WIFI Congestion?–Help Requested?

Update: See bottom

I have a problem – and I’m going to shamelessly use this blog to see if someone can help.

In the midst of my home control nirvana – I’m having some WIFI issues that I simply don’t understand.

So -  my setup:  Broadband coming in is maybe 6Mb in each direction so no issues of lack of speed.  It comes into a Draytek (PPOE) router from a small window mounted dish to a large aerial in the middle of the village. I have no problem streaming TV or downloading big programs etc.

Indeed I have very few problems with hardwired use – our PCs always connect to the web. I’ve had my little wifi ESP boards talking to a Raspberry Pi for months without issue.. and then a couple of weeks ago it started…

The units would log in but not stay logged in, or be unable to login. I thought it might be my little WIFI control units – but then my wife started griping here phone would not log into the network. The mobile phones suggest this is not an issue with my little WIFI boards.   Sure enough – try to log in – the phone would say “connecting” then just go back to “saved” .  After a while it would connect for no good reason.

I switched the WIFI off for a couple of hours and back on. All was well for maybe 30 hours – and then slowly but surely my little control units would have trouble logging into the WIFI – and so would our mobile phones.  Strangely, though our PCS which are hardwired would have no apparent issues, I’d find it might take seconds to connect to the router with the web interface, something that should be virtually instant. WHY? The router is on a wire as is the computer. What can be so jamming up the network?

So – I reduced the router to factory settings. I altered the DHCP to start at 192.168.1.100 and did some IP-MAC fixing for things like the PI (well, two or three of them) and my PC. This was to ensure that no devices with fixed IPs were in the same range with possible overlap with devices working on DHCP. We’re taking maybe 20 devices all in – some controllers, a couple of computers, 2 phones, 2 tablets, 2 Raspberry Pis… and some ESP8266 WIFI units.

All worked well, my devices operated flawlessly as did the phones and the TV for DAYS on end. Yesterday a couple of times I heard devices logging in to the Pi (I have speech warnings). Might be an issue, might be not.

But this morning, whatever is wrong is back full steam, devices logging in every few minutes and generally not taking any notice of remote instructions – or taking ages to respond.  It is as if something is absolutely hogging the radio network.

And yet… I put checks in to see how many messages were  being passed back and forth between my PI and the devices… one a second if that – no big deal. I’ve looked at bandwidth use in the router – nothing stands out . BUT SOMETHING is gumming up the works.  I have 3 routers acting as access points and yes I’ve spread channels so no overlaps.

I’m missing something here, I have no way to see what it is that is actually going wrong here.  If anyone considers themselves a whiz at wireless networks – feel free to get in touch – pete@scargill.org – I could do with some fresh thinking.

Update 28/10 – A couple of days with almost no problems – tonight back with a vengeance – WIFI feels “clogged” – attempts to get Wireshark invariably gave me an incomplete download on my otherwise very fast laptop. Went to the (wired) PC and no problem.

Here’s a heatmap of the place with very rough positions of kit.. only the two WIFI access points (actually identical position as both on the Draytek) are mine – others are WAY further away than they look and so unlikely to cause problems.  All but one is nearly a street away.

heatmap

 

So after having difficulty loading Wireshark, I downloaded it on my WIRED machine and ran on the laptop in the main room – I’ve no idea how to run Wireshark but I’m pretty sure that screen was showing problems – failures, some stuff in red, retries… After saving the file from Wireshark I turned off the external WIFI coming in – and turned the Draytek off. I left them both off for maybe 10 minutes. I turned them back on – hardly any units logged in and would not communicate (but one did as the Pi replayed the login message that comes from the unit). So any ideas of overheating went out the window (it’s cooler here today anyway than normal).

At this point, I’m sitting with a perfectly working wired machine, a laptop that seems ok and ESP units that are failing to get through – and resetting – after many hours of operating perfectly.  The laptop is seeing my two access points and next doors, intermittently – they are actually disappearing and re-appearing. Next I started to turn off all the ESPs, no difference, the mobile phones – no difference. My wife who by now was in bed reading said that the WIFI said it was disabled.  I even disconnected the Raspberry Pis from the network – no difference, our neighbour’s weak signal and our two strong signals were turning on and off, on and off on my Windows 10 tablet.  I turned off the dish and the router – this left the neighbour’s signal – still turning on and off about once a second. Turned the router back on – turned the phone on and turned on WIFI Analyser on the phone – I could watch the two WIFI signals come on full strength – and turn off – once every few seconds (I suspect even though it has a 1 second refresh it’s response is slower). I turned the entire Pergola off (PWM solar powered lights – actually disconnected the battery – no difference.

