Fluorescent Lamp Coil Resistance

The answer to this can be found in the "Homework answers" section of CR4.

I sincerely doubt you can find the answer to these questions in "Homework answers". But, I think this real world problem would make an excellent homework or test question because the obvious answer is not always so obvious.

The assumption that a lot of people make about Ohm meters is that they are accurate over a wide range of readings. That assumption can cause you some serious headaches. Consider the typical ohm meter circuit which is really just a voltmeter circuit in conjunction with a current limited power source (i.e. battery).

When choosing the values of internal "meter" resistance to be used with the battery one must consider what the full range must be for the most sensitive scale. As you consider the problem, it becomes apparent that your choice for one end of the scale causes problems with the other end of the scale. So, one way to compensate for this problem is to distort the little tick marks (for an analog meter), to approximate the effect. From a digital perspective, it becomes a mathematical estimate issue with special compensation factors that come pretty close. Actually a few specialized meters offer more accuracy.

Invariably, a phenomenon known as "contact resistance" becomes a very important issue. Contact resistance causes problems with your accuracy and your precision. And this is part of the issue I am trying to get everyone to think about. Because if you can't eliminate contact resistance, you must figure out how to compensate for it. And, it is always different from one connection to the next. So, when you need to know what the value of R(cold) is to within 1% of the actual value, how are you going to get it? If you measure it 10 times, will you get the same answer within 1% every time? I sincerely doubt it!

But there is a way! Put on your thinking cap and try again. Oh, buy the way, if you put too much energy into the attempt to read the cold resistance value, it will heat up the coil (resistor) and change the value. So, you need to know exactly what your specialized meter is doing to the load or R(test).

See post #3 for the answer to the contact resistance problem - and indeed the method of temperature measurement and the problem of overdrive*. (The off-topic vote makes you wonder...)
*Even the cheapest meters only use 100mV for full scale readings on on non-diode resistance ranges. For the last umpteen years, even basic units use a Voltage-limited low-conductance current source for resistor excitation. Full scale on a 1-Ohm range would give a maximum dissipation in the coil of 10mW, but for this class of meter it would probably be more sensible to use the 10-Ohm range for the preliminary measurement (maximum dissipation is 2.5-mW at 4-Ohms).
However, in practice this will be part of a measurement of temperature vs power, so you would more likely use a variable Voltage source with series current monitor and a separate Volt-meter attached around the coil (integrated versions exist for DC, but you would probably wish to do this AC to provide optimum match to the conditions of use)

Perhaps I failed to explain the problem completely. Before one applies high voltage starting potential to the lamp, it is best to bring the coil up to temperature (i.e. R(hot) = R(cold)*4+ meaning a little more than 4 times R(cold)). This helps to start the lamp but also helps preserve the electron emitting coating which extends lamp life.

Also, it is important to know how fast the coil can achieve this temperature. So, the current meter and voltmeter don't quiet solve the problem.

In my study, I used a digital oscilloscope and a step function application of a given DC voltage to compare characteristics between coils. The o'scope recorded both voltage and current(voltage across a 1 ohm shunt resistor) and generated about 25K points of data for each. I used MS Excel to drive the instruments and collect the data. But, I'm sure there are other ways to do this as well. Your ideas?

Joe, that's not your ordinary question!

You asked us to think about cold resistance but you already have the only answer possible to gain the info you need about the dynamic behaviour of the coil in real time, real life conditions.

Chas

Current would need to be measured across a series resistor, not a shunt one.

Assuming that you are using the information to design the ballast, you would do well to ensure that the impedance of your test set-up is as similar to the expected lamp ballast as possible, because the dynamic response depends on this as well as on the final excitation Voltage. (Assuming that the supply goes straight to its final value, a series resistance will slow the warm-up. On the other hand, a suitable series resistance will enable the ballast to provide appropriate operating conditions over a range of ballasts - but the relationship may not be obvious, because the required power may vary between different lamp designs).

We don't measure lamp ballasts - generally the response times we need are very much faster than heater coils (>10Msamples/sec). We also need multiple systems for high-volume testing. So we use A/D converters, because they are sufficiently cheap, fast and accurate for the purpose (though we have to match the sensitivity to the measurement function using amplifiers and attenuators as appropriate). The control also ties in to the mechanical handler, and is written in C++. Because we've been running the system for a long time, we use FPGAs to provide the interface. It's almost certainly design overkill for what you are trying to do, as just about everything you need is already in the scope.

The output voltage & impedance of a typical VOM is not high enough to match the task you desire. In the resistance measurement mode it tries to create a known current through the resistance to be measured and measures the voltage drop. If the resistance to be measured is too high and insufficient current is drawn to generate a measurable voltage drop, it will read (display) as an open circuit.

We all know that coil resistance changes with temperature, so (apart from any corollary to the final question) where's the problem?

And, in answer to your final question - any modern 4T meter should do the job nicely when set to a non-diode range.

Next?

We all know that coil resistance changes with temperature, so (apart from any corollary to the final question) where's the problem?

And, in answer to your final question - any modern 4T meter should do the job nicely when set to a non-diode range.

Next?

(put in wrong place originally)

For a reasonable but not fully scientific test / comparison use a resistive ballast to produce an effective constant current source, thus

making V_R directly proportional to R_coil ?

Chas