At the wiki modem, all channels are “2” and the ppp-client info function is permited when ppp-client is down.
On E156 (and perhaps on E220, E620 etc) , using:
Command Channel: 2
Data Channel: 0
Info Channel: 1
is possible to use ppp-client info with ppp-client up or down typing “int ppp-client info ppp-interface”. I’m not sure if this channels numbers are really valid to E156. Those numbers was simply the combination which allowed me to connect - and - use the ppp-client info at the same time. I did not understood why this function is restricted to a ppp-cliend down state since we have 3 different channels as explained on wiki. If all channels was the same I agree. If they’re different they can or must be used simultaneously.
Now the “issues”:
My init string is ATZ only and it’s running ok. If I try - using Winbox - to type the standard AT+CGATT=1 etc the modem freezes. Is necessary to dettach and attach the modem from/to USB to bring it back to life. It’s not enough to “undo” the action neither reboot the router. Modem hangs.
Every time ppp-client disconnect and reconnect, there is a speed peak on graphing (reaching 480Mbps even on a USB 1.1 interface) messing up the graph. I cleaned ppp interface graphing data at first because I thought this peak was due to sucessive USB attach/dettach durying my tries to figure the USB channels but it was not. The peaks occur when ppp-client reconnect normally.
The modem is locked to 3G only, the frequency band is locked too. But ppp-info shows some wrong data with ppp up (my tries) or down (wiki instruction):
[mramos@WRAP-3G] > int ppp-client info 3G_NET
pin-status: “READY”
manufacturer: “huawei”
model: “E156”
revision: “11.604.09.01.149”
serial-number: “356368020595709”
activity-status: “call in progress” (no, it’s already connected or ready if disconected)
functionality-status: “minimum”
gprs-mode: “B - EDGE” (no, should be (some letter but A or B) - HSDPA)
current-operator: “3G_NET”
signal-level: “-95 dBm” (this is the RX noise floor on 850 or 900 MHz band, not the true 3G 2100MHz received signal, which in my case is -78dBm)
And now with the different aproach to 3G USB modems on ROS, it’s not possible anymore sniff modem info using serial-terminal on USB nor send AT commands to do some tests.
BTW, with this “ppp-client revival” after - Dial up => GPRS, EDGE, EVDO, 3G upgrades - is it possible to add more init string lines to gave us more flexibility? Or there is any character I can use to put several AT commands at the same line?
I have been eagely waiting for the RB433UAH and am currently doing some testing with USB cards for HSDPA/CDMA-2000, but there are some issues that prevents the solution from being good enough. Some of my problems/wishes are close to yours.
I have also written to support about this.
I am now testing with the Huawei E200 and the Motech CNU-680 (CDMA-200 EVDO)
For the Huawei I use init string from wiki, for the Motech this doesn’t work so I use simply at&f.
I experience the following:
Sometimes one of the ports get invalid without reason and the corresponding PPP client therefore stops working.
For the Motech, like I told you it sticks with status “reseting link”. Most times the cure for this is pulling out and reconnecting the Motech, and then i re-initializes.
For the Huawei, a router reboot is required.
After power-cycling, the port for the Motech is often invalid, and a re-plugging (sometimes twice) is needed.
I have seen that after a reboot, one of the ports is back at 115200 bps, and the command baudrate has to be set manually to 9600 for the PPP client to work.
From time to time everything seems ok, but trafffic cannot be sent on any of the interfaces. Router reboot required.
These issues stops the solution from being something useful for non-technical end-users. They simply want to connect “the box” to power and they expect it to work at all times without touching it.
Therefore it should recover from all error situations itself!
A method for detecting issues like I describe and re-initializing the ports/devices is badly needed. What about the possibility of pinging the respective gateways and then re-initializing the ports if not reachable for a while? It should be possible to “pull the plug”, simulating a power-cycle or total reset of the USB devices?
Can something of this be scripted?
Regarding AT command set: I guess that apart from standard Hayes, these devices has their own extensive AT command set, including a bunch of proprietary commands? Any command that can be used for monitoring/detecting issues?
I am very interesting in some kind of “technical reference” for the mentioned USB cards if anybody can point me to it.
Perhaps the USB virtual serial channels are brand specific and since mine “cut-and-try” procedure worked I left it this way from 3.23 to 3.24. So I still able to watch the modem status without hung it up. I agree that’s not MT mission to keep a ww modem database 'cause you already said, besides standard AT commands each manufacturer has some proprietary ones and this means 100% compatibility an endless task.
Ok, may be each AT query that MT uses could be “user defined” somewhere … but as I pointed somewere on my original post the “EDGE” result on E156: may be E156 answers to this query (the debug log shows it) in a uncomplete format. So if MT followed the standards, the “EDGE” translation to some number that E156 sent back is ok and E156 is the problem.
Mine used to freeze but since I reach this virtual ports setup it did not froze anymore, except for some 3G providor error on their network core. Since my post I had only 2 of them.
At the start I used to feed the modem with the regular USB power and it hangs often. Now I’m using an external 5V power injector with a linear voltage regulator at one female USB connector, plus a mix of capacitors to really have at the +5V bus a clean voltage and low AC of any frequency. This voltage regulator have 12Vdc at its input (my config).
My init string still ATZ only. Most of the info the wiki sugests in my case is stored at the E156 flash so I don’t need to send it each time I need to connect. But depending on the modem more init lines may be usefull to set up the modem to an specific condition (e.g. WCDMA only, 1.9GHz band, automatic network registration, etc).
I don’t have this “invalid port” behaviour on E156.
