Callback Systems



Information about our Callback systems.



A paper about our experiences as callback service providers, and how our callback systems were developed.








OUR CUSTOM TELEPHONE CALLBACK SYSTEM



This inexpensive customizable system allows one user to make long distance phone calls to anywhere in the world, from anywhere in the world, using a US dial tone. The system is set up in the US on a regular telephone line that has the "3-way calling" feature provided by the local telephone company. When the system is called and rang once or twice only, it will call back the user at a pre-programmed phone number. When the user answers, he will be able to dial the party that he wants to reach using "touch tones". The system records the tones. Then using the 3-way calling feature, the system puts the user on hold while it calls the number the user entered. Finally, it takes the user off hold so he can talk with the party he wants to reach. Since the system does not answer when the user rang it once or twice to trigger the callback, he is not charged for that "trigger" call. When the system calls him back, the call is originated from the US at US rates. Then the call to the party the user wants to talk to is also originated from the US at US rates. The user is charged for two calls at the same time, but the total is still expected to be less than the high rate that would have been paid if the call originated from the user's non-US location.

[The system can optionally be set up to provide "call through". This allows you to call the system and after entering a passcode, you can call out from where the system is located.]

Here is a list of standard special features:

1) Last Number Redial: The user can redial the last party he called by dialing # 0 instead of the entire phone number.

2) Change Number to Call Back: After the user is called back, he can change the number that the system will call back the next time it is triggered. This allows the user to access the system even if he changes location.

3) Error Correction: If user makes a mistake while entering the number he wants to call, he can hit * and start over again.

4) Speed Dialing - The user can store 9 numbers that can be speed dialed by hitting n # where n is 1 to 9. Speed numbers are programmed in advance and can not be changed.

5) Multiple Calls: The user can hit * when the first party he called hangs up in order to make another call.

6) Multiple Callback Numbers: The user can supply up to 3 numbers to be called back in rotation if first number is not answered and used to dial a party. The system rotates through all 3 numbers twice before giving up and resetting.

7) Exit System: The user should hit # # to exit the system. This immediately terminates the call and resets the system. If the user forgets to do this, it will take up to a minute for the system to hang up and reset causing unnecessary extra phone charges.

The system is available in voice or non-voice versions. The voice version provides the user with four customizable voice prompts. Voice prompts are given when the user answers the callback, when the user exits the system, if the Error Correction feature is used, or if the Multiple Calls feature is used. The non-voice system uses tones instead of voice prompts.

The system comes in a computer mini-tower case. The system is connected to the phone line through a FCC Part 68 approved interface. A modular phone cord and power supply cord are included.

The software that controls the system is customizable for special orders.

Retail prices as of 10/01 are as follows:

Non-voice Board ....................................... $250.00 each
Voice Board ............................................... $500.00 each

Shipping, handling, and any applicable taxes are extra. Prices are subject to change without notice. All equipment is sold with a 90 day warranty.

To order or for further information, contact B K Services Inc. at bkservices@compuserve.com or bkserv@bellsouth.net.

Back to the top of the Callback page.








Information about our experiences in the International Callback, or re-origination telephone business. Custom inexpensive switching system hardware designed for this business is described. Suggestions for the system's use in other applications is discussed.

Copyright 1995 by B K Services Inc. Written 8/95

DISCLAIMER: All information in this text file is the opinion of the author. Every effort was made to present accurate information, but the reader should verify this before acting on it. The author assumes no responsibility resulting from the use of this information.

Back in the winter of 1992-3, I heard from a family member about a man who was receiving income of $100,000 a month in a new telephone based business. He had a piece of switching equipment that let him offer phone service to people in Argentina at lower rates than they could get from the state owned Argentinean phone company. The equipment somehow allowed the Argentinean clients to originate their calls from the U.S. instead of from their homeland. It was cheaper to call Argentina from the U.S. than the other way around. So this man could put in phone lines at his office in the U.S. that people in Argentina could use to re-originate phone calls. Then he would bill them for the service at a rate lower than what they would pay at home but higher than what he was charged by whatever long distance company he used.

I became interested in what seemed to be a lucrative business. People are going to use the phone anyway, all this business does is give them service for a lower rate. So it seemed that getting clients should be easy. I began to research this business opportunity. I found out that most countries, not just Argentina, had a state run phone company - a monopoly. These state owned companies are called PTTs for Postal, Telegraph, and Telephone. They are the only game in town so you are forced to use them for your phone service. As a result, they charge any rate that they want. In the U.S., we have many companies competing for people's long distance business, so the rates are lower. Therefore, for example, it costs over $3 a minute to place a call from Brazil to the U.S. but only around $1 to call from here to there. With a disparity like this, the spread provides good profit potential.

