1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Wireless Driver Issues & Settings for N speeds

Discussion in 'Tomato Firmware' started by Toastman, Sep 24, 2011.

  1. Toastman

    Toastman Super Moderator Staff Member Member

    NEW WIRELESS DRIVERS

    Now, some interesting and welcome news. Teddy Bear has posted a LOT of updates and changes to the git repository, there is a hint that there may be a new "official" tomatoUSB build soon. Thanks Teddy!!

    One VERY important thing to note - the wireless driver is changed to a slightly different version 5.10.142.0 from 5.10.147.0

    Looks like there's now support for F7D3301, F7D3302, F7D4302, F5D8235v3

    From Fedor's notes:

    So far I have found this driver to be somewhat unstable and have reverted to 5.10.147.0

    Now - also very interesting, there is also a new RT-N branch from Teddy Bear. It incorporates a newer wireless driver - 5.100.138.9 which is referred to as the "N" driver . It appears pretty stable even on my E3000 and RT-N16's. Since we'd all like to test it and see what it does, I'm also posting this as a BETA. Again, if it doesn't do what you expected, be patient.

    NOTE - I wouldn't recommend this "N" driver for earlier routers, I think it should only be treated as one that is necessary to get the 5GHz wireless working in the E4200 - it does have some issues, which we'll see later in this thread. If you do want to test it on an E3000 or RT-N16 etc. all well and good, but please don't post complaints on the forum when it gives poor performance!

    Any E4200 users who wish to try it please report back on your findings.

    This thread will concentrate on the different experiences people have with these different wireless drivers and settings to get the best performance out of them.

    Have fun!
     
  2. pharma

    pharma Network Guru Member

    I just switched to .0486 and also noticed that wireless speeds were faster and more consistent with build .0486. It's possible that the newer wireless driver needs additional tweaking/troubleshooting.
     
  3. Solace

    Solace Networkin' Nut Member

    give it about another month or less and something stable with 5ghz should be polished off, quite impressed with the current activity and quality of the builds in comparison to ddwrt
     
  4. pharma

    pharma Network Guru Member

    Agreed! IMO Tomato beta firmware for the E4200 runs better and more stable than when I had beta dd-wrt firmware for same router.
     
  5. Jacques

    Jacques Addicted to LI Member

    Hi Toastman.
    After re-loading tomato-E3000USB-NVRAM60K-1.28.7486.2MIPSR2-Toastman-RT-VPN.bin everything is ok (one day's work)
    Before when I had problems with WiFi 5 GHz the RSSI value was = +120 dBm and Quality = +206 dBm.
    i count 206 - 120 donne the value of the noise floor = 86 dBm.
    And during the normal work RSSI value was = -55dBm
    and Quality = -31 dBm 55 +31 = -86 dBm is the value of the noise floor.
    something bad happened in the tomato that there are these strange values.
     
  6. mraneri

    mraneri LI Guru Member

    These are nonsensical numbers. +120dBm = 1,000,000,000 Watts. (Yes, 1 billion watts)..
    If your router was receiving a 1 Billion Watt signal, you'd probably be dead...
    Something's not working right...
     
  7. Toastman

    Toastman Super Moderator Staff Member Member

    DRIVER FOR THE RT-N16:

    The RT-N wireless driver(s) is somewhat experimental in nature. It is intended for use with the Linksys E4200, the ASUS RT-N10/12/66 models. The driver is large and takes up a lot of room, so it's not very useful for earlier routers. Experience also tells us that for the majority of users, it has been rather slow and buggy. However, at least one person has had good results using it on an RT-N16.​

    My recommendation for the RT-N16 is to use the normal (STD) RT version. If you try the RT-N version and have problems with it, please don't post complaints in the forums, as you aren't recommended to use it in the first place. And there's nothing we can do about anyway.​

    FYI:

    We are trying to use one wireless driver for a multitude of different routers. That has always worked in the past when the drivers were relatively simple. We tried several drivers from different manufacturers and by trial and error found one that seemed to work with the majority of routers. As you will all know, there have always been various problems with encryption, WDS, and so on, from time to time.​

    Now the situation has got much worse.​

    We now have many new routers on the market using wireless N. That necessitates rather complex wireless drivers. Each manufacturer commissions Broadcom to come up with a wireless driver for their own hardware model. Each model will often/usually have a different driver! There are many, many different driver versions used in the various models around these days. They are in the form of precompiled modules and we have no source code, and even if we did have it, without the knowledge of how each function worked and what parameters it used, the code wouldn't be of much use to us. There are also hundreds, perhaps thousands, of changes in other files for some of the drivers. Some drivers have additional NEW support files (by which I mean any file using or related to the wireless driver) and others deleted. We have very little idea what those changes mean.​

    Now, we may see the same "version" wireless driver used on different routers. Fine. But we don't know if or how many parameters have been "tuned" in that driver for a particular model by a particular manufacturer, you see?​

    Just one wrongly set parameter may be all it needs to wreck the wireless performance, cause dropouts or reboots.​

    Do you see the problem?​

    Last year, for example, Teddy Bear changed the wireless driver for the RT builds from 5.10.147.0 to 5.10.142.0. He found it worked better with certain Belkin models. However, he never released a version for others to test, so we ended up with that task. But tests over several months proved that it was responsible for a lot of rather obscure and hard to pin down problems on many more common router models. The driver was clearly not suitable for hardware other than that which it was designed for.​

    I therefore reverted my builds to 5.10.147.0 and the results have been better.​

    There are even more complex drivers coming out now, and for the most part, they are bulky and don't work so well on older models. They are also exceedingly difficult to incorporate into Tomato - which is why there's nobody working on this at the moment. We really do need someone with expert knowledge of these drivers and how to use them, but sadly, we aren't likely to get anyone with that knowledge to do it. In fact probably the only people with the required knowledge work for Broadcom.​

    So support for newer models is usually limited to the existing drivers (RT and N) from Teddy Bear. Which may have (mostly) worked for (most of) the functions on (most of) the models that were around a year ago but won't do so much longer.​

    You should therefore be beginning to see the problems - we have cobbled together a few (kind-of) working wireless driver(s) into Tomato, but they may support a particular model well, others less so, and may be unstable with yet another. Even the original manufacturers often had unsolved bugs, which we inherited.​

    The fact is - we don't have any "One Fit For All" solution. It almost certainly doesn't exist. In future, some models are probably going to need very special builds.​

    .
     
