This involves the construction of a custom circuit that is effectively a master/slave loop following circuit. The master end is the actual phone line, and the slave end repeats hook state (such as dial pulses) into an FXS port. This way, the FXS port can be used for dial pulse accounting of outgoing calls, and this data can be used to fill in gaps in the SMDR data from the WHOZZ Calling? unit, since it does not support dial pulse.
I believe all that's required is to get a cable that takes a 2.5mm female audio connector and breaks it out into two wires, which will be tip and ring inside a jack, that can then connect to the FXS used for processing dial pulses. From here, the module will need to be modified to accept a "dial pulse interface" for pulse support, and process accordingly. The 43-228A then goes inline between the line and phones.
Notes from Steph:
- The Radio Shack device I mentioned previously (43-228) has a modular plug and jack and inserts between a phone and line. It will follow dial pulses produced by a device connected behind it and deliver them as an isolated contact closure
- It will also deliver audio from the line via a separate plug. You will need a 1/8" AKA 2.5mm micro phone jack to connect to the DP output plug or have to cut the plug off and connect to the wires.
- A common 100 ohm ¼W 5 or 10% tolerance carbon composition resistor will withstand the 24mA you measured without overheating but if the device were to be connected to a real electromechanical or electronic exchange capable of delivering 100mA it would overheat. A common 2W resistor could tolerate the 100mA without overheating. I'm assuming there is enough internal space within the housing to fit the 2W resistor but don't know that for a fact. ½W and 1W resistors are also common and would provide more tolerance for connection to a real exchange line. Unless one were very close to the exchange (a PBX on the premises or very nearby CO), the theoretical 100mA value would not be experienced anyway. A 5W wire-wound resistor would be much more compact than a 2W carbon composition one, so even though the dissipation would only be 1W at 100mA, a 5W part would be a good choice simply because it's much more compact. It's necessary to derate any resistor's power rating, so where 1W could be dissipated, the appropriate part would have a 2W rating. Similarly, at the 24mA limit, the actual power dissipation is only 0.0576W, so a ¼W resistor is well capable. Even a 1/10W resistor has lots of headroom. So depending on what's available, I'd suggest using the highest wattage rating among these you can find.
- On the board: It turns out that there is already a 4.7 ohm resistor, R4, in series with the control output. That can easily be replaced with a 100 ohm resistor to provide the appropriate seizure condition towards the FXS.
- Another alternative which comes to mind with respect to connecting to the FXS is to remove the existing output cord by unsoldering its leads at the AUD. OUT and START/STOP OUT solder pads and connecting the R & G leads of a half-modular cord to the START/STOP OUTPUT solder pads instead of making an adapter cord consisting of 2.5mm jack connected to a half modular cord.
- The existing lower resistance certainly WILL cause a seizure but it presents a near short-circuit condition which the FXS will probably tolerate without damage but since the software seems to monitor T-R voltages there might be some other adverse functional effect, so it's always best to maintain normal operating conditions.
- My note: I don't think we need to mess with the resistor at all, at least not on the board. If needed, we could add resistance in series by using a "resistor box" like the one in series with the door sensor. But really, this could be connected to an FXS on a channel bank with no adverse effects, so why bother?
- The AUD. OUT solder pads provide microphone level audio intended to connect to the microphone input of a cassette record. A mike input is typically around 1-10mV audio level whereas a phone line is typically around 100mV-1V.
- The transformer appears to have a 1:1 ratio so it could be suitable for coupling audio to an FXO port were it not for a resistor attenuator consisting of the series 10K resistors R1 & R3 and 1K shunt resistor R2, but capacitor C2's value would also need to be increased, so it seems easier to just bridge the FXO across the input line cord or output line jack. However after removing the existing 3.5mm AUDIO & 2.5mm START STOP OUT plug cord you could add a second half-modular cord connected to the R&G solder pads of the LINE cord or LINE jack to avoid the need for a modular Tee adapter to provide the FXO connection. It would provide a neater installation with fewer "barnacles" (the Tee adapter) sticking out of jacks.
- You will need a "subminiature" 3/32" AKA 2.5mm 2-conductor jack. Ideally, a cable mounted one would be best since it would have an insulated shell. A panel mounted one would work
- The Radio Shack recording controller has a modular plug and jack which will connect in tandem between your phones and/or 5-way splitter and the ONT.
- Assuming the FXS is connected through a modular jack, you will need a short cord with modular 6-pos. 2 contact line plug on one end and stripped leads to connect to the 3/32" jack.
- There are half modular cords that only have a plug on one end. The cord must have stranded rather than tinsel conductors since tinsel while more flexible, is problematic to solder to the terminals of the jack. If you cannot come up with something, I have bulk modular cordage and plugs and can make up a short 6"-1' long half-mod cord and send it.
- I'm talking about the FXS port on your * system which is going to recognize the seizure of your ONT line to capture the CDR data using the companion FXO. The submini plug on the controller needs to connect to that FXS using an adapter cord you need to assemble, equipped on one end with the 3/32" jack and on the other end with whatever is required to connect to the FXS.
Comments
You must be logged in to leave a comment.