-
Notifications
You must be signed in to change notification settings - Fork 191
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
(info) Very low TX power on Alfa AWUS036ACHM #34
Comments
Hi @silanosa I have noticed the same thing but haven't said much about it because that is almost certainly a bug in what is being reported and not in the actually txpower. I say that because the performance in the 2.4 GHz band is very good. Would you interested in helping me look around in the source to see if we can figure out what the problem is. If we can find it, we can submit a PR. Regards |
Yeah, I would be happy to. Could you point me in the right direction, so I have a lead on where to start looking? :) |
I was thinking maybe if we compile code from this link could fix tx-power issue. |
While that Mt76 repo is not exactly the same as the code as is in the kernel, it is close as it serves as a downstream for the code in the kernel. Patches seem to go upstream and come downstream between the two. Check this out: That is quite a discussion. I had not thought much about this because the performance of the ACHM is so good. I actually use two of the drivers in this repo with openWRT. They are available to install on OpenWRT as packages so you can take usb adapters that have mt761xu chipsets and plug them into your router running OpenWRT and you get extra radios. It is cool. |
Nick, as discussed, I'm also having the same issue with my new ACHM. I haven't had much of a chance to run some tests on real-world transmission distance but I intend to in the next day or so. I'm no longer having DC power issues thanks to the powered hub I picked up (Jeaxin 4 Port 3.0 hub, GL3510 chipset). Some troubleshooting I've done is changing country codes between 00-WORLD, US, DE, JP, BO etc. with no change. I tried manually setting the tx power via iw dev wlanX set txpower fixed 2000 with no change. I verified the Country Code that's set in Luci matches what's displayed in uci show wireless. As would be expected, however many adapters (MediaTek or otherwise) are plugged into the raspberry pi plays no part in the displayed tx power of another adapter (I didn't think it would although this was a good time to confirm). I'm only getting 2mW of power in managed mode and 15mW max tx power in AP mode. Abysmal if the numbers are accurate. How do you suggest we proceed? Here is my uci show wireless output:
Here is my iwinfo output:
Here is my iw dev info output
Here is my iw phy3 info output
Here is my hwinfo output:
|
I generally don't pay much attention to the txpower numbers. I probably have adapters where I have never checked the numbers. I have used a lot of adapters and cards over the years. Maybe experience has taught me to check results, not numbers. I'm not an EE but I have read a lot of things written by EE's. Fully understanding radio waves and electricity is not easy. It is far more complicated than a simple number. Here is an interesting article: https://metis.fi/en/2017/10/txpower/ I have other links if you are interested. My recommendation of the Alfa ACHM is not based on a txpower number. It is based on my use of the adapter and my testing. The results of testing show that it has impressive capabilities. The only adapter that comes close to it as far as range is concerned is the Alfa ACH. I have adapters that are made by companies besides Alfa. I just happen to be impressed with Alfa's products. Alfa has a history of producing adapters that are durable, have long range and last a long time. I look forward to the results of your tests. Regards |
I agree. How far away do you think I should be able to go and still achieve a good signal?
Thank you for the article, I appreciate it. Anymore I can learn is always good. If this were a home installation I can understand trying to be neighborly by not bathing my neighbors in rf energy although that is not the case for this application. I want to melt Hershey's bars.
I have absolutely no doubt. Its speed is awesome in N and it keeps up with my other AC adapters too (at least for my 100Mb conn.) but transmission distance seems to be a significant issue in OpenWrt? Is my experience anecdotal and the discrepancy in txpower displayed in OpenWrt isn't actually impacting the real world transmission distance? Maybe. I still don't know although it seems like it's not performing as some had expected. So, bug or not? If it's a bug, where is it? Any suggestions for additional testing so we can rule out variables? I think it's helpful to have multiple people complete similar tests to rule out any subjectivity.
