To many people, connecting to the NBN infers giving up their previously used phone(s), whether DTMF (tone dialling), or decadic (rotary/pulse/loop-disconnect) dialling. This, in actual fact, is not true, and this article explains how I was able to restore full operation of all my phones.
But first, for non Australian readers, what is the NBN? It's a national fibre optic broadband system (National Broadband Network) that is intended to replace the existing copper telephone cabling which also supports the ADSL internet. There is so much written about it, I won't go into further detail, but the important point is that once fully implemented, the existing copper cable from the exchange to the house will no longer be available. In other words, one is forced to subscribe to the NBN, with no choice to retain the existing exchange line.
The NBN reaches the consumer in a number of different ways, depending on the technology available at the time, cost, and existing infrastructure. Typically, there is FTTN - Fibre To The Node, FTTC - Fibre To The Curb, FTTP - Fibre To The Premises, and HFC - Hybrid Fibre Coaxial. The initial plan was for all premises to be fed with fibre optic cable (FTTP), but politics (cost and speed of installation) caused the more outdated methods to be implemented instead, for most installations. As a result, the NBN is the laughing stock of Australia and is seen as already outdated. Problems abound with dissatisfied customers; reports of which are readily available in any relevant forum, social media, etc.

Telephones and the NBN.
Knowing I would have to face up to it one day, I had already devised plans of how I'd deal with the loss of my exchange line, and how to keep all my existing telephones working. The NBN service provides a telephone service as well as the internet. It uses VOIP (Voice Over Internet Protocol) for the telephone; something which has been around for some years now. Effectively, telephone communication is done over the internet, and as such, VOIP phones have an IP address etc. You can use a conventional domestic phone with VOIP too, by use of an ATA (Analog Telephone Adaptor). This small box is set up the same way as a VOIP phone, but provides an output which simulates a normal exchange line; e.g. 50V line and 75V ring voltages. Outgoing calls are made in the normal way, and the phone rings normally on incoming calls.
The NBN modems supplied by the ISP actually work on VDSL, and typically, as well as providing a 4 port router for the internet, also have an inbuilt ATA which provides an analog telephone output. At this point, one may plug in a phone to this socket, and a computer to one of the LAN sockets, and everything will function as it did before.
However, a major stumbling block is the ATA built into the modems. These only respond to DTMF dialling. If you connect a decadic dialling phone, the ATA will ignore the dial pulses and it is impossible to make outgoing calls. The phone still rings on incoming calls, and the speech circuits function normally. There are a few stand alone ATA's that do work with decadic dialling, and if the VOIP settings are known, one could be used instead. Some ISP's, such as iinet, will not reveal the VOIP settings, so one is forced to use their modem for the phone.
What is the difference between DTMF and decadic dialling? DTMF is Dual Tone Multiple Frequency, and each time one of the keypad buttons is pressed, two tones are generated. The tone frequencies were carefully chosen to be outside those used in normal speech. The system was developed in the 1960's in the U.S. but did not become commonplace in Australia until the late 1980's. Tone decoders in the exchange determine the digit transmitted. Decadic dialling has been around since the first automatic exchanges at the beginning of the 20th century. It transmits one of ten pulses by repeatedly interrupting the line current. It is also known as loop-disconnect, or sometimes pulse dialling. 'Rotary dialling' is a term used by less technical people, since it relates to a mechanical dial, but it is not entirely correct, since there are many push button diallers that are decadic dialling only. The advantages of DTMF apart from being faster to dial, is that dialling information can be transmitted where it is difficult to retain the DC component, such as over a radio link.

The Dialgizmo and DTMF to Decadic conversion.

Entire unit is built into an RJ11 wall socket.

Obviously, to use a decadic dialling phone on the NBN or any other VOIP network, a decadic to DTMF converter is required. While I have contemplated designing and building one myself, I had a look at what is already available. There are two different methods used with commercially available devices. One is the Rotatone or similar device, which connects directly to the telephone dial inside the phone, and produces the DTMF signal. In effect, the telephone becomes a DTMF type despite the mechanical dial. There are also plans on the internet for homemade devices based around an Arduino or other microprocessors, which operate the same way.
The other method is to convert the decadic pulses to DTMF externally to the phone. This is used by the Dialgizmo and a similar Chinese device available from eBay.
Where multiple phones are in use, each would require its own Rotatone, whereas with a true decadic to DTMF converter, only one is required to feed all the phones. Furthermore, installation of a Rotatone requires modifying the phone and detracts from originality. The line is still DTMF only.
In my situation, with eight phones in use, the choice was obvious and a Dialgizmo was purchased and tested. The idea was to connect it at the phone output of the modem, and thence into the phone wiring of the house. This would restore the ability of using decadic dialling on any phone as previously.

