OpenWrt Forum Archive

Topic: TP-Link 703n external antenna modification

The content of this topic has been archived between 12 Apr 2018 and 20 Apr 2018. There are no obvious gaps in this topic, but there may still be some posts missing at the end.

The Chinese market TP-Link703n 150n USB router is a robust, small, low power and reliable router with many applications.

It is well loved around the world. It is truly diminutive in size and power needs.

It may be flashed from its TP-LINK Chinese firmware to TP-Links 3020 English firmware or DD-WRT or OpenWRT or OpenWRT based firmwares like Gargoyle and ROOter etc.

It is powered by 5 volts and supports a single 100Mb LAN/WAN port, a USB 2.0 port and a micro USB port for power connection.

The 703n has an internal inverted F ¼ wave length PCB etched antenna. There is no external antenna connection.

This tutorial documents the necessary modifications to add a standard RP-SMA antenna socket to the 703n.

This modification increases it's application and versatility considerably

The inspiration for this modification largely comes from these two blog articles:

http://blagg.tadkom.net/2012/09/01/wr-7 … a-mod-diy/

    See comments by luno

http://blagg.tadkom.net/2012/09/15/bett … tenna-mod/

    See comments by Diarmaid Ó Cualain

Tutorial (pdf) https://app.box.com/s/cispknq8b9zgog8k5vxn

The

PunBB bbcode test

finished product

Caution: It is very tiny inside and prior soldering skill on micro circuits is advisable; however these little things are tuff and can cope with a fair amount of wilful abuse.

Note: WIKI updated with this mod.

http://wiki.openwrt.org/toh/tp-link/tl- … dware.mods

Why don't you just remove the big "J1" jumper on the top side, solder the coax center feed to the C28 side, and the coax braid to the nearby large round pad?

Then if you want to revert to the PCB antenna, you just have to unsolder your pigtail cable and put a large solder bridge over both J1 pads...

Because this way ... it works. smile

It is very important to maintain the Pi impedance matching circuit next to the strip line antenna.

At the point of this connection (the antenna) the impedance is 50 ohms (RF impedance), which matches an external 1/2 wave or 5/8 wave dipole perfectly and full energy is transferred.

Pretty easy to undo it; but why would you if it works?

Cheers

Wahroonga Farm wrote:

Because this way ... it works. smile

It is very important to maintain the Pi impedance matching circuit next to the strip line antenna.

At the point of this connection (the antenna) the impedance is 50 ohms (RF impedance), which matches an external 1/2 wave or 5/8 wave dipole perfectly and full energy is transferred.

I thought that the matching network was to match the PIFA antenna to the 50 ohm source, not matching the source to a (supposedly) 50 ohm antenna?

http://www.ti.com/lit/an/snoa519b/snoa519b.pdf#page=15

Do you have RF experience on this subject?

Squonk wrote:

Do you have RF experience on this subject?

Enough to know when a transmitter is exiting an antenna. smile

That is the object of doing Tx testing using inSSIDer on a lappy and Android phone and setting up a wireless Lan.

An antenna that receives only, is less critical to match than one that is transmitting. A receive only mod is relatively easy to create, by mistake.

It's transmitting and transmitting well. It transmits stronger than an Asus RT-N12 300n that sits next to it in any signal survey I can throw at it.

If it's transmitting, sure as hell it's receiving, which is also confirmed by the WDS link set-up where both ends of the link report good and equivalent receive levels and where it runs at 80Mbps on 150n wireless over 10m through stud walls.

I can't make any guarantees as that is the limit of my testing capability; but doing the mod in this manner does appear to deliver a working device by all the tests I can throw at it.

If you can think of some other way to confirm it's performance I'm happy to give it a go.

Cheers

John k

PS Have you external antenna modified the 703n?

Wahroonga Farm wrote:

PS Have you external antenna modified the 703n?

No, but the famous J1 jumper strongly looks like a way to calibrate the PA output over the full Wifi band during production...

The Wifi band is too wide to be able to guarantee that you have a flat response over the whole frequency range just because of the component tolerances.

Thus, a calibration step is required (well, except for some low quality Chinese routers, which is certainly not the case with TP-Link), where a test software is run on each individual router on the production line in order to calibrate the PA register setting required to get a flat response over the whole band.

The result of this test is stored into the ART ("Atheros Radio Test") partition and used by the WiFi driver, see a process description here.

Of course, this requires an RF test point to get the output signal, and my guess is that the J1 jumper and nearby GND pad just serve this purpose, the jumper being not mounted during the test, then soldered afterwards, thus is its rather large 0603 (!!!) size. This is a very common practice for RF production tests.