I need sleep.

Facebooktwittergoogle_pluspinterestlinkedin

54 thoughts on “WIFI Congestion?–Help Requested?

  1. Hi Pete, I have suffered exactly the same problem in my home network. I have 3 PCs, 3 phones, 2 Rasp, 1 NAS, and around 5 ESP8266 reporting every 10 seconds.
    In my case I would first loose internet connection and then the network would start to go really slow with transfer rates of 500kb/s.
    I have recently added a TPLink 841 in order to remove all the internal network traffic and work to this router acting as AP, and leave the ISP router only to provide Internet and DHCP.
    So far it has improved the situation quite a lot. I see you already have several AP so I am not sure if this will help you, I did just leave the original ISP router only for internet.
    I have been pushing ESP8266 development quite a lot by integrating it in many of my projects, however this WiFi issues are starting to pull me away from them for certain usages. They are great for standalone projects but I am finding the RFM69 to have many more advantages for home networks like yours or mine. Check https://lowpowerlab.com amazing work done by Felix.
    Let us know if you solve or improve your Wifi issues, I just installed the AP one week ago so I dont know how it will do in the long run, its seems better definately.
    Oh, something I did when I started to have this connection issues, I developed a small ESP + arduino + relay board that would reset the router if connection to my NAS was lost via the DDNS. As you have posted before, this always happens when you leave home for holidays or some days. Works great !

    Un saludo !

    1. I think this is worth progessing - could the ESPs actually be causing some kind of WIFI blockage? And if so, how? I'm not seeing horrendous traffic. The stuff you've pointed to is very good - but I am having so much luck generally with the ESPs and my current work on the PI I don't relish the idea of a restart. I can do just about anything I want with the Pi + MQTT + Node-Red... so narrowing down the WIFI issue and eliminating it is the preferred way forward... thanks ever so for the feedback. The router reset solution sometimes works, sometimes I'm back to square one almost straight away.

      1. I do understand what you mean regarding progressing with the ESP8266, in fact I just got a couple of NodeMCU and ESP12 in the mail. I still want to learn much more and develop much more with them. But I am working in parallel with the RFM69 as i have really been sold after much reading.

        After reading your latest message, I did noticed also that the router was getting extremely hot, seems to be running cooler now with the new AP.

        I think you best bet right now is to sacrifice the system for 2 or 3 days by disconnecting all the ESP devices and see if your cellphone/Pc connection suffers the same problems of slowness and such. At least this way you can confirm that the ESP and their additional wifi load are related to the issues you are having. Maybe you find out that the problem is still there after disconnecting all the ESP so we should look in another direction.

        Btw: I have also experience the problems with the microwave running and the Wifi dropping.

      2. Serge, indeed the RFM69 is interesting but I disagree with you on your 'many more advantages' statement. I can't see any advantages over the ESP devices, only disadvantages. You need to add the radio boards to arduinos, making them much more expensive and less versatile. I know there are integrated versions but it's still way more expensive. And the frequencies used can be just as congested as the wifi range without the inherent clash avoidance in wifi etc.

        But that said here's another interesting link -> https://www.wirelessthings.net/iot-devices 🙂

        1. ESP8266 is just much more versatile, it is easier to install anywhere with WiFi coverage, it can work standalone, etc etc.... I completely agree on this, and that is why I keep developing with ESP quite happily. For now it has been via AT commands and an Arduino so you can imagine that this is not the best way, just now I am going to start with LUA so long way to go.

          But I do see some nice features using the RFM69. I just started with them also so it is very new to me but I do like the concept and just wanted to point Felix's website as he has done a very good job with this modules.

          I guess my WiFi issues are one of the main weakness I was having with the ESP and that is why I started looking around. Nevertheless, I am sure I will end up with ESP and RFM working depending on the project. I even have 2 XBee connected wirelessly to my RPi. So its all about tinkering I guess.

