Connecting a pair of Dialup Modems

[@rt]

Hi Guys

I’m trying to connect two terminals together, each with their own 56k dialup modem, and an ordinary phone cable connected between them.
One is an old XP laptop running Win XP using the Hyperterminal program, and it’s internal modem, and the other is a 56k portable travel modem.

Both modems can echo back to their respective terminal programs, and both also respond to AT commands with “OK” etc.

I’ve had no luck getting them to connect though. There are a few small articles about, and a YouTube video demo of the same,
and they all use ATA for the terminal doing the answering. The command used to dial varies from a simple “ATD” to something that might cause
the modem to ignore the fact that there is no dial tone without a real phone exchange. Some examples are “ATX3D” or “ATX1D”.
I hear that the function of the “X” command can vary between modems. I can get both modems to make noises by answering,
and one modem to make noises by dialling with some variations of the “ATD” command. It doesn’t matter if I have to dial with one of them all the time.

Ultimately, they both quit with a “NO CARRIER” message, and I have never heard anything that suggests to me they are responding to each other.
Is there a physical (electrical) reason for this (assuming everything is working), or a possible incompatibility that would prevent them even connecting to each other?
Cheers, Brek.

[CJay]

Seem to remember 56K modems won't work back to back, at least not at 56K, as they rely on some trickery with the digital telephone exchanges so you might need to persuade them to connect at a slower speed.

But, first, are you expecting them to connect over an actual phone line or just a piece of phone wire?

If they've not got the relevant voltages on the line then they may just not work at all.

[@rt]

With a straight phone wire. I have seen this done on YouTube, and know the fellow who did it.

56k should connect at 33.6 or so if that’s the uplink speed of each I assume.

[bd139]

You can do this with older Hayes modems. I had a 28.8k building to building ppp link running for several years on some Hayes Optima modems over one spare pair. ATX1D (blind dial no busy detect) Needs to be configured or it’ll wait for a dial tone. That had Pegasus mail running over it (yuck) as part of a supposed loose air gap.

It should negotiate 33.6k max but no guarantees with 56k/v90 cack. Depending on the chipset you might be able to turn that off and negotiate 33.6k max.

[Berni]

The phone line injects power into the wires. That's where phones get there power from.

You can get two phones to talk to each other by connecting a 9V battery in series with one of the phones terminals before connecting them with together with a phone cable. The battery pushes current around the loop to power both of them, munch like the telephone exchange would. Audio is transmitted by varying how much current is drawn from the line.

Perhaps the battery trick works with modems too in order to trick them into thinking they are on a real line. Hopefully they don,t also expect to hear a dial tone too, but perhaps that can also be bypassed by simulating the line ringing by injecting 100V AC on the line for a short moment. When you pick up a ringing line you don't get any dial tone, it's straight to audio(apart from the caller id signal nowadays)

[CJay]

They should be able to connect at 33.6K if they support V34, I'm pretty sure V.90 and V.92 rely on the exchange trickery and won't connect without seeing an exchange between them, maybe see if you can rate/standard limit them?

Which models do you  have and how are they connected to the hosts?

Some modems relied on having line voltage to power the DAA, most often laptop modems with an external 'lump' to connect to the telephone line which the 'travel modem' sounds like so it might never work if you just sling wire between them but it's been a *long* time...

[@rt]

When I first tried the laptop on it’s own it complained it had no dial tone and did nothing,
but I assume one of the commands to turn that off worked, because now it will just go ahead.

The travel modem on the other side might still be waiting for a dial tone, but it doesn’t complain that there isn’t one.
As I said though, it would be fine with me if the modem that will dial is always used to dial.

I have seen a trick to put a 9V battery, or a pair of them in series with one of the phone wires.
It supposedly doesn’t matter which wire, or the polarity of the battery.
I’ve heard of this, but always assumed it was just to get a phone to ring. Thinking about it though, DC shouldn’t get a phone to ring.

[CJay]

There are AT commands to persuade modems to dial without dialtone and answer without hearing a ring so I don't think it's that, you've got round the dial problem and ATA should make the other modem answer, the reference manual for the modems should detail the 'advanced' AT commands to limit protocol, allow you to monitor the line via the speaker etc.

