digital data through cell phone

[bearcreek]

I am trying to send data through a digital cell phone using a microcontroller. My phone's data mode has been disabled by the phone company and I would prefer not upgrading to a more expensive service plan and/or phone. I have a miniature dialup modem connected to it and have successfully connected directly to other modems, but when I acoustically couple it to the cell phone it makes unusal chirping noises like it can't reliably connect. I suppose I could program the microcontroller to do a slow simplex type communication, but I am hoping for full-duplex. I hate the new digital phones. Packet delays and such. Is it possible to use a dialup modem over a digital voice channel, or should I just try for simplex?

[TMorita]

Ummm...do you understand the difference between simplex and duplex?

Duplex means the data is sent in both directions, and simplex means the data is sent in only one direction. Strictly speaking, duplex and simplex do not imply a change in modulation technique, and you're probably having problems with the compression algorithm used for phone communications interfering with the modem's data encoding scheme.

It's been a while since I played with modems, but if I remember correctly, the faster modem speeds use a phase-shift encoding scheme which is very isochronally sensitive. This is obviously very bad if the phone is using a time-domain multiplexing scheme.

If you use a slower modem baud rate which uses a frequency-shift modulation scheme instead you'll probably have a better chance of it working.

Toshi

[bearcreek]

I know sorta what duplex is, but simplex I wasn't sure if it was one direction only or one direction at a time. The reason I mention simplex, is that I am considering just using a really slow tone generator from a microcontroller at say 300 baud to send a small amount of data like gps coordinates over a prepaid CDMA cell phone in one direction (simplex).

I was hoping to use a modem, but I guess I'll build something.

[evdo]

All digital cell phones use some type of speech coder to compress speech, it's not a simple analog connection like land lines. Since the speech coders are optimized for speech and not modems, it would be tricky to get any type of reliable data transmission.

This said, there are some frequencies as well as lengths of tones that would be passed fairly well by the speech coder depending on the system you are using. If you were to build your own modem and experiment with frequencies/lengths of tones you may have some success. A simple modulation scheme such as FSK would need to be used- phase information is trashed when coded. Single frequencies are also preferably to multiple tones (some speech coders don't do a good job passing DTMF tones, for example)

I'd guess that 300 baud would be pushing it, unless it's 300 baud + a ton of error correction on top of that.

-Tom

[jeffs555]

You should google for APRS and TNC. There are lots of designs for these low speed AFSK simplex and half-duplex modems. Some are done with just a microcontroller and a few resistors, and there are also programs for the PC that just use the sound card. APRS (amateur packet radio service) generally uses these low speed modems to communicate weather info, GPS coordinates and other low speed data over ham radio. Haven't seen of any mention of them being used over cell phones so not sure they would work, but I don't know why they wouldn't.

[mjsas]

A few years ago I did something like this with a cordless phone using it as a robot controller. The base station was the transmitter and the handset was the receiver in the robot. Since the phone cannot send a DC signal the encoding cannot have a DC level or component. I used Manchester encoding, each 0 is replaced by 01 and each 1 is replaced by 10. I forget the data rate but it was higher than 300 baud, maybe 9600 baud. This is very close to what auto remote entry systems use. Also cheap RC toys may use this type of encoding so the chips are available.

[lazy-b]

Guys, for a very slow data rate.......I would rather use a Touch-Tone system.....Touch-Tone is the Most reliable way......can be use in Cellphone, or even in any two-way radio.....very easy to interface,just though a Speaker and Mic......and a Touch-Tone Encoder and Decoder ICs.

If your Cellphone has a Built-in Modem and a Serial Interface.....its much easier to interface, But you need to know how to control that Modem using "AT command set"

You can but a Finished Product Cellphone with serial interface. just insert a SIM card, and you can use any controller or even a Window XP PC using a simple dumb terminal to control Cellphone.

[bearcreek]

lazy-b,
my phone does have an internal AT serial modem and I tried to use it, but my telephone company is not supporting the data service. I am sure that for a hundred dollars or so I could get a cell service with the serial modem, but I think it would be fun to build a very simple modem like tinytrac. Plus I have over a 1000 unused minutes that I need to get rid of. The phone is CDMA, which I have heard will not work reliably with DTMF tones.

As an experiment I connected an old amateur slow scan television unit to the microphone of the cell phone and received it on my landline. The picture was okay though it had a few specks of noise. Not bad, but I was impressed that all the colors and detail was there considering the digital compression. There are over 1000 pixels per second being sent with SSTV, so it seems that 300 baud shouldn't be difficult with small packets and error checking.

[mjsas]

There is a good book available from www.rfsolutions.co.uk, titled "An Introduction to Low Power Radio". It talks about how to encode data to transmit it over a link that cannot pass a DC level which is any telephone or radio link. There is not much difference between a telephone link and a radio link.
Another good source is www.microchip.com - search for their KeeLoq products.
You should be able to interface to the phone via the headphone plug.

[lazy-b]

Bearcreek: That very Strange that you can not send a Touch-Tone in a Voice Band, I know some of our CellPhone Network here in the Philippines, They Intentionally suppressed the Touch-Tone due to some people use it for Data communication or use it for Phone Patching to a Land Line.........But If you use a Regular Rate Cellphone Network you should able to send Touch Tone.....Like in Telephone Bank Inquiry they all use Touch-tone for getting the inquiry.

I know there are Cellphone Network Service that give free unlimited for Cellphone to Cellphone...........But there are some clever guy made a Phone Patch , by connecting a Cellphone to a Landline.....Actually I have Done it.....by connection it thru a Speaker and Microphone........Now, any people can call any Landline for Free....People have to use cellphone to call another cellphone, once its connected, you will get another dial tone, then you just keyin a new no. to call any other landline for free.

[evdo]

Some cellphone systems can pass touch tone (DTMF), some cannot- CDMA happens to be one that has issues. The typical problem is the DTMF detector either detects what is a single tone multiple times or sometimes misses tones completely. It works sometimes, but certainly isn't reliable..

A digital cell phone conversation only uses around 4 to 5 kbits/sec on average to transport the compressed speech, as opposed to a cordless home phone which has much higher data rates (usually >=32kbps) With digital cellular the high amount of compression will lead to nasty distortions depending on the source.

Onstar did have some success in the past with in-band FSK modems, so your on the right track with the SSTV-type modem, it will just need some tweeks to get it working reliably.

[jeffs555]

This paper describes a method for carrying TDD/TTY communication for the deaf over cellular phones. It describes some of the problems with TDMA and CDMA, and says that frequencies above 1kHz are problematic. The paper describes coding 2 bits into four frequencies - 400,600,800,1000hz. TDD is only about 45 bps. http://www.ee.columbia.edu/ln/labrosa/proceeds/icassp/2001/main/papers/pap651.pdf