Model T owners have often come up against
the difficulty of correctly testing and adjusting their ignition coils.
A fortunate few own a hand cranked tester of the kind illustrated in the
Model T Ford Service Manual. Others try and get by with simple buzz box
testers. There are even some who use completely non electronic methods,
such as measuring spring tension, or listening for a certain "musical note"
as the coils buzz. To any technically competent person, these latter two
methods are extremely crude, and they cannot give known or repeatable results.
Hand cranked coil testers (HCCT) are few and far between, and when they do turn up are extremely expensive - hence the crude methods used instead. However, there is still a limitation with the HCCT, in that there is no way of setting the coil's firing time so it is the same as all the others.
Electronic Coil Testing.
At this point I recommend reading the article here if you have not already done so. It goes into further detail on why setting coils for equal firing time is important, and describes an electronic tester that performs this function. While the tester so described works very well, it requires an oscilloscope, and the ability to interpret its display. Thus, it is only suitable for the Model T owner with an electronics background.
The good news is there is now a test instrument that anyone can use, the ECCT. It works on the same principles as my CRO based tester, but it has been carefully designed so that any Model T owner can use it. I have no affiliation with the design or manufacture of this instrument, but I have had the opportunity to try one out.
It immediately became very clear that the ECCT fulfills the need for an instrument that anyone can use, regardless of technical ability. It is not only very inexpensive compared to a HCCT (if you can get one!), but allows for the coils to be set for the all important equal firing time. Additionally, facilities to test the coil capacitor are provided, and for those with a working magneto, the instrument can test this also.
What is the ECCT?
The "Electronically Cranked Coil Tester" is a very compact unit, hardly bigger than the coil itself. The unit is made using printed circuit boards for its enclosure and has a very simple, yet elegant look. It runs from an external 12V 8A switchmode power supply, which has the usual 100-240VAC input via an IEC socket. Thus, it is a simple matter to substitute a local power cable wherever you might be in the world. Provided care is taken with polarity, it can also be used with a 12V battery, giving complete portability.
On the front panel is a LED display and testing switches.
The coil to be tested is simply inserted and then the relevant buttons pressed to perform the required test.
In its basic form, the instrument will
perform coil and capacitor tests, with the LED's displaying the results.
The "Excellent", "Good", and "Poor" indicators mean that no technical interpretation
is required. When setting the vibrator spring tension, one simply observes
the timing error LED's and reduces or increases the tension for a zero
degree timing error.
It really is that easy to use. However, don't be deceived by the simplicity, because internally, it is a very accurate instrument. It is this simplicity plus accuracy aspect that makes it the perfect coil tester for everyone.
Of course, Model T owners who have a technical understanding of the ignition system will be keen to see more than a LED display. In view of this, a USB port is provided for interface to a computer. This opens up a whole new world because results are now displayed graphically, and numerically.
It must be pointed out that the computer interface is an option, and for the non technical user who wants to keep the cost down, it is not actually required. The simple front panel LED display will allow coils to be adjusted just as effectively.
Trying out the ECCT.
The earliest computer I have with USB support has Windows 2000. The ECCT software functioned perfectly on this, as well as other machines with XP and Windows 7. What I did find was that a minimum screen resolution of 1280 x 1024 is needed, otherwise the right hand side of the software window is cropped. This does not impede functionality, however. I found the ECCT to be intuitive right from the start, probably because of my experience with this method of coil adjustment. Nevertheless, the supplied instructions are very well written in a step by step format. Directions are given on how to correct the coil settings if they are not as they should be. Now, let's look at the tests the ECCT can perform:
This is the first test that should be performed,
because if the capacitor is excessively leaky, firing performance may be
reduced regardless of how well the points are set. Capacitors can also
fail short circuit or open circuit, and no amount of point adjustment will
give a proper spark.
Here, the capacitor measures over 92M which is an extremely low leakage. This is to be expected if there is a new capacitor installed. Old leaky paper capacitors may indicate 100K ohms or less. The test fails with under 10M leakage. That is not to infer a coil with, say 2M leakage, is defective and won't perform adequately, but it indicates something is not as good as it could be. Typically, this would be because the original capacitor is still in place, and/or the coil box wood has been painted with something conductive. My recommendation is to routinely replace original coil capacitors, because it eliminates future problems, ensuring the spark is always as strong as it can be. Modern plastic capacitors do not deteriorate, so once installed they will last a lifetime. Of course, the correct capacitors must be used or short life will result. See here for further details.