  8. kthaddock

    kthaddock Network Guru Member

    @Toastman
    I use 0407.1-nocat and have 17 days uptime with no issues. It's working really good.
    Is't possible that you can make 0407.4 version, vlan-rt-n with all new fixes.
    New RT-N wifi drivers wotking very good on my RT-N16.
    EDIT: I have tested and I get 8-9MBps when dowload from samba with laptop Atheros AR5007 card.
    I have "WLAN Auto with Noise Reduction" turned on.

    Thank you.
    kthaddock
     
  9. Jacques

    Jacques Addicted to LI Member

    Hi Toastman,
    tomato-E3000USB-NVRAM60K-1.28.7486.4MIPSR2-Toastman-RT-VPN.bin
    in device list RSSI of user WiFi 5GHz going slowly from -54 dBm to -108 dBm during one day of work
    and broken connection internet
    I'm back to tomato-E3000USB-NVRAM60K-1.28.7483.1MIPSR2-Toastman-RT-VPN.bin then
    RSSI is stable -54 dBm +/- few dBm (WiFi 5GHz)
    I will be able to say anything more for a working day of 7483.1.
     
  10. Jacques

    Jacques Addicted to LI Member

    Hi Toastman,
    tomato-E3000USB-NVRAM60K-1.28.7483.1MIPSR2-Toastman-RT-VPN.bin
    in device list RSSI of user WiFi 5GHz
    30/10/2011 at 12:51 -56dBm
    30/10/2011 at 15:03 -62dBm
    30/10/2011 at 15:21 -59dBm
    30/10/2011 at 16:09 -56dBm
    30/10/2011 at 17:07 -59dBm
    30/10/2011 at 17:56 -54dbm
    30/10/2011 at 23:01 -56dBm
    30/10/2011 at 23:50 -59dBm
    31/10/2011 at 10:16 -54dBm
    31/10/2011 at 10:50 -55dBm
    31/10/2011 at 11:22 -55dBm
    31/10/2011 at 11:50 -54dBm
    31/10/2011 at 12:20 -54dBm
    31/10/2011 at 13:05 -53dBm
    31/10/2011 at 14:04 -56dBm
    7483.1 is stable never RSSI exceed the value of noise floor (-86dBm)
    as do the tomato-E3000USB-NVRAM60K-1.28.7486.4MIPSR2-Toastman-RT-VPN.bin
    and broken connection internet.
     
  11. Toastman

    Toastman Super Moderator Staff Member Member

    I'm still running the E3000 here next to me, and wireless readings have been stable since we last spoke several days ago.

    OK, try the latest - 7483.2 - which brings 7483 up to date with the latest small updates - but does not have the latest git updates from Teddy Bear. That should help you.
     
  12. Jacques

    Jacques Addicted to LI Member

    Hi Toastman,
    OK, i try the 7483.2 now...
     
  13. Jacques

    Jacques Addicted to LI Member

    Hi Toastman,
    tomato-E3000USB-NVRAM60K-1.28.7483.2MIPSR2-Toastman-RT-VPN.bin
    in device list RSSI of user WiFi 5GHz
    31/10/2011 at 14:54 -56dBm
    31/10/2011 at 15:54 -57dBm
    31/10/2011 at 16:54 -58dBm
    31/10/2011 at 17:54 -58dBm
    31/10/2011 at 18:54 -57dBm
    31/10/2011 at 19:54 -58dbm
    31/10/2011 at 20:54 -56dBm
    31/10/2011 at 21:54 -60dBm
    31/10/2011 at 22:54 -55dBm
    01/11/2011 at 15:54 -55dBm
    7483.2 is stable
    thanks Toastman....
     
  14. pharma

    pharma Network Guru Member

    Toastman,

    I know 0406 and 0407 are basically the same except for some small changes, but with 6 wireless clients (4 laptops (N & G), 1 iPhone (G), 1 TV (N/G), I noticed the signals stronger and varying less in terms of quality readings and RSSI. With 0406 the highest wireless quality reading I get is about 88 with a RSSI reading of about -12 dbm. The other wireless quality readings are less (between 42 and 72) but much stronger with 0406 than with 0407 and less RSSI variation as time goes on. With 0407 I noticed immediately the lower quality/RSSI and more signal variation as time went on. I've loaded 0407 a few times but with no improvement -- when switching to either 0407 or 0406 I cleared NVRAM and loaded a saved config file. I've also loaded different firmware mods for the same version (ie., VPN, EXT, NOCAT), and once loaded one of the .trx files but noticed no change.

    In all cases wireless Channel is set to Auto, so it automatically determines the appropriate Channel Width and Sideband to use .... varies throughout the day.

    With both 0407 and 0406 I noticed clients choosing automatically whether to use N or G -- not sure how it makes the decision but N capable clients seem to prefer 5ghz even though the signal wasn't as strong. Some times all clients switch to using 2.4 ghz but I really don't know why. Once I set the 5 ghz radio to 2.4 ghz and after a period of time my iPhone started using it, but 5ghz signal with both 0406 and 0407 was not as strong as the 2.4 radio in this scenario.

    Not sure if this helps, but feel free to ask if you have any questions or need something tested.

