One of the challenges often faced by the Model T owner is a faulty magneto. Unfortunately, this usually leads to the installation of some other form of ignition system, usually a distributor. There seems to be an underlying assumption that a Model T can only be used with its original four coil ignition system, if it has a functioning magneto. This, in actual fact is not correct, and the existing 6 volt electrical system is quite sufficient.

Here we see speeds of 70 and 75km/h are attainable in a Model T Ford with the original four coils powered from the 6V electrical system. This car does not have a magneto.

Background to the Model T ignition system.
The Ford ignition system comprises four separate ignition coils, each with its own vibrator contacts. These are connected directly to each of the four spark plugs. As to which spark plug is firing at any given time, that is determined by switching the low voltage supply to each coil in turn. This is done by a device called a "commutator" or "timer". It is mounted at the end of the camshaft and is essentially a four way switch. Similar to a distributor, but it switches the low voltage.
The supply for the ignition coils comes from either a magneto or a battery.
Much has been written on the subject in various Model T literature and on the Model T Forum. Suffice to say, the magneto provides about 6 to 30V AC, the voltage and frequency dependent on engine speed. Because of how the coils operate, and the waveform produced by the magneto, the ignition timing operates differently when operating on the magneto than when on battery.
The Ford coils also can operate on a battery. This is the 6 volt accumulator in the 1919-1927 cars. Prior to this, the battery was optional and could consist of a number of different types, a dry cell "Hot Shot" (similar to a modern day 6V lantern battery), a collection of No.6 dry cells making a 6 or 9V battery, or a 6 or 12V storage battery - this being more likely if the car had been updated to include electric lighting.
The battery option was provided in the 1909 -1919 cars simply because unless the magneto is perfect the car is much easier to start on battery. The continuous spark helps apart from the battery voltage being higher than what comes from the magneto when turning the engine by hand.

Prior to the Model T, cars like the Model N which used the four coil and commutator system ran on battery only, although there was an accessory dynamo which was available from 3rd party suppliers.
However, the problem was dry cells can't be recharged and no means was provided to charge storage batteries in these cars. So, it was an inconvenience having to replace dry cells all the time, or to take the storage battery to a garage to be recharged. Needless to say, some drivers were stranded with flat batteries.
One of the improvements that the Model T was to include, was not to have to rely on a battery at all. This is where the flywheel magneto comes into the picture. 16 magnets mounted around the flywheel induced current to flow in adjacent coils mounted on the back of the engine block.

Magneto problems.
Unfortunately, the idealistic design has a number of disadvantages. Over time the magnets weaken, with the vibration and hot temperatures assisting with their deterioration. So, although not approved of by Ford, methods were developed to recharge the magnets in the car. The distance between the magnets and field coils is critical. Too close and they hit each other at 2000 rpm resulting in broken magnets and damaged field coils. Too far apart, as happens when the 3rd crankshaft bearing wears, and the voltage decreases making for weaker ignition. Sometimes the magnets are cracked and eventually fly apart.
Then there's the field coils themselves. These are prone to short circuits. All that hot oil is hard on the insulation. Sometimes it flakes off and metallic particles become embedded between turns of the copper coils, reducing output in varying degrees. In some cases, the coils detach from the bobbin plate and once they come into contact with the magnets, there's several metres of copper ribbon wrapped around the flywheel and transmission.
The worst aspect of all this is that the engine has to be removed to service most faults with the magneto.
In my opinion, it is one aspect of the Model T design that could have been improved. The external dynamo of the Model N would seem to be a better idea.