Coil & Dwell.
Having ensured the capacitor is satisfactory, the next test to perform is the coil firing time (dwell). What is dwell you might ask? In an inductive circuit, such as a Ford coil (or even a modern ignition coil), when current is applied it takes a certain amount of time for the iron core to fully build up its magnetic field. When fully built up, it is "saturated" and continuing to apply current has no further effect. Most efficient coil performance will occur when the coil fires just as the saturation point is reached. Here, the magnetic field is strongest without wasting power. Any further application of current is simply wasted as heat. With a Ford coil, the ideal dwell time happens to be 3.5ms if the supply is 6V, and 2ms if it is 12V. With magneto operation, the dwell time varies with the change in voltage. The important thing is that each coil has the same dwell time to ensure that no cylinder fires advanced or retarded relative to the others. This is the basis of ECCT testing.
As can be seen, the primary coil inductance
is measured, as is the peak firing current. This is the current right at
the firing time, and is not to be confused with the lower, average, current
often quoted for coil adjustment (1.3A at 6V). From this, the ECCT microprocessor
calculates the actual spark energy.
The all important coil dwell time is measured to an accuracy of three decimal places. While such accuracy is not actually necessary, it does show the instrument can actually measure it. With the coil test voltage performed at 12V, the ideal firing time has been determined to be 2ms, so this corresponds to a zero degree timing error.
This corresponds to 3.5ms on my own CRO based tester, when the test is performed at 6V. While it is perfectly satisfactory to set firing time at 6V, it is really necessary to use 12V to check for multiple sparking, as this does not always show up on the lower voltage. My own tester can do this, and the ECCT does so by default.
While the dwell test is being performed, any signs of multiple sparking are detected. This results in a message to adjust the upper cushion spring tension.
In this test, the coil is fired at up to 100 times with and equivalent engine speed of up to 2000rpm. These parameters are user selectable.
The purpose is to examine the firing consistency. While one single fire may appear to show a good result, is this always consistent? Any deficiency with the coil points will show up when rapidly fired, like they would be in a car at normal engine speeds. Problems like contact bounce and dirty contacts will show up here.
This is an additional option. Given that not all Model T's have a working magneto, further cost saving is possible by not opting in for this feature.
Magneto voltage is measured along with engine rpm.
Load testing is performed by a simulated St. Louis Magneto Tester, an original design from the T era, renowned for its ability to test a magneto’s power output regardless of engine speed.
Testing The ECCT.
With all features enabled in the ECCT I was testing, it was an ideal opportunity to do some detailed tests and compare with my own electronic tester. As my tester had been developed completely independently, it was fascinating to see how it compared, and to confirm the legitimacy of the testing method.
As expected, capacitor values tested within tolerance, when compared to measurements taken from an Escort ELC-131D hand held LCR meter. It was noted that accuracy decreased with an increase of capacitor leakage. However, this in itself was not a limitation because capacitors with high leakage should be replaced anyway.
One thing that did cause confusion was that the capacitor test would fail, with a reading of 0uF with some original paper capacitors. This was found to be because of a small voltage being generated from within the capacitor, most likely due to electrolytic action occurring because of moisture. I measured up to 320mV into a 10M DVM with the set of original coils I used for the test. Again, such readings indicate the capacitor should be replaced.
It was found also that incorrect results would be given when the coil points mounting nuts were touched during the test. Evidently, there was enough 50c/s radiation from the power mains being picked up (1.2V p-p into 1M) to upset the testing. This was confirmed when the ECCT was operated from a battery away from the power mains. Here it worked normally even if the nuts were touched. The capacitance test calls for holding down the vibrator spring so that the capacitor is not being short circuited. However, if the vibrator spring touches the coil core underneath, then leakage to this point is measured as well, giving a not entirely true capacitance leakage result.
Ultimately, the most accurate way to measure the capacitor is by inserting a piece of plastic between the contact points and not touching any part of the coil during the test.
Coil firing time.
This test actually showed up some inaccuracies with my own tester. While the results were sometimes identical, at other times they were not. It became clear that using a low cost CRO was one limitation. For one thing, the timebase calibration was found to have drifted, and was not completely linear. There was also the possibility of Parallax error because of the graticule's distance from the phosphor screen of the CRT. These problems have since been eliminated since I performed the tests by changing to a decent CRO; a Tektronix TAS475.
Once that was cleared up, the ECCT was found to be completely accurate with regards to the coil firing time. It also was found to reliably detect multiple sparking. If it showed up on the CRO, then the ECCT also detected it, and vice versa. From this test alone, it is clear that the ECCT can substitute for mechanical testers such as the HCCT.
