News:

"The phone is a remarkably complex, simple device,
and very rarely ever needs repairs, once you fix them." - Dan/Panther

Main Menu

Spanish Payphone, I need help to make a controller

Started by Spanish_phones, December 03, 2013, 08:14:12 PM

Previous topic - Next topic

G-Man

The Karlsson System for subscriber billing is still in use around the world, albeit the payphones and other equipment are now microprocessor controlled instead of relay based technology.

Searches for newly manufactured payphones reveal that they are commonly available with Karlsson Systems relying on 50KHz, Polarity reversal, 12kHz, and 16kHz pulses.

One such product is the T597 Outdoor Coin Payphone:

•   http://www.weiku.com/products/3546964/T597_Outdoor_Coin_Payphone_Powder_coated_zinc_plated_steel_.html

Note that in the listing for "Supported Charging Signal" it list Polarity Reversal, 12kHz, 16kHz pulses for coin/token collection.

dsk

Great, this makes me even more sure that the normal state of conversation are reversed polarity, with the relay engaged, and 5000 ohms winding shorted. When the last token are ejected, the resistance of the 5000 ohms coil will prohibit any conversation.

A short of the line will let the springloaded relay return to rest position, this will collect the coin. When line short are removed. the 5000 ohms coil engaged the relay, next token falls down, and shorts the 5000 ohms coil, while the 100 ohms coil holds it engaged.

dsk

PS if this was a normal operation in Europe???  Not sure. but it would fit right in here in Oslo. As far as I recall local calls were only one coin, no time limit. Did not call  long distance before the 16kHz system was introduced.  DS

Jack Ryan

Quote from: G-Man on December 10, 2013, 03:23:30 AM
The Karlsson System for subscriber billing is still in use around the world, albeit the payphones and other equipment are now microprocessor controlled instead of relay based technology.

I still haven't found any documentation for a Karlsson System.

Quote
Searches for newly manufactured payphones reveal that they are commonly available with Karlsson Systems relying on 50KHz, Polarity reversal, 12kHz, and 16kHz pulses.

It must be a broad specification because of the wide variations between each of these.

Any pointers to documentation?

Thanks

Jack
(Only recently interested in payphones)

Jack Ryan

Quote from: dsk on December 10, 2013, 04:01:21 AM
PS if this was a normal operation in Europe???  Not sure. but it would fit right in here in Oslo. As far as I recall local calls were only one coin, no time limit. Did not call  long distance before the 16kHz system was introduced.  DS

It's interesting that STC had much to do with payphones in Australia but none of ours work like this one.

Jack

Spanish_phones

#34
Wow!!! How many answers!!!!! I think we almost guessed how it works!!

Unfortunately, I have no clue if this system was usual in Europe, but I'm completely sure it was the system used in Spain since payphones appeared in the 50's until the 80's, when touch tone payphones arrived (I don't know if touch tone also continued with this system or not)

Well, so the only thing that I don't know how to make is collect the first token when party answers. I'm afraid is going to work like USA controllers, collecting the first token when a certain period of time is passed.

I know the collecting system works with reversals, so I made a reversal circuit with a double relay (8 contacts). Anyway, I would like to test if the payphone also works with pulses as some of you have said. Any ideas of what circuit it would be? I measured with the multimeter and the payphone works with 6V DC and 21,5mA

Last, but not least, yesterday at the university, the professor taught us electronic components, and one of those was the timer (NE555), maybe one of those could control the time between pulses/reversals.

One more time, thank you so much!!!!

Iñaki ;)

G-Man

Quote"Last, but not least, yesterday at the university, the professor taught us electronic components, and one of those was the timer (NE555), maybe one of those could control the time between pulses/reversals."

Yes, to repeat myself, I already suggested using a 555 timer along to generate the pulses. Coupled with a relay it would comprise a simple and inexpensive circuit to emulate the original polarity reversal pulses. The initial collection could be initiated by polarity of the first reversal of polarity of answer-supervision from the VoIP modem.

It would be a very, very elemental circuit so it should be extremely easy and inexpensive to construct.

Karlsson Message Rate Pulses generated message unit pulses for message rate billing which included answer pulse recognition signals and answer supervision from a central office for providing a variable number of pulses based on call duration, time of day and distance of call.

These pulses were optionally extended to payphones for coin collection as well as to subscribers for call-accounting purposes.

It did not matter whether the pulses were delivered by d.c. or high-frequency pulses, it was the same system that also drove the messages registers located in the central offices.

If I recall correctly, Karlsson developed the system in Norway for WECo International/ITT-Standard Electric but I would have to research it once more to refresh my memory to be certain.

This system was widely used throughout Europe and Asia and apparently in Australia as well, allowing what you folks call STD toll calling. This system was not used in the U.S. because it did not provide enough billing detail required by American subscribers and their regulatory agencies.