Now I had to find out the details of how a person in another country could originate a call here. I went to the library and did a magazine search using the keywords: (international telecommunication discount) and found out about a company that provided this service. I asked for their brochure which described how their system worked. A client would call a certain phone number in the U.S. at this company's office but they would only let it ring once, then they would hang up. The line would not be picked up here so there would be no charge to the client for this call. This call is used to signal the system that the client wants to use the system to make a call. There is a unique phone number given to each client so the system knows who is calling when that number rings once. The system calls back the client in their home country, thereby originating the call in the U.S. at the lower rate. The client answers and is given a dialtone, letting him know that this is the system calling him back. He then dials the number he wants to reach as if he were dialing it from where the system is. The system is actually giving the client a U.S. phone line to use, which he is connected to by the system.

OK, now where could I get a system like this and what would it cost? I did a little research and found that a switching system could be had for around $10,000. I didn't have $10,000. There was a possibility of leasing it, and at the present you can rent space on someone else's system. I had very little money to put into this venture but I had a bit of extra time. Since I have a BSEE from Polytechnic University, New York (class of '83), I figured it might not be too difficult to whip up my own system.

The systems that could be bought had some disadvantages. 1) They cost a lot. 2) If the system goes down, all of your clients are affected. If you stay down for long, they will go elsewhere. 3) Everyone gets the same service, it would be hard to customize. 4) It seemed that 2 lines would have to be put in for each client, one to call them back and one to let them make their call on. If you have less than 2 lines per client, you must use some of the "incoming" lines for "outgoing" calls. That means there is a possibility that the system may be unavailable for use at peak times. I decided to make a simple circuit for each client that used the 3-way calling capability of the local phone company. Basically it works by listening for one ring on the line, calling back the client and listening to the tones from the numbers he hits, doing a "hookflash" to get another dialtone, repeating the tones from the client, and then doing another "hookflash" to connect the client back into the call. This would overcome all 4 disadvantages of the switching systems that other companies used. 1) An individual inexpensive circuit could be built for each client as soon as they signed up. This keeps down the "up front" costs. 2) Since each client gets his own "system", a problem with one doesn't affect the others. 3) Each system can be set up to suit the individual needs of that client. 4) There is only 1 line per client, with 3-way calling service. This keeps down overhead and prevents a congested switching network.

By the spring of 1993 I had an idea of the profit potential of the callback business, and an outline of how I could get into it on a shoestring. My next step was to find customers and design my circuit. I had a friend who was from Brazil and had some contacts down there. I told him about this business and he quickly became interested. We became partners. He would get the clients and take care of the finances, I would make the circuits.

I chose the Intel 8749 microcontroller to be the brains of my circuit. It sells for $10 or less, and I have the necessary software and hardware to write programs for it and write them into the chip. As a first cut, I came up with a circuit that listened for and decoded tone inputs with an SSI 202 tone decoder. The dialing would be done by the relay that would be used as the hookswitch. In order to legally hook anything up to the phone lines, it must be FCC Part 68 approved. I read in an electronics magazine [Electronics Now - March and May 1993] about a company that made a Part 68 approved interface. It could be used between my circuit and the phone line to make sure nothing funky or destructive got on the line from my circuit. As a bonus, the interface came with a built-in hookswitch relay. The interfaces now cost $39.95 each. They are sold by CircuitWerkes. Their phone number is (904) 331-5999.

About the time I got version 1 of my circuit designed, built, and debugged in the summer of 1993, my partner got us our first customer. He had the client sign a contract promising to pay for service used, which would be billed to his VISA card. The client paid us a set up fee for the line and equipment. I ordered an extra telephone line for my 1 bedroom apartment whose dining room table served as my lab and gave the phone number to our client. Every time I heard the relay clicking, I knew we were making money!

There were still bugs to work out and improvements to be made. First on the list of improvements was to have the tones that the client generated in dialing the number he wanted to reach echoed back out as tones and not hookswitch relay pulses. That would reduce the connect time to 4 to 6 seconds instead of around 30 seconds. The long delay was OK for our first customer because we made it clear to him that he was our "guinea pig" and things would only get better. The first bug we found had to do with the nature of 3-way calling. The problem occurred when it was time to end the call. If the client and the party he called both hung up, the MPC-2 line detection output tells the 8749 which also hangs up and resets to wait for the next call. But if the client called a line that kept ringing and wasn't answered when he hung up, our system had no way to detect this since the called party never hung up (because it never answered!). We told the customer to end each call by hitting # #. When our system hears # # it hangs up and resets.