          1. The issue I have with the esp8266 is the fact they are not really suitable for battery powered devices, and having to run a mains cable to the 'box on the wall' is a big turn off.

            Regarding the router issue, this could be down to memory leaks in the router firmware, I've seen this happen quite often, when more than a few devices are connected to wifi. As the memory of the AP fills up, connections start to drop.

  2. You found already, that there is a difference between the wired and wireless connections. This means you need to look at lower layers. There are many things which have influence on your wireless connectivity, which is not the case in the wired environment.
    Anything which works in the same frequency range in your neighborhood will pose a problem. Some issues you can control, some not. The ESP boards have lousy antennas. First thing is to check this:
    - are there any other wireless routers in your neighborhood. Do they use the same or different channels. There is software around for laptops to check this. I'm an apple user, so others might be in the better position to help here
    - is there equipment like microwave ovens which run in your neighborhood? This can seriously mess up your wireless.
    Things to look at:
    - better antennas for the ESP boards
    - do a channel planning together with your neighbors.
    Kind regards,
    Urs Eppenberger, Switzerland

    1. More good feedback and thanks for that. Ok, my "neighbourhood" is a mountain in Andalucia - right now the nearest active neighbour is almost a street away and nothing they could be doing would affect me - WIFI sniffer shows no other access points than my own. Household Microwave not on that much and no correlation between issues and it being on.

      The ESPs have built in antenna and the only thing I'm aware of that could improve them is replacing them with ESP-12F - which IT IS CLAIMED have better antenna but I've yet to see any evidence....

      I hope that's given folk a better idea of where I am right now... SO following the last blog, I turned the router off - I mean I was SURROUNDED by login messages - really the boards were not having it at all - just not responding or very slowly. 10 minutes after power cycling the Draytek router, all was fine. But I know it won't last. The boards are programmed to retry connections for maybe 10 minutes - then reboot. After turning the power off then on - 15 minutes later all is still well, very occasional logins but boards are responding... but then I know it won't last.. already I'm hearing the odd board re-login. MEANWHILE another fellow suggested he had an issue with the Draytek overheating.. so many possibilities. ESP antennas, dodgy Draytek, it's not interference from others - ok someone could be down in the village with a dish firing stuff our way, not the most plausible and surely that would show up on a mobile phone WIFI analyser....

      Keep em coming...

  3. Pete hi,

    This sounds like the problems I have had over the years with routers. A sort of gradual loss of function. I replaced one router many years ago and was sure it was the wifi side of the router that had given up the ghost (intermittent misbehaviour) but when I replaced the last one I'm pretty sure that it was actually the power supply for the router which was failing. Obviously I'm no expert on routers but I'd think the wifi would be the first to suffer from a failing power supply. I'm sure someone will correct me if I'm wrong.

    So I'd try replacing the power supplies to the routers or at least testing them.

    The behaviour you describe is one I have seen on a couple of campsites on our recent vacation. Intermittent connections etc. I found a neat app for android called wifianalyzer which shows the strength of each wifi in the area. Might be an idea to try this out. With this app I could see the campsite wifi signal coming and going. If I try it at home my home signal is good and strong with some other signals in the area coming and going which is what you would expect.

    I'm sure there are other tools other people will suggest as well. I'd be interested to see them.

    Simon

  4. So some feedback. One of you suggested power or heat problems - so - I took the top off the Draytek (which to be fair has pretty ok ventilation - and changed the 1.5amp power supply to a 2 amp power supply. Now, it occurs to me and this may have nothing to do with anything - that we've had power ussues here the last couple of weeks - the odd power cut for a second - and two deliberate cuts lasting hours - so clearly something is up there.

    Anyway, I put the new power supply in and put the Draytek back on top of the shelf (lots of open air). I turned it on. That was half an hour ago. I've not hear a PEEP out of my little boards - which are operating perfectly - as is the streaming TV and my phone. Now, this has happened before - turning the router off then on improving things but when I did it this morning it was good for all of a couple of minutes.

    If I get to a week with no issues I'll consider that power was the issue - right now I'm just thankful for the peace (having the Pi give verbal announcements of logins was a good idea but when they're happening every minute or so it gets a tad annoying). What I'm getting right now is utter silence...