A quick look at the Dialgizmo.
The unit is tiny. It's actually built into an RJ11 wall mounted socket and has a short lead to plug into the DTMF line.

Inside the Dialgizmo. Note the configuration switches.

Inside is the expected microprocessor and crystal to count the pulses and generate the tones. There's also a set of DIP switches to set the configuration for different dials and to enable or disable some of the other features. Not all dials in use around the world have the numerals from 0 to 1 in a clockwise direction. For example, NZ dials are the reverse. The dip switches allow for these differences. Conveniently, it's also possible to dial # and * by holding 1 or 2 against the dial stop for two seconds. There is also a last number redial, and it's also possible to store numbers. Unless one is always using these latter features, it would be to easy forget the sequence since it is not intuitive. In any case, they didn't exist before so I do not feel I am missing out on anything.
The Dialgizmo certainly works, and works well!

• See me testing the Dialgizmo with a series 400 bakelite phone here.

• From reports I've seen so far, the Chinese converter has been problematic. Even though the Dialgizmo costs more, it is my preference. Postscript Feb. 2023: The Dialgizmo has now been in operation for over two years without fault. In fact, having had the modem damaged twice during storms, and the NBN box damaged once, the Dialgizmo remained unaffected.

An Important Note about the Dialgizmo.
The Dialgizmo requires a certain minimum line voltage to function when the phone is off hook. The line voltage appearing across the Dialgizmo is a function of the line current and the resistance of the phone. It appears that the minimum voltage across the phone must be 2V, or the Dialgizmo will not produce the DTMF tones. The problem may arise when a modem/ATA is used which provides a lower than normal line current, and a load with a very low resistance, such as a line isolating transformer. If this happens, insert a resistor in series with the load; typically 100R 1W. See the article on the Gold Phone interface for more information.

NBN Connection Device.

The NCD feeds 60V into the copper pair out to the street pit and powers the optic fibre transceiver installed therein. Disadvantage - the NCD operation adds to my power bill, and unless I use a UPS or power it from my solar DC system, it stops working during a power failure.
From the NCD comes a VDSL signal which then feeds the modem. I was supplied with a TP Link Archer VR1600v. It has four LAN ports as well as both 2.4GHz and 5GHz Wi-Fi.
I can only assume it's for political reasons the NCD and modem are separate boxes. There is no technical reason they could not be in the one unit, which would be a lot tidier.

Modem and Dialgizmo.

The cut-over is an interesting process. While I was away and did not see it, it seems the optic fibre transceiver was installed in the pit on a Saturday. On the Sunday morning I received an SMS to say I could connect the NCD and modem. The existing phone and ADSL were still working normally. The warning is given that once the NCD is connected, the exchange line is permanently disabled. My knowledge of this technology is not great, but my guess is that there are fuses in the optic fibre transceiver, between the exchange line and my house line from the pit, and once the NCD is plugged in, the 60V blows the fuses, thus disconnecting the exchange, and powering up the transceiver. So, cut-over is done by the consumer at their convenience.
Anyway, as soon as the NCD and modem booted up, I had full internet and a dial tone. I could make outgoing calls, but incoming calls had not been enabled until the Tuesday. Temporarily, I was using an 800 series phone with the Dialgizmo, until I sorted out the rest of the cabling.
Apart from deletion of my website hosting by iinet, without warning, it was actually a very easy and straightforward process.

Phones at the House of Cool386.
Now to the interesting part - how did I get all my old phones working? I have eight phones in use, and an answering machine. Here's an introduction to them all:

Workshop phone. 332AT. Three wire operation, bell disconnected.

Hall phone. An ersatz 37AW. Three wire operation. Bell capacitor supplies the other three wire phones.

Living room phone. Telecom Gold Phone. Two wire working.

Bedroom phone. 400APKO. Three wire working; bell disconnected.

Computer room phone. 332AT. Three wire working; bell disconnected.

Kitchen phone. DTMF only 2500 from the U.S. Two wire working.

Answering machine.