Another problem specific to 3-way calling was what happens when the client calls a number that is busy. The circuit is set up to do a hookflash after it echoes out the number the client dialed. Normally this brings the client into the 3-way loop. But if the number dialed is busy, the hookflash brings the first party (the client) back in with the calling party (the circuit), and the busy line is dropped. So the client hears silence and doesn't know what happened! Luckily, the IC I chose to improve the circuit with the ability to produce as well as receive tones also did call progress.

Version 2 of my circuit uses an M-8888 DTMF transceiver to create and listen to touchtones. It also uses a 741 op-amp to match input and output impedances and provide isolation. The original circuit used separate 5 volt power supplies for the digital and the analog parts of the system. Version 2 uses those plus a -5 volt analog supply to the op-amp. The M-8888 also provides a call progress function. It allows a microcontroller to follow the cadence of signals on the phone line. For example, a dial tone is a steady signal. A ring is two seconds of 1 or 2 frequencies and four seconds of silence, repeating. Busy signals are 1/2 second of 1 or 2 frequencies and 1/2 second of silence, repeating. The call progress output from the M-8888 goes high whenever it detects frequencies in the commonly used range of signal frequencies. So with the proper software, I could tell if the number echoed out resulted in ringing or a busy signal. If it was busy, after the second hookflash I would create an artificial busy signal for 6 seconds after which I would wait for input from the client. He could hit # # to hang up or * to place another call.

By the fall of 1993 we had a couple of customers and more on the way. We got some feedback that showed that there was some confusion when using the system. The user didn't always know what to dial and when to do it. I designed version 3 of my circuit which uses an ISD chip to store 20 seconds of speech. That time period allowed us to record 4 different voice prompts to guide the clients in the correct use of our system.

The software for the system was continually being upgraded, debugged, and improved. I fixed bugs related to the use of the * key to place another call, and the placing of international calls by our clients. The first bug was due to the way 3-way calling responds to a hookflash when the intention is to drop the third party on the line. Different things happen depending on if the third party has hung up before the hookflash, how long before, or if he is still on the line. The second bug was due to how long it takes to connect to an international line and the type of feedback you get during the process. Details about these problems and their solutions will be given to interested persons upon request.

Software upgrades were also made to add new features. I gave clients the ability to change the number that we called them back at. This let them travel and still use the system. Some users transferred the callback number to their cellular phones. I added a speed dialing feature to hold 10 commonly called numbers that a client would supply to us. One of these was our customer service number. I also implemented a hunting feature at the request of one client. He wanted us to call back a number that might be in use by other co-workers. If the line called back was busy or if nobody answered the callback within 3 rings, he wanted us to try another number, and if nobody answered that he had a third number for us to try. I would cycle through the 3 numbers twice before giving up. This insured that most times the callback got to him.

My research into long distance rates found that the first minute or the first 30 seconds is billed at a higher rate than additional time units. That is why I created the * feature. Hitting * after finishing one call puts the system in a mode where it listens again for a number to connect to. When the user ends the number he wants to reach with the #, the system does a hookflash, echoes out the number, and does another hookflash to bring back the client. This is just what happens for the first call. The difference is that the client has already been called back and the initial higher rated time period was used during the first call. All of the subsequent call is billed at the additional time rate. The callback company can decide whether or not to pass this savings on to the customer.

As time went on, we were learning that a lot of other companies were springing up to compete with us. We learned what they were charging and we were forced to match their prices. We came up with our pricing structure. We charged the official AT&T rates for international calls for the callback part of the connection and nothing extra for calling anywhere in the U.S. This meant we would have to eat up to 20 cents a minute depending on where the client was calling in the U.S. If the client was calling outside of the U.S., we would charge him for the callback and the call he connected to at the AT&T rate. We charged the additional minutes rate even for the first minute, which made short calls less profitable for us. But we had a 3 minute call minimum, which helped.

Another issue we had to decide what to do about was busy lines. If the client was called back and tried to call a number that was busy, we were still charged for the call to him. We decided to charge a 1 minute call to the client for busy calls.

