[Lonny Knudson]

Hi Matt,
Here is the ‘quick answer’ to your questions:

1. I’m not sure what you are asking. The SocketModem pins are designed to plug into a socket or be soldered directly to thru-hole pads on the host PCB. (Not sure what you mean by ‘standard jumper wires’)
2. This is the manufacturer/part number for the sockets we use on our developer boards: Neltron Industrial Co. (http://www.neltron.com.tw/)
2209S-XXG.
3. I believe Q4 2015 or Q1 2016, contact sales for the most current information regarding Dragonfly release dates.
4. This depends on how the Dragonfly will be used. If you are just using the LTE radio to bring up a ppp link on the Raspberry Pi you only need the USB connection (and power).

If you need more detailed answers, I recommend opening a support ticket at http://support.multitech.com

[Lonny Knudson]

Hi Eric,
Try using AT&K0 to disable flow control. If this doesn’t work, I believe early firmware on the -H5 series modems needed RTS forced on even when flow control was disabled. Try forcing RTS active and see if this resolves the problem. If so, you can either use a custom cable that forces RTS active or update the firmware on your modem. If you would like to update, open a support ticket at http://support.multitech.com and we can get you the firmware and Windows update utility.

[Lonny Knudson]

Hi John,
The most common cause of reading garbage back from a serial port is serial port parameter mismatch. The first thing that comes to mind, if I remember correctly the F2 and F4 series modems shipped with autobaud enabled (+IPR=0). The -H5 modems are set to 115200. Is it possible that you had putty set to 115200, but your custom application was setting the serial port to a different rate?

[Lonny Knudson]

Hi Herraiz,
The command to check registration on the -EV3 is AT+CREG?
A response of +CREG: 0,1 or +CREG: 0,5 indicates that the modem is registered.

Here is a link to the Verizon activation procedure for -EV3 products:
http://www.multitech.com/activation-guides/verizon-wireless-activation-procedure-ev3-and-c2-devices-s000606.pdf

Here is a link to the AT command manual for -EV3 products:
http://www.multitech.com/manuals/s000546.pdf

[Lonny Knudson]

Hi Brian,
It’s possible that something changed as the F2 and F4 modems used a radio from a different vendor than the h5 modems. If you still have the source code for the custom application, could you try making the application force the RTS signal on, then see if you can then read data back from the modem? If this doesn’t help, I suggest opening a support ticket, include a description of the custom application including any initialization commands the application sends.

[Lonny Knudson]

Hi Andrei,
There is a known issue with some Rogers SIMs in the -F4 radios. Please open a ticket in our support system here:
https://support.multitech.com

[Lonny Knudson]

Hi Siva,
The MTSMC-H5-IP is intended to be used as a simple device server using an internal ppp implementation and IP stack. By default the device server application will respond with ERROR to the ATD command. There is a pass-through option (AT+WOPEN=5) that may allow you to bring up a network link using pppd, however we don’t recommend using the MTSMC-H5-IP in a Linux environment (the MTSMC-H5 is the correct choice).

[Lonny Knudson]

Hi Sol,
Have you considered establishing the network link using pppd and using python on the OCG to manage the SSL link ?

[Lonny Knudson]

Hi Sol,
With the current radio firmware the answer is no. May I ask why you are using the Telit IP stack rather than the Linux stack to open the SSL connection?

[Lonny Knudson]

Hi Sol,
The python implementation in the current Telit firmware does not include a module for python to communicate with the USB interface on the radio module. The OCG hardware does not connect to either serial interface on the radio (SER or SER2). So, currently there is no way for a python script running on the radio to communicate with any interface shared with the OCG hardware.

I have been told that future firmware will include a module to allow python to communicate with the USB interface, however I don’t have any idea when this will be available.