My own car.
The magneto in my own car was partially operative when I got it, but output was too low to run at idle. I had noticed damage to one of the field coils where attempts had been made to remove the starter motor without removing the Bendix first. Apart from this, attempts to remagnetise the magnets in the car did not work.  This was right at the start of my Model T learning curve back in 2002. I had seen messages in the Model T Forum that the coils wouldn't operate properly on 6V at driving speed, and that a 12V supply would have to be used or a distributor ignition system used. Neither option were acceptable for my car.
But, as it turned out, my car seemed to operate perfectly well with 6V powering the coils, and other club members who drove it said that it "drives just like a typical T". And so, I stopped worrying about the inoperative magneto.
When I rebuilt the engine in January 2010, I removed the magneto and replaced the magnets with commercially made paddles to keep the oil circulating function. As it was, the coil insulation had deteriorated on a number of coils and there was miscellaneous damage from various objects that had fallen into the transmission. It was no wonder that recharging attempts failed.
I thought long and hard about restoring the magneto, but ultimately, the magnets would still weaken over time, they could still break off at 2000 rpm, and the coils could still detach themselves making a mess. Plus, I'd driven the car for over seven years without it and wouldn't miss it. Not only is the flywheel so much easier to work on, but service access in the transmission is so much better. The lighter flywheel should also prolong the life of the 3rd bearing.
Regarding the paddles, I used the ones from Texas T Parts as they were braced against bending. However, the supplied ring gear bolts cannot be used as they will only jam up and break. Someone has stuffed up somewhere and there's now three sizes of repro ring gear bolts doing the rounds.

No magnets to break off, no field coils to come apart, and much better service access!

The original bolts are #14-24, but one of the repro parts manufacturers interpreted this as 1/4" -24. Hence we have two kinds of repro ring gear screw in circulation. Some repro ring gears are tapped for the 1/4" thread. Some of the oil paddle kits come with 6 mm screws, and this is what I've used. It is important that the paddles are made of steel and braced as shown. There have been stories of unbraced paddles, especially aluminium ones, breaking off. There is a lot of force against them when the oil is cold.
6 mm x 1 mm screws fit perfectly for the ring gear and will self lock; the difference in pitch being just enough. However, they must be high tensile screws or this self locking function won't work. I notice the kits providing 6 mm screws also provide a 6 mm tap. I'd be hesitant to use the tap as the screws won't be as tight, and I'd also check the screws are in fact high tensile.

Running on battery.
Why have some owners said that Ford coils won't work well on 6V at speed? Invariably, these claims have come from Model T owners who have used a "timing gauge" to set up their initial timing, or from those who are using non rebuilt coils, with an ignition system in average condition, or poor performing spark plugs.

The importance of ignition timing is that with a low supply voltage, the points take longer to open at 6V than they do at 12V, or the higher voltage provided by the magneto. The result can be retarded timing. At 6V, the coils take 3.5ms to fire. On 12V, they take 2ms. 12V has been found to be, from dynamometer testing, to be equivalent to magneto operation in terms of the car's performance. So, it can be seen that at high speeds where advance is needed, 6V performance will lag behind, if the same timing settings are used. For this reason, there is an assumption that if battery operation is desired, the battery needs to be 12V. However, my testing and experience has proven this not to be so.
It needs to be pointed out that when the car is running at speed, that the coil supply voltage is actually closer to 7V because of the generator charging.

To operate Ford coils successfully on 6V, several requirements must be met, and having said that, everything must be 100%. First, the coils must have new capacitors in them, and they must be the right kind or they will fail. It is amazing what I have seen installed in coils, from 1uF electrolytics to .1uF 400V paper. Needless to say, the owners of them claimed the Ford coil system was incapable of operating properly, and subsequently fitted distributor ignition. 12V operation will often mask faults in defective or misadjusted coils simply by brute force. Once the owner has found that the ignition system appears to work properly with 12V, having masked a defective set of coils, poor timer, or weak spark plugs, the conclusion is come to that 6V will not operate the coils properly.

The coils must to be adjusted correctly, for equal firing time.  For this, I designed an electronic coil tester as described here. The ECCT is a commercially available coil tester that also tests for equal firing time. If the coils are set using a "buzz box", or hand cranked coil tester, there is no way of knowing exactly when the coil fires after the timer makes contact. Thus, there is the possibility that the cylinder to cylinder timing is not consistent.