And I doubt that they would have put anything but a 50 ohms impedance there, but this is just a guess!

Squonk wrote:

No, but the famous J1 jumper strongly looks like a way to calibrate the PA output over the full Wifi band during production...

I wholeheartedly agree that is why J1 is there and like most factory test points, they don't serve modifiers very well. smile

Squonk wrote:

And I doubt that they would have put anything but a 50 ohms impedance there, but this is just a guess!

And I guess I guess differently; because the inverted F antenna, like a J-pole or any similar antennas are tapped to be fed by 50 ohm or there-about. This is the way us hams see it. (pdf)

I agree that it may not be exactly 50 ohm; but it must be a relatively low impedance by design. 25 ohm? 75ohm? It really doesn't matter. If you do the maths, the energy gets transferred.

The other interesting observation I have recently made, is the power drawn under WIFI load.

This particular 703n is running a USB 3/4G Sierra 320U modem. It is configured as an AP(WDS) on Channel 11 at 40Mhz. I run this device (as I do for all 703ns) on a 5v @ 2A supply. The USB port is good for 1.5A.

Idling it draws 2.4watts.

Whilst dragging a large file over the wireless LAN link at 80Mb/s, the external antenna 703n power consumption jumps to 5.1 watts.

There is nothing in the 703n that could possibly disperse 3 watts of energy, apart from ... the antenna.

I'm happy to be proven wrong, or for someone to point out a better way; but I am doing as much due diligence on this modification as I can; so as not to spray bad information around the place. smile

To re-cap, here is a test of my antenna modification referenced to a standard MR3420 using inSSIDer on a laptop at 10m distance. Both routers are co-located

PunBB bbcode test

-----

Here is a 703n external antenna modification using J1.

PunBB bbcode test

PunBB bbcode test

Here is a test of the J1 modification referenced to a standard MR3420 using inSSIDer on a laptop at 2m distance. Both routers are co-located.


PunBB bbcode test

It is transmitting very poorly.

ref 1 - http://forums.whirlpool.net.au/forum-re … p=54#r1065
ref 2 - http://forums.whirlpool.net.au/forum-re … amp;#r1038

I am not sure to understand fully the results of your tests, but it looks like the J1 modification is performing very poorly indeed.

However, it looks like you don't get a better result compared to a stock unit using the external antenna, do you? They look they are both at -46/-47 dBm, right?

What would be interesting too is to make a test without cutting the PCB antenna and soldering the pigtail to J1 (still mounted), then cutting the PCB antenna and see if we get worse/similar/better results.

I suspect that the PCB antenna matching is not ideal for the pigtail antenna, and that we may have to find to correct values to get the best range.

I may be able to get access to a VNA (Vector Network Analyzer) some time, so I can get the Smith chart plotted and experiment to get the best matching possible.

Squonk wrote:

However, it looks like you don't get a better result compared to a stock unit using the external antenna, do you? They look they are both at -46/-47 dBm, right?

If you have a look at the pdf write up, it compares a stock internal 703n to a modified external 703n.

The modified external 703n performs better ie from -64dBm up to -48dBm

Squonk wrote:

What would be interesting too is to make a test without cutting the PCB antenna and soldering the pigtail to J1 (still mounted), then cutting the PCB antenna and see if we get worse/similar/better results.

That has been suggested elsewhere??

I'll leave that for others to test. smile

Squonk wrote:

I suspect that the PCB antenna matching is not ideal for the pigtail antenna, and that we may have to find to correct values to get the best range.

I may be able to get access to a VNA (Vector Network Analyzer) some time, so I can get the Smith chart plotted and experiment to get the best matching possible.

That would be excellent.

There's always room for improvement.

I have a TL-WR702N.

Do you think the antenna mod is possible on this device?

Which would be the best method finally to get stronger signal?

Thank you.

omiez wrote:

I have a TL-WR702N.

Do you think the antenna mod is possible on this device?

It's almost identical to the 703n, so it is the same mod.

Have a look at the board snaps of the 702n here.

http://wiki.openwrt.org/toh/tp-link/tl-wr702n

PunBB bbcode test

Wahroonga Farm wrote:

Because this way ... it works. smile

It is very important to maintain the Pi impedance matching circuit next to the strip line antenna.

At the point of this connection (the antenna) the impedance is 50 ohms (RF impedance), which matches an external 1/2 wave or 5/8 wave dipole perfectly and full energy is transferred.