The “reseting link” endless state I had twice but here was a dirty USB connector at the hill top (wet place). So a contact cleaner spray gave me some weeks of stability (I’m changing this HW connections and sealing them accordingly using and old Scientific Atlanta CATV amp aluminum housing with a WRAP + modem inside).
Router needed to be rebooted once I guess due to a problem at the 3G providor core network. I don’t know what causes that but trace route show that suddenly providor network stopped answering pings from an specific HW on the path and then I wait 5 minutes and decided to reboot the router. Than things came back to life (5 minutes is a loooong time for a network which suposes to run over an SDH ring or a metro ethernet with rapid spanning tree).
I don’t have this port problem neither the speed (as far as I know E156 have some auto-config for serial comms from 2400 to 115200). Neither do I have this traffic hang.
Mea culpa, I’m vy far from 3G cells and even using a 1.5m dish (29dBi) I know that the chance of this antenna receive more than one cell exists at the same azimuth or reflected here and there, knife effects at hill tops, propagation ducting and even rain or airplain scatter. So may be the modem sometimes “see” more than one signal with similar field strenght. But I guess 99,9% of the 3G users are in regular covereage area so may be some problems I face here are particular ones.
Yep. I guess need to focus on 2 or 3 modem models, stress them up and only uses this modems untill you “certified” other brands. E156, for instance, still on the market and have external antenna connector. But its costs are prohibitive here. 3G providor bonus programs offers a ZTE one, without external antenna jack. Enough for most users, useless for me.
The main problem is the modem power cycle. USB voltages are always there (I asked somewhere for a HW pin, may be a USB to paralell adaptor tied to some device (relays) that can cut the +5Vcc to modems and set it back. And then using netwatch or any other script we can simulate the modem unplug-replug. Not elegant but sometimes …
Actually my modem is away from home. I guess I will build a HW using a microcontroller + some IP interface and put it at the same cabinet (something like a parallel over IP interface) to send commands to HW via TCP-IP). Much easier if MT creates a parallel interface resource activated manually, via script commands or using scheduler).
I use part of netwatch sugested by MRZ at wiki to disable and re-enable ppp-client if modem stops receiveing pings back from a router at the core of 3G net (2 or 3 hops away) and it’s working fine. My initial target was something based on the ROS traffic-monitor feature because the problem I observed was not the ROS stop sending data to modem but the modem not sending anything back. So I wanted to watch only the receiveing path avereage data, put a trigger (near zero) and when this condition was ok starts a counter which can be reset by any traffic. If not, it disable ppp-client and re-enable ir again. I simply want to avoid pinging all the time some address that can be down by any reason so “any” received traffic means that ppp-client is up.
I tried at Huawei forums … the manufacturer said: “its proprietary” … lol … so I tested most common ones that was enough for my needs (e.g. do not use windows gui to lock the modem at 3G on 2.1GHz band).
The signal level issue: in fact the signal I read (-95dBm) is the real signal +/- 6dB. Modem stops working with -113dBm. And most modems does not tells signals in 1dB steps but in ranges, e.g. some number means something between -113 … -105, then -104 … -99, etc …
Besides that they update this readings from time to time which does not means an update per second. Useless for my needs (aim the antenna). My workaround: I picked up and old 25dB gain MMDS downconverter, re-tunned its filters to 2.13GHz and connect its output (now 108 to 168MHz) to a home made 0.1 to 500MHz power meter (-80dBm to +15dBm). Associated to the downconverter gain this setup is able to measure the 3G cell downlink signal if bigger than -105dBm. Practical, isn’t?
Well, actually ppp-client is up 24h/day … except 2 or 3 outages along the day that costs me 30 seconds each (hung up => dial out => auth => connected) without internet connection. And the spikes at the graphing tool has gone on 3.24.
do you have some configuration informations about the C-Motech CNU-680 USB-Modem for me?
I have the 3.27 software running on a x86 system and have connected the CNU-680 to USB 3.
With the command “system resource usb print” I see that the modem is connected.
But I need such configuration data like:
I guess first off all must see if this modem is suported by MT. If not, send a support file to MT guys.
Second: at system > ports (e.g. using winbox) take a look to see if there is a serial one named USB3. Set it to 9600 8 N 1 , no flow control (mine have not flow control).
If the info channel is correct using a terminal you’ll be able to see some modem info, like this:
[mramos@WISP-3G-PTP] > int ppp-client info PPP
status: “call in progress”
pin-status: “READY”
functionality: “minimum”
gprs-mode: “D - WCDMA”
manufacturer: “huawei”
model: “E156”
revision: “11.604.09.01.149”
serial-number: “356368020595709”
current-operator: “3GPPP”
signal-strengh: -87
If not, try some Data Channel / Info Channel changes (e.g 0/1, 1/0, 0/0, 1,1 etc). My modem data channel is 0 and info channel is 1. On this way I can check modem status even when connected.
Modem init, in my case is simply ATZ
Dial command is ATDT and the number vary from system operator.
I don’t need username nor password (all built in at GSM chip) neither pap/chap. But this can vary from op to op. Take a look under any modem Windows app (proprietary dialer, if any): normally there is some “profile config” that gave some info abt pap/chap, APN, dial info, etc.
The Modem ist supported till 3.22.
The USB3 Port is also OK.
I have tried now different values for the Data and Info Channel (0/0, 0/1, 1/1, 1/0) and with some I get no output with the "int ppp-client info ppp" command, and with some I get this:
[admin@MikroTik] > int ppp-clie info ppp
status: "unknown"
pin-status: "ERROR"
functionality: "unknown"
gprs-mode: ""
manufacturer: "unknown"
model: "unknown"
revision: "unknown"
serial-number: "unknown"
current-operator: ""
signal-strengh: 17367040
Here the system resource usb:
[admin@MikroTik] /system resource usb> print