It is a very robust system since various incarnations of it appears to still be in use as evidenced by a reference to it in a 2003, CISCO publication:

Quote"The associated meter pulses are transmitted at the end of each timing interval. The asychronous method (also referred to as Karlsson) starts at the first timing interval at a random value..."
Cisco MGC Software Release 9.5(2)
Meter Pulse Messages Support
Document Release History
Feature History
The Metering Pulse Messages (MPM) Support Feature is described in the following sections:
• Feature Overview, page 1
• Supported Platforms, page 4
• Prerequisites, page 4
• Configuration Tasks, page 4
• Charging Data Structures, page 4
• Provisioning Examples, page 12
• Command Reference, page 13
• Reference Information, page 15
• Glossary, page 17
Feature Overview
The Metering Pulse Feature enables the handling of meter pulse message pass through, modification,
and generation. Billing information is derived from and provided to the billing mediator using Call
Detail Records (CDRs)).
Publication Date Comments
06/02/03 Initial version of the document.
Release Modification

G-Man

Though there are various incarnations of how they are generated, Karlsson Message Register pulses were delivered by means of D.C. polarity reversals, and High Frequency formats including 50 KHz, 12 KHz and 16 KHz.

Certainly not any different than voice transmissions being delivered via d.c. over copper wires or superimposed upon those same wires by using carrier frequencies; still the same voice-information, merely delivered by different formats.


Quote from: Jack Ryan on December 10, 2013, 05:05:27 AM
Quote from: G-Man on December 10, 2013, 03:23:30 AM
The Karlsson System for subscriber billing is still in use around the world, albeit the payphones and other equipment are now microprocessor controlled instead of relay based technology.

I still haven't found any documentation for a Karlsson System.

Quote
Searches for newly manufactured payphones reveal that they are commonly available with Karlsson Systems relying on 50KHz, Polarity reversal, 12kHz, and 16kHz pulses.

It must be a broad specification because of the wide variations between each of these.

Any pointers to documentation?

Thanks

Jack
(Only recently interested in payphones)


G-Man

Further research shows that this system was used throughout the world and "was developed by S.A. Karlsson, technical director of the Helsinki Telephone Company, who championed the technically simplest solution of keeping the metering pulse sequence out of phase with the theoretical beginning of the call (sending of the called subscriber's answer signal).

The first pulse in the series was transmitted within a certain margin of exact time determined in respect of the answer criterion.

After much initial hesitation, the Karlsson method swiftly became a great success."


Some of the countries adopting this system included:
Argentina, Australia, Austria, China, Denmark, Ethiopia, Germany, France, Ireland, Italy, Japan, The Netherlands, Norway, Poland, Portugal, Spain, Sweden, Switzerland, United Kingdom, Yugoslavia.

Countries that DID NOT adopt it include:
The United States, Canada, USSR, Belgium who favored detailed billing information.



dsk

Quote from: Spanish_phones on December 09, 2013, 08:07:23 PM
Well, here you are a few answers to your posts.

First, a photo of the contacts L and C. L permits to dial, I think C should end the call when the last token is collected. They work separately, but together, in other words, the position of L doesn't affect on C and viceversa, the work independently.

I've also made a video of how the system should work:

http://youtu.be/bSDMtUCeMBs

As you can see, collecting one token requires two movements, only one pulse/reversal, which make the first movement, then when the current returns to normal polarity, a spring pulls the big coil back. In the video, the last token collected gets stuck, leaving C contact closed and L contact open, so this combination should end the call ( when there are more token left, L contact is open by the next token)

I'm afraid reversing the polarity is going to affect on the quality of the call, or even end it. Maybe I have to develop a device that sends negative pulses to the telephone, in order to control the collect system without ending the call meanwhile? Anyway, I think I'll make a test with the two relays reversing polarity and see if it works or not. Does anybody know were can I buy relays?

Thank you so much ;)

Great film. and it is fantastic to be able to do this in the forum, not that many years ago it would have been just fantasy.

I  would like to ask you to do one more test, and film the relay as you did.

1) Connect the phone with reversed polarity from the start, and leave it that way. (no dial tone when you go off hook)
2) Have 2 tokens ready, go off hook and insert 2 tokens (within 15 seconds)
3) Hang up.

4) Do 1 and 2 again, and Dial a number, your own number = OK

What I expect to happen:
1) You connect the phone, and the relay moves when going off hook, and releases when you hang up.
2)The relay operates immediately when going off hook, first token passes directly to contact C and relay stays but you get dial tone. The next token stops at contact L.
3) One coin is collected, and one coin is returned.

4) 2 possibilities: A it just works and it rings, or you get a busy signal, or the pulses forces the relay to act, and the tokens are collected. (the last at hangup).

dsk

Spanish_phones

Hi dsk!

First of all, I assumed that "off hook" means lift the handset and go for listen to the dial tone.

The first step works as you said, I lift the handset and you don't have dial tone
The rest of them don't work, because when the phone is connected always on reversed polarity, the tokens are stopped BEFORE contact L, so you don't have never dial tone. Then both of the tokens are returned, because when you hung up, the mechanism open the "door/lever" and it collects both.