Next is the timer. Roller timers do not provide a reliable low resistance path, especially at 6V. There are multiple connections between moving parts, and the roller is prone to bounce as the contacts and insulating material in the timer case wears. An Anderson timer provides positive and reliable contact. Also available is the TW timer. Both these timers make a positive contact throughout the switching time. However, the Anderson, while of much lower maintenance, does suffer from timing variation as it wears. I have not tried a New Day timer, but it would appear to be suitable.
For the spark plugs, Champion X or similar are essential. I have used Lodge spark plugs with excellent results. The worst were KLG's and modern plugs with 14mm adaptors.

Initial Timing most important.
An important thing to note is the initial timing. That is, the firing time ATDC with the timing lever fully up (retarded). It appears that some owners in the U.S. set the initial timing to 15 degrees ATDC, because of the availability of a so called "timing gauge". The purpose of so doing to prevent any kickback when starting on magneto.
For the Australian assembled Canadian cars, being right hand drive and with the timing adjusting linkage being quite different, no such tool is known to be available and it has been customary to set the initial timing for just after (2 degrees) top dead centre. In the absence of any defined adjustment, it is the most logical because it provides the full range of timing control. It is mainly for this reason I had immediate success with 6V operation.
This effectively gives 13 degrees extra advance over cars that have been adjusted with the timing gauge. Conveniently, this happens to compensate for the slower firing time of coils on 6V.
With everything set up thus, I can often get over 75km/h. This is quite sufficient for a Model T. While higher speeds are possible at times, the risk of breaking something worries me, so I keep to 75 or below. Also of course, are the extra stresses on the car if it has to be stopped quickly at that speed.

Requirements for successful 6V operation:
 

• TW, Anderson, or E-Timer. Do not use roller (Tiger) timers.
• Rebuilt coils with new high dv/dt capacitors.
• Coils adjusted for equal firing time (3.5ms when tested on 6V, or 2ms when tested on the 12V operated ECCT).
• Champion X spark plugs.
• Initial timing so coil buzzes just as the piston starts to leave TDC, with the timing lever fully up (retarded).

Further Experiments.
In learning about coils running from battery alone, I have experimented with running them from 9V DC and have spent some time driving this way. I was not keen on 12V as this causes a shortened point life, as the blackened points I've replaced on coils so used demonstrate. Nevertheless, for the sake of completeness in testing, I have tested performance on 12V short term,  by means of a SLA battery to power the coils separate from the car's 6V electrical system. No more will be said about 12V operation because I found no real difference between it and 9V. 9V effectively gives 12V performance, but without the points damage.

For long term testing, I created two adaptors that provide 9V, which I describe here. The first one was generator based, making use of one of the interesting properties of the 3rd brush generator used by Ford.

Obtaining 9V from the Generator:
The Model T generator, being of the 3rd brush type is a constant current source and will put out much more than 6V if it's allowed to. So was there some way I could use this property?
It's really quite simple. Let's say the generator is set to 5A. That 5A remains fairly constant over a reasonable voltage range; certainly up to 9V which is all I'm after. 9V was chosen because it gives ample advance, but without the amount of points arcing and deterioration which occurs on 12V.
Simply by connecting a resistor between the generator and battery we can increase the generator voltage while still charging the 6V battery at 5A.
Rough calculations indicate the resistor should be 3V/5A, or about 0.6 Ohms. Simply putting this resistor in the wire between the ammeter and cut out will provide us with 9V when the engine is above idle speed. So easy! There's a few refinements however:

The circuit and how it works.

Only two rectifiers and one resistor make up the circuit. It keeps in with the theme of simplicity inherent to the Model T.

While a simple resistor will work, the problem is the value is critical, and the charge rate is not perfectly constant. So, what we need is a device with a constant voltage drop, regardless of current flow. Silicon diodes are ideal, with a drop of around 600-700mV. They're easy to get, and for high power applications like this, you can use 35A bridge rectifiers. Used as voltage dropping devices, a bridge rectifier can provide the drop of two diodes, or about 1.2V. Note that diodes also provide a fail safe function. If only a resistor is used and it burns out, the generator could be damaged as well as anything connected to the boosted supply. Diodes fail short circuit, so if this happens no harm will come to anything, and the 9V supply will simply drop but not fail altogether.
SImply add further rectifiers to get the required voltage. If you only need a 600mV drop, then use only half the rectifier bridge.

