Home AMX User Forum AMX Control Products

MVP-5200i Wifi indications

Am wondering if anybody else is having problems with the Wifi links as the Touch Screen Roams? We have 4 access points around the house. All channels are seperated as best we can (11, 1,6,11). The annoying part is the 5200i indicates it has reasonable signal level and signal quality (about half on the levels bars) but it stalls and seems to display old DATA and info when buttons are selected. Worse is that the NI-4100 seems to stall trying to require the Touch Screen. When the link is broken between controller and 5200i it takes 30 - 40 sec to recover and while that isn't long - when the client is standing there trying to do something it really doesn't look good. It appears that signal and quality bars need to indicate more than 75% for it to work well or aren't these very accurate?

I've wandered around with network stumbler and we have reasonably good coverage.

The frustrating thing is that if we leave the 5200i in the drawer (where coverage is good) and link into it using an iPhone, we have much better reliability. In other words the iPhone seems to have better Wifi reception.

Am I expecting too much here?

Is there a way of indicating the link has dropped out on the Touch Screen?

Comments

  • viningvining Posts: 4,368
    Well I was going to send you to this link:
    http://amxforums.com/showthread.php?t=6627
    but it appears you've already been there, done that.

    Unfortunately for some reason beyond my comprehesion these will never work as good as a laptops, iPhones or any other significantly less expensive device. God only knows why. The best you can do is try and give them the strongest signal in the areas of prime use so that in these areas there will always be a clearly stronger AP to connect to. Then as you roam which for the most part they won't do to much of there will be fringe areas that are weak or where signal strenght of 2 or more APs will at the same levels in which case the TP's connection logic just farts and can't decide who to connect to. I think it then runs the fnInyMenyMinyMo(INTEGER nLevel) function but without fingers or toes it has a hard time.

    You really need to get a WiSpy or some other spectrum analyzer and see what evil 2.4ghz devices are in the house and possibly stepping on your wi-fi too. Pana Cells, Microwaves and a host of other devices could be really screwing things up. So why does the iPhone work so well or the laptop?
  • I thought I'd done really well sorting out the wireless network after your advice a few weeks back - mobile devices worked well. The 5200i and NI-4100 seem to have a problem 'getting it back together' after the link is broken (well in this case anyway). It may well be that if we'd waited a couple of minutes they would have sorted themselves out. Unfortunately I given the client technician free setup - basically the ability to adjust and tweak everything (lighting circuits and other stuff) via the touch screen, so of course when it kept on selecting the wrong room and circuits (probably because that's what was buffered in th TS) it was deemed unreliable at this stage. :( And yet when I tested it, it worked perfectly!!

    I'm working on it though - I figure if I can proved a better indication of a failing link (PopUp) at least they can see theres an issue.
  • DHawthorneDHawthorne Posts: 4,584
    Even laptops and iPhones take a little bit to reconnect when jumping access points; you just don't notice it because the nature of the communications. Your laptop isn't constantly connected to some server, usually; it goes out when it needs to, gets its data, and drops the connection. A panel connects and stays connected, and all sorts of stuff goes on when that connection is re-established. The connecting process itself is probably about the same, its the overhead due to the nature of the connection that causes the issue.
  • viningvining Posts: 4,368
    DHawthorne wrote:
    you just don't notice it because the nature of the communications. Your laptop isn't constantly connected to some server, usually; it goes out when it needs to, gets its data, and drops the connection.
    That's actually probably very true now that I think about it. Unless you look at the wi-fi connection icon in the systray while walking about the job how would you really know how often the connection drops and for how long before it associates with another AP. It's very apprent when walking with a TP looking at the connection button. I know I've never done that and even then who knows what amount of time then signal is lost before the searching sysmbol starts and using NetStumbler might just allow to you to see the APs at levels far below the level necassary for a proper connection.

    Maybe we've just been giving AMX a bad rap for no real valid reason.
  • a_riot42a_riot42 Posts: 1,624
    vining wrote: »
    DHawthorne wrote:

    That's actually probably very true now that I think about it. Unless you look at the wi-fi connection icon in the systray while walking about the job how would you really know how often the connection drops and for how long before it associates with another AP. It's very apprent when walking with a TP looking at the connection button. I know I've never done that and even then who knows what amount of time then signal is lost before the searching sysmbol starts and using NetStumbler might just allow to you to see the APs at levels far below the level necassary for a proper connection.

    Maybe we've just been giving AMX a bad rap for no real valid reason.

    This is one thing I have never understood. When the TP loses its connection why do all the channels and levels go back to their default values rather than just keeping them wherever they are? Then when the TP loses its connection briefly it doesn't reset everything and can continue along without much disruption.
    Paul
  • I wasn't meaning to give AMX a hard time about this. The gear is great! TCP/IP connections are extremely complicated things and am constantly amazed that they work as well as they do. Toss in the Wifi aspect and eek! (I used to do Radars and stuff in the NZ Navy so have a good idea about RF)

    My questions should have been;

    - Are the TS strength and quality level indications accurate?