[@rt]

Does it sound correct that it wouldn’t matter what polarity the battery was connected?
Looking at an open PCMCIA modem I have here, it has tantalum caps on the isolated phone side. This isn’t the one I’m using, but might be similar to one of them.

[CJay]

Tip should be positive, Ring should be negative but it probably doesn't matter.


Sorry, that's would be better said as Tip is usually positive, Ring is usually negative.

[@rt]

It ends up that one modem will get the wrong polarity anyway.
If you look at the diagram here: https://electronics.howstuffworks.com/telephone4.htm
There’s obviously no way to make the circuit where both red wires are positive for both phones.

[Berni]

There is nothing special about the 9V. It just happens to push about the right amount of current around the loop.

A real phone line actually has 48V DC on it, but with a high impedance so when you pick up the phone the speaker in the earpiece loads it down with its internal resistance to bring it down in to a few volts range with a few miliamps of DC bias current flowing trough the line. Phone lines don't have a defined polarity so phones are designed to work when connected backwards.

The ring is a >100V AC signal applied on the line and is designed to push enough current trough the high impedance electromagnet of the ringer bell mechanism to make it ring. This signal can be pretty low impedance so it has plenty of grunt to give you a nasty shock if you happened to hold the line when it rings. You could recreate this signal by simply connecting a mains transformer with a 100V winding right across the phone line, it won't be spot on but it would make phones ring and likely fool a modem too. Tho don't attempt this is you are not experienced with handling high voltage as you can really shock yourself.

It also might be possible to trick a modem into thinking the line is ringing by modding the circuit that detects the ring and wiring a switch in to make it "ring" by a push of a button.

[CJay]

It does, you're right, but the 'standard' is as I said above, in practice it rarely ever matters and if the DAA needs power they usually have a bridge rectifier which sorts out the polarity for you anyway.

[@rt]

The ringing I’m not so interested in since the “ATA” command works with both modems to answer. Easier than modding the modem and pushing a button.
I’ve used two 9V in series with 330R resistor. Here goes nothing.

[@rt]

That totally worked!!!

Since photo bucket started asking for money to host images, I can’t easily explain why that is so exciting.

The laptop modem is ordinary, but the travel modem takes 2xAA batteries, plugs into the phone line on one side, and it’s only other interface is IRDA.
So using that as a dongle hanging out of a phone socket on my laptop is very neat. One small thing hanging off one short wire to a phone socket flush with the side of the laptop,
for an IRDA serial port that I needed.

Thanks for all the replies, and of course, regarding my setup, I’d appreciate any suggestions. I assume 330R resistor is closest real value I’ll get to 300R/ does 300 Ohm even exist?
and maybe I could do the 48V DC supply, but if there’s no issues with errors is that really a problem?

BTW it was just “ATD” & “ATA”, but previous setup has obviously removed the need for a dial tone. Probably one of the ATXn commands.

[bd139]

 

Use imgur.com for image hosting.

[SeanB]

Typical telephone system is designed for a 48V central office battery, a 600R coil of the off hook relay doing triple duty as both transformer, off hook detection and a line choke. The phone itself is designed to emulate a 600R load ( which is why you have 600R transformers in audio use) and the line itself will allow over 1k before it degrades too far. Long lines fix the capacitance between the 2 wires with 88mH coils every few miles, which turn it into a rough 600R transmission line.

Thus with 9V and a 330R resistor you probably are emulating the line as if it is a long line, but there is enough headroom on most phone interfaces to handle it fine, and it will not do any harm, as the opposite can fry the line interfaces. You probably are drawing around 5mA from the battery, which is fine, though you probably should place a 100uF-470uF 16V capacitor across the battery terminals to lower the audio impedance, which will slightly increase the operating life of the battery before it's rising ESR causes the modem to drop out.

[@rt]

Hi Thanks for the replies.

Please check back Sean, I’ll have another question for you about your explanation, but it’s 2:20am here and I’m fighting to stay awake.
Goodnight!

[CJay]

What speeds do you get?

[@rt]

I’m in bed but the sleep thing isn’t working.
They connected at the 33 something, and I could type,
but I haven’t tried a modem protocol to move a file yet.

There’s enough grey matter here to put the last few bits
together to finish something off if I ask the right questions.
Now is the time I’ll screw it up.

