• Please review our updated Terms and Rules here

Acoustic Coupler Help

I seem to remember early fax machines had acoustic couplers in the mid-70s. Fax machines became more common late 70s (Vodafone started life as Vodafax in UK). Our then-monopoly, now Telstra, insisted on dedicated lines for fax machines. As I recall it, the audio-coupled dialling standards were applied to faxes a good ten years before data modems. The telcos hung on grimly to their profitable monopolies in 50-baud telex services and kept data services at bay as long as they could hold out. Ma Bell invented generations of excuses about "protecting network integrity", and monopoly telcos worldwide happily emulated them. It was an earlier version of the net neutrality arguments - except that the consumers who wanted more liberalised network services were far less organised and wealthy than Google and friends.

It was Japan that forced fax into the mass market - their script was much less friendly to keyboards and limited mechanical font sets, they had the demand, the capital, the manufacturing scale and the ability to enforce standards that drove fax penetration first across Asia, then into the West. IMHO that is what accelerated extinction of the audio-coupler.

Rick
 
That is really interesting. I never really thought of fax machines being related to the evolution of modems. It seems that computer history is linked to telephone history, especially in regard to networking.
 
Yes, and also remember that the first networks, both telegraph and wireless, were "digital" using Morse code. I believe a good operator could do about 3 "bytes" (actually 3-bit letters) per second.
 
I seem to remember early fax machines had acoustic couplers in the mid-70s. Fax machines became more common late 70s (Vodafone started life as Vodafax in UK). Our then-monopoly, now Telstra, insisted on dedicated lines for fax machines. As I recall it, the audio-coupled dialling standards were applied to faxes a good ten years before data modems. The telcos hung on grimly to their profitable monopolies in 50-baud telex services and kept data services at bay as long as they could hold out. Ma Bell invented generations of excuses about "protecting network integrity", and monopoly telcos worldwide happily emulated them. It was an earlier version of the net neutrality arguments - except that the consumers who wanted more liberalised network services were far less organised and wealthy than Google and friends.

It was Japan that forced fax into the mass market - their script was much less friendly to keyboards and limited mechanical font sets, they had the demand, the capital, the manufacturing scale and the ability to enforce standards that drove fax penetration first across Asia, then into the West. IMHO that is what accelerated extinction of the audio-coupler.

I recall early FAXes using photosensitive paper wrapped on a spinning drum. There was also a gizmo called a telautograph that wasn't a FAX, but a pantograph arrangement for a sender to write and have his motions traced out. I think today's Autopen is a direct descendent of that technology, which dates back to Elisha Gray. When I worked in steel mills sometime after the discovery of fire, they were used extensively to telegraph information between the shipping office, QA lab and the production line. Kind of cool to watch working.

And then there was the Picturephone--I remember it being presented as bleeding-edge technology in the 1960s. The idea was that you'd have Picturephone locations where you and the called party could to to videoconference. It was one of the major attention-getters at the New York World's Fair.

Are there still US Weather Service map FAX broadcasts on shortwave? It's been a long time since I've listened...
 
All of these technologies sound very impressive. I can see why some of the old technologies are displaced by modern computers, but at the same time I think that it is somewhat sad that they are pushed aside. It is amazing just what analogue can do.

As for Morse code, that is a really good point. I never really think of it as digital, but since it has a limited number of states it is.

I do believe that there are still weather faxes on shortwave. I like to listen to shortwave sometimes, and the last time I looked into finding some software to decode digital/non-voice transmissions, the software mentioned compatibility with fax as well. According to Wikipedia they are still used, and a Japanese news service still transmits the news this way. http://en.wikipedia.org/wiki/Radiofax
 
I wonder how far past 56k modems we would have gotten if they hadn't come up with cable modems and DSL lines. Would we all be on USR Sportster 1Mbps dialup? LOL
 
I was fortunate enough to maintain one of the very first computer-computer networks, the US Navy NTDS system. The modem (AN/SSQ-29) was initially designed in 1959 and was in service by the time I began my 18 months of training in 1966. The data link was over HF radio (2-30 mhz), and once when skip conditions were good our ship in the Gulf of Tonkin linked up with a ship nearing San Francisco. We got around 2500 bps over the link.

Here is a very lengthy excellent write-up on NTDS (with lots of pictures of old equipment): http://www.ieeeghn.org/wiki/index.p...Story_of_the_Naval_Tactical_Data_System,_NTDS
 
Thanks for the article. I haven't read it completely yet, but I saved it in my favorites.

How reliable were those transmissions on HF? It seems like a lot could be messed up with the signal with all the possible sources of interference.
 
Thanks for the article. I haven't read it completely yet, but I saved it in my favorites.