Unit before the perforated cover installed.

Unit ready for installation in the car.

The bridge rectifiers I used have a diode drop of 750mV at 5A, so two rectifiers give 3V drop. Perfect. Obviously, the diodes must be rated at a higher current than the charge rate. For reliability, 35A ones are the best with ample reserve. Mounting is easy with a single 4mm hole.
Now, let's analyse the operation but first assume the coils are running off the 6V battery as normal. The car starts, and as engine speed increases, the generator voltage increases. At about 8V, the cut out closes and current starts flowing through the 1R resistor. Not the full 5A, but still some current. As generator output comes up to 9V, the diodes start conducting and there is now a constant 3V drop between the generator and battery. 5A now flows into the 6V battery via the ammeter, and the battery charges normally. Now, look at the "Gen" terminal of the circuit. Because we've added 3V onto the battery voltage, we have 9V at this point. The red switch wire (that used to go to the magneto) is connected here, and with the switch in "Mag" position, the coils now receive 9V DC.

Engine above idle. The reason we have 10V is because the battery is 7V when charging. This drops back to 9V when the headlights are on. The brilliance of the 12V test lamp is further indication of the success of this circuit.

So, we switch to "Mag" and happily drive along at high speed with 9V to the coils. At some point we have to slow down and let the car idle.
What happens now is the generator voltage is less than the battery voltage. The cut out has opened, so no supply from the generator. At this point the coils would be deprived of power, as the diodes are not conducting in the reverse direction. This is where the resistor comes into operation.
Current from the battery flows through the resistor and to the coils. Because of the resistor's presence the coil voltage will now be a little less than 6V, but as the engine is only idling that's not important.
The value of resistor is not super critical. It must be high enough to allow at least 3V drop at the required charge rate, but it must also reduce coil voltage at idle by the minimum amount. I can't see anyone charging at less than 5A, so 0.6 Ohm is ideal. 1 Ohm is close enough.
Next thing is what kind of resistor to use? Ignition coil ballast resistors are ideal. Easy to mount, high wattage, and very rugged. They are also of the low resistance required. I found that what the label says isn't always accurate. I bought several RU51 resistors which were claimed to be 1.6 Ohm. One measured 1 Ohm which is what I used in this circuit. It is possible to reduce the value as the resistance wire is just screwed under the terminal bolts.
Keep in mind that the wattage rating decreases proportionally when doing this.
And what power does the resistor dissipate? P=IV. Assuming 5A charge rate, it's 15W.

Limitations.
The circuit will work only with the electromechanical or diode cut outs. It will not work with the Fun Projects or any other voltage regulator. Voltage regulated generators (or alternators) will not work, as they function as constant voltage sources.
Keep in mind the power rating of a Model T generator is about 100W. The fact it's now outputting 9V (10V on charge) derates the current that can be drawn.
Power dissipation of the resistor and diodes needs to be considered and efficiency reduces as charge current increases. As an example, for a 10A charge rate, about 30W minus the coil power is going up as heat. I find 5A quite sufficient and have never had a flat battery despite driving at night or using the radio.

Installation.

The unit installed under the floor. The two pin socket is for the brake light switch attached to the upper floorboard.

Given the lengths of wires from the harness, I mounted the unit on the right hand side body rail under one of the existing bolts. This is a RHD car.
The only new wiring is one wire to go from the adaptor to the ammeter/battery terminal on the terminal block. Because I'm using a LHD wiring harness mirror image fashion, the original generator wire couldn't be used anyway so I'd already made one up previously. The length of this was already suitable for the adaptor.

Switchmode 6 to 9V converter.
Some problems were observed with the adaptor circuit previously described. One limitation is that the coils can run on insufficient voltage at low engine speed. What happens is that when the engine is going slow enough that the cutout opens, the coils are running on the battery voltage, minus the voltage drop across the resistor. This can cause erratic firing. In the worst case scenario, going up a hill in top gear with the headlights on, the voltage fed into the coils could be as low as 5V.
Another limitation is that I was given a Fun Projects voltage regulator which is incompatible with this circuit. The correct thing to do was to make a regulated 6 to 9V converter to run the coils. This would have the advantage that the coils would be fed with 9V regardless of what the generator was doing. Even with the battery down to 5V, coils would still receive 9V.
The circuit I used is a very typical design based around the MC34063 switchmode IC.