    Pharma
     
  15. csm746

    csm746 Network Guru Member

    Toastman

    My experience mirrors Pharma's. I now rolled back to 0486 build finding the 0486.1 and to a greater degree .2 exhibiting slower and erratic wireless performance. I use the 5ghz mode for routine wireless, but my Droid and when my daughters are home use the 2.4. I have two E4200's and can test this back to back and have done so. I believe somewhere in the thread you indicated Teaman had made revisions to the wireless driver implementation. Thanks again for your work.....

    CSM746
     
  16. kthaddock

    kthaddock Network Guru Member

    I can't say anything about 0407-nocat build then is the best build regarding wireless performance and stability I ever tested.
    Have "Uptime 23 days, 05:26:22" and still counting. My wifi is, laptop with Atheros AR5007 (1x1 mimo and 40Mhz. Android with (1x1 mimo 20Mhz) and Cisco wusb600n v2 mimo 2x2 40Mhz). I can measure up/down speed to Samba server 8-9 MBps.
    I think RSSI vary over time is when you have G equipment attached then affect N divice display speed. But when I do speed test I get 8-9 MBps. But this is only on my asus RT-N16 which do not have 5Ghz.
    Btw I have Interference Mitigation "WLAN auto with Noise reduction On" and I feeling signal is more stable and don't affect speed.

    kthaddock
     
  17. Jacques

    Jacques Addicted to LI Member

    Hi Toastman,
    7488 has again a problem with WiFi 5 GHz RSSI +120dBm
    i'm back to 7483.2
     
  18. Toastman

    Toastman Super Moderator Staff Member Member

    Yep. Your RSSI problem seems to happen with the reverted wireless driver 5.10.142.0
    from Teddy bear's latest code. Since it is clearly a display problem, it doesn't affect operation in any way. Funny thing is that you are the only person who has reported this as far as I recall, and the E3000 here never shows it either, it's quite stable. Can't do much about that I'm afraid.
     
  19. kthaddock

    kthaddock Network Guru Member

    Yes but that disappear after awhile, wait a couple of minutes.
    Wifi displays shows right values after a couple of minutes.
     
  20. Jacques

    Jacques Addicted to LI Member

    Hi Toastman,
    Why 7487 7483.2 and others working ok and 7488 not working well ! ! !
     
  21. Jacques

    Jacques Addicted to LI Member

    Hi Toastman,
    7488 RSSI on 5GHz = +120dBm then i click on Noise Floor (eth2) button mesure then RSSI becomes -55dBm.
    This can will help you understand what is happening...
     
  22. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Will try the builds with the newer wireless driver. Clearly had stalls with the RT version of xx88.2.
     
  23. jsmiddleton4

    jsmiddleton4 Network Guru Member

    No stalls by the way.
     
  24. Toastman

    Toastman Super Moderator Staff Member Member

    So was this the situation with the wireless drivers, as you found it? (We're assuming that any problems are actually to do with the driver, though this isn't actually proven it seems a good hypothesis).

    EDITED:

    7483.2 ---- 5.10.147.0 --- No stalls
    7488 ------ 5.10.142.0 --- Stalls
    0488.2 ---- 5.100.138.9 -- No Stalls

    (I am keeping a list, so far inconclusive).
     
  25. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Had stalls with 7488 No stalls with 7483.2. No stalls with 0488.2.

    Have no equipment with older wireless cards. My older HP laptop that does, I am using Netgear mini-usb wireless N adapter. I can try the older wireless chipset and see.
     
  26. Jacques

    Jacques Addicted to LI Member

  27. Toastman

    Toastman Super Moderator Staff Member Member

    7488.2 ------ 5.10.142.0 --- RSSI Jitter
     
  28. jsmiddleton4

    jsmiddleton4 Network Guru Member

    No stalls with my Broadcom internal 54G wireless card for my laptop on 0488.2. But now with xx89 coming out.....
     
  29. pharma

    pharma Network Guru Member

    Initially loaded the VLAN version of Tomato "tomato-E4200USB-NVRAM60K-1.28.2488.2MIPSR2-Toastman-VLAN-RT-N-VPN" with no problems. Performance was very good and all wireless clients had no problem connecting at fairly decent speeds.

    Even though I was satisfied, I decided to try the non-VLAN version "tomato-E4200USB-NVRAM60K-1.28.0488.2MIPSR2-Toastman-RT-N-VPN". Performance was extremely good with all wireless clients and as I surfed the web pages "seemed to render faster" than the VLAN version, but this is most likely my imagination. For now I'll keep using 0488.2 on my E4200 since it meets all my needs and (in my mind) seems to be more responsive.

    Thanks to the developers and Toastman for your tireless support!

    Pharma
     
  30. csm746

    csm746 Network Guru Member

    Pharma, I don't believe it is time to sign yourself into the asylum. I have the benefit of having a couple E4200s and I agree with your perception that the .0488.2 version has better wireless performance. I have tried both routers loaded with the .2488.2 and .0488.2 back to back and can see the difference. I am going to try .0489 now.

    I also picked up a NetGear WNDR4500 and even with stock firmware 1.0.0.58, it is blazing fast. I am using all the same features I use on Tomato and I believe it may be the best router I have used. It is my first NetGear router after using Linksys for well over 10 years, but I am impressed. In my garage, I have a NetGear WNDAP350 dual band access point sitting on CAT5E from the house and I have not been overly impressed with that device. Maybe someday, we will see Tomato for the WNDR4500 and it will get even better.
     
  31. jsmiddleton4

    jsmiddleton4 Network Guru Member

    0489 in WDS mode reports the master and slave router's speed incorrectly OR they are not connected to each other at full speed. Should be 300. 0489 reports 54, 108 and 217. Never 300. Most of the time says 108 despite reboot. 7489 however reports properly up to 300.

    Could be cosmetic but something isn't right with rx/tx report in Device listing for 0489 in WDS mode.