Numbers will be wrong here, I’m making it up.
I made a ringer from scratch some years ago starting with a
555 timed for the 20Hz to ring a phone, modulated with some
gated outputs of a decade counter clocked with another 555
timed to pulse at a derivative of the Australian ring pattern.
Select outputs of the decade counter were ORed such that they
handled the ring timing, and the counter’s own reset.

From there, the twice modulated ring was used to drive some
current from another supply into the primary of a 1:10 transformer
(with some heavy transistors),
and it’s output is the 90-100V standard Australian phone ring.

But that only was intended to make an old Bakelite phone ring.

To superimpose the ringer straight over the DC circuit, you’d need to
either switch between one or the other, or protect the battery
pack with diodes or something, is that right?
There’s also the mater of what causes the phone to stop ringing, and
connect the DC circuit when it’s taken off the hook.

For this modem situation, I have one of the old hand cranked dynamos
which would be funny to try getting a modem to pick up on its own.
But at that moment, when something takes the modem off hook,
 surely you must immediately prevent the 100V AC
ringer circuit from working don’t you?

[Berni]

The battery will be fine since you should connect the ringing voltage in parallel acros one of the two phones. And yes the moment you pick up the phone the ringing waveform stops. Telephones are electrically pretty tough things tho(especially the old rotary ones as there are no semiconductors or even active devices)

It's amazing what a layering of legacy technologies the phone network is. And yet the same 90V AC signal is being used today by modern voip enabeled router boxes at home to tell a cordless phone that it's time to ring.(While the said router box gets to the Internet via ADSL on the old telephone wires that no longer carry the analog signal)

[@rt]

Technology rotating one bit to the right

For a phone line the ringing might stop when you pick up the phone,
but if I were to connect the ringing signal across mine,
what makes that stop when a modem picks up?
and even if that doesn’t damage a modem,
The event of it going off hook would need to be used to turn it
off anyway, or it will still make the first modem to hang up start ringing again.

[Berni]

Yeah you don't have detection for that. But if you only apply it for a few 100s of a millisecond then even if the on hook resistor can't handle the full waveform it not have time to get hot. The modem will proabobly still see it as a ring tho as sometimes the waveform from the exchange starts half way trough the ring

[2N3055]

http://www.freeinfosociety.com/electronics/schematics/sensor-based/pictures/telephonelinesimulator.gif
EDIT : this link is broken, image attached

[SeanB]

Trying to emulate a complete phone line is going to be more effort. first you will need a 48VDC supply, but it only has to supply 100mA at most. Each phone will need a 48V 600R coil relay to provide it with power, and you will need a 600R 1:1 audio transformer to couple the audio between them. Ringing is simply 100VAC 25Hz applied via a 2uF capacitor to the phone line, and turning off the ring voltage is when the phone is off hook and the relay contacts close.

However just to connect 2 modems this is a little overkill, though you can make a nice intercom using regular phones with this. There are circuits online (IIRC Don Lancaster published a few) top do this as well.

[@rt]

That links comes up “Forbidden” for me.
I have been told the 48V is present when the line is idle,
and drops to 8-12V for the voice circuit. Is this correct?
It does seem to explain why other diagrams use a single 9V battery.

[2N3055]

That links comes up “Forbidden” for me.
I have been told the 48V is present when the line is idle,
and drops to 8-12V for the voice circuit. Is this correct?
It does seem to explain why other diagrams use a single 9V battery.

Sorry it is broken to me now too.. I uploaded image up there.

[2N3055]

Originally it was like this:

1. 48V power
2. 48V high sensitivity relay. It serves as a current source, an AC block (high inductance), and also as an Off-hook detector
3. Voltage was 48-60V open 10-12V with phone off hook. Current draw was 10-20mA.

Good results can be had with 24V (many small inexpensive Panasonic in house switches had 24 V, better ones had 48V).
Also 12V/220Ohm combination should work quite well, it is lower impedance that in should be for the voice signal, but that is not a problem when locally connected.
You can create dial tone with arduino pin, lowpass filter it and attenuate to get 100mV ish level, couple it with a capacitor to the line.

It can be done, but it deceptively complicated to do it right..