100% this. I've owned an AWUS036NH for years and years (ran it with Atheros drivers) and it's such a beast that when I started building this Frankenstein Pi I had to pick up another Alfa adapter and ended up with the 2 ACMs and the ACHM on your recommendation. All is working beautifully and I credit Alfa for being exactly as you say. Both of these MediaTek chipsets are awesome, along with the Ralink 3070 in the NH. |
As noted in the article, it depends on the client. If you are using the ACHM as the AP and a phone to test, you are really mostly testing the range of the phone and not the ACHM. To get a good idea of what is performing better, you need multiple clients and at least one additional AP in the same spot as the ACHM. I have tested the ACHM against other adapters and a good WiFi router and it beat all of them in both bands. The actually distances varied depending on the client.
Tell me what you are doing in OpenWRT and I will try to duplicate it.
The more, the better. And it we can find and fix bugs, that is good also. The following issue has some interesting data in that we were posting test data that included data from an ACHM: If you read that issue, you will see that the ACHM and ACH performed best is what is basically a range test in monitor mode. |
I'll test it further today with a Windows laptop & the TP-Link 8812BU adapter I have.
Ok. Not a problem, that's the current configuration I have. Asus RT-ACRH13 is sitting about 6 feet from where the OpenWRT router is. The AWUS036ACHM setup I have is also capable of running off battery, it may be helpful in testing as location isn't much of a problem anymore. My current testing is being completed while the router(s) are indoors but an outdoor test with no obstructions (windows/front doors) and less interference is possible. I do happen to be in a highly rf congested neighborhood.
On that note I saw that you recently helped someone reconfigure their /etc/config/wireless in order to achieve better speeds with the ACHM although I don't understand all that you did (VHT settings, etc) and I can't find the post anymore. Would you mind giving me that info so I can use it when not testing other things? Regarding the thread you linked, it looks like the testing was done in something other than OpenWrt where we're seeing the (perceived?) restriction? Another thread that you linked went further in depth as to how the call was being made by OpenWrt to the adapter's firmware defined frequency settings and I though that was an interesting direction, although I could be ignorant of other variables. There's so much good stuff in that thread though, I'll go through it further this evening, thanks for linking it. I very much like how the findings/data is presented. |
Generally speaking, congestion should not hurt your range too much wifi is designed to share. What is will do is slow how fast data can be transferred. You might see if up to 200 Mb/s with iperf3 with the ACHM in 5 GHZ, 80 MHz channel width mode if there is no congestion. On the other hand, with congestion, your wifi has to share with others on the same channel so you will see slower transfer rates. I can't tell you what you will see because it depends on the level of congestion.
hostapd was never designed as a normal user application. Normal users are generally lost at a terminal. hostapd was designed to do what it does... setup an AP. It is very widely used in wifi routers and in many situations. Since it is not and end-user application, we need to take our time and learn it. I will be glad to help you tune your AP. It would be best if you post your hostapd.conf and then let me reply with a optimized version. In fact, so that others can see it better, would you mind starting a new issue titled something like "Need help optimizing hostapd.conf for Alfa ACHM".
Testing was done with a Dell Vostro desktop computer running Linux Mint 20.2. I'm hoping to expand on that testing as time allows as that specific type of testing is very telling about the capability of various adapters. Some things to know: The MT76 driver maintained at OpenWRT is not the same driver as the one in Linux kernel MT76. The one at OpenWRT is a downstream of the Linux kernel MT76. They are very close to be the same though. Patches flow back and forth. OpenWRT have very demanding requirements and it cannot allow bloat as the devs at OpenWRT are dealing with hardware that has extremely limited ram and storage capability as well as slow processors so they have to cut out all the unneeded stuff. That does not apply much to MT76 as the MT76 drivers are lean to start with. Is it possible we see a bug in one driver and not the other? Certainly. It is also possible to report mt76 bugs at the OpenWRT repo and if that is taking too long, then bugs can be reported in accordance with Linux-Wireless guidelines. Regards |
Ok. Well, I'm certainly experiencing some throughput issues due to the congestion (~60-90Mbps). I see 24 access points in Windows and upwards of 3 dozen if I use nirsoft WifiInfoView. That's with either the ACMs or the ACHM as AP, doesn't matter. Interestingly the ACHM has no problem picking up a whole bunch of access points when scanning but when selecting tx power only allows 2mW - I'm under the impression tx power and rx power are symmetrical thus if it's capable of scanning many access points the tx power can't be as low as it displays in OpenWrt.