    The connection is fast and solid. Which makes me think its cosmetic.

    My WDS channel is the 5 ghz by the way.
     
  32. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Stalls are back with 7489 however.

    Even though I'm not certain about the Device tx/rx display I'm back to 0489 as I have no stalls with it. The tx/rx display for the wds slave and master did seem correct with 0488.2 by the way.
     
  33. Toastman

    Toastman Super Moderator Staff Member Member

    It's really hard to get a picture of these drivers, isn't it?
     
  34. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Yep.

    No stalls at all with 0489 or 0488. Can't tell for sure but tx/rx display for wds going back and forth between both 2.4 and 5 ghz channels? My wds is setup for only 5 ghz but the rx/tx numbers would make sense if it was hitting the 2.4 channel. The 2.4 ghz channels are AP mode only so the two routers should "see" eachother's 2.4. But wds data is only on 5 ghz. 0488 reported correctly.
     
  35. kthaddock

    kthaddock Network Guru Member

    This my thought about new wifi drivers in toastman build(s):
    I have discovered when I don't have any traffic on wifi then seems that wifi goes into "good neighbour" mode or some similar.
    Eg rx/tx go to 1x1 mimo and when using wifi again than speed goes to 2x2, 1x2 or 2x1 mimo.
    So there is some new function to "save" space, good neighbour function.
    When having some traffic wifi can recognize then it goes to that mode,
    but what I have seen so far is that always come back to full speed (2x2 mimo) after awhile using it.

    kthaddock
     
  36. Toastman

    Toastman Super Moderator Staff Member Member

    Something may be happening. I am just wondering why some people have problems while others don't. It's happened with every driver going back right to when I started using Tomato several years ago, but there's just too many variables now to get a good picture of what is happening. I'm sure there are power reduction features in the newer drivers that are probably getting more and more agressive, because of low power devices now around. They have the capability of changing parameters in the wireless client, sending them to sleep for a period before supposedly waking up - but some don't wake up or are erratic. But every wireless client is different, may or may not support a particular feature, or may not implement it in the same way. Some win, some lose. For some the latest drivers are better, others find the opposite. The exercise is probably not going to prove much. Oh well.
     
  37. jsmiddleton4

    jsmiddleton4 Network Guru Member

    The wireless driver is the same for 0088 and 0089 version isn't it? This morning after no traffic during the night my WDS 5 ghz channel is reporting 54/108 no matter what I do. It seems like its set back to a "G" standard and not "N" some how? I'm going to try 0088.2 again just to see.

    0088.2 reports 213-300 between slave and master in WDS mode, 5 ghz channel. Which is as it should be.
     
  38. Toastman

    Toastman Super Moderator Staff Member Member

    Did you mean 0488/9 ? Yes. All the RT-N builds have the same driver, (which has become known as the "N" driver). There has only been one for the E4200 series, 5.100.138.9 . You can check in the "About" page to avoid confusion, what's there is definitive.

    Something similar to that behaviour was reported before. I'll see if I can find a link.
     
  39. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Yes, 04xx. On second cup of coffee. Just tried something with 0489. I set the Mode manually in both routers to "N" instead of "Auto". Worked in terms of connection speed display, now showing 217-300 for tx/rx. So I'm thinking the display is correct. Why its falling back to the slower modes I have no idea. As noted did not do this with 0488.

    I can't keep playing as the folks who use my little WDS system are waking up around me and they have been quite patient.
     
  40. Liudvikas Mikalauskas

    Liudvikas Mikalauskas Networkin' Nut Member

    Hi, first of all thanks for great firmware. ;)
    Today I updated my two linksys routers (E4200 and wrt320n) to 0488.2 fw:
    E4200 with MIPSR2-Toastman-RT-N K26 USB VPN
    WRT320N with MIPSR2-Toastman-RT-N K26 Std

    Everything seems to work fine, I use wrt320n as wireless ethernet bridge in 5ghz mode. However I noticed that on E4200 device list gui my wrt320n reports fluctuating rssi (jumping between -48dBm and -67dBm) and quality (jumping between 40 and 24). However in WRT320N device list gui E4200 reports steady rssi of approx -48dBm and quality approx 40. I did not notice any change or fluctuation in wrt320 performance using speedtest and some usual streaming. Other wifi clients on E4200 report steady rssi and quality.
    Don't know if this could be fixed, just wanted to report this issue. ;)

    Thanks again, will keep on testing.
    br
     
  41. Toastman

    Toastman Super Moderator Staff Member Member

  42. pharma

    pharma Network Guru Member

    I'm getting very good stability and performance with .2489 on my E4200. This time I flashed the generic .trx version "tomato-K26USB-NVRAM60K-1.28.2489MIPSR2-Toastman-VLAN-RT-N-VPN.trx" and followed with "NVRAM clear". Connected wireless clients show a fairly consistant and high Tx/Rx rates. What did surprise me was web browsing ... in the past the non-vlan version seemed to render web pages faster, but at least for now the vlan and non-vlan versions experience similar browsing characteristics.

    I'll be running this since it seems to be fast, reliable and offer some of the highest wireless speeds when running in a dual radio mode (2.4 ghz & 5 ghz).

    Pharma
     
  43. Damarr

    Damarr Addicted to LI Member

    Hello Toastman, many thanks for all of your hard work on the builds. I just thought I'd report what seems like an unusual bug on my E4200:

    I upgraded from a fairly old build to tomato-E4200USB-NVRAM60K-1.28.0489MIPSR2-Toastman-RT-N-Ext, so I could enable both radios at once and have a 5Ghz network for one of my housemates. When I did this, throughput on the 2.4GHz wireless connection plummeted (from 30Mbps to 2-3Mbps) - Windows reported an excellent connection with 72Mbps, but the connection rate on the Tomato overview dropped from 78Mbps to 5.5Mbps. Upload speeds weren't affected (they peaked at the ISP limit of 3Mbps) and there was no issue with packet loss. The connection was totally stable, but the downstream was screwed.

