E400 first impressions

Been working with an E400 for a few days and wanted to offer some observations.

firmware 1.6-r20, local UI control only (no cnMaestro)
poor tablet/phone layout, menu missing in portrait mode. no apparent refresh of dashboard data unless I reload browser tab. first time I rebooted from webui led to webui connections being refused (tried 10 mins)
until power-cycled (not reproduced) client list needs to indicate radio/band/channel 2.4 ACS kept picking the worst channel about 50% of the time(chan1 at -45 to -50 per
wifi analyzer, now locked to chan 11) manual 2.4 channel complains that default 30dbm transmit power "exceeds regulatory
max vlaue 25" (yeah 'vlaue', though that's just nit-picking, my main point is
the default power being not allowed) At one point throughput graph showed two spikes apparently reaching
500 billion Mbps... "5000000000 M"... WOW this AP is fast... 'maximized' panels (IE 'wireless clients') need to respond to resize events
(and did I mention refresh would be sorta nice??) came home from work today (3rd day) and 5ghz was MIA...
Dashboard showed radio 'off', config said 'on', tried analyzer in troubleshooting
and was told radio was already scanning. (5ghz set to ACS when no clients so I
figured that was why) 10 mins later still MIA so downloaded support tarball,
disabled 5ghz, restored 5ghz, back to working without reboot.

Just to balance the above concerns, on 5GHz my cellphone (HTC One m9) clocks
235Mbps download via speedtest.net. :)

j

Sorry about layout, I pre-composed that post in a text editor then pasted it, and it looks like it embedded it in code or quotation tags or something.

j


@newkirk wrote:
Been working with an E400 for a few days and wanted to offer some observations.

firmware 1.6-r20, local UI control only (no cnMaestro)

poor tablet/phone layout, menu missing in portrait mode.

no apparent refresh of dashboard data unless I reload browser tab.

first time I rebooted from webui led to webui connections being refused (tried 10 mins)
until power-cycled (not reproduced)

client list needs to indicate radio/band/channel

2.4 ACS kept picking the worst channel about 50% of the time(chan1 at -45 to -50 per
wifi analyzer, now locked to chan 11)

manual 2.4 channel complains that default 30dbm transmit power “exceeds regulatory
max vlaue 25” (yeah ‘vlaue’, though that’s just nit-picking, my main point is
the default power being not allowed)

At one point throughput graph showed two spikes apparently reaching
500 billion Mbps… “5000000000 M”… WOW this AP is fast…

‘maximized’ panels (IE ‘wireless clients’) need to respond to resize events
(and did I mention refresh would be sorta nice??)

came home from work today (3rd day) and 5ghz was MIA…
Dashboard showed radio ‘off’, config said ‘on’, tried analyzer in troubleshooting
and was told radio was already scanning. (5ghz set to ACS when no clients so I
figured that was why) 10 mins later still MIA so downloaded support tarball,
disabled 5ghz, restored 5ghz, back to working without reboot.


Just to balance the above concerns, on 5GHz my cellphone (HTC One m9) clocks
235Mbps download via speedtest.net. :slight_smile:

j


- the local UI not working from a phone is currently a known limitation, we are planning to fix this early next year.

- thanks for the feedback on needing auto-refresh on dashboard (& logs) and the additional info in client list, we'll add this to the planned features roadmap. Channel in the client-list seems redundant though, as it would be in the radio info table and same for all clients on a particular band. We do need to add radio/band there.

- ACS improvements are also planned for the next release in Q1.

- we have seen that throughput spike issue when a radio is disabled and re-enabled (a bug in how we were wrapping around counters). This has been fixed for future releases. Were you disabling radios or wlans when this occured? After some time this should have cleared out (once the APs rolling buffers wrap around), and on a steady/running system the values hsould be good. Is that what you see as well?

- Thanks for taking the tech-support tarball for the 5GHz MIA issue, will reach out separately to you for that so we can analyze what happened.

Nice to see the good throughput number there.

Yes, the radio/band/channel comment didn't necessarily mean all three, just some way to tell from that table where a given client is connected.

The graph spikes did indeed wrap away, and likely coincided with radio restarts when I was configuring.

The bit about maximized panels is a direct observation from chrome on a tablet, where changing portrait/landscape orientation doesn't redraw the panel to fit.  I've not actually tried but expect the same on a desktop if browser is resized.

j

I just noted that if i set any TX-Power value over and including 25db, there is a significant decrease in troughput (AP-To-Client), (RX in Mtik Btest),

And an increase of 100ms in average ping time from the test PC to the RB450G that is right behind the E400.

I took some measurements running mikrotik's btest and pinging the rb450g:

TXPower    Troughput    Average Ping Time
                  (Peak)    
30db          41m/s           130ms
28db          41m/s           130ms
26db          41m/s           130ms
25db          41m/s           130ms


24db          54m/s           20ms
23db          55m/s           15ms
22db          56m/s           15ms

18db          56m/s           15ms

10db          56m/s           15ms

----------------------------------------------

What could be causing that?

Is that on the 2.4ghz or the 5ghz radio?  I wasn't able to set a value over 25dB for the 2.4ghz radio (US FCC), and the 5ghz radio it lets me set it up to 30dB but never goes over 25dB...

j

2.4Ghz radio, Brazil country code in a ROW unit. Channel 13.

Wierd thing is, i went to re-test it now, about 6 hours later so i could try to get some RSSI data from my test laptop, and the interface gives me a "cannot go over 25db" message. I swear i didn't change anything.

Just for clarification:

I don't see the point in using that much power in a bi-directional environment, in which the PC/Laptop upload power maxes at 15dbm@mcs0, and 10dbm@mcs7. (ar9565).

But for the sake of testing i played a little with the power slider while running a Btest.

I think 23db is in the high-end of what's usefull, so that's what i intend to let set for now.