Thanks much! Below is my hostapd-phy3.conf for the ACHM adapter in 2.4GHZ & 5GHZ configurations.
Thank you for going into further detail how this all works and how to get this rectified if it's determined it's actually an issue. |
Sorry, @silanosa I mucked up and hijacked your thread. I'll start the "Need help optimizing hostapd.conf for Alfa ACHM" as requested. |
Ill tell you why its only 2dBm its because that adapter ACHM has internal Power amp where the TP link one dosent. 2dBm powered thru an Amp gives 30dBi thats why you see amazing coverage at only 2dBm |
Which tp-link adapter has |
TP-Link Archer T2UH? |
@usama7628674 check deviwiki for all devices or source file for USB vid pid |
I don't know off the top of my head but I can look around if you want. It will hurt because I normally don't have anything to do with TP-Link due to their abysmal Linux support. |
I was skeptical at first but having had the opportunity to use this adapter over the past few weeks has convinced me of its capabilities. @morrownr said to pay attention to real world usage instead of just numbers and so far that's panned out for me with the ACHM. |
After I initially purchased my ACHM, I just used it for general purpose things and did not pay attention to things like txpower. It just worked. Then about a year ago someone had ask if I could do a performance test that included a few adapters so that we could see how they compared. I happened to include the mild-mannered ACHM in the test. The results opened my eyes. I repeated the tests several times. As @pauly617 points out, you need a good amp to get the results you get from the ACHM but it also takes a good, high quality antenna because it takes both to extend range. If I had to guess why the twpower reading is low is that an error in firmware may be causing us to see only txpower as it relates to the main chipset and is not taking into account what the amp is doing. |
amazon.com/TP-Link-Archer-T2UH-Wireless-Adapter/dp/B00P115WMY Ouch. There seem to be versions with Realtek and Mediatek chipsets so good luck getting the right one. TP-LINK Archer T2U http://en.techinfodepot.shoutwiki.com/wiki/TP-LINK_Archer_T2U There are several versions of this model. Some have Realtek chipsets and some have Mediatek chipsets. I'm not sure if this is a good product even if you can figure out how to get the version with the chipset you want. TP-LINK Archer T2UHP This is probably the one I would get if bought TP-Link wifi products, which I don't. amazon.com/TP-Link-Wireless-Adapter-Archer-T2UHP/dp/B072FK6V9M It is end of life according to TP-Link and it is not clear how many versions were released so good luck with that. Overall: I realize people in different parts of the world do not have access to all of the products that are available. As you can see from my remarks about TP-Link in the main README, I try not to buy TP-Link products and one of the big reasons is that it is VERY difficult to be sure of what you are getting. TP-Link has a long history of changing the chipsets while keeping the same model number and packaging. If I were to include TP-Link products in the adapter links here, I would spent most of my time sorting out the problems people would have. I don't have time to deal with that so TP-Link is blacklisted. |
What about this one for a mt7610u chipset? I have no knowledge of its quality etc, just found it. Edit: lol. nevermind, just tried checking out and they want $22 shipping and handling |
Ouch! I've seen that adapter on several sites. I don't own one but I have not run across complaints. I'm hoping that ALFA bring a new 7610u adapter on the market this year. When I contacted them with ideas for new products they seemed very receptive and a new adapter like the ALFA ACS but uses a 7610u was on the list I sent in. |
I like that idea - 7610u capabilities in a smaller package. In regards to the ACHM and yet to be released Alfa adapters - Although it's moot if they come out with a 7610u similar to the ACS.. |
Hi, silanosa! |
@Antozzz - I messaged Alfa Technical Support ~2 months ago and they said it's a known bug. They said this:
Someone said somewhere that it's 2mW connected to a booster/internal amp that raises the tx power to 1000mW. To be fair, my TP-Link Archer T4U Plus (rtl8812bu) gets quite a few more access points when I do a scan than the ACHM. But the TP-Link Archer T4U doesn't work for my distro and its capabilities are lacking compared to the ACHM (AP, injection, etc) I've been tempted to hook the ACHM up to a 2500mW powered booster but you have to set tx power for the ACHM from 31mW to 100mW and I can't do that with the current drivers and I don't want to blow up an expensive booster. So, no go there too. |