How reliable were those transmissions on HF? It seems like a lot could be messed up with the signal with all the possible sources of interference.
When the SSQ29 was properly adjusted the reliability was quite good. Among other things there was a forerunner of ECC on the data sent, which could correct up to triple bit errors as I recall. The entire HF radio (the collins SRC-16) was a class A double sideband. A 5KW Class A power amp was quite a unit, about 7000 VDC at around 2A driving the 4CX5000A water-cooled tube. Everything in the radio was phase-locked to a 100khz NBS calibrated frequency standard.

The data bits were sent over 15 quadrature phase modulated audio tones giving 30 bits per frame. I don't remember the frame time but I believe it was around every 100ms or so. The SSQ29 had delay-line "resonators" per tone which would "remember" the previous frame so it could compare phase with the incoming frame and decode the data bits. The resonators were a pain to adjust properly, which kept me busy whenever we were in port and I could work on the beast, but once adjusted it worked very reliably. Out at sea when the link was up I had nothing to do unless there was a problem.

It was really fun stuff to work on in those days, complete with surface-mount chips and a copy of http://en.wikipedia.org/wiki/Spacewar! someone ported over to Univac computers. We had to do extensive display console testing using the Spacewar game when in port :)

This was all very similar to working over telephone lines, the available bandwidth was very similar to the bandwidth of a standard POTS line. I suspect if I knew more about modern modem encoding I'd see a lot of similarities in the techniques used. I had a serious flashback recently when I was reading up on QAM tv tuning.
 
The main constraint on modems designed for the POTS at consumer level was that the whole system originally had to fit within the POTS audio frequency response standards, calibrated for the human ear, which could be 4Khz or less. DSL systems using inaudible frequencies took a long time to develop, required huge network re-investment, and are still limited by the performance of the oldest copper network segments, though clever ECC coding has kept pushing the limits a bit higher with each generation.

Fortunately, digital mobile telephony systems since GSM have all been designed with data carriage in mind, though still constrained by bandwidth sharing on any given base transceiver. Analog mobile systems were even more constrained in bandwidth and all acoustic fidelity standards than the POTS - good riddance to them.

Rick
 
@Doug G Sounds awesome that you got to play a game to test the computer. :) It really impresses me that the signals would work so well on HF. When I listen to the HF freqencies I can barely hear some audio transmissions, but I didn't think about error correction. That sounds like it was extremely robust.

@Rick So data has probably reached its limit with POTS, or could a little bit more throughput be achieved? Beside cell phone use it seems like the telephone companies would start phasing in better lines since data usage is more prevalant. The long distance lines could probably handle in now because aren't a lot of those fiber optic.
 
The issue with POTS handling anything faster than 56K is that it would require an overhaul of the network. When your analog POTS call gets to the CO, it's converted to PCM (8KHz sampling rate with 256 levels). This gives an absolute maximum for a digital connection of 64KHz (256/8 bit levels at 8KHz). For miles of copper pair, that's pretty remarkable, given that some of those pairs can be over 50 years old, run through a virtual maze of compensation coils, splices, leaky cables, etc. It's a wonder that it works at all. But you don't ever get 64K samples/second--the telco steals some of the bandwidth for signalling. So, 56K is a reasonable maximum. The further you get from the CO, the more the analog signal is distorted. Eventually, the modem determines that there's nothing to be gained and falls back to plain old quadrature amplitude modulation (QAM) and you top out at 33.6K, or whatever the line will permit.

On voice, many telco have an informal standard of "hear and be heard" for voice-grade lines. It used to be that if you asked pretty please and were friends with a lineman, the telco would tweak the equalization on the line for you to squeeze a bit more out of it. But telcos hate sending data using voice-grade lines, particularly when they'd love to sell you ADSL service.

So, no, if the line is good enough for voice, it's extremely unlikely that there are equipment upgrades in the future to enable faster rates over voice-grade lines.
 
It really impresses me that the signals would work so well on HF. When I listen to the HF freqencies I can barely hear some audio transmissions,
When one listens to voice over HF radio, one is listening to a wide range of frequency. The amount of received noise is proportional to bandwidth. If instead of using voice, one switches to say, dual-tone FSK, then that is only two specific audio tones being sent. The FSK receiver therefore uses a narrower bandwidth and thus receives with much less noise in comparison to a receiver designed for audio.

For that reason, many of the HF radio links that I was involved in (many moons ago) used teletype-to-teletype-via-FSK rather than voice. Whenever the quality of a HF radio link headed in a downwards direction (due to weather, atmosphere, etc.), FSK continued to be useable long well after voice comms on the same link became impossible.
 
