I have a situation where I'm trying to connect two computers via their dial-up modems. I've built a line voltage simulator (9v @ 40ma) into the line that connects them, and I've tested with a telephone to listen and make sure the modems generate DTMF tones, and by pressing keys on the phone, that the modems can receive DTMF tones.

The problem is that when I connect both the modems together, the receiving modem doesn't detect anything from the transmitting modem. There must be some difference between the tones/voltage/frequencies generated by a normal telephone, and those generated by the source modem.

I'd like to measure the DTMF tones generated by the phone, and those generated by the source modem so I can compare them and perhaps isolate the problem. But I'm not sure how to measure a telephone line with an oscilloscope.

The line is made up of a tip-and-ring pair, how can I get a nice measurement of the audio on the line?


From what I understand it should work just hooking up the wires to the probes - at least by using the scope's math function and subtracting one channel from the other. The problem is, I never get any signal at all, except background AC noise. The main things I don't understand is what to do with the ground probe, as the cable doesn't have a ground, and if there can even be any signal without a device attached on the other side of the cable to complete the circuit between the two cables

  • 1
    \$\begingroup\$ Back in the days, we connected speakers to modems and could listen to all the various stages of handshaking and if it sounded right... \$\endgroup\$
    – PlasmaHH
    Jul 22, 2015 at 8:38
  • 1
    \$\begingroup\$ Will just hooking the probe up across the two wires not work? \$\endgroup\$ Jul 22, 2015 at 9:36
  • \$\begingroup\$ I've updated my post :) \$\endgroup\$
    – Kazade
    Jul 22, 2015 at 9:56
  • \$\begingroup\$ Do you have the scope set for AC coupling and a reasonable sample rate for DTMF tones, say 500uS / div? I don't really follow why you'd need two channels, just connect the two devices together and connect one scope lead between the tip and ring. \$\endgroup\$
    – PeterJ
    Jul 22, 2015 at 10:37
  • \$\begingroup\$ Are you connecting the two modems back to back without a phone exchange in the middle that applies the correct line voltages, dial and ring tones? Also what response do you get from the AT commands? \$\endgroup\$
    – jippie
    Jul 22, 2015 at 13:13

2 Answers 2


DTMF tones consist of a Row and Column tone. On many handsets, pressing two of the row buttons simultaneously will generate only the row tone, and pressing two of the column buttons will generate only the column tone.

The columns are assigned to the high frequency group and the rows are assigned to the low frequency group.

The matrix in Hertz can be found at: this Wikipedia page.

In addition, DTMF tones exhibit Twist, which is a difference in amplitude between the row and column tones. If this difference is incorrect, the receiver may not interpret the tones properly.

The high frequency tones are transmitted at a nominal -6dB relative to the low frequency group; i.e. the voltage of the high frequency tones is 50% of the voltage of the low frequency tones.



Having just 9V may be a bit low for your exchange simulator, it leaves very little headroom for the line impedance. Traditionally it was 48V but I have seen PABX units that work on 24V. I would also DC isolate and AC couple the two modems or one may prevent the other from detecting the line voltage when it goes off-hook and the AC signal may be too attenuated.

To monitor the DTMF tones you could insert a filter in the monitoring line that you can switch between low and high pass at around 1KHz so that you can check the high and low group alternately. The combined signal while pretty and somewhat distinctive and of limited high harmonics (smooth) is not visually decodable on a scope.

In the linked page may be the simplest circuit for connecting, it is polarity agnostic and assumes the 12V is adequate for both devices with the 680 loop resistance. I would increase to 24V and then it looks like an adequate design.


If you want to support dial and ring tones you are looking at a lot more work.

The page below has a lot of interesting POTS circuits that include some for monitoring the line audio.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.