Hi @Antozzz Like @amisix says. This is a known issue. I am pushing to see if we can find out where the problem as that would determine who needs to fix it. The problem could be in firmware (closed part of the driver) which means Mediatek needs to fix it. The problem could be in the open source part of the driver which means the Mediatek kernel devs need to fix it. Or it could be an issue created by ALFA when they added an internal amp. The bottom line is that the issue is cosmetic because performance in my experience is excellent. I ignore twpower here. @amisix You said you have seen better scan performance on another adapter. Would you mind providing details so that I can do a little testing when I have time... what band, what app did the scanning, etc. Regards |
Amisix, morrownr ! Thanks! I also made a request to Alfa Network, they wrote me the instructions that I had previously tried! |
@Antozzz I use this adapter with great success and encourage you give it a shot. Are there other issues, or is it just I would encourage you to make your judgment based on the receivers data transfer performance in your target physical configurations. There is much more to the system then just transmit power in terms of quality and long range radio. |
@morrownr Yeah, sure. The TP-Link Archer T4U Plus did identify more APs but that's where it ends. The Alfa ACHM did significantly better in regards to RSSI and signal quality (in blue highlight). Obviously those numbers are not attainable with just 2dB. It's clear to me the Alfa ACHM is a far better adapter despite concerns regarding inaccurate tx power display. |
@morrownr I re-ran the tests in Raspbian Bullseye with airodrump and came out with different numbers. I used your driver for the TP-Link Archer T4U Plus and it worked great, thanks! The Alfa ACHM found 6 more APs than the TP-Link Archer T4U Plus, different than my first tests in Windows. OS: Raspbian Bullseye |
Hey @amisix , how is it going? I was still pondering your previous test results. I've never tested in Windows. The results of this test are very close to what I have seen on my tests. Generally speaking, the ACHM detects more APs and is able to inject more at a higher rate. There are so many adapters on the market with the 8812bu chipset that results will end up all over the place. This is mostly due to there being a lot of cheap, low end products with the 8812bu chipset. The results I see with you T4U tells me it is a good adapter.
The ACHM has a mt7610u chipset. Regards |
Thanks for the correction, I've become such a fan of both MediaTek chipsets they've become jumbled in my head (edited as necessary). I was glad to see the airodump results, they make more sense to me and what we've discussed. I found it odd that in Windows the TP-Link adapter found so many more APs but the Alfa ACHM had far better RSSI & signal quality. But it's Windows. Now that I've got some good system images of Raspbian and OpenWrt I can play around a bit more. Thanks |
Hello , abalmos! |
hello morrownr and all other contributors, did anyone find a solution to change the txpower on the AWUS036ACHM adapter? Got the same problem, latest arch, mt76/mt76x0u driver. even if I change the region to for example BZ (iw reg set BZ), there is no difference, iw still reports txpower to be 2.0 dbm. Still i got the same experience, that this reported value cant be true - the adapter performs quite well. best regards |
Hi @katzenheinz
This has to be a reporting error as tests that I have completed on the ACHM show that its range is much longer than that of most other adapters. |
It's worth noting that there are still problems with |
This does need to be investigated but it is an indication problem, not an actual txpower problem. I'm testing the following PCIe card right now: #261 $ iw dev
$ iperf3 -c 192.168.1.1
So, the above card is showing over 600 Mb/s and txpower of 3 dBm through 3 walls and about 10 meters of distance. I don't think so. I wish this was fixed but it is just an indicator problem. How is the range with the ACHM? Regards |
I still have yet to test the actual range, but if I do, I'll do my best to report back. I saw it suggested elsewhere that the ACHM has an additional amplifier which may have something to do with what's indicated vs behavior seen in reality. |
I think you will find that the ACHM has excllent range. In my testing, the ACHM has shown the longest range of any modern dual band adapter. Let me know what you find. |