For that reason, many of the HF radio links that I was involved in (many moons ago) used teletype-to-teletype-via-FSK rather than voice
Our NTDS Link-11 didn't use fsk, rather 15 audio tones covering about 2K of bandwidth. The radio system was also used as general purpost HF radio for voice, etc.

But we had a pretty spiffy radio! It took 7 or 8 full-size cabinets and had it's own room on our ship. Here is a good pic, scroll down and view the AN/SRC-16 which was my baby to keep running. http://www.virhistory.com/navy/xmtr-ship.htm
 
You can't actually get 56k POTS to POTS. Remember that 56k is only 56k download, its only 28.8 up. Thats because 56K modems rely on the answering end having direct access to the digital sysem via ISDN and can by-pass the filters.

The next pahse was ISDN telehony. Its a pity this has largely disappeared because TELCOs din't want to offer ISDN and ADSL but ISDN allows 2x64k which can e bonded to give 128k. There is also a 9.6k signalling circuit. Going back to POTS from my old ISDN system were the signalling was done digitally back to a system where its done by ringing current is a real step backwards.

Perhaps in the UK where BT is switching its network over to be entirley IP based:-

http://en.wikipedia.org/wiki/BT_21CN

http://www.btplc.com/Sharesandperformance/Industryanalysts/Newsletter/Issue15/Feature/index.htm

we might get back to some of the nice features you got from ISDN ....
 
Thanks for all of the info guys. So dial-up has pretty much reached its limit now.

@Doug G: That is an awesome radio system. I bet it was really cool to operate.
 
You can't actually get 56k POTS to POTS. Remember that 56k is only 56k download, its only 28.8 up. Thats because 56K modems rely on the answering end having direct access to the digital sysem via ISDN and can by-pass the filters.

I neglected to mention that, but I knew that from my own experimenting. So why was 56K only 28.8K up, when 33.6K was possible?
 
Remember that 56k is only 56k download, its only 28.8 up.
With V.90. V.92 fixes this problem. It still require the answering modem be connected via a digital connection like ISDN.
 
The issue with POTS handling anything faster than 56K is that it would require an overhaul of the network. When your analog POTS call gets to the CO, it's converted to PCM (8KHz sampling rate with 256 levels). This gives an absolute maximum for a digital connection of 64KHz (256/8 bit levels at 8KHz). For miles of copper pair, that's pretty remarkable, given that some of those pairs can be over 50 years old, run through a virtual maze of compensation coils, splices, leaky cables, etc. It's a wonder that it works at all. But you don't ever get 64K samples/second--the telco steals some of the bandwidth for signalling. So, 56K is a reasonable maximum. The further you get from the CO, the more the analog signal is distorted. Eventually, the modem determines that there's nothing to be gained and falls back to plain old quadrature amplitude modulation (QAM) and you top out at 33.6K, or whatever the line will permit.
I wonder what would be the theoretical limit if the phone system was still analog from end to end.
 
I wonder what would be the theoretical limit if the phone system was still analog from end to end.

Back in the day when the only (seeming) way to get high-speed service was leased-line, you could order up a bare copper pair (offered to burglar and fire alarm companies). It was expensive, but not as expensive as a leased line. I suspect some telcos still offer this local service, but probably don't offer new leases--and Ma Bell frowned on using the direct connect for data (between, say, two corporate offices). That was about as close as you could get. I think that direct microwave links turned out to be much cheaper if you could get line-of-sight path.

I recall that a 9600 bps (synchronous) leased line between Minneapolis and Santa Clara ran about $5K/month, but that included the 208 modems at both ends.

Well, bandwidth isn't everything. Was it Andy Tannenbaum who said "Never underestimate the bandwidth of a station wagon full of tapes hurtling down the freeway"? I used to fill my big Samsonite attache case with 9 track tapes (I think I could squeeze in 6 tapes) and hop a plane to whatever installation was having problems to deliver a new system release. Probably more bandwidth over,say, 8 hours than could be gotten any other way at the time.

I remember that Lockheed used to send data to its test site in the Santa Cruz mountains (at the end of Felton Empire Road) from the Blue Cube in Sunnyvale via carrier pigeon. It was much cheaper than using a data link and faster than hand-carrying the data.

But purely analog telephone was horrible, particularly for long distance. As the line length went up, so did the noise. Many larger companies had access to Centrex telephone and their own private set of tie lines. (After hours, this was handy if your relatives lived in a city served by the company Centrex--you'd simply dial a string of numbers and then get an outside line at the far end. Bingo--free phone calls). I used to have fun while waiting for things to happen by figuring out the most circuitous path through the system, making the phone on the next desk ring. Someone would pick it up and it would sound like I was calling from the other side of Mars, barely audible over the line noise.
 
Back
Top