As was expected, it was found that firing current varies slightly amongst individual coils, even though the firing time is identical. This is a trap for those who have always set by current on other kinds of tester. The golden rule is to set for equal firing time, not equal current.
Coil Firing Consistency.
My CRO based tester does not have the ability to log firing consistency, but one can subjectively get an idea of how good or bad this is by listening to and observing the coil sparking. The ECCT can log this over a period of up to 100 sparks. I found with coils which had a less than perfect firing consistency, on the CRO based tester, that the ECCT would detect this also, and vice versa.
This was something I could not test in a real world situation because I do not have a magneto in my Model T. However, using a BWD Mini Lab 604 function generator, I was able to determine that the tachometer and magneto voltmeter were accurate. The latter is actually a true rms voltmeter, and gave consistent readings regardless of input waveform. I used sine, triangle, and square wave inputs for the test.
The tachometer readings were calculated from the input frequency, and again were found to be accurate.
A challenge was simulating the magneto
field strength test. I largely overcame this by using an audio power amplifier
in between the function generator and the ECCT.
This was necessary to get the low output impedance and voltage swing to simulate a real magneto. It was found that the pass/fail for the internal resistance test was 3 ohms. A high resistance reading alerts the user to a fault such as a magneto post with poor contact.
While I couldn't actually do the field strength test and compare the reading against anything, it certainly appeared to function as it should.
Overall Impressions of the ECCT.
As the designer of the CRO based tester, I related to the ECCT straight away, and understood the results with no further interpretation required. This is because the coil testing feature of the ECCT works on the same principle of measuring dwell.
A welcome option would be to include a dwell time (ms) calibration for the graphical and LED display. However, it is understood the ECCT is designed for a market more familiar with mechanical engineering, hence the preference for displaying degrees. This is a very minor point, because anyone who wants a millisecond reading would most likely be using the software, which does display this.
Using the ECCT showed it was much more than just a coil tester. It is in fact a comprehensive Model T ignition system analyser, suitable for coils used either on battery or magneto.
It is hard to imagine a more compact, attractive, and user friendly design with the simple, yet effective and rugged method of construction used.
I found the instructions to be well written, in logical order, and in terminology anyone can understand. How to interpret the results and take corrective action is made particularly clear.
Versatility is also a welcome part of the ECCT which should appeal to prospective purchasers. The user is not obliged to purchase the software interface, if this is seen as too technical. The essential results are visible on the basic tester, and are just as accurate. Likewise, those without magnetos do not need to purchase the magneto testing option.
Naturally, all the extra features which
the ECCT has are interesting in themselves. If you have an interest in
the Model T ignition system you will find it addictive, wanting to spend
hours with the ECCT learning all about coil adjustments and characteristics.
In fact, I was so impressed with the ECCT that I bought one!
In terms of my own personal use, I use the ECCT in conjunction with my CRO based tester. The CRO based tester is good for the initial coil test because it allows quick real time adjustment. Then the ECCT is used for the final firing time accuracy, and to look at the other coil characteristics. To provide a CRO interface on the ECCT can be done during manufacture, but it would be lost on most users. However, once I did purchase my own ECCT, I did fit it with connections to trigger a CRO, and provide a current waveform. Now, I have the perfect coil tester - a visual current waveform of the coil firing and the superior accuracy that the ECCT offers.
Why the Model T owner will benefit
from the ECCT.
It should be clear by now that there is a lot more to the Model T ignition system than simply making the coils produce a spark. A car with an ignition system set up that way will actually run, but the full potential of its performance will be impossible to achieve. These are the cars that struggle to get up to 60km/h and run rough.
Improperly set coils may suffer from:
The HCCT is a vast improvement in that it can test for multiple sparking as well as coil current. Again, because of the simple analog meter and differing coil characteristics, it still cannot test dwell time.
For this reason, the Model T owner who
wants optimum ignition performance will benefit from the ECCT. It performs
all these tests.
Clearly, electronic testing, such as that provided by the ECCT, is the future of Model T Ford coil testing. All aspects of coil operation, invisible on other testers, can now be displayed, and corrected where necessary.
The ECCT site is here. More information is available there on how the ECCT was developed, and you can also get a feel for the instrument by downloading the instructions.
I really do urge all Model T owners using the original ignition system to think about this, because properly adjusted coils and a good timer will give performance as good as a distributor system.