[pamperchu]

Set Up A Dial-In Server ,

h++ps://www.howtoforge.com/linux_dialin_server

[Berni]

Oh and there is also the easy way out of simply buying one of those household telephone exchanges that connects 2 to 10 phones to one POTS line. Connect all your modems to it as phones and simply dial the internal number of the other one. You get the dial tone, ring signal and perhaps even a caller id signal to both of the modems.

By the way 2N3055 I'm in your country on holiday. Good luck with the big soccer match.

[2N3055]

Oh and there is also the easy way out of simply buying one of those household telephone exchanges that connects 2 to 10 phones to one POTS line. Connect all your modems to it as phones and simply dial the internal number of the other one. You get the dial tone, ring signal and perhaps even a caller id signal to both of the modems.

By the way 2N3055 I'm in your country on holiday. Good luck with the big soccer match.

Thanks a lot! I hope you are having a good time, enjoy the holidays... Lijep pozdrav!!

[kakabouras]

Back in 1989 I was assigned the task to make two dial-up modems to talk to each other over a leased line.

After a few experiments the following solution was implemented :

- I used one modem as a call initiator by using a command like ATX0DT5555555.

- X0 means do not detect dial or busy tone.

- DT5555555 means send DTMF tones corresponding to number five ie 1336Hz and 770Hz.

- At the other end I implemented a dual tone detector using two LM567 with their outputs tied in a logic AND configuration. I could have used a special DTMF detection IC like  CMD CM8870CSI  but those were not easily available in Greece then.

- When the DTMF signal was detected a hardware interrupt signal was used by the receiving end computer to send an ATA command to the receiving end modem.

It worked like charm for this telemetry project for many years until it was replaced by internet DSL lines in 2007.


When I described the idea to my boss he said  : "That's too simple to work".
Afterwards when we delivered the project he called the whole idea "the Columbus'  egg" .

[2N3055]

You don't have to detect dialed number, just play dial-tone after off hook, insert few second pause, ring other side, and after it goes off hook connect them..
Or if you put ignore dial-tone in dial string for modem, no need for dial-tone at all.

[@rt]

Those telephone line simulators are neat,
but still surprisingly expensive on eBay for technology on the very edge of obsolescence.

[CJay]

Those telephone line simulators are neat,
but still surprisingly expensive on eBay for technology on the very edge of obsolescence.

Because that stuff is on the verge of being obsolete it's possible to buy office PABX boxes for cheap, they'd do just fine but they don't tend to be particularly small.

[Peabody]

This is going back a long way, but my memory is that the DAA, which is the part of the modem circuit that connects to the phone line, usually runs the lines through a FWBR because often Tip and Ring are wired backwards.  The FWBR makes the modem work with either polarity.  But with a direct connection with no voltage on the line, nothing gets through the two diode drops of the FWBR.  So you need to put some DC voltage on the line so the audio will get through.  But I'm confused that this was described earlier as putting the 9V battery "in series".  The battery would be *across* tip and ring.  Right?

And I also thought there were AT commands to get around the absence of dial tone and ringing.

[@rt]

The ring signal goes across the ring and tip, which puts the phone’s ringer solenoid in series I presume.
The circuit for speech/data is a series circuit (in series with either the ring or tip).

Just in case this thread is searched, the current arrangement worked well until I tried a PCMCIA card modem.
To get this working I use a pair of 12V SLA batteries in series with a 510 Ohm resistor.

Funny enough though, this is also a Fax modem, and when I tried it with Telstra’s fax test number,
it couldn’t interpret the fax to return it. The report was as if my modem was silent after connecting.
I had to go to a friend’s house for this, as he still has a POT line which is getting pretty rare in Australia, as they are being phased out.

[NiHaoMike]

I had a 28.8k building to building ppp link running for several years on some Hayes Optima modems over one spare pair.

Nowadays, it's pretty easy to hack some Homeplug adapters and send data at a few hundred Mbps through a few hundred feet of telephone wire, basically a cheap way to get a pair of really fast "short haul modems". Isn't technology amazing?

[Rasz]

I had a 28.8k building to building ppp link running for several years on some Hayes Optima modems over one spare pair.

Nowadays, it's pretty easy to hack some Homeplug adapters and send data at a few hundred Mbps through a few hundred feet of telephone wire, basically a cheap way to get a pair of really fast "short haul modems". Isn't technology amazing?

dont you run into the same problem of needing to pump AC voltage into the line first?
do those Homeplug devices work over dead wire?