Hello All, My ALFA AWUS036ACHM is no longer working after using it on hcxdumptool. I have recently bought an ALFA AWUS036ACHM (Already have ALFA AWUS036H and AWUS036NHA, TPLINK TL-WN722N V2/V3). I have just started learning about hcxdumptool / hashcat and am now familiar with basic usage. After running for a few hours (working on hcxdumptool as expected) it stopped working (unfortunately I could have the logs) and I noticed it became very (very) hot. And after that, it stopped working. The day I started using it, it stopped working. I believe it's a hardware failure in my card. As I read about this adaptor on the internet and was not able to find any such abnormality (the adaptor becomes super hot and stopped working). All I read about this card were only positive reviews, and now not able to figure out what went wrong. I use ALFA AWUS036H, AWUS036NHA, and TPLINK TL-WN722N in the same environment and never noticed any abnormality. Just wanted to know if a similar issue was reported with ALFA AWUS036ACHM ? Thanks. |
I've had my ACHM for a few years and I cannot remember it getting warm, even when pushed with extended testing with iperf3. I can't recall any reports about the ACHM getting hot. Manufacturing issues happen. You might want to talk to your dealer or to Alfa. It could be a faulty adapter. |
@morrownr Thanks for your quick response. I was very confused and annoyed that how this could happen with such a renowned brand adaptor. You are right, this seems to be a manufacturing defect. I will contact the dealer for a replacement. BTW, can you tell me how long I can run ACHM adaptor on hcxdumptool monitoring? Is it okay to run continuously for long hours (24 hours....)? I have never noticed any issue with ALFA AWUS036H and AWUS036NHA while running 24/7 on hcxdumptool. Is ACHM reliable in such a similar performance? Thanks |
I understand. I have several Alfa adapters and not a single one has failed over the years. I think Alfa does a better job than most manufacturers and they seem to use better quality parts than most makers but it would be cost prohibitive to do lengthy full function tests on each product so it is common to do sampling and adjust/investigate based on the number of bad samples and returns. No maker produces 100% perfect products. Not Intel. Not AMD. Nobody. It is the nature of the business.
I am not familiar with hcxdumptool but probably should be. My recommendation is that you contact the author of hcxdumptool. In fact, you might want to do that before asking the dealer for a replacement. Good luck and keep us posted. |
@morrowner I recently purchased an Alfa ACHM, and the Tx power is 3 dBm. Is there any solution or command that can help me fix this issue in Linux, please? What is the actual range of Alfa ACHM? Last question, please, I notic that there is a noisy sound when I am close to Alfa ACHM. Is this normal or not? Please help me. |
Hi @John23ee < Does this issue still exist Yes < is there any solution for it? No, let me see about elevating this issue to see if we can get some action. The Mediatek devs have been extremely busy but they managed to get the new mt7925 driver in kernel 6.7 so maybe they can find time to work this issue.
It seems you are assuming there is a real problem. There is not. It is simply an indication problem.
Very long. The following is a link to a range test I did a few years ago: https://github.com/morrownr/USB-WiFi/blob/main/home/Performance_Comparison.md Note that the ACHM won.
It is very rare but I have seen two reports of this over the last few years. I think the resolution was Alfa sending a new adapter but it has been a while since I saw this. You should probably contact Alfa with the details and see how they want to handle it: |
@morrownr Thanks for your quick response and your time. I just mean if I want to decrease or edit TX power, I can't because I cannot see the actual TX power. |
Hi morrownr,
First of all, my sincere thanks for your dedication and commitment for the Linux community, I appreciate it a lot!
After reading your article about recommended USB WiFi adapters, I purchased a new Alfa AWUS036ACHM adapter, which works really well out-of-the-box on my 5.15.12-arch1-1 kernel, monitor mode included.
I was wondering about one peculiar thing though and as I could not find any information about it online, I thought that maybe this could be something in your area of expertise:
As displayed on your
iw list
output here, the output TX power for the AWUS036ACHM adapter is extremely low, only 2.0 dBm for the 2.4 GHz band. I get the exact same values on my laptop.I have found TX power values online for other mt7610u adapters which were much higher, e.g. a TP-Link Archer T2UH with 20.0 dBm.
I know that lower TX power does not necessarily mean worse performance of the WiFi adapter, but I was still interested if you have an explanation on why the Alfa AWUS036ACHM has such low TX power values and if this could have a negative impact.
The text was updated successfully, but these errors were encountered: