Using AR9580 in 2.4G

Hi guys.
I have TL-WDR4300 which has 2 radios.
The radio AR9580 is only used for 5 GHZ although it is multiband.
This is also discussed here -> https://github.com/xieyaxiongfly/Atheros_CSI_tool_OpenWRT_src/issues/3

My idea is to use the router 2x 2.4G radios, that means make the AR9580 run at 2.4G. Both radios use ath9k. I doubt that AR9580 can run 2.4 and 5G simultaneously so only 2.4 is OK to me.
I suppose somewhere it is decided that AR9580 would be used at 5G.

In regd.c we have

	switch (band) {

	case NL80211_BAND_2GHZ:

		return reg->regpair->reg_2ghz_ctl;

	case NL80211_BAND_5GHZ:

		return reg->regpair->reg_5ghz_ctl;

	default:

		return NO_CTL;

	}

Quick check on ar9580_1p0_initvals.h I see available instructions for 2.4 and 5G.
Maybe a simple change in "common.c" file somewhere can flip the usage for 5G to 2.4G and will work ?

Maybe @jow knows ?

Thanks :slight_smile:

Maby you dont try, it is attempt to breach local radio regulations, quickly punishable by frying half of routers PCB.

I am aware of the radio regulations and having 2 radios at different 2.4G freq (channels) is totally fine.
It is no different than having 2 routers next to each other at 2.4G with different channels.
Actually the emitted EMP on 2 routers is much greater as they have separate PA and antennas. But since they operate on different channels it is even still totally fine.

Your router has one 2.4-GHz-only radio (ar9344) and one 5-GHz-only radio (ar9580) - which is typical for most routers, the later does not support 2.4 GHz at all. This is a design limitation decided at the manufacturing/ PCB level, it's neither a bug nor changeable.

1 Like

In the github link we have this
"Thank you, I just thought that AR9580 was a dual-band radio transceiver, but it seems like EEPROM has 2.4GHz disabled."

My first thought was that EEPROM/driver is as well editable with openwrt when building the image and that regdb and athk9 is all editable to make it work at 2.4G

A wireless card consists of multiple components:

  • MAC
  • PHY
  • rf amplifiers
  • rf bandfilters
  • and all the other small analogue components in the signal path
  • the physical antennas connected are single-band as well
    (in this case it's more about the multiplexer, than the antenna itself, as the antennas on this device are rather basic, but that are implementation details)
  • the wifi calibration data needs to have data points for both bands
    (it doesn't for the AR9580 radio on the tl-wdr4300)

For a typical router, you have one dedicated radio per band - it makes no sense for the vendor who designs the PCB to account for dual-band functionality (more components, more expensive, harder PCB routing - and even detrimental to operations on the desired band, as dual-band functionality enforces a compromise you don't have to do if you optimize for a single band).

Yes, AR9580 as a chipset (MAC side) is a dual-band chipset. The actual card on your router however is explicitly not dual-band enabled, at least the later five (and probably the PHY) of the list above are 5-GHz-only on the tl-wdr4300 (and the same goes for most other concurrent dual-band routers). It's explicitly designed to be single-band/ 5-GHz-only. Even if you would force-enable 2.4 GHz operations (which you can't on this device, because of the calibration data, but could on some other routers), you would end up with two problems - the signal would be weak and and mis-tuned, with a very high risk of physically damaging the radio (SWR (standing wave)).

This device simply can't do it, it's not designed that way. Most other devices can't do it either, for the same reasons, there are only very few devices with two 2.4 GHz radios (two 5 GHz radios, one low-band (ch36-64), one high-band (>=ch100) is more common. Apart from the envisioned use case, there's also another reasons - antennas for different radios in the same band do cause interference to each other, different channels aren't enough for a clear separation (there are only 80 MHz in total on 2.4 GHz - while on 5 GHz you are much farther spread and have a sizable gap between ch64 and ch100). And even on devices with two 5 GHz radios the individual radios are locked into low-band and high-band and can't be used for the other (sub-)band, as the rf components and bandfilters would be mistuned.

2 Likes

I may be wrong but there is calibration data inside ar9580_1p0_initvals.h for 5G and 2.4G.

static const u32 ar9580_1p0_low_ob_db_tx_gain_table[][5] = {

	/* Addr      5G_HT20     5G_HT40     2G_HT40     2G_HT20   */

	{0x0000a2dc, 0x0380c7fc, 0x0380c7fc, 0x03aaa352, 0x03aaa352},

Use case could be channel 1 and channel 13 which is enough gap. Can have use case of 20Mhz + 40 Mhz at least.
Same antennas ? No problem, actually on this device they are detachable to whatever you want.

"The signal would be weak and and mis-tuned, with a very high risk of physically damaging the radio."
I guess you are referring to something practical that you did in the past? Otherwise if the AR9580 is supporting 2.4G physically I see no reasonable problem of damage or low performance.

"as dual-band functionality enforces a compromise you don't have to do if you optimize for a single band"
Well as I said in my post..
"I doubt that AR9580 can run 2.4 and 5G simultaneously so only 2.4 is OK to me.".
So my idea is to run AR9580 only on 2.4G - single band for that particular radio.

Some of your explanations are good and I appreciate that but still not fully convinced.
I believe this wasn't implemented on tp-link software side as nobody would expect usual customer to use 2x 2.4G radios at once.

You are wrong.
EEPROM at 0x5000 explicitly sets "Allow 2.4GHz : 0" and has only 5 GHz data points.

Incorrect, for the reasons quoted above.

Did you read the reasons given above, the whole rf path is physically not capable of 2.4 GHz, nor tuned for it. I hinted you at SWR problems, but that isn't the only potential for damage by forcing the power amps and bandfilters out of spec.


There is a whole lot of 'chicken food' around the AR9580 chip on the top right, which is 5-Ghz-only and not 2.4 GHz capable (among others, particularly the Skyworks power amplifiers and the diplexer, but also a lot of passive rf components).

You don't need to doubt it, that is completely out of the question (but so is 2.4 GHz operations on this ar9580 radio).

You have a twelve year old 802.11n/ wifi device that sells for ~5-8 EUR on the second hand market these days, you got what you paid for. TP-Link didn't advertize 2*2.4 GHz operations and the hardware can't do it. Even if you could force it (and you can't on this hardware), doing so would be out of spec and void your regulatory certification.

If you find me this puppy for 8 EUR I will buy 8 of these :smiley:
Don't ask me why.
I've bought it last year for 35 EUR.

Band pass filter isn't damaged that easily as it is doing its job to bypass only the requested frequencies.
AFAIK it should be built inside the radio itself. Obviously for AR9580 it should bypass 2.4G and 5G.

It really comes to the PA/LNA if it is single for both radios or there's PA/LNA for each radio and I believe they would put explicitly 5G PA after the radio, making the radio useless at 2.4G.
But if it is a single PA for both radios they've tuned it to amplify both 2.4 and 5G.
And yeah the "chicken food" is a lot and as this isn't ever tested I guess it isn't worth it.

There are five power amplifiers, one per chain (2*2.4 GHz, 3*5GHz) and two diplexers involved (and more).

TP-Link (and most other router manufacturers) don't do this out of spite or malice, they do it to save money and improve signal quality for the desired band, it's a basic hardware design decision.

Oh, oh I never saw I posted 4300.. it is 4900 the N900 one.
TL-WDR4900 (N900)
This is the pumped version of WDR4300 but I guess nothing changes in the use case except the price.

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.