[SeanB]

dont you run into the same problem of needing to pump AC voltage into the line first?
do those Homeplug devices work over dead wire?

You separate the power pins from the RF injection side, and run the RF over phone cable or CAT3 cable to the other side, where you have done the same. The units do not check for AC power, as they are powered by the AC line, so if there is AC power they work, sending out a query RF pulse every so often to see if there is another unit on the line, and listening for a query back. Then they use Rf to negotiate the link, they do not sense line voltage at all, and will after modification work well over a dry cable link.

[Rasz]

dont you run into the same problem of needing to pump AC voltage into the line first?
do those Homeplug devices work over dead wire?

You separate the power pins from the RF injection side, and run the RF over phone cable or CAT3 cable to the other side, where you have done the same. The units do not check for AC power, as they are powered by the AC line, so if there is AC power they work, sending out a query RF pulse every so often to see if there is another unit on the line, and listening for a query back. Then they use Rf to negotiate the link, they do not sense line voltage at all, and will after modification work well over a dry cable link.

interesting, found this https://forum.openrov.com/t/teardown-of-a-homeplug-adapter/305/32
I remember from reading up on X10 20 years ago that it detected power phase and only transmitted inside zero crossing window, somehow expected something similar here.

[tooki]

Seem to remember 56K modems won't work back to back, at least not at 56K, as they rely on some trickery with the digital telephone exchanges so you might need to persuade them to connect at a slower speed.

Yep. Pre-56K, your ISP had banks of modems, and when you called, you got a switched circuit (in theory; read on) to the ISP's incoming phone lines, where a free modem would then take the call and connect. It was very much like how a pre-VoIP call center worked, with banks of operators waiting to take your call.

But of course, we haven't actually used true circuit-switched lines for phones in most places for decades. Instead, at the local exchange (where your phone line physically terminates), an A/D converter would digitize your audio, compress it down to 64 or 56Kbps, and then send it digitally through the phone network, and then at the local exchange at the receiving end, it would be converted back to analog.

Of course, the A/D conversion and compression algorithms were designed for human voice, not for modem noises (or music, which is why hold music sounds so awful). Somebody realized this is silly for a modem; you were taking digital, converting it to analog, converting it to digital, compressing it, decompressing it, converting it back to analog, and then converting that back to digital!

Enter the 56K modem.

In essence, the 56K speeds worked by not actually connecting the client modem through to a modem at the ISP. Instead, the ISP modem is actually at the local phone exchange, and from there it went digitally to the ISP itself. By avoiding the voice-centric ADCs and compression algorithms and instead using the 56Kbps channel bandwidth for the digital data directly, it increased bandwidth and slightly reduced latency. (56K was chosen because not all phone networks supported 64Kbps channels — some used 8Kbps for signaling, leaving 56Kbps payload.)

The other advantage is that only the "last mile" (aka the local loop: the wire between your phone jack and the local exchange) were analog, compared to potentially much, much longer on a true analog circuit-switched line, where it might be many miles of analog signaling between you and the ISP.

As an aside, ISDN essentially connected you directly to the 64Kbps digital channel. The total absence of A/D conversion is why ISDN was a lot faster for Internet use, despite having similar bandwidth: far less latency.


ADSL (ADSL, VDSL, etc) works the same way as 56K modems, in that the ISP equipment (the DSLAM) is actually in the phone exchange. The only-the-last-mile-is-analog advantage applies here, too. In fact, more so, since ADSL depends on using only frequencies beyond what POTS lines were designed for (since ADSL leaves the voice frequencies free so that it can piggyback on an active voice line). POTS lines' frequency bandwidth plummets with distance, taking data rate "bandwidth" with it. Because of this design, the uplink channels are in one block of frequencies, and the downlink channels are in a different, much larger block of frequencies, so the modem can only send one block of frequencies and receive the other, and the opposite for the DSLAM.