    This happened on 0489 and 0488.2 after clearing the NVRAM. I eventually disabled the 5Ghz radio, but the problem on the 2.4Ghz-only radio remained. I have now switched over to tomato-E4200USB-NVRAM60K-1.28.7489.1MIPSR2-Toastman-RT-Ext and all is well again.
     
  44. Damarr

    Damarr Addicted to LI Member

    It definitely wasn't just cosmetic - I tested three devices and none could download at more than 3.5Mbps. Switching to the single radio firmware solved the problem.

    Perhaps it's a hardware problem - is power drawn from the 2.4GHz radio when the other is active in any way?
     
  45. pharma

    pharma Network Guru Member

    Damarr,

    I also have a E4200 and see excellent dual radio performance using "tomato-K26USB-NVRAM60K-1.28.2489MIPSR2-Toastman-VLAN-RT-N-VPN.trx". You might want to try this to see if you get similar performance. Connections down and up are very stable and consistent. Remember to use the same firmware file (.trx) as noted above.

    Tomato Device Status image:

    https://docs.google.com/leaf?id=0B7...YTM0MS00N2Q1LWI1NTMtNTVlZTI5MmMwNTE0&hl=en_US
    Pharma
     
  46. marciton

    marciton Network Guru Member

    @Pharma
    stupid question, do you have a link to download it?
    thanks
     
  47. kthaddock

    kthaddock Network Guru Member

  48. wasp87

    wasp87 Network Guru Member

    I have similar issues, sometimes the router will not put out more than ~3.5Mbps download speeds. What "single radio firmware" did you switch to?
     
  49. Damarr

    Damarr Addicted to LI Member

    I switched to this one: tomato-E4200USB-NVRAM60K-1.28.7489.1MIPSR2-Toastman-RT-Ext

    Pharma, I've tried the firmware you recommended and had the same problem. Under any dual radio firmware, even with the 5Ghz radio disabled, the 2.4Ghz connection will not exceed 3.5Mbps. If I have both radios set to 2.4Ghz (at opposite ends of the spectrum), I struggle to connect to the 2.4Ghz-only radio (which is still limited to about 3.5) and can't connect to the other radio at all.
     
  50. pharma

    pharma Network Guru Member

    Something interesting happened. I disabled the wireless setting "Afterburner" for my 2.4 ghz radio and noticed all wireless clients decided to use the 2.4 radio on Eth1. However the tx/rx rate for some of the wireless clients still approached 300 with the 2.4 radio. I should mention Channel is set to Auto and Channel Width is set to 40 for both 2.4 and 5 ghz.
    See link below :

    https://docs.google.com/leaf?id=0B7...ZWEyOC00YTFjLTkxMDEtMmIwYWI4ZTgzZWRj&hl=en_US

    First time this has happened but I definitely notice an improvement in Wireless N client performance when compared to my earlier post.

    Pharma
     
  51. pharma

    pharma Network Guru Member

    Damarr,

    Have to ask just in case, but after flashing you did the nvram clear "stuff". Also are you reading in a saved config file to restore settings?

    It might have something to do with your Advanced wireless settings, but not sure...
     
  52. Damarr

    Damarr Addicted to LI Member

    Yup, I run the nvram clear every time - I tried various firmwares with absolutely everything on default and had no joy.
     
  53. pharma

    pharma Network Guru Member

    I tested my wireless upload/download speeds at SpeedTest.net. With QOS enabled my wireless download/upload speeds were 5 mbps/ 3 mbps -- this confirms my QOS settings are working with regard to how they have been setup. With QOS disabled the wireless download/upload speeds were 24.41 mbps / 5 mbps -- this is the speeds I should be getting from my IP provider.

    You also might try disabling QOS and testing again since you are using the default settings. I'm not sure whether QOS is enabled or disabled by default.

    Pharma
     
  54. Damarr

    Damarr Addicted to LI Member

    QOS is disabled by default, so it couldn't be that. Hmmm, I hate technology sometimes.

    Edit: Just to add, I put the Cisco firmware back on and had no problems getting full speed via 2.4Ghz with the 5Ghz radio enabled, so the problem isn't with the hardware.
     
  55. Damarr

    Damarr Addicted to LI Member

    Thanks for the suggestion. I tried it and got some moderate gains (up from 3.5Mbps down to 6Mbps on 2.4Ghz), but I couldn't find any combination of frequencies that provided full performance.
     
  56. jsmiddleton4

    jsmiddleton4 Network Guru Member

    0490 still with tx/rx speed error OR its not connecting in WDS mode at the higher 5.4ghz N protocol speed and the display is correct.

    7490 reports speeds up to 300 for the 5.4ghz N protocol channel in WDS mode. Which is correct.
     
  57. Toastman

    Toastman Super Moderator Staff Member Member

    I think that there's simply some quirks that we have to put up with if we want to use the "N" wireless driver. Probably that won't change until Teddy Bear gets around to exploring why or changes to another one.
     
  58. jsmiddleton4

    jsmiddleton4 Network Guru Member

    "quirks that we have to put up with if we want to use the "N" wireless driver."

    You mean the newest "N" driver? Cause 74xx uses "N" driver yes?
     
  59. Jacques

    Jacques Addicted to LI Member

    did you have also the problems with 1.28.7487 wich uses Broadcom Wireless Driver 5.10.147.0
    for me this is the last stable version what's working on WiFi 5GHz
    next firmware uses Broadcom Wireless Driver 5.10.142.0.
     