[Lonny Knudson]

Hi Thierry,
The MTSMC-G2-IP does not support autobaud so you must set your host system’s baudrate to match the MTSMC-G2-IP. The factory default setting is 115200 and this can be changed using the AT+IPR command.
Regarding voltage levels, you should be able to successfully interface the 5V MTSMC-G2-IP with 3.3V host system logic levels. The modem inputs (TXD, RTX, DTR) need to be pulled to >.8V for a logic low, and only need to be driven to 2V for a logic high. The voltage divider you describe on the modem outputs (RXD, CTS, DSR, DCD, RI) looks OK as long as there is no significant load presented by your device. You may want to verify signal levels with a scope to make sure they are within tolerances of the modem hardware and your host system hardware.

• This reply was modified 9 years, 10 months ago by Lonny Knudson.

[Lonny Knudson]

Hi Lucas,
Looks like you posted a 2nd question while I was typing my last post. The H5 modems do support 2G and 3G services. By default the -H5 modems are shipped with 2G and 3G both enabled. If you are having trouble connecting with one SIM card and not another, you may want to open a support ticket.

http://support.multitech.com

[Lonny Knudson]

Hi Lucas,
In the AT Command manual search for the #PSNT command. This command returns data on the type of network service currently in use.

Search for the +WS46 command if you want to disallow either 2G or 3G services.

[Lonny Knudson]

Hi Mark,
As I understand the issue, periodic re-registration will prevent the SIM from being barred. This will need to be done as long as a SIM is installed and the cellular radio is powered up.

[Lonny Knudson]

Hi Mark,
So the automated reset is to circumvent ‘fixed terminal’ detection algorithm? If re-registration is the event that prevents your SIM from being barred, then you will need to manage this even if you aren’t using the cellular modem as the primary network link (the cellular radio maintains registration even when not actively used for a data link). You can force the cellular modem to re-register with a reset or by issuing AT+COPS=2, then AT+COPS=0.
If you are in contact with Campbell, I would recommend following his advice regarding this issue, I know he has worked on projects where the cellular carrier has implemented ‘fixed terminal detection’ in the past (I’m in the US and carriers don’t seem to do this here).

[Lonny Knudson]

I don’t think it matters much if you are using the cellular service as a primary or secondary network link. In general, I wouldn’t recommend rebooting the wireless modem unless there is real problem detected (no AT command response, unable to obtain cellular service for an extended period of time, etc…). What was the reason for implementing an automated radio reboot every 4 hours?

[Lonny Knudson]

I don’t recall any reports of similar behavior for this product, however this model has been obsolete for several years. A hardware problem could explain the events captured by the serial port monitor software.
I noticed that the software appears to be monitoring ‘COM25’ which I would assume is a virtual COM port of some type (maybe a USB to serial converter?). You may want to try connecting the modem to a PC with a standard serial port (COM1 or COM2) and see if you can duplicate the behavior, this could help further isolate the problem between the modem itself and the virtual COM device. This would be a good test if the problem goes away when you reboot the computer without removing power to the modem.

[Lonny Knudson]

Hi Ben,
In your peer file, could you try changing ipcp-max-failure to 10 and see if this allows the ppp negotiation to complete successfully?

[Lonny Knudson]

The default image should include the file /etc/ppp/peers/cdma. This will most likely work without modification if you specify /dev/ttyUSB0 on the command line.

Example:
pppd /dev/ttyUSB0 call cdma

[Lonny Knudson]

Hi Billy,
It looks like the cellular provider is assigning your mobile device a private IP address. Addresses in the 10.x.x.x range are not accessible from the internet. I would recommend consulting your cellular provider and telling them that you need to support a mobile terminated data application and see if they can assign your device a public IP address or provide you with a VPN connection that would allow access to the private IP.

[Lonny Knudson]

Hi Billy,
I’m assuming that you are able to successfully establish a ppp link using the MT100UCC-G2. If this is correct could you issue the following commands to the linux box while the ppp link is up and post the output:
ifconfig
route

If the above is not correct (you are not able to establish a ppp link using the MT100UCC-G2), then we would want to look at your chat script, peer file and any command line options you are using when you start ppp.