Garage phone. 400AW. Three wire working; 2.2k in series with bell to reduce volume.

Shed phone. 891. Three wire working. 4.7k in series with bell to reduce volume.

Boosting the Ring Current.
Theoretical plans were immediately thought of and drawn up for a unit to boost the ring current from the modem. Officially, PMG and Telecom literature stated that a maximum of three bells were allowable. Modern terminology would say this is an REN of 3, where REN stands for Ringer Equivalent Number. If this number of bells was increased, there would be no guarantee they would ring properly due to increased ring current.
In practice, I've never had any problem with six bells on the line; I've not tried more, and since I'm not far from the exchange, the line resistance would have little effect.
In my present installation, there are three bells fed  directly from the line, with another two at reduced power through series resistors, and one opto-coupler in the answering machine. The main bell in the house is the one in the 237AW in the hall. Since the gongs are exposed, and it's mounted on a hollow wooden wall, the ring is very loud. For this reason, all the other three wire phones in the house have the bells disconnected (and it lightens the ring current).
I was under the impression that the REN of the modem would probably be only 1 or 2, with the assumption electronic ringers with their low current would be used, hence my plans for an REN booster. Such things are commercially available, but cost a ridiculous amount. Hence, the decision to design my own.
So, just what did we need to ring my system? I fed an audio oscillator into the line and obtained a normal ring volume with a 50Vrms sine wave at about 20Hz.
Out of curiosity, I then measured the output of a stand alone LinkSys ATA and found it provided about 97Vp-p, or 34Vrms at 25Hz. Thinking it would have a fairly low REN and probably go into current limiting, I connected it into my phone system and was quite surprised to find it had no trouble ringing all the phones.
Calculating the actual ring power required, with a current of 22mA, it came to about 750mW. Surely the modem could supply that amount of power I thought.
Thus inspired, I tried with the modem, and likewise all bells rang successfully. I had just saved myself a considerable amount of work! Evidently, the REN is a conservative rating.

What about the Dialgizmo and ring current?
The instructions quote an REN of 2. Again, I could not see how only 22mA of ring current could do any harm, and tried it - yes, complete success! It seemed that the recabling of my house would actually not be very involved after all.

Cabling Modifications.
The original first socket now feeds the NCD only. The wiring to the rest of this house was disconnected from this socket, and instead connected to another socket. All that needed to be done was feed the modem output, via the Dialgizmo, into this new socket (or any of the other sockets throughout the house). The sockets are, of course, the Australian 610 type, patented by AWA in the early 1960's in preparation for the release of the series 800 Colorfone. The wall phones are permanently connected without plugs and sockets.
I had previously had an ADSL filter/splitter connected before the first socket, and this was removed.

The Gold Phone.
Everything was seemingly back to normal, until I tried the Gold Phone. Two problems - the Dialgizmo did not respond to its dialler, and it would disconnect after three seconds. Since the Gold Phone has lived on my coffee table for 23 years, I was not going to give up with it.
The Gold Phone is more computer than phone, and has a number of peculiarities. Truth be known, it was never meant to be used as a domestic phone, which is what I'd been doing for all these years. Rather, it is meant to have a special exchange line, an earth connection, and the line must be of a certain polarity. Nevertheless, I had successfully used it, albeit with no coin collection and metering pulses, of course.
As it happened, I was able to design and build an interface which was completely successful. Due to the complexity, it's necessary to devote a separate article to it which is available here.

Typical circuit of an Australian telephone of the bakelite era.

Brief Operation.
In the above circuit of a 400 series phone, we can see the bell connects across the line via a 1.8uF capacitor. Thus, the bell responds only to the AC ring current, and ignores the 50V DC supply. Furthermore, the capacitor prevents the resistance of the bell looping the line when the phone is on-hook (i.e. handset in cradle).
Now a quick look at the speech and dialling circuit. When the phone is off-hook, the gravity switch contacts make. Tracing through from the L1 terminal, current flows through the anti side tone induction coil, between its terminals 1 and 2. From terminal 2, current flows via the carbon transmitter across terminals 3 and 4, through the dialling contacts 4 and 5, and out to the L2 terminal. Thus, the phone loops the line and speech current flows.
The dial contacts at 4 and 5 are normally closed of course, but open and close when a number is dialled. Thus, the loop is disconnected a certain number of times depending on the number dialled. To prevent loud clicks being heard in the during dialling, contacts 1 and 2 of the dial short circuit the receiver.
When dialling occurs, it can be seen that the line voltage will go from almost zero to the full 50V. Since the 1.8uF capacitor is charged to 50V, during dialling when the line is effectively shorted out, the capacitor discharges through the bell, causing it to tinkle as the capacitor charges and discharges. This would happen, except for the bell being loaded heavily by the induction coil resistance, because of the previously mentioned contacts 1 and 2.
Contacts 3 and 4 short circuit the speech circuit, so that during dialling the phone presents no resistance when the dial contacts are closed. Leaving the resistance of the speech circuit in series with the line would cause unreliable dialling on long lines.
An important feature is that the 1.8uF and resistance of the speech circuit form an RC time constant across the pulsing dialling contacts. This gives the correct dialling waveform, and due to the inductive nature of the supply from the exchange, prevents damaging voltages building up each time the dial contacts open.