    Also - and I hope it's just not because if trivial changes of positioning - I did my best to put the Draytek in exactly the same spot.... I was in the next room where my office is - getting a -75db signal - I'm pretty sure the signal is around 5db higher... I realise that's a bit vague.

    1. Good idea changing the power supply too.
      Maybe it's a problem from the wireless provider too, as we have bad weather in Spain and worse in your place (Andalucía).

      One of the things to test when slowness appears is to do a tracert to a known address (google.es). You may see delays through hops and maybe spot the problem.

      Ps. Registered successfully!

  5. Don't some routers have a default limit of how many client's they'll allow to connect ? Sure you're not bumping up against something like this ?

    A quick google shows some Draytek routers with configurable bandwidth and session limit settings, does your unit have these ?

  6. Hi Peter.
    I have similar issue with cheap router.
    Some devices fail to authenticate (bad signal/power/firmware/hw or noise) and this caused a big amount of broken sessions on the router. Each retry creates the new session, while the old one remains active on the router. So it quickly goes into high CPU and/or RAM usage. As well, as the number of devices waiting authentication increases. Now I have two routers. The one is much more expensive for managing ISP connection and important devices: laptops, phones, NAS and player (last two are connected with ethernet cable). All the 'iot' devices are connected to the dedicated router (the same cheap one).

    In You case I would try to turn off several devices. Maybe the amount is critical for the router and this causes the avalanche of broken session. Or try to figure out a failing device by logs or turning off them using a 'binary search' if possible.

  7. An RFM69 based solution should last a lot longer on batteries than an ESP8266. If powered by mains, though, the ESP8266 has obvious advantages.

  8. so next I fixed ALL the IP addresses to MACs (doesnt' stop anyone else coming in. Didn't seem to make a difference, some devices would still not communicate or would be re-logging in. Then I put a limit on any device of no more than half the bandwidth. It would be easy to say "that worked" - not no - fixing the IP addresses made no difference at all.

    OH Power, that was a nightmare. I took out the 1.5amp supply and put in 2amp supply. That lasted half an hour and packed in !! I thought I recognised that model of rubbish Chinese supply - that's in the bin. I now have a 12v laptop supply capable of 5amps - that kills any argument about power.

    No noticeable difference - losing the ability to stay connected- reconnecting - failling - eventually getting through for a while. I've never come across anything like this in my life. And yet my hardwired PC seems to be having no problem - while wifi devices cannot connect - the PC (which comes via the Draytek, through a 16 way switch to the PC - no problem at all ?!!??!!?

  9. Hi,
    To discard degradation due to wifi low level traffic (i.e. handshaking, ....) you may try to "sniff" wifi activity by using a linux distro like Wifiway. You can use a live version so no install needed and should be easy to see if there is abnormal traffic.

    1. I need something like this - looked to see if there was anything better than WIFI ANALYSER on the Android phones but found nothing yet).... a set of glasses I could see radio waves with would be nice 🙂

  10. Most consumer access points max out at around 8-16 devices connected depending on their wireless chipset and resources available (sometimes just crippled in firmware to force enterprise products onto customers). I'm guessing this problem started when you added another (or more than one) device to the 802.11a/b/n network.

    Get a separate access point for your data acquisition devices and keep this in mind moving forward, you could even put this into a different VLAN to keep things separate from your clients on the Draytek Vigor 2830.

  11. It definitely sounds like a router problem.
    You should be able to see in logs what happens in the router.
    Is it possible to install custom firmware on your router?
    Being a techy guy like you are, don't you have another router you can test?
    Also, all AP points can affect.
    What I would do is:
    1 Change router.
    2 If problem persists, switch off APs and then turn them on one by one to see if they interfere.
    3 Switch off all ESPs and turn them on one by one to see if one of them could be the culprit (try various scenarios, ie ESP1, ESP2, ESP4 one time and ESP1, ESP3 ESP4 etc so you can pinpoint if it is due to bad ESP or "congestion" in router.)
    Good luck!

    1. I have another router back in the UK but I can't get to it - I have one ONE other router able to plug into a network PPOE connection (10.10.10.1 or similar - from a WIMAX-type disk in the village - quite high speed both ways) - and then my router turns that into the 192.168.1.x range)... I am going at some point to have to turn off ALL the ESPs and leave them off to see if it is THEM or something else - but this problem has been known to go away completely for a day. It's not a short term test. I've tried turning everything off - and on - and on one occasion all was well for 30 hours!!! Then it just slows down and clogs up.... units are slow to respond, mobile phone gets logged out and can't log back in... etc.

