with an ohm meter. I did not have a short and connected the center wire again.

The biggest improvent I had was not hardware but software. Using the MK808B with the stock 4.2 ROM, Finless 2.1 and the Hybrid ROM all gave a low speed reading and problems with streaming. However the Finless 1.7 ROM works perfectly and I get much higher speeds. And also with the 1.7 ROM I see access points which I do not see at all with the other ROM's.

So this proofs the antenna's actually work fine and the signal is not the problem. There is enough signal but the driver in ROM has to play nice with the Wifi chip and have enough priority to handle the chip in time. But since the driver source is kept secret there is no way to analyse this further or to get some statistics out of it.

So before trying a lot of weird modifications first try the Finless 1.7 ROM (or 1.7c for the MK808); if it works with that ROM than modification of the antenna is just trying to fix something which is not the root cause and is therefore futile.

I confirm dude. After going down to 1.7 my signal was fine.

[crystawth]

If I'm not mistaken, 50 ohm cable and antennas are used for WiFi. 75 ohm is used for video services such as cable and satellite.

[pe1dnn]

If I'm not mistaken, 50 ohm cable and antennas are used for WiFi. 75 ohm is used for video services such as cable and satellite.

I used 75 ohm as example because that's an open dipole load. I first wrote 50/75 ohms but then chose to simplify it a bit and settled for one value. It doesn't matter much for the story unless you want to design an antenna... In fact the antenna is adjustable, it just a matter of connecting the center wire at the correct distance from the ground base. I assume they measured the SWR when they designed the antenna and made the connection at the right spot. But this topic is not a HAM radio course...