In contrast, the much rarer SDSL (SDSL, SHDSL, etc) (which doesn't piggyback on an active voice line, and uses voice frequencies) can theoretically have the endpoint in the local exchange, but was more commonly used with a leased pair (a true pair of wires permanently wired through from you to the other end, onto which you can place whatever signal you want, within reason). As such, SDSL modems must be able to both send and receive all the same frequencies. Consequently, two ADSL modems are categorically incapable of directly communicating with each other, while two SDSL modems will happily connect over a copper pair.

[Beamin]

The phone line injects power into the wires. That's where phones get there power from.

You can get two phones to talk to each other by connecting a 9V battery in series with one of the phones terminals before connecting them with together with a phone cable. The battery pushes current around the loop to power both of them, munch like the telephone exchange would. Audio is transmitted by varying how much current is drawn from the line.

Perhaps the battery trick works with modems too in order to trick them into thinking they are on a real line. Hopefully they don,t also expect to hear a dial tone too, but perhaps that can also be bypassed by simulating the line ringing by injecting 100V AC on the line for a short moment. When you pick up a ringing line you don't get any dial tone, it's straight to audio(apart from the caller id signal nowadays)

Why don't cell phones have the text caller ID like land lines did? Its like we went a step backwards. Also phones should sound crystal clear but they don't you really notice this when you face time someone then call them right away. Im sure it cuts into profit but a voice would be such little data compared to everything else.

Doesn't the phone run at 50 or 80 volts? with the ringer at 110 ac? I remember getting shocked from the red wire but not the green.

[SeanB]

Caller ID is a network operator setting, often they turn off the number reception unless you pay for it, though here that is only on contract SIM services, where there is a $1 per month charge for the "privilege" of getting it, but the prepaid market has it by default turned on. The voice quality is simply because they compress the heck out of the data, and add a little noise so that they can send it at a very low data rate in the individual packets. The original spec had a roughly 4kb/s data rate, and only roughly 6 bit resolution, so the data could be packed densely to save transmit power in the phones. Just had to pass the phone bandwidth of roughly 300Hz to 6kHz for bandwidth, and as the compression was logarithmic it was very compressed amplitude wise, with a 2 bit dither added so that silence still gave a data stream of some minimum rate, so the cell network would not think the phone was out of range of the tower and drop the call thinking it was ended, instead of simply having a quiet background and thus almost no output from the CODEC in the phone.

[tooki]

Why don't cell phones have the text caller ID like land lines did?

I have wondered this for years, too!!!!

I have a theory on this. Land-line caller ID was a kluge, built onto the analog POTS network. It worked by using a (IIRC) 2400 baud modem signal between rings to send the Caller ID information. This was introduced in 1987-88, but didn’t catch on for a few more years. Meanwhile, also in 1988, ISDN (native digital phone lines; exceedingly rare in USA for domestic use, somewhat more common in Europe, for example) was born, and to the best of my knowledge, only included numerical Caller ID. Now, if you look at the feature sets of GSM and ISDN, they are practically identical. One could easily assume that when they were designing GSM, someone said “GSM should just be wireless ISDN”. This is absolute speculation on my part, but there’s some plausibilty to it. Letting ISDN and GSM use the same functions would certainly simplify the phone network to some extent. And with ISDN having only numerical caller ID, so did GSM.

With that said, at least modern mobile networks do support alphanumeric caller ID — at least to some extent, namely in SMS text messaging. SMS seems to allow alpha caller ID, as witnessed by commercial SMS messages, like the specials I get from IKEA occasionally, whose caller ID is “IKEAFAMILY”. (And no, you can’t reply to them, in case you were wondering!)

But I’ve never seen a mobile network enable alphanumeric caller ID for calls.

Oh yeah. It may be worth explaining to European (and perhaps other non-North American) readers what we even mean, since it was only ever available in North America as far as I know: Caller ID on land-lines used a little box with an LCD, and it would show not only the number, but also the name of the caller (and the name was indeed being transmitted from the phone company, it wasn’t just looking it up in the customer terminal). In contrast, cellphones only receive the number, so if you see the name, it’s because the phone has a matching name in your saved contacts.

Also phones should sound crystal clear but they don't you really notice this when you face time someone then call them right away. Im sure it cuts into profit but a voice would be such little data compared to everything else.

Actually some mobile phones and networks do have improved-quality mode. (The official name is “wideband audio, but most carriers market it as “HD Voice”.) But both handset and network support is so spotty that almost no calls seem to actually use it.