  12. Peter, Do you have another spare access point/wifi router ? If not to resolve this go and buy a cheap one.

    Set it up hard wired to the Draytek but with a Different SSID and channel etc. Setup the closest ESP to login to that with a hard coded IP - thus removing DHCP from the equation.

    By doing this you have isolated a single unit and can then start a reasonable process of elimination.

    I also do not really understand why you are focusing on the Draytek - you mentioned multiple channels and no overlap - so i therefore assume you already have multiple access points - i would therefore assume that the ESPs are not talking to the Draytek but (at leasts some of them) to the Access Points ?

    have you tried to be tricky and have all the APs have the same SSID ? (this is how my house is setup and i know that Android has a long standing issue of not letting go of a weak access point. Maybe the ESPs do not handle hopping between APs with the same SSID and lose it somehow ? ?

    Anyway a new cheap (different) brand of AP to your existing - hardwired to the Draytek, and have a single (closest) ESP talking to it with Hardcoded IP and see what happens.

    Craig

    1. Ok... so The Draytek has the ability to have multiple wireless access points (on the same channel) - that didn't help. I also tried bandwidth limiting the ESPs - that didn't help.

      The Draytek seemed to be dropping to bits - so I took the ONE other router I have that has PPOE (the older ones only take ADSL and I don't have that - we have a small dish here - with an ethernet connection that then comes in as PPOE to my router) - I tried that on it's own - and it seemed to have similar issues - but it was only an ebay £15 spare. I even thought of having that allocate DHCP for say 191.168.1.2 to 192.168.1.50 (for people and the TV) and wire another router to it - as a slave - and having that do its own DHCP from say 192.168.1.51 upwards for the ESPs). Well, all seemed to work except that no-one was taking any notice of the allocations of the second router - I have NO idea why. I'm not therefore convinced this is all down to the Draytek - but as my other routers are just old rubbish and the only one I have that handles PPOE seems to have a similar problem - I'm left with little choice but to go out and buy another decent router like the Draytek.. I want some devices bound by MAC to fixed addresses (so I can get external access by port redirect) and the rest can be any old address. I did find some severe limitations with the second router - no way to bind IP addresses to MAC, only one wireless channel with very little control over bandwidth etc.. I guess I'm just too used to having lots of control (Draytek 2830 series - but they are expensive). For now I've taken other routers all out - and put the Draytek back in with 2 wireless signals - one for ESP, one for phones etc... but with no special DHCP - I have to be able to work. Shops open Friday I'll go try to find something - not quite as easy here in rural Spain as it might be in central London or Newcastle. Problem is I'm not entirely sure WHAT it is - if it's the Draytek - why am I getting similar effects with another cheaper router - and how come this has worked for weeks without issue and then in the space of a week I'm almost reduced to the stone age. I have two Rasberry PIs hardwired rather than one (a test one) - I can't see that affecting the WIFI... most confusing. Oh this morning on the Draytek - I did try just having the one Access point - and ticking every device to be IP-MAC bound... so there was no dynamic addressing going on - made no difference.

  13. Peter,

    You are obviously getting frustrated but are now starting to shotgun with your approach to solving this which will take you a long time.

    Take your spare router and put it in Access Point mode, Hardwire one of its LAN ports to the Draytek, put it on a seperate wifi channel and SSID to the rest of the network. Turn off DHCP on this device

    Reconfigure one of the ESP to have static IP addressing - not MAC based DHCP static but actually hard code the IP in the config of the ESP. Have it attach to the 2nd AP and see what happens.

    By doing this you have eliminated the wireless from the Draytek, DHCP and a host of other small things. If this is stable you can then move a 2nd ESP across etc - until they all attach to this new AP and everything is stable.

    You can then start turning things back on - Static DHCP assigment through MAC address etc.

    The last thing i would be doing is buying a big expensive new router

    I suspect you have an issue with duplicate MAC addresses (can you manually assign these on the wireless interface of the ESP ?) or duplicate IP

    The reason your 2nd unit did not give out IP addresses through DHCP was that if you have two DHCP servers on a single network - it is random which one will answer first to provide a client with the request.