It was fantastic to find your detailed PDF about this proper mod for an external antenna on the WR703N.  I had been toying around with an external antenna mod for the similar device TL-MR3020, which also took the approach of de-popping a 0ohm resistor and soldering the pigtail there.  However, I found this mod may have rendered the device more susceptible to ESD damage, as two MR3020 so modified appeared to lose much of their TX power after lightning storms passed over.  I am wondering if attaching the pigtail further upstream on the antenna path was also bypassing some filter caps needed for proper DC block and/or ESD protection.

At any rate, your mod does not bypass the filter circuits (and possible ESD protection elements), so thank you again for posting this!  I hope to try it out soon.

westbywest wrote:

At any rate, your mod does not bypass the filter circuits (and possible ESD protection elements), so thank you again for posting this!  I hope to try it out soon.

There is no ESD protection element, only filter/impedance matching: the reason is that it is not required for an internal antenna.

But if you modify the device to get an external pigtail antenna, you need to add a proper RF ESD protection diode between the RF signal and ground, as close as possible to the antenna base and with a good ground connection.

ESD is not only to ptrotect from ligthning storms, but also just when your hand touches (or even approaches) the antenna.

Squonk wrote:

There is no ESD protection element, only filter/impedance matching:

Thanks Squonk,

Do you have suggestions for suitable ESD protection?

Cheers

Wahroonga Farm wrote:
Squonk wrote:

There is no ESD protection element, only filter/impedance matching:

Thanks Squonk,

Do you have suggestions for suitable ESD protection?

Cheers

Maybe some dedicated RF ESD chips like these:
http://www.avx.com/docs/catalogs/aguard.pdf

Thanks

Or perhaps the same question another way. smile

Are consumer indoor routers, with external antennas, typically provided with antenna factory ESD protection?

Clearly those intended for outdoor use have ESD protection and are marketed as such.

Typically, the cheap indoor routers do have footprints for ESD protection on all their externally visible ports populated for certifications (including USB, Ethernet and RF pigtail antennas), but they are seldom mounted on stock devices;)

Of course, this is different for outdoor routers which are typically more exposed to ESD, but you pay for it.

However, ESD protection is not only against lighning storms, you may have more casual ESD hazards  up to 8 kV just by human body contact on or near a metallic part that is accessible outside the enclosure.

Squonk wrote:

Typically, the cheap indoor routers do have footprints for ESD protection on all their externally visible ports populated for certifications (including USB, Ethernet and RF pigtail antennas), but they are seldom mounted on stock devices;)

I should clarify that my experience with ESD apparently damaging units was actually for a couple MR3020's, not WR703Ns.  These units were modded by depopping a 0ohm resistor rather far upstream on the antenna path (per these instructions), possibly ahead of most of the filter / matching circuits.  That is, the MR3020 modification had the pigtail attached at a different location in the schematic than what Wahroonga has illustrated for the WR703N above.

Anyway, I do agree this WR703N mod does not include anything that would constitute dedicated ESD protection.  My suspicion is that the MR3020 mod I tried was possibly lacking even filter caps that could provide some DC block.  Besides, I'm not seeking protection substantially more robust than what's already found on cheap indoor routers with external antennas (e.g. the MR3220).  Indeed, the old Open Mesh OM1P 802.11g access point I have sitting nearby, which does have an RP-SMA jack, likewise appears to have no obvious ESD protection, and those units survived the aforementioned lightning storms OK.

I now have a guinea pig WR703N modded per Wahroonga instructions, sitting at the same (indoor) location where a MR3020 previously had failed.  I can report back after the next electrical storm. wink

westbywest wrote:

... I now have a guinea pig WR703N modded per Wahroonga instructions, sitting at the same (indoor) location where a MR3020 previously had failed.  I can report back after the next electrical storm. wink

Good stuff.

We await any reports.

PS Mine are still going strong. I've done a couple.

Cheers

PS are you happy with the antenna perfomance?

I have been looking for more information about how to mod this router as i intent to use a few for some solar powered projects but so far i get a mixed feeling about how well these mods work.
A friend tried one of these "popular" mods and did not find, it was worthy.
Whats the update with your experiences and what is the current recommendation ? Has there been any progress, changes or improvements ?

Since the alternate modifiers have not posted WIFI Tx/ Rx levels or throughput, it's hard to know how they perform?

A friend did an alternate mod and it's performance was very poor.

This one works and is still going strong. I've modded several.

PunBB bbcode test

Note: Wahroonga Farm is a 300n Asus RT N12

PunBB bbcode test

Hello friends

Which is the right way/place to solder the antenna cable?

I am on my cellphone so can somebody link me the Pdf or Link with the right mod please.