Doesn't the phone run at 50 or 80 volts? with the ringer at 110 ac? I remember getting shocked from the red wire but not the green.

No, a POTS line runs at 48V DC, but when the phone picks up and loads the line, it drops to 3-9V. The ring signal is 90V AC.

Caller ID is a network operator setting, often they turn off the number reception unless you pay for it, though here that is only on contract SIM services, where there is a $1 per month charge for the "privilege" of getting it, but the prepaid market has it by default turned on.

Really? Caller ID was always a premium option for POTS lines, but I’ve never ever heard of a mobile network charging for it!

The voice quality is simply because they compress the heck out of the data, and add a little noise so that they can send it at a very low data rate in the individual packets. The original spec had a roughly 4kb/s data rate, and only roughly 6 bit resolution, so the data could be packed densely to save transmit power in the phones. Just had to pass the phone bandwidth of roughly 300Hz to 6kHz for bandwidth, and as the compression was logarithmic it was very compressed amplitude wise, with a 2 bit dither added so that silence still gave a data stream of some minimum rate, so the cell network would not think the phone was out of range of the tower and drop the call thinking it was ended, instead of simply having a quiet background and thus almost no output from the CODEC in the phone.

POTS bandwidth is specified as 300-3400Hz. Analog local loop circuits often did much better (hence why DSL is even possible, but also why old phones with large speakers and microphones could sound so clear), but the amplifiers, filters, and multiplexers (in the purely analog days) and the ADCs and compressed codecs (once we went digital) adhered to those limits to maintain efficiency on the trunk lines. As such, the mobile codecs had no reason to exceed that — and often do worse, since they have to adapt to poor link conditions.


The modern HD voice standards extend the analog bandwidth to 50-7000Hz, with some taking the upper bounds all the way to the edges of human hearing, at 20KHz. They also sometimes use higher bit depths to reduce noise.

[janoc]

Really? Caller ID was always a premium option for POTS lines, but I’ve never ever heard of a mobile network charging for it!

I seem to recall that back in the early 90s when GSM was coming to Slovakia we had CLIP (caller ID) and CLIR (blocking of display of your number when calling someone) as a paid service. It was stupid and the telcos quickly abandoned it. I believe only CLIR was kept as some sort of a premium service for a while.

[tooki]

Weird. I’ve had SIM cards (or mobile accounts for networks that don’t use SIMs) from many countries (Switzerland, USA, UK, Guatemala, Thailand) and I’ve never had one that charged for any of those services! Not saying that I don’t believe you, of course... the price models for mobile calls have varied wildly over the years (e.g. how in USA, the mobile user pays for incoming minutes, too, or how they charge extra to enable tethering)!!

[Beamin]

Really? Caller ID was always a premium option for POTS lines, but I’ve never ever heard of a mobile network charging for it!

I seem to recall that back in the early 90s when GSM was coming to Slovakia we had CLIP (caller ID) and CLIR (blocking of display of your number when calling someone) as a paid service. It was stupid and the telcos quickly abandoned it. I believe only CLIR was kept as some sort of a premium service for a while.

. Why was it stupid? I loved it. Plus our 900 MHz digital cordless phone supported it . Best phone ever long range too. People thought it was a cell phone.

[tooki]

I’m pretty sure it’s the charging extra part that he’s calling stupid.

[janoc]

I’m pretty sure it’s the charging extra part that he’s calling stupid.

Yes, sorry for not being clear. It was abandoned because it was actually more costly for the telcos to not provide the service (the caller's number is transmitted by default and they had to block it for everyone not paying for it).

[SeanB]

On POTS caller ID is there in the central office frame for billing purposes, but sending it to the customer side is a simple bit flip in the configuration for the port in the card in the frame. Was generally here turned off unless you paid for the service, but the cellular providers turn it on for prepaid so that they get return calls off the prepaid service, otherwise no calls unless you leave a voice message, and most owners on prepaid will not call back to unknown numbers, or reply at all to no number calls. postpaid (contract) numbers they charge a fee as part of the monthly subscription cost, which is just a money making thing for them, the $1 per bill is just extra profit off the business customer, who in any case is paying off the inflated handset price anyway over a 24 or 36 month period. They just get some sop of free minutes, a small data bundle or a certain number of SMS sendings, that is almost always going to be used anyway. No operator these days here offer unlimited calls or data, though there are some grandfathered plans where you get actual value if you have a massive call volume or data use, but they do not offer that to those who can use it.