The MC34063 drives a MOSFET in order to obtain the required output current. I designed it for a continuous 1.5A output. The MOSFET must be one with a gate voltage of 5V or less given the low supply voltage. If the MOSFET does not saturate it will overheat and efficiency will be poor. I used a SNP60NE0-16.
The choke is wound on a Jaycar LO-1244 toroid core. The rectifier is a Schottky type as used in computer power supplies. Because of the high frequency ripple currents, low ESR electrolytics are used for input and output filtering. Feedback is provided by the voltage divider consisting of the 15K and 2.2K + 180R resistors. Changing any of these resistors will actually allow the circuit to function as a 6 to 12V converter if desired to run low power 12V appliances.
The BC327 circuit is to speed up the switching off of the MOSFET to minimise power dissipation.
Protection of the circuit is by a 5A polyswitch on the incoming 6V supply. This is superior than a fuse, given the likely possibility of inadvertent short circuits when working on the electrical system, etc. It has the advantage of being self resetting once the short is removed. It is necessary in case the MOSFET shorts out, which is the normal mode of failure. To guard against the spiky back EMF of the coils getting back into the converter, an 18V MOV is connected across the 9V output.

Inside the converter. I used the PCB from the June 2003 Silicon Chip DC-DC converter article and changed components to suit my version.

Enclosed in an aluminium box, the converter easily fits under the seat.

Performance is excellent. Efficiency is 85%. Typically, for an output of 1.5A at 9V, the input is 2.7A at 5.8V. Note that the coils are not drawing current continuously when the engine is running, so that the average input current is less than this. Standby current is only 3.6mA which means the input can be permanently connected to the battery, provided a maintainer is connected when the car is not in use.

Conclusions - what difference does 9V operation make?

The biggest difference was when I was experimenting with modern 14mm spark plugs with adaptors. I found running on 6V poor, with the exhaust pipe glowing red (visible through the handbrake hole in the floor). Maximum speed was about 65km/h. This was quite different to the original Lodge spark plugs which performed very well with 6V coil operation.
Alas, the Lodges gradually broke and had to be replaced. I had used Motorcraft F11 types with some success, but it turned out Champion X was the best, performing like the Lodge.
So, golden rule is to use Champion X plugs, or old pink Lodges if you can find them.

Around this time, the car also had carburettor problems (worn mixture needle and seat). Again, I found it necessary to use higher voltage, but this was just masking the problem. Once the mixture needle was turned down to remove the ridge worn into the needle, and the seat was replaced, the car once again worked perfectly with 6V for the coils.
Next golden rule is that you won't need more than 6V if the carburettor is 100%.

With everything in the car 100%, what is the difference you may ask? Well, not a lot as it turns out. I do a lot of freeway driving in the Model T because of where I live, so I'm well acquainted with how fast it goes. What makes more difference than anything else is the weather - cool moist mornings are the best. There is then no real noticeable difference between 6,9 or 12V. Once the temperatures warm up for the afternoon, and the winds come up, there is a drop off in performance. Switching to 9V will give perhaps a slight improvement, maybe a couple of extra km/h above the existing 78 odd km/h. For suburban driving at 60km/h there is no difference at all regardless of weather conditions.
On the down side, coil operation rapidly deteriorates below 6V. If the headlamps are operating for example, and the battery is defective, then erratic operation starts to occur. With the generator putting out only 5A, and the headlights drawing 10A, the battery voltage can not come up to 6V. However, this is merely indicative of a faulty battery.

In conclusion therefore, 6V has been proven sufficient to run the ignition coils with good performance. The photos at the top of the page are proof of this. If you like experimenting, by all means try the 9V adaptors presented, preferably the switchmode one, but they're certainly not essential.