  60. Toastman

    Toastman Super Moderator Staff Member Member

    I was thinking of the "N" driver - 5.100.138.9

    For the newcomers, if we were "nicknaming" these drivers traditionally, the original driver used by Jon for the WRT54 was the "normal" driver. It was simple and fast but had some problems. Then came the "ND" drivers (new drivers) and they fixed the problems we had with Intel wireless cards causing router reboots. The current ones also keep this "ND" title, except the one for the E4200 which seems to have acquired the "N" tag.

    Anyway, these are what we have now:

    5.10.147.0 - ND driver used in previous Teddy Bear builds. Superseded by:
    5.10.142.0 - ND driver in latest Teddy Bear code, supports some extra Belkin routers. Used in latest Toastman versions.
    5.100.138.9 - "N" driver used for E4200 builds. (RT-N)
     
  61. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Just making sure I understand this thing we're seeing with router to router speeds in WDS mode and 04xx versions.

    Whatever the cause is, that cause is in the core of the code for versions 04xx that is handed off from a different source and then Toastman adds the custom features he and the other custom features he is over seeing. Whatever the cause is, again not sure what is the problem, has to be looked at from the originator of the 04xx core code.

    Is that correct?

    I find it odd that IF its the wireless driver why its having a problem talking between the same two versions of the driver, router to router. I was speculating that it has something to do with the way WDS traffic flows and the tx/rx speed is being cut in half due to that traffic flow. But the tx/rx speeds shown do not get to 150, half of the 300, but rather show the 54/108 speeds.

    Maybe today I'll play with wireless ethernet bridge mode and see what speeds show.
     
  62. jsmiddleton4

    jsmiddleton4 Network Guru Member

    First impressions. 0490.2. Went to wireless ethernet bridge mode. Streaming internet radio through bridged router. tx/rx speeds show correct at the up to 300 rate. Only into it a few minutes so not 100% sure. But thinking tx/rx reporting issues is WDS related and perhaps the past issues with WDS have resurfaced?
     
  63. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Wireless ethernet bridge mode showed promise but it is not reporting tx/rx between routers properly either. One shows tx/rx at 162/162 and the other shows 243/162. Can't be. One or both of them are wrong. Will need to see if when activity is down if they reset to 54/108 speeds.

    Now one is showing 162/81 and the othr 243/162. That almost makes sense but not quite.

    Don't know enough to know what part of the code is amiss but something isn't quite right.

    Since the 74xx versions report correctly reasonable to say its the wireless driver but again that doesn't make full sense either given the two routers are talking to each other.

    There is no apparent performance slow down. This is why I'm thinking it is a reporting matter, not an actual function matter. Functioning quite fast.
     
  64. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Ok with 7490 on slave, 0490 on master, router to router speeds report correctly.

    Its hard not to believe its the wireless driver when both are any of the 04xx versions. Hard to understand though.

    Holding 243/300-300/243 tx/rx on both routers while streaming internet radio between them out to client attached to slave/bridged router with 7490.2 on slave/bridged 0490.2 on master.
     
  65. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Because I am suspicious of whatever is happening with the 04xx "N" versions with router to router communication I'm going back to 74xx versions. Hopefully the folks with the appropriate know-how will take a look at the symptoms I have reported and track them down and address them.

    With 7490.2 all tx/rx reporting correctly and at 300 tx/rx for both routers. Clients on at their relevant highest speed as well. Newer laptops at 150 or 300. Older ones at 54.
     
  66. pharma

    pharma Network Guru Member

    Short note about 2490.2 -- seem to get stronger signal between remote wireless devices and router using same configuration. No issues so far.

    Pharma
     
  67. buggage

    buggage LI Guru Member

    Wow...what a difference the RT builds made over the RT-N on my device. Just picked up an E4200 recently and been thinking of returning it, as the wireless speeds in the RT-N builds have just been horrible (either band and with g, n, etc) and extremely strange Tx/Rx rates shown (though I confirmed with actual testing the speeds were actually bad). Thought the router was either defective or I was loosing my mind, as couldn't even get a signal close to what I should have been getting. Tried the RT build tonight, and the wireless speeds are great and signal stable again, and reporting pretty much exactly what I would expect. The only downside now is I can't test to see what improvement I would get @ n-300 speeds, as I need the 5Ghz band for that (too congested on 2.4), but at least n-144 speeds - and my few non-n device speeds - are great. Anyway, feeling much better about my purchase now. Hopefully one day the RT-N builds will be just as good (at least on my E4200 anyway) so i can fully push it's limits.
     
  68. eahm

    eahm LI Guru Member

    You can get an E3000 to have 5Ghz 300mbps.
     
  69. jsmiddleton4

    jsmiddleton4 Network Guru Member

    I have 300 speeds on my E3000's and RT versions.
     
  70. bripab007

    bripab007 Network Guru Member

    Buggage was referring to poor wi-fi speeds using the RT-N, simultaneous dual-band build (at least, I believe that's what the RT-N build is).

    It sounds like he, too, could probably get 300mbps @ 5ghz on RT build, but then he wouldn't have the simultaneous 2.4ghz broadcast.

    This is somewhat disheartening, as I was hoping to consolide wi-fi APs with an E3000 I just ordered :(
     
  71. jsmiddleton4

    jsmiddleton4 Network Guru Member

    "but then he wouldn't have the simultaneous 2.4ghz broadcast."

    Huh? RT versions do dual band just fine. Nothing about RT versions are crippled as compared to "N" versions. Basically the only difference is the wireless driver.
     
  72. DanielCoffey

    DanielCoffey Networkin' Nut Member

    I am finding the 7490.2 version perfect for my E3000 since I use it at home as my only router with a few 2.4GHz g or n devices and one 5GHz device. I am getting 30.35Mbs wired on my 30MB Cable service and around 17Mb/s on both the 2.4 and 5GHz on my main wifi device.
     