[Beamin]

I’m pretty sure it’s the charging extra part that he’s calling stupid.

Yes, sorry for not being clear. It was abandoned because it was actually more costly for the telcos to not provide the service (the caller's number is transmitted by default and they had to block it for everyone not paying for it).

Thats bad marketing, they should have said " Sick of lossing a limb running accross a mine field to get home to your phone only to realize its a tele arms marketer? Well no more because: FREE CALLER ID AT SLOVAKTEL !!! A VALUE WORTH TWO LANDMINES OR 1 WEEK OF RATIONS!!! YOURS FREE!!! It doesn't cost an arm and your wooden leg. So cheap even a Chek could afford it! Nikoli chochesko and slobodon milosevich approved!!! Call now!!

[ebastler]

FREE CALLER ID AT SLOVAKTEL !!! A VALUE WORTH TWO LANDMINES OR 1 WEEK OF RATIONS!!! YOURS FREE!!! It doesn't cost an arm and your wooden leg. So cheap even a Chek could afford it! Nikoli chochesko and slobodon milosevich approved!!! Call now!!

Mixing up four Eastern European countries in just a two-line post is quite a feat... 

[Beamin]

FREE CALLER ID AT SLOVAKTEL !!! A VALUE WORTH TWO LANDMINES OR 1 WEEK OF RATIONS!!! YOURS FREE!!! It doesn't cost an arm and your wooden leg. So cheap even a Chek could afford it! Nikoli chochesko and slobodon milosevich approved!!! Call now!!

Mixing up four Eastern European countries in just a two-line post is quite a feat... 

I can't keep up, weren't they all one country at one time? That was my joke now they are like 7 or 8 defacto countries because no one could get along. Most creative but pure evil thing I saw was a hand grenade placed in a glass jar with the spoon resting on the glass, placed on a door knob, so you open it the glass breaks and you are a victim of a forgotten war. That's almost impossible to clean up.


BTW did he get them working? We used to play doom two like this we had two phone lines in my house and two computers so all my friends would come over and just using doom you could dial out and back into to the next room. Deathmatching it was called. We were spoiled but didn't realize it. 

[janoc]

I can't keep up, weren't they all one country at one time?

Well, the last time that this was one country was perhaps under Franz Joseph in the Austro-Hungarian empire before the WWI.

Czechoslovakia has split in 1993, Milosevic is Serbian ( FYI that's down south in Balkans, almost 1000km from Prague, former Yugoslavia) and Ceausescu ruled Romania which is (and was) a totally different country, some 600km farther east from Serbia, bordering the Black Sea ... You need to up your geography game, methinks.

That was my joke now they are like 7 or 8 defacto countries because no one could get along.

Sorry but your joke is neither funny nor correct. Only shows your ignorance. Czechoslovakia had never had anything in common (not even border) with the former Yugoslavia countries - for your information, there is still Austria and Hungary between the two regions - and about 500km straight line distance.

And many of these "defacto" countries are both EU and NATO members.

One would think that 20+ years after the events (both the fall of communism, split of Czechoslovakia and the bloody Balkan wars) people would at least know where those places are on the map. Especially given that the US (and NATO allies) has bombed Milosevic's regime in the past.

[ebastler]

Thanks for the great summary, Janoc. Just one additional comment: The split into Slovakia and the Czech Republic was a peaceful one; no bombs or such involved at all. https://en.wikipedia.org/wiki/Dissolution_of_Czechoslovakia

And now, back to modems?

[janoc]

Thanks for the great summary, Janoc. Just one additional comment: The split into Slovakia and the Czech Republic was a peaceful one; no bombs or such involved at all. https://en.wikipedia.org/wiki/Dissolution_of_Czechoslovakia

Yes, fortunately. People weren't happy about it (not counting the nationalists on either side) but we have managed to do it in a civilized manner. On the other hand, the nationalist fervor or the historical baggage we had between ourselves (I am Slovak) was nowhere near as bad as in the Balkans.

[Rasz]

coming back to original topic, stumbled today onto this quite funny video