    Craig

    Craig

    1. Frustrated - I'm having difficulty not throwing the cats and the computer out of the window - this has had me demented - and you're right I am starting to shotgun. Ok, well, it's pain to dismantle but the ships are closed tomorrow anyway - I don't THINK there are duplicate MACs as the chips have a unique mac - you don't program it in - though you can.... I will firstly check that very point in the morning - one on at a time and check it's up and and mac - and then look to making them fixed (actually that makes for a useful command - once a unit is in place and known working - fire a command at it to store it's IP in FLASH and use fixed IP from then on... good idea)... and yes, I'll do it your way. More tomorrow. If you hear nothing - I've no internet at all 🙂

      1. Peter just making sure we are onthe same page (and i hope you are not reading this until the morning when you are fresh !!)

        DO NOT dismantle anything !!! Simply connect the spare AP to one of the LAN ports on your Draytek

        In order of sequence

        1) Power up spare router and connect your PC to it through hard wire.
        2) Configure it to have a static IP address on your network
        3) turn off DHCP and put it in Access point mode (what router is it BTW ??)
        4) Setup SSID and channel different from your current Draytek
        5) Cable it to the Draytek (LAN port on Router to LAN port on Draytek) and preferably position it a reasonable distance away from the Draytek
        6) Reconfigure one of your ESP to use a STATIC IP (obviously within your 192.168.1.x range and that is not used by another device or allocated via DHCP) and the new SSID/Password Combo
        7) see what happens from here

        Craig

  14. Peter, In terms of analysing

    Ekahau Heatmapper for Windows - is free and gives you a respresentation of what your wireless looks like

    Wireshark will enable you to capture all the packets (assuming you have a decent wireless chipset) and see what is actually travelling around the Wireless network

    Craig

  15. Also - i doubt the problem is the Draytek - as you say they are a damn good router and you have proven that throughput is good with all your other devices.

    You have also intimated that even when the Wireless units are problematic you still have good connectivity from your hard coded devices.

    Maybe setup a NODE-RED flow to continually request a page from google on one of the PIs as see is this is still working OK when the wireless devices go belly up.

    Craig

  16. Keep em coming Craig - but it's 1.30am here now - this would not be a good time to start experimenting. Will do in the morning and will let you know.

    Thanks to EVERYONE who has posted here. Most of the time I give my knowledge away, it's great to know I'm getting help the one time I'm utterly stumped.

  17. Hi

    Not ignoring anyone - just taking my time (and we're away overnight tomorrow.

    I got up this morning and could not shake something Maureen my wife had said "it'll be your Pis". I have 3 Pis on the network and one is a direct clone of the other. Something in the back of my head said "I wonder if they have the same mac? So I got up this morning - all was quiet - and took them out of the network.

    Everything worked for half an hour or so... I noted one of my test ESPs was disconnected, could not remember at which point I disconnected it - I plugged it in - and immediately started ot have issues. I took it out - no better, power cycled the router.... all well. I waited (doing stuff with comms so using the network heavily) for 2 hours - not a hitch. I turned back on an ESP house light I'd disconnected yesterday, logged in no problem. I had another ESP on the desk - plugged that in.. still ok - the upshot is - I've not had a single failure ALL DAY. This could be just sheer fluke, or the PI's might have had something to do with it or that single little ESP.... I'm leaving it until I have time to investigate over the weekend without being rushed - meanwhile I found a bug in one of my ESPs - nothing network wise - I'm turning it on and of and on and off and...no network problems at all. Indeed as you'll see in the blog it's given me space to do more work on the promising BLYNX app which is using up even MORE bandwidth without issue.

    HOWEVER - I DID try the EXCELLENT Ekahau Heatmapper for Windows that Craig suggested - what a winner - next job probably tonight - Powerpoint up a map of our place and do a proper WIFI survey - in a QUICK shot last night despite being dead tired I did a 20 point survey of part of the place - not only did it accurately place our routers - it accurately placed the routers of our neighbours.... as soon as time permits I'm giving this one a proper try-out.

    More in time and thanks for the feedback without which I'd probably be breaking things with a hammer by now.