With the polarity reversed:

-When you lift the handset, have no dial tone, main coil acts
-When you hung up, the coil returns to rest position
-If you open the phone and close contact C, yo have dial tone but you cannot dial
-if you open the phone, close C and open L, you have dial tone and are able to dial, but this never happens inserting tokens with polarity reversed

Jack Ryan

Karlsson Message Register pulses
Quote from: G-Man on December 10, 2013, 09:53:49 AM
Though there are various incarnations of how they are generated, Karlsson Message Register pulses were delivered by means of D.C. polarity reversals, and High Frequency formats including 50 KHz, 12 KHz and 16 KHz.

Certainly not any different than voice transmissions being delivered via d.c. over copper wires or superimposed upon those same wires by using carrier frequencies; still the same voice-information, merely delivered by different formats.

Still without a reference, I searched for "Karlsson Message Register pulses" and found a Cisco document. The subject of this document is meter pulse pass-through for CDR generation. Payphones are not mentioned.

In addition, the reference to Karlsson applies to a subtle difference in the way the pass-through meter pulses start (synchronously or asynchronously), not to the general method.

With respect, without a reference to a clear definition, I can't see how using a reference to Karlsson helps.

Jack

Jack Ryan

Iñaki,

Quote from: Spanish_phones on December 10, 2013, 07:04:09 AM
Well, so the only thing that I don't know how to make is collect the first token when party answers. I'm afraid is going to work like USA controllers, collecting the first token when a certain period of time is passed.

I'd like to know what the problem is with the first collection (on answer).

Originally, you said:
· You [pick] up, and have dial tone but yo cannot dial (it works like this right now)
· You put at least one token, you still have dial tone, then you can dial (so far, so good)
· When the phone you're calling to answers, the payphone collects the first token (this doesn't happen)

On answer, instead of applying a reversal for answer supervision, apply a fleeting reversal to collect the first token. In other words, apply a reversal, pause and then apply another reversal so that the line voltage is back to normal.

Does this collect the first token?

In Australia, at about this time or perhaps a bit later, payphones no longer required a reversal on answer; they required a charge pulse (of whatever type).

If there are too many cooks (too many people asking questions) let me know.

Thanks
Jack


Spanish_phones

Jack, the problem here in Spain is that when party answers, nothing happens on the line voltage, there is no reversal for answer supervision, line voltage doesn't change unless for the dial pulses. Do you see now the difficulty of collecting the first token just when party answers?  That's why I thought of collecting the first token with the timer also, giving you enough time to dial and the party to answer, but then it wouldn't work exactly as it used to do :(

Jack Ryan

Iñaki,

I understand that on the PSTN, there is no longer a line reversal for answer supervision. I am suggesting that one was never needed for your payphone to operate.

To collect a token, your payphone requires a "fleeting" reversal; that is, a reversal, a pause and than another reversal to return the line voltage to normal. I am suggesting that on answer, your payphone requires a fleeting reversal to collect the first token.

If the call is timed (is charged per unit time), additional fleeting reversals would be generated by the exchange to collect additional tokens.

If you supply a fleeting reversal on answer, the first token will be collected on answer.

If this is, indeed, how the phone works, a simple reversal on answer from an ATA won't help. You will need to use the reversal from the ATA to:
- generate the first (on answer) fleeting reversal to collect the first token
- enable a timer to generate additional fleeting reversals to collect additional tokens.

Your controller will need to perform those tasks and, in addition, detect the no token situation and disconnect the call.

I am talking ATA here because it will give you an indication that the call was answered. This no longer happens on most PSTN lines and that makes the controller more complicated.

Jack


dsk

#44
Quote from: Spanish_phones on December 10, 2013, 04:11:13 PM
Hi dsk!

First of all, I assumed that "off hook" means lift the handset and go for listen to the dial tone.

The first step works as you said, I lift the handset and you don't have dial tone
The rest of them don't work, because when the phone is connected always on reversed polarity, the tokens are stopped BEFORE contact L, so you don't have never dial tone. Then both of the tokens are returned, because when you hung up, the mechanism open the "door/lever" and it collects both.

With the polarity reversed:

-When you lift the handset, have no dial tone, main coil acts
-When you hung up, the coil returns to rest position
-If you open the phone and close contact C, yo have dial tone but you cannot dial
-if you open the phone, close C and open L, you have dial tone and are able to dial, but this never happens inserting tokens with polarity reversed

Thank you.
I have modified your drawing, slightly,
The circuit are just for testing, and will be slightly different when you have put in timers.
For testing, the relay circuit could even be your polarity changing switch.

What I expect to happen are:
-Lifting handset (= going off-hook), and getting dial tone,
-inserting a token (ore more). (Contact L open dial short )
-Dialing and getting an answer.
-Polarity reverses (manually/by timer/by  ATA)
-(token falls one step) and you may talk to the other party.
-I the call lasts you may get a signal to collect one more token. (pressing button for 1/2 a second, or activited by timer)
- The relay collects one token, and when short are removed the relay acts, next token are closing the contact C and you may speak for one more timer interval.

If you run out of tokens, the line will hold for a short time, and you might be lucky to put in a new token and keep on (not for sure).

When you hang up, the line returns to normal, (by manually or automatically change polarity to normal) You may not place call again before the polarity has returned to normal.

dsk