Parallel operation.
Now, what if more than one phone is connected across the line? When ring current is applied, the bells will all ring as before, each supplied via their own capacitor. When the handset is taken off-hook, the speech circuit will function normally. The dialling will also function - but now the waveform will have changed shape because of the extra bells and capacitors across the line. This is termed "impulse distortion". Depending on circumstances, dialling might become unreliable.
A more obvious problem is bell tinkle. As one phone is taken off hook and dialled, the bells in all the other phones will tinkle in time with the dialling pulses. This is because the dialling is shorting and opening the line, causing the bell capacitors in each of the other phones to charge and discharge through their associated bell.
This is not the correct way to connect parallel telephones, and when bell tinkling is heard, it usually indicates the extra phones were installed by someone unfamiliar with the correct procedure.

Three Wire Operation.
The way around the problem is to supply all the bells from one capacitor only. This requires a third wire.

The extension bell is optional and shorted out when not required.

One of the phones is wired conventionally, but the bell capacitor in all the other phones is disconnected. All bells are linked by the third wire, and since this third wire is also loaded down during dialling, bell tinkle is prevented, regardless of which phone is being dialled from. Furthermore, since the capacitance across the line is that of one capacitor only, the correct dialling waveform is maintained.

Wiring plan for parallel 300 or 400 series phones.

The above wiring plan shows in practice how it's done. The individual phones can be configured various ways with links on their internal terminal blocks. Here we see the bell capacitors disconnected in the 2nd and 3rd phone, by removing the link between terminal 10 and 11.
Thus, the capacitor in the first phone drives all three bells in parallel. As mentioned previously, with my installation, some bells have been disconnected. This is done by removing the link between terminals 11 and 12. In some phones, I have connected a resistor across these terminals so that the bell works at reduced volume. In any case, this does not affect the impulse distortion or bell tinkle suppression.

The previous description applies to Australian telephone wiring used up until the 800 series phones were phased out in the mid 1980's. Since Australian telephone practice was initially based on that in Britain, it will be noted that British telephones were used in the same way. However, when they introduced a plug and socket system for portable telephones, the bell capacitor was mounted in the first socket, with all telephones wired the same, with their bell capacitor disconnected.

Once electronic telephones came into being, their ringers were by default not responsive to bell tinkle, and did not cause impulse distortion. Of course, impulse distortion does not occur with DTMF. Thus, all modern phones are two wire only. In my installation there are two phones with two wire configuration; the Gold Phone and the 2500 series.
Since the Gold Phone was only intended to operate as a single phone, three wire operation was not required when it was designed. And, by this time, a clever Telecom technician had developed an "Anti Tinkle Module" for the purpose of using newer two wire phones with existing 800 series phones. Phones from the U.S., such as the 2500, have always been two wire. By means of mechanical arrangements, their bells are made unresponsive to bell tinkle.
Prior to the NBN, in my installation, bell tinkle was heard if the Gold Phone was dialled from, and similarly, the Gold Phone would tinkle when other phones were dialled from. However, the bell in the Gold Phone is a soft and subdued sound, and was not annoying. Since the 2500 is DTMF, it does not cause bell tinkle on dialling; only when the handset is lifted. And, as the Dialgizmo keeps the line looped when dialling, bell tinkle is no longer heard from, or caused by the Gold Phone. The occasional "ding" from the 237AW bell is heard when a two wire phone handset is lifted, but that is all. The bell capacitor for three wire working is that in the 37AW; all others in the three wire phones are disconnected.