6 March 2012
I also left place holders for filter components RA1 and CA1. This setup allows adjustment of the transformer side impedance to limit frequencies and levels. As this function is not necessary in this application, I used jumpers for RA1 and CA1 and left RA3 unpopulated.
|Table 2 – Bill of Materials from the Avnet Express Website.|
|Fig. 5 –A little floor planning helps when prototyping, debugging, probing, testing, and hooking up.|
|Fig. 6 – Construction using perfboard is a little tougher to do, but in my opinion, it yields more rugged prototypes.|
|Fig. 7 – To mount the RJ-11 jacks, I pre-drilled for the mounting holes and where the wires solder in place. Since these parts are subject to stress, I also hot glued them in the other side.|
|Fig. 8 – The solderless breadboard allowed easy tests of the relay and the Ring and Off-Hook detection circuits.|
|Fig. 9 – With the intercepted phone line On-Hook, I was able to test the signal recovery through the bleed capacitors.|
|Fig. 10 – With the intercepted phone Off-Hook, the Off-Hook Detect LED illuminates, and the audio is available for our DTMF detector.|
The impedance on both input legs of the transformer should be the same to maintain a balanced line that wants the same impedance relative to earth ground. Any variation or imbalance could cause noise, typically 60 Hz hum.
Since the transformer I have is center tap, I ground referenced one side and passed both full tap and center tap out through my eight-pin interface connector (CNX1). This option provides more flexibility when interfacing to this board. For example, we can pass one output to a DTMF decoder audio stage (e.g., a speaker phone) and use the other tap to send voice or data out. It is a simple 2-to-4 wire interface.
Some Zeners (ZD4 and ZD5) protect the audio circuitry down the line. A spike squelching diode (D1) also helps protect the relay driving line from EMF back-spikes from the relay coil. An eight-pin 100 mil standoff header interfaces the complete phone interface from our next stage.
In Search of Parts
Using the Bill of Materials (BOM) tool online at the Avnet Express website, I was able to build a BOM for this project. (See Table 2.)
I chose to prototype the RJ-11 with a right angle through-hole part similar to the Corcom 6609214-5. These parts are pretty straightforward and a little more difficult to deal with than the discrete wire styles. I did not want dangling wires and connectors, so I decided to deal with the little extra work needed to use the through-hole soldered connectors.
I was able to find a 130V MOV similar to the Weidmuller 94015200 and use it as my surge suppressor. Once an MOV absorbs its rated energy, it acts as an open circuit. You may want to socket this part and replace it every so often if you live in a place that gets constant lighting storms.
I found the Panasonic ECW-F6135JL high-voltage capacitor to use for the coupling stage to the ring detector. The voltage does not have to be this high, but it does not hurt. This part is reasonably priced and will take the high-voltage ring signal with room to spare.
For the .047uF 150V signal bleed capacitors, I found a Panasonic CQ-B2473JF 250V polyester cap at 5%. The 10% could be used, but since they were the same price, I chose the 5%.
The bridge rectifiers are the popular through-hole button bridges. I chose the Zetex WO8G part. This part is a member of the family that ranges in voltages. For example, I could use the 400V WO4G, but I found that I can get the 800V parts for only $0.10 more. Again, the higher voltage rating is overkill, but it will not hurt.
I chose a generic 1N4758 for the 56V Zener diode — in this case, a Fairchild part. I also chose a 1N4744 as my 15V Zener — this time from Vishay. For a good general purpose opto-isolator, I chose an Everlight 4N25. These parts are six-pin though-hole and are easy to prototype with and replace if socketed. I decided to socket them because I never know how they will be mistreated, and I do not like to unsolder to replace a part.
I chose the Vishay H11AA1 for the Ring indicator opto-isolator. This part is specifically designed for isolated telephone applications and features a bidirectional emitter that engages the NPN transistor for positive and negative going waveforms at its input. Since the Off-Hook detector is a positive voltage that we control with the bridge rectifier, we do not need this functionality with its opto-isolator.
Note that I isolated both sides of the telephone line when not in use. Therefore, I chose a double-pole, double-throw (DPDT) relay. Again, a small (form C) through-hole part, such as the TE Connectivity 1-1462033-4, provides more than enough current capacity (2A per contact) and a 250V rating. In addition, the 5V contact fit nicely into our design.
I used a 600:600 Ohm transformer from parts I had lying around. The Vishay TA10EB07 series also will work for this project.