  73. Toastman

    Toastman Super Moderator Staff Member Member

    I'd think I'd pretty much agree with that. The 7483.2 build used the 5.10.147.0 wireless driver. Also, worth a look at 7487 which was a trial build also using the same wireless driver but with better IP Monitor.
     
    TrueBlueBlooded likes this.
  74. buggage

    buggage LI Guru Member

    Actually, on any RT-N build I tried (Toastman or other - I tried a few that were available), I had bad speeds all around no matter what I was trying to connect, and no matter what options I tried changing. Worse than my WRT54GL that I replaced. Same settings with the RT worked perfectly. So looks like the only way to get at the 5Ghz with the E4200 would be with the RT-N...no chance with the RT, correct?
     
  75. jsmiddleton4

    jsmiddleton4 Network Guru Member

    With the 4200's its a different ball game as compared to the E3000's in terms of wireless driver. I'm trying the "test" builds. Tried the wrong one first. Am trying the Tomato Firmware v1.28.999998 MIPSR2-Toastman-RT-WL5.10.147.0_TEST_BUILD K26 USB Ext one now. Holding TX/RX speeds between routers up to 300 and no stalls.
     
  76. jsmiddleton4

    jsmiddleton4 Network Guru Member

    I know its subjective but this version is snappier. Connected WDS without the usual second warm reboot on part of master router too. RX/TX speeds still around and up to 300 while streaming internet radio. No stalls.
     
  77. bripab007

    bripab007 Network Guru Member

    Sorry for my malinformed post. Just getting back into Tomato again and was under the impression RT-N build was necessary for simultaneous 2.4 and 5Ghz on any of the newer routers. Apparently that only applies to the oddball E4200?

    So the RT builds should provide simultaneous dual-band support on the E3000 no problem?
     
  78. pharma

    pharma Network Guru Member

    I was asked to test the Tomato Firmware v1.28.999998 MIPSR2-Toastman-RT-WL5.10.147.0_TEST_BUILD K26 USB VPN for my E4200 and had fairly good signal strength (but nothing special). Since it is a RT version I only had access to the 2.4 ghz radio and none of my wireless clients had any problem connecting, though none connected at a rx/tx rate of 300 (this would probably change over time). It should be noted I have comparable or better rates with the RT-N version in my signature.

    One thing I notice is it's easier to get good signal strength (after flashing) with the RT version than the RT-N version without "playing around with settings". With the RT-N version and my E4200 I had to make changes for both 2.4 & 5 ghz radios" either in the 1.) Basic --> Network --> Wireless (ie, Channel Width, Sideband, Channel) or 2.) Advanced --> Wireless (DTIM Interval, Fragmentation Threshold, etc...) settings. I also noticed after flashing and reloading my config file it helps to do another router "reboot"- kind of like 30-30-30 router reset (especially with the RT-N version). Right now since I have pretty good "RT-N configuration" it's simply a matter of loading the same config until I feel that "need to play" again. Since I only have clients capable of transmitting up to 300 mbps, it will be interesting once I get a client with a tx/rx of 450 mbps since that is what the 5 ghz radio is currently transmitting at.

    My guess is most people who experience "performance" issues really haven't taken the time or have the patience to try different settings (especially people with simultaneous dual band routers). In hindsight I guess I was the same way when I initially loaded the RT-N version on my router. I'm just glad I took the time to play around, and hopefully people will. If they did they would see the router is performing well above spec with Tomato firmware and the newer driver, but requires more setup time.

    Pharma
     
  79. bucher

    bucher Networkin' Nut Member

    Toastman, first I want to say thanks for making this awesome firmware. I am running a WRT320n CFE flashed to E2000 using "tomato-E2000-NVRAM60K-1.28.7483.2MIPSR2-Toastman-RT-Std". I've noticed that my signal bars are a little lower with Tomato than when using DD-WRT, cannot seem to figure out the reason, same tx power of 50. One other issue that I cannot seem to figure out is when using Grabit newsgroup reader with SSL, once I queue up files and try to download, it sometimes stalls for a good 15 seconds or so before things get going. I think it may have something to do with preamble, but I'm not sure. If I am using 2.4ghz N mode only, does the "Preamble" setting do anything, or is it strictly the "802.11n Preamble" setting that affects N? Can anyone point me to whether or not I should be using "Greenfield" or "GF-BRCM"?
     
  80. Toastman

    Toastman Super Moderator Staff Member Member

    I always see pretty much the same strength from these routers here no matter what version I try. Dunno why. If you use Windows you might find it interesting to install InSSIDer (free) which will give you a nice way to look at the band and signal strength of the router.

    Max power is usually with a setting of 60. Since you are already using 50 you won't see much change though, as it is less than a dB it probably won't even be noticeable. Make sure the country is set to something that really permits you to change power (U.S. and Singapore are good). Vary the setting to suit your own circumstances. The rest of it is down to you, there are many so settings to twiddle. This would be a really nice topic for another thread, keep this one clear. [If you look at the source code for the wireless configuration etc, you will end up in a mental institution, because there are HUNDREDS of things that *could* be changed and absolutely no documentation at all].

    Here are some places to look:

    http://www.dd-wrt.com/wiki/index.php/Advanced_wireless_settings
    http://informatione.gmxhome.de/DDWRT/Standard/V23final/help/HWirelessAdvanced.html

    If you look at the "common Tomato Topics" thread (sticky) you will find several links that may help you to fiddle around. Other members here will also offer help, what worked for them may or may not help your particular useage. There are no "best" settings for many things, you have to make your own decisions. Google and the search facility on the forum will help you.

    This post, although not specifically about your router, may be informative as the same principles apply.

    http://www.linksysinfo.org/index.ph...ction-on-e3000-with-tomato.35877/#post-174761
     
  81. bucher

    bucher Networkin' Nut Member

    Oh yeah been using InSSIDer to figure out which channel to use. I even logged into a neighbor's router that was not protected and moved them to another channel so I could get a better signal :) I also stumbled upon a SSID "Dorothy's Hotspot", sounds a little scary.
     