[Lonny Knudson]

Hi David,
If you need to access your device from the internet, you will need to work with the cellular provider. The log showing your ppp connection shows that your provider is assigning your device a private IP address which is not available on the public internet. Your provider may be able to assign your account a public and/or static IP for remote access applications.

[Lonny Knudson]

Hi David,
For the -H5 devices you should be using the gsm peer file and chat script. You will need to modify the gsm chat script so that the +CGDCONT command references the APN you need to use for the Vodafone account.
If you have trouble when using the gsm files, please post a copy of syslog events during the attempt to start pppd.

[Lonny Knudson]

Hi Terry,

Here are a couple of suggestions:

After setting #NITZ=7, save it to memory with &W.

Also, keep in mind that the cellular radio needs to receive a NITZ frame with the ‘Universal time and local time zone’ field populated before it can update the RTC. The network controls when NITZ frames are sent and which fields are populated within the NITZ frame.

In the US, the carriers appear to always send a NITZ frame about the time the radio registers on the network, and then periodically (~ once an hour or so) after that.

[Lonny Knudson]

Hi Anshuman,

Regarding the items in your post:

1. If you are using an IP stack and ppp implementation that is running on your microcontroller, you will need to learn the IP from your IP/PPP software when the ppp link is up. If you are using the IP stack that is integrated in the cellular radio you can query the IP address using the command AT+WIPBR=3,6,15 after you start the GPRS bearer service.

2. For GPRS connections, IP assignment is performed by the network. If you want to poll your application over the GPRS link, you will need to use the address assigned by your service provider.

[Lonny Knudson]

Hi Ron,

There’s not much you can do to make pppd manage itself, the persist option will make pppd automatically re-connect if the ppp layer goes down, and failed lcp-echo requests will tear down a presumably bad link. With cellular devices, lcp-echo requests are not likely to fail unless the cellular radio itself is in a bad state.

Your approach (having some process check the IP layer) is very common for cellular applications. It’s a good idea to keep in mind that the IP layer checks use metered data and could have a significant data usage impact on low bandwidth accounts. What action does your ping check take in the event that it doesn’t get a response from your server?

[Lonny Knudson]

Hi Steve,

The only known resolution for this issue is to get Rogers to issue you a 3G (non-LTE) SIM card. The cellular radio (SierraWireless/Wavecom Q24Plus) used on the -F4 products was obsolete before this issue was discovered. If you have any trouble getting someone at Rogers to issue a 3G SIM, please open a support ticket ( http://support.multitech.com ) and we should be able to provide some additional assistance.

[Lonny Knudson]

Hi Toyo,

When you see the CONNECT response, this means that the interface has switched to data mode. When in data mode, the modem does not echo characters back to the host. This is expected behavior.

The HTTP request should be terminated by 2 <cr><lf> pairs. A single <cr><lf> pair is used to delimit the http command and individual headers. I suspect that the http server is closing the socket (producing the NO CARRIER response from the modem) because it was expecting more headers. The ‘Host:’ line in your request appears to be terminated with one <cr><lf> pair rather than 2.

[Lonny Knudson]

Hi Toyo,

If you get a CONNECT response to the #SD command, the socket to the http server should be open.

Could you post the contents of your HTTP request?

[Lonny Knudson]

Hi Toyo,

Could you capture a log of your terminal session showing all commands you are sending and the responses from the Socket modem. I recommend issuing AT+CMEE=1 at the beginning of your command sequence, this should give us more detail regarding the ERROR.

The IP stack has a firewall built in, by default inbound connections are blocked. You will need to add a firewall rule to allow inbound connections (see documentation for the #FRWL command for more details). I would recommend getting the client connections working before moving on to the listen mode issue.

[Author]

Posts