  18. Keep us updated Peter - the other thing to also do is install wireshark on your laptop. Then capture a good working startup sequence from a single ESP to the Router - you can then store this away - next time you have an issue you can do the same thing and compare them side by side to see if there is a discrepancy in the packet sequencing to enable you to get to a lower level

    Craig

  19. I wouldn't be surprised if the ESP8266 has never had true interoperability testing of the kind needed to earn an official WiFi badge. So, even if your router is good, or even great, by normal standards, maybe a router with a different chipset would react differently in the presence of a plethora of ESP8266's. If you remember, it took literally years to iron out the interoperability problems between 802.11n chipsets from different vendors, even though the chips from the same vendor could communicate with each other just fine. The non-profit group that owns the "Wi-Fi" trademark was created and gained traction because it was proven over and over again that without it vendors wouldn't coordinate on their own to ensure interoperability.

    1. Good point Whitehare...
      Craig - will do - we're off to the coast overnight so I'm awol today and tomorrow - but the "problem" is certainly not resolved so I'll be going through your list. I had ALMOST no issues at all yesterday so I think you're right, it is premature to blame the Draytek.

    2. My tests last night suggest this has nothing to do with ESP. The WIFI was going off and on and clogging up constantly - long after I turned off all the ESPs and Raspberry PIs. Right now the WIFI is off on the Draytek - a TPLINK is supplying a single WIFI channel while the Draytek just handles the DHCP and wired - and an hour later all is well - but of course an hour means nothing. I took WIRESHARK readings before and after... and they mean as much to me as encrypted satellite signals. I'm guessing the village doesn't have a wireshark quick course.

  20. An update: When I came back from our short break, I had more to worry about than WIFI - the power had gone off. Everything back on - but left the two Raspberry Pis out of the picture. All seemed to work - and has done since. I thought at one point I was seeing a re-occurance as changing the RGB on one of my boards caused it to stop responding - turns out I was firing out info too fast and the miraculous DELAY node in Node-Red took care of that. Today I've had ONE of the other RPis turned on to do some work - and for a short while an Orange Pi (see review - no further forward) - and I'm NOT getting any problems. Now, to be fair at one point I did get 30 hours with no problems so it may not be history.

    Next I'm going to try that ESP8266 that SEEMED to be giving problems - and then BOTH of my test PIs but I'm under no delusion that this will be a short term test - without leaving everything on for a couple of days there's no way to tell if this problem has "gone away" and I still don't really understand why it all seems to be working.

  21. Peter,

    before you fire up the suspect ESP - run wireshark and capture some conversations to get a feel for what the network looks like when everything is running OK - it is amazing how quickly you will be able to see the packet flows and get a feel for something being wrong.

    Craig

    1. Having woken up in the early hours I'm going back to bed, but sometime today I'm going to have to get to grips with this.

      So after the problem seemingly went away and all my gadgets were just fine (I was demonstrating to a neighbour last night BLYNK dimming the pergola lights etc) I went to bed and decided to finish watching something I have stored on my NAS. Could not access the WIFI on the phone... gave up.

      This morning I woke up to hear a relay clicking - as the various ESPs (6 of the, running) were trying to log back in - on 5 of them the solution is to reboot - I've abandoned that in latest software as rebooting really isn't necessary). I tried rebooting the router - made no difference after they initially logged in but when I tried controlling my shelf light it felt "clogged" and slow to respond. I'm sitting here typing this on my (hardwired) PC no problem with connection there... but just now I've just checked the phone - so last night - the WIFI merely would not connect - you could see it on the phone - but it would say connecting - then "saved" repeatedly and would never actually connect.. This morning I'm sitting watching the phone as the two access points (same router has options for multiple points so I have one for phones etc and the other for ESPs - didn't help) actually DISAPPEAR and re-appear on the phone screen as if someone was putting a metal screen over my router and taking it off again. Yet nothing has changed since yesterday when all was working just fine all day.

    2. Craig, I'm not sure wireshark is a wise idea. I tried it and found that my PC and HP wireless printer are having a very intense relationship behind my back!! What are they talking about I wonder.... Should my ears be burning.

      Seriously though 95% of the traffic over a few minutes seem to be just between these two machines and I wasn't even trying to print anything!!

      I'll try again later to see if they've agreed something between themselves and have shut up!