  82. jsmiddleton4

    jsmiddleton4 Network Guru Member

    "Interference Mitigation" and "Wlan Auto with Noise Reduction"

    Those options are in the RT non-N versions.
     
  83. kthaddock

    kthaddock Network Guru Member

    Nope 2492-nocat is VLAN-RT-N with "Interference Mitigation" and "Wlan Auto with Noise Reduction"
    Advanced =>Wireless => Interference Mitigation => Wlan Auto with Noise Reduction
     

    Attached Files:

  84. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Just checked. Interference mitigation is in RT version. Don't see "WLAN Auto with noise reduction" now.

    Interference mitigation option is WLAN Auto.
     
  85. kthaddock

    kthaddock Network Guru Member

    See pic above....
     
  86. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Let me know the measurable difference the two options have on performance will you kt...

    I'm sure since you can see a difference you must be able to measure the difference in performance between WLAN Auto and WLAN Auto with noise reduction.
     
  87. pharma

    pharma Network Guru Member

    Like most things Tomato especially with the RT-N, most measureable differences comes down tweaking various settings till you get best performance. There are a few visible other differences between the RT-N and RT driver that are tweakable, but I guess you would know if you tried an RT-N driver in the past. Like Toastman mentioned in a previous post about there being a tremendous number of settings in the driver code related to wireless settings that we can't tweak, I'm sure there will be differences between the drivers. Best scenario is to use the driver best suited to your router and be happy! :)

    Pharma
     
  88. jsmiddleton4

    jsmiddleton4 Network Guru Member

    If the improvement can't be objectified it makes it rather difficult to persuade the developers to add any of the RT-N options to the RT versions.
     
  89. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Did the 04xx N versions get new broadcom driver?
     
  90. bucher

    bucher Networkin' Nut Member

    Per "Toastman-RT-N-1.28.0486-RT MIPSR2 K2.6 for E4200 etc."

    "BCM SDK: 5.100.138.9: wireless driver"

    I do not see any changes to the more recent versions for RT-N that change driver version, while the 7xxx versions were changed at 7491:

    December 2 2011 - Toastman-RT-1.28.7491

    Reverted Wireless Driver to 5.10.147.0 in all RT builds
     
  91. pharma

    pharma Network Guru Member

    The developer may have tried but found RT driver versions don't support certain wireless features - probably the reason you don't see the full list of options. Since this is Linksys's code the differences you now see are driver specific and don't think it's as easy as porting RT-N wireless options to RT versions. Even if there was benefit that task would be monumental and would involve Fedor or (Jon) --it'd probably be more effective buying a router that supports the features you need. Or the other option being offer here is load the RT-N driver on your router, but it may come with disadvantages and not be as easy to configure or perform as well as the RT driver with your hardware.
     
  92. jsmiddleton4

    jsmiddleton4 Network Guru Member

    I need to watch the N version in WDS mode during the day. The issue with not having tx/rx speeds at full N speed didn't show up until there had been no traffic for awhile. It was as if when going into some kind of sleep mode due to no traffic when awakened the drivers only spoke to each other in G mode. I know that isn't the best technical explanation. It does describe what I observed however.

    I'm not one to use all the graphs, charts, etc., for utilization. I'm sure your network needs much more observation regarding client usage than mine.
     
  93. jsmiddleton4

    jsmiddleton4 Network Guru Member

    After awhile of no activity I checked tx/rx. Stuck at G speeds no matter what I do. Going to 7492.2.
     
  94. kthaddock

    kthaddock Network Guru Member

    Are you stucket at G-speed on 20/40 Mhz? Have you tried "traffic" that channel and se if N-speed comes back ?
    Mine can stucked but after awhile eg download/upload speed always comes back to N-speed.
    Speed is always changing depending on environment.
     
  95. jsmiddleton4

    jsmiddleton4 Network Guru Member

    Yep. Seems to wake up at nothing more than G speed. Does not do same thing with the 74xx driver set. This is router to router in WDS mode 5 ghz channel so its pretty easy to isolate. It is not client to router.
     
  96. bucher

    bucher Networkin' Nut Member

    I thought I read something that WDS cuts the speed in half?
     
  97. pharma

    pharma Network Guru Member

    I believe that is normal. My TV has build in wireless and connects at tx/rx of 300 --- when the it isn't using the router it show a tx/rx speed like "300/6" or "6/6". It could be "N" code which reflects the speed when the wireless device is "hibernating". Once it starts to use the router again the speeds increase back to "N" speeds but not always at 300/300.

    Same is true with two routers connected -- doing nothing you should not expect to see 300/300 continously. What speed it reconnects should depend on the Wireless Network mode you choose -- "N Only", "G Only", "B/G Mixed", etc... as well as some other settings like whether Channel is set to "Auto" or a specific channel. If you have only "N" Devices then try setting your routers to "N Only".

    Pharma
     
  98. jsmiddleton4

    jsmiddleton4 Network Guru Member

    "Once it starts to use the router again the speeds increase back to "N" speeds but not always at 300/300."

    That would be great if it ever returned to 300. Never does. Stays stuck at G speeds. The non-N version works properly between routers.

    I've been down this road several times now. Done everything suggested many times.
     
  99. jsmiddleton4

    jsmiddleton4 Network Guru Member

    "WDS cuts the speed in half?"

    It can add an extra route between routers. Thought that was part of what I was seeing on tx/rx speeds. problem is gets to 300 with non-n versions and the speeds the N version will wake back up to are not half speed.
     
  100. Toastman

    Toastman Super Moderator Staff Member Member

    Because of the long-running discussion on the various new wireless drivers in the various threads, I decided to start this one. Perhaps the tests can be continued here.
     

Share This Page