    - How would I display a PopUp to indicate as soon as the link is compromised? e.g. you can stop pressing buttons now - nothing is going to happen....

    - Is there some thing the Touch Screen can display to show the status of the link when it's retrying with the Controller telling to?

    Client is very technical - he founded a very large computer company early on so want's all the answers now - if not sooner!
  • viningvining Posts: 4,368
    Nice to see a fellow ex squid and being from NZ you probably ran into a few of the big "arcitoothus" type. I'm ex USN myself "electronic warfare" EW1.

    You could put the copy the connection status ball from the set up pages or the Link quality and signal level buttons from the protectedSeconadryIP page (TPD4 >file>Open Sytem Page Template).

    Or create your own buttons with the values used in the set up pages as shown in the pics. Now these level are set to display so maybe if you create your own (or copy/paste) you can change the Level Function to active and then write code to handle these levels.

    If you just put an obvious connection icon on the panels I think the customers will get the idea. Yolu could also make the offline button state the full screen. You would have to make it pass through so the other button on the page would work and you could still push buttons, you just wouldn't see what you're pushing and those button pushes would get queued and run when it comes back online so that might not be a good idea.
  • viningvining Posts: 4,368
    Now these level are set to display so maybe if you create your own (or copy/paste) you can change the Level Function to active and then write code to handle these levels.
    Disregard this statement, it makes no sense at all!
  • We have been hearing reports of wireless problems with some 5200 panels in the field. We are in the process of investigating this. The new 8400 firmware should be available on the web (we released last week) and the new setup pages will give you an idea of what we are trying to do from a configuration/diagnostics standpoint. These same changes will be applied to the 5200 so that will be a start to helping. We will also be trying to duplicate the wireless problems that are being reported and weed out the environmental issues from the real ones. I won't promise a firmware upgrade in a month because it will take some time to duplicate and chase these things. But I will promise that we are seriously looking at it and will have a firmware upgrade at some time.
  • To address the question of resetting values back to default when the system reconnects to the master....
    since the master is "in charge" and updates the values on the panels, if the panel disconnects from the master it has no idea what those values should be. So as a matter of protocol all devices are required to reinitialize their values upon connection and wait for the master to update those values (that's what the online: event is for). For example, if a panel shows the light in the room on and then disconnects from the master and the light is turned off during that time, when the panel comes online we would rather reset all values then display a value that is incorrect. One could argue that most values are incorrect anyway when the panel resets them (the light could still be on, right?) but in most cases it is more obvious that the panel is disconnected when you see these values. It keeps things much cleaner from a protocol standpoint and the NetLinx code should really be updating those values anyway. The master will update all channels and levels automatically without any NetLinx code. But the NetLinx code needs to update the variable text that appears on the panel (the master doesn't maintain these).
  • a_riot42a_riot42 Posts: 1,624
    rgelling wrote: »
    but in most cases it is more obvious that the panel is disconnected when you see these values..

    Yes, I always prefer obviously wrong to just slightly wrong, but it would be nice to be able to turn that feature off when a less than optimal connection takes the panel offline for a few seconds every once in a while.
    rgelling wrote: »
    The master will update all channels and levels automatically without any NetLinx code.
    .

    What do you mean? The master doesn't reinit the panel once its back online without code to tell it what to do in the online event as you mentioned.
    Paul
  • A Site Survey Approach

    Brent, I don't know if RF coverage is the issue for this installation but I thought I'd pass this along. Perhaps there's something here you can use to add to your current Site Survey procedure or to use for comparison. It's a simple technique for doing a Site Survey that "should" give pretty good results:

    (NOTE: You should not have a master connected to the panel for this procedure.)

    1. Start with a single AP, put it in a "good" position that you expect would give good coverage.
    2. Get the 5200 associated to the AP.
    3. Go to the Tools::Statistics setup page and monitor the Signal level value. While walking the room and adjacent hallways (and other rooms, above and below room with AP) define a perimeter where the Signal level is -65 to -70 dBm. You will need to push the "refresh" button to update the Signal level as you walk.
    4. Move the AP to a new position in the house (good guess work is valuable here) and take another survey. Make sure you get some overlap (unless dead zones are in areas where you don't need the panel to operate).
    5. The less APs you use to get coverage the better, more than 3 and you have to deal with interference.
    6. Watch out for "hidden node" issue. This occurs when the panel is between 2 APs on the same channel and the panel can "hear" both APs but the APs cannot hear each other. This will cause interference.

    Hope that helps...
  • a_riot42 wrote: »
    Yes, I always prefer obviously wrong to just slightly wrong, but it would be nice to be able to turn that feature off when a less than optimal connection takes the panel offline for a few seconds every once in a while.



    What do you mean? The master doesn't reinit the panel once its back online without code to tell it what to do in the online event as you mentioned.
    Paul

    Ok. That was a dumb mistake on my part. The master also reinitializes all of its state values when the panel come online. You are right. An online event would need to send these back to the panel. Sorry for the brain fart.
Sign In or Register to comment.