      Simon

      Will also see if I can filter out only the ip addresses I'm interested in.

      1. Yep you will find that there is a lot of handshaking and hello traffic between various devices. You will probabyl find you have the HP App for monitoring printer health on your PC and hence it is continually chatting to the printer asking it how its ink levels, etc etc are.

        Craig

  22. Good one Simon - yes you can filter on source and desitination IP, MAC address, Protocol etc etc

    In Peters case what would be interesting to start with would be to capture from a single ESP when in "good working" mode.

    He can then set a filter on the ESP vendor portion of the MAC address and capture all the ESP' talking back and forwards and see if there is any that is remarkably different as a starting point for some serious troubleshooting.

    It is quite fascinating when you do some random captures on a network and see some of the conversations that are going on behind the scenes between different devices.

    Craig

  23. Pete, I saw exactly the same thing recently when we were on vacation. At the campsite we were staying at there were 2 access points about 100 metres apart and some more in the campsite next door.

    We had terrible trouble getting decent access and when I viewed things on my tablet I saw exactly what you described. The signals going off and on. Makes me wonder if there isn't something in the algorithms to shift band when there is a clash that gets stuck somehow. Having said that you'd expect that a power down and restart would get the router out of that kind of trap.

    I'm no expert on routers, but it makes you wonder if that functionality isn't in a wireless chip and if so would be present in many types of router.

    One other thought - could you set up a Pi as your wifi router? Connect it to the wired hub and make it an access point. I think I've seen instructions somewhere on how to do this.

    Simon

    PS my PC is still in an intense relationship with the HP printer - pretty bizarre when they are both idle...

  24. Having a lazy afternoon here in rainy cold Blighty Pete and thought I'd do a bit of Googling...

    A number of sites from router manufacturers suggest first setting a fixed channel number instead of using auto, not sure what you are using without re-reading all the comments etc.

    Then there are a few that mention changing MTU size and fragmentation threshold. In fact here's a site that gives a few other settings to change as well. Not sure about the veracity of this of course but worth trying these out. (I also found a method for determining MTU size on one of the sites, Netgear or Belkin etc. and one of them had a different value for fragmentation)

    Anyway here's the site and the relevant settings section:

    http://www.werockyourweb.com/your-wireless-connection-dropping/
    Adjust your wireless router settings
    Lower the MTU (Maximum Transmission Unit) from 1500 to 1492 or less (usually found on your router’s main/ basic setup page)
    Adjust the following advanced wireless settings:
    Lower the beacon interval from 100 to 50
    Lower the fragmentation threshold from 2346 to 2306
    Lower the RTS threshold from 2347 to 2304
    That should solve your wireless connectivity problems and should prevent your wireless connection from dropping.

    Worth trying some of these for a while to see if things improve.

    Simon

  25. Before I do any more, chaps.... mid-day I turned off the WIFI on the Draytek - I then set up a TP/Link to run WIFI "wififorus" only - fixed address outside of the Draytek dhcp range. Turned off TPLink DHCP..... started up - not one failure since - with everything on..... that means NOTHING as I've had a couple of days without issue - but fingers crossed. If it falls over in the next 24 hours or so I'll start on Simon's comments.

    This does however lead me to some conclusions - no problem with ESPs, no problem with Raspberry PIs... as the problem was occurring with them all turned off. Good job really as I've spent months mastering the ESPs!!

    More as it happens.... or not.

  26. Pete, in your latest blog on node red restarting you mention that you seem to have solved the WiFi issues.

    I'd be interested to know what you think fixed it and the steps you took. Could be useful for the future. Thx.

    1. Well, not so much solved as it went away. So what I did was, took the Draytek back to factory settings - that didn't seem to solve the problem. I then turned off the WIFI on the Draytek altogether while keeping it as the DHCP controller... and put the WIFI on a separate TP-LINK WIFI access point. I made a DHCP pool of 50 (starting at 192.168.1.25) and put the access point, PC and other stuff on fixed addresses (by MAC address) in that 1-24 range. That seemed to do the trick - as far as I'm aware I've had no WIFI issues since then. Certainly now I'm back in the UK, the Spanish unit is continuing to grab temperatures etc from remote ESP8266 units.

Leave a Reply

Your email address will not be published. Required fields are marked *