We do not need a center tap, but the transformer I found had it, so I decided to route it to the connector. This way, for our simple DTMF encoder, we do not need a 2-to-4 wire converter because we have two audio points.
Floor-Planning and Construction
It’s nice to have good, clean documentation when debugging and trying to reproduce a design. A good floor plan is one of those tools that helps when building, testing, and maintaining a design.
I created a simple floor plan and laid out the main components. (See Fig. 5.) These components include the larger high-voltage cap, the relay, the phone jacks, the transformer, the bridge rectifiers, the opto-amplifiers, the power supply, and the signal header.
I used plain perfboard with holes at 100 mil spacings for this project. (See Fig. 6.) This option is a little tougher to do, but in my opinion, it yields a more rugged and durable board. I used banana jacks for the 12 Volt inputs.
From experience, I know that the RJ-11 jacks are subject to a lot of stress during use. To mount the RJ-11 jacks, I first precut and drilled the perfboard, snapped the jacks in (See Fig. 7), and hot glued the connector so that the stresses would not break the delicate soldering. I also used sockets for the opto-isolators.
The bridge rectifiers went right in, as did the Zeners, caps, and resistors. I had to drill for the transformer mounts and banana jacks. I placed the voltage regulator along one side and folded it over. Later on, if I draw more current, I can heatsink the regulator. I used a standard SIP 100 mil header for the signals and made a little cable that I can connect to a solderless breadboard for testing. (See Fig. 8.)
The solderless breadboard will make it easy to test the Ring and Off-Hook detector. Since the prototype board generates a regulated 5V, I placed a LED in series with a 1K resistor for both the Ring indicator and Off-Hook detect signals. The LED should light for each condition.
I also used the ground strip on the solderless breadboard to let me test the relay. By bringing the relay Pull signal to ground, I should be able to engage the relay and take the phone line into an Off-Hook condition.
The solderless breadboard also served as a ground point for the oscilloscope, so I will be able to observe the signaling, voice, and DTMF tones as audio waveforms on the scope.
When using a POTS line from the telephone company, use an isolation transformer for your oscilloscope. The oscilloscope may use earth ground as a reference and so does the phone company but at a different location. Differences in voltages can occur.
Powering Up for the First Time
I like to do things sequentially, so when it came time to power up, the first thing I did was to check the 5V regulator circuit. With 12V applied to the banana jacks, I measured a clean 5V everywhere except on the eight-pin interface header. A quick examination showed that I had forgotten to bridge the ground signal to the header. With a quick solder connection, power was clean and full everywhere.
With 5V on the prototype board, I tested the relay. Bringing the Pull signal to ground on the solderless breadboard let me hear the distinctive click of the relay engaging and disengaging. A pulse train on this signal would let it function as a rudimentary pulse dialer once the phone line was connected. So far, so good.
My next test would be the Off-Hook detect. I plugged in the phone line cable and an older style analog telephone to the telephone jack. With power on, I lifted the handset, but there was no LED.
After examining the schematic and board, I found I had the wrong resistor value for the opto-isolator LED driver. I had also placed the Zener in series instead of parallel. Both were limiting the current to the opto LED emitter. A quick update on the schematic and a swap on the board, and I was ready for a retest. This time, when I picked up the handset, the Off-Hook LED illuminated and went out when I replaced the handset. To test the Ring Detect, I called into this phone line and watched as the Ring Detect LED flickered.
It was now time to test the audio. With the handset On-Hook, I picked up a parallel handset and punched in a DTMF digit. I could see the waveform for the DTMF codes on the scope as I dialed. (See Fig. 9.) This test verified that the signal bleed capacitors were passing the signal, even when the phone was On-Hook locally and the relay was not engaged.
Next I picked up the local handset and punched in some DTMF digits. The LED Off-Hook indicator lit, and I could see the waveforms on the scope. (See Fig. 10.) We now have our control signals, monitoring signals, and audio links in place. Next month, we will hang a DTMF decoder on this setup and make a home remote control interface that you can access with your cell phone.
Next month, we will design and build the 2-to-4 wire interface that will allow us to route extracted audio to several different stages, such as caller ID, speaker phone, and speech record/playback modules.
We will also set up the DTMF decoder to allow security code access and control.
While this project is not easy to build, it has a lot of potential for learning and use. Please let us know what you think about this project. If you like it, I will design a printed circuit board for easier construction, and we will make kit parts available in a bundle.
Enjoy, and have fun.
– Dr. Gizmology