How did we make money if we were charging the listed AT&T rates? We committed to $5000 a month in volume. Then they gave us a discount package which totaled about 60% over the course of a year. But we soon found we had a misunderstanding with AT&T. They would give us no discount in months when we did not meet our target volume. We thought we could work up to our target and hit it on average over a year. We cancelled with AT&T and went to MCI. There we committed to $60,000 a year of volume and got a discount package of about 73%.

By the winter of 1993-4, we opened our office and put in 13 lines. One was for voice for us, one was for our fax. The other 11 were for customers we had signed up. Our volume was over $1000 from the moment we went to MCI, and growing. We collected our billing charges from our customers by billing their VISA or Mastercards. My partner was signing up bigger clients who would be spending over $1000 each a month. Some of these would be our representatives in South America. The most promising client he signed up gave us his American Express number, he didn't have VISA. We thought we could become an American Express merchant but it was not to be. We couldn't collect from him, and we couldn't pay our MCI bill. By mid summer 1994, we were out of the callback business!

Early in 1994, I heard about a man who might be able to use one of our systems. He lived in a town with only 1 long distance carrier. He made a lot of phone calls and the one company charged high rates. He had access to office space in a town a few miles away where all of the long distance services did business. He asked if I could set up a system that he could call in the other town and originate his long distance calls from there. I modified my callback system to answer on the first ring, then listen for tones. After the pound sign, it would do a hookflash, echo out the tones, and do another hookflash to connect the user with the party he wanted to reach. I also added a user definable time out period. This would check the line periodically to see if the user was still talking. This was necessary for this user incase he forgot to end the call with # #. He would have to drive 15 minutes each way to go and reset the system if he forgot to end the call properly in certain circumstances. So I put in a feature where he could set it to reset after a user selected number of hours after a call started. But before it would reset it would beep every 6 seconds for a minute. If the user hit any key, the system would know the user was still using the system and would check again later.

We have also sold one of our systems to an export company. They do business in South America and want a way to allow some of their better customers to call them and maybe others in the states at discount rates. The owner of the company may let his sister in South America use the system too. The system can be set up to recognize ringmaster rings, so the same system could call back different users.

Another use for my system would be for small business owners. They and their employees could call a system back at the office when they are home or on the road which would would answer and originate the call from the office (this is call-through). This allows better tracking of phone calls, higher volume discounts, and savings over calling card calls.

Yet another use for the system is to set up a telephone co-op. Join with some of your friends or business owners and get one of my systems. Everyone would call into and use the system for their long distance. The volume would earn a discount rate greater than what each user could get individually. When the bill comes, each user picks out his calls and chips in his share.

For those of you in the South Florida area [or other areas with a new area code whose middle number is NOT 0 or 1], my system can help with problems caused by the new Broward County [Ft. Lauderdale] area code. On September 11, 1995 we are getting a new area code in Broward: 954. Now we have 305 in Broward and Dade [Miami] counties. But with the proliferation of Fax machines and PCs and new arrivals in town, we are running out of phone numbers. So the phone company decided to let Dade keep 305 while giving Broward the new 954 area code. But businesses are worried about losing the South Florida identity that comes with 305. Plus, many old PBX systems won't call the new area code because it's middle digit is not 0 or 1. Until now, all area codes in the US had a 0 or 1 for a second digit.

My company is planning to set up an office in Dade county where the area code will remain 305. Businesses that want to have a 305 area code can set up an account with us. Their customers will dial a 305 number in Dade that will be answered by our system on the first ring. The caller will hear a message - Connecting to XYZ Co. Then the system will call XYZ Co. and connect in the customer in a 3-way calling configuration in about 5 seconds. Customers can sign up at one of 3 price/feature points. They could share a phone number with other companies. This would require the caller to enter a code to specify which company they want to reach, and there could be busy lines. Or they could have their own system and share the power supply and back up power supply with other systems. Or they could have their own systems and power supplies. We may also offer the system for sale for a company to own and operate themselves. Contact us for more information.

You may wonder we need any equipment at all to take a call in the 305 area and transfer it to 954 land. It would be easier just to order call forwarding service on each line we connect and just forward the calls across the county line from Dade to Broward. But the phone company says they may not allow calls to be forwarded from a 305 line to a 954 number. So some sort of system, like ours, will probably be needed to make a service like this work.

I hope you found this to be informative. Any feedback or questions can be directed to me by e-mail. Our e-mail address is bkservices@compuserve.com.

Back to the top of the Callback page.
Back to B K Services, Inc. Homepage.