UnintegratedCircuit
Part 2 - Hardware Pulsing Circuit
​
The Aim
The aim of this circuit is to generate a signal that can be applied to the control grid of the magic eye tube. The signal should be able to ramp smoothly up and down, this means no simple square waves. Alternatives could be a sawtooth wave (such as that produced by the charging and discharging of a capacitor), a triangular wave, or a sinusoidal wave. Images of all these waves can be seen below in figures 1 through 4 respectively.
​
​
As can be seen from these images, the square wave has very sharp transitions in voltage (the y-axis) with respect to time (the x-axis). This is fine for a simple blinking LED but for a truly analogue signal such as this where there are absolutely no voltage thresholds, it makes sense to use a smoother transition to create a more pleasing visual effect. In more mathematical terms, a smooth transition is one that minimises both the first and second derivatives of voltage with respect to time (i.e. the gradient and the change in gradient, respectively); however, minimising this first derivative is the most important - at low frequencies, a sine, sawtooth, and triangle wave all look fairly similar.
​
To summarise, all this circuit needs to be able to do is produce either a sine, sawtooth, or triangle wave. Choosing which one, in this instance, boils down to whichever is easiest, cheapest, and uses the most readily available parts.
​
The Circuit
The circuit I came up with uses the typical 'jellybean' 555 timer IC and LM358 op-amp. The exact inner-workings of these chips do not need to be delved into here, although their functional role in the circuit will be discussed. Along with these two chips are some resistors and capacitors and that is essentially it. The only other building blocks required are, obviously, the tube itself and some form of high-voltage supply, in this case, I'll be using my very own USB-input Nixie tube power supply which will provide 180V to the relevant high-voltage connections. Details of this power supply are available elsewhere on this website.
​
The NE555 Timer Block
The 555 timer IC is commonly referred to as "the most common IC in the world" so suffice to say that finding images, calculators, datasheets, and articles/tutorials on how to use it will not be difficult. In this circuit, it will be configured in an astable configuration (i.e. the output never idles in one state, but rather it will constantly change between a 'HI' and 'LO' logic level). This oscillating behaviour produces a square wave at the output... Which is nice... But, as already discussed, a square wave is not desirable here at all.
​
Instead of taking the intended output signal from pin 3 of the 555, the signal at pins 2 and 6 (since they are tied together in this configuration) will become the 'output' here. This signal represents the charging and discharging of the capacitor... Which produces a sawtooth waveform as in figure 2... Which, also as previously discussed, is perfectly acceptable for this application.
​
The only downside to the produced waveform is that it has a peak-to-peak voltage of only 1/3 of the input voltage, this is due to the internal design of the 555 IC. Taking a look at the datasheet for the EM80 magic eye tube, it is easy to see that, to get a respectable range of 'openness' and 'closedness' of the shadow, a minimum peak-to-peak signal voltage of about 5V is needed. The next paragraph provides the solution to this and below, in figure 5, is a schematic for the 555 circuit block.
​
The values of R1 and C1 here have been found based on a combination of what I had laying around, and trial & error. These values are not critical and can be adjusted to personal availability and preference.
​
One thing to note about this schematic is how the output signal is actually feeding back so as to charge AND discharge capacitor C1 through resistor R1. This works because the 555 has an inverted output relative to its input. An odd multiple of inversions (1 inversion, 3 inversions, etc. etc.) will produce a phase difference of 180° between the input and output thus satisfying the condition of an oscillating circuit. This configuration of the 555 is a simple way to obtain a symmetrical waveform at both pin 3 and pins 2 & 6 - again, symmetry is good here because symmetry is pleasing to the human eye.
​
For those who understand circuits through seeing them actually working, a simulation of this circuit block is available here.
​
The LM358 Op-Amp Block
The other building block circuit revolves around the LM358 op-amp. This circuit acts so as to amplify the signal coming from the pins 2 & 6 of the 555 so that the peak-to-peak voltage is greater (and therefore has more of an impact on the 'openness' and 'closedness' of the magic eye shadow). This block also exploits a property of any decent op-amp: its high input impedance. This means that the current going into charging the capacitor (which is already quite small because of the 100kΩ resistor) actually charges the capacitor instead of being lost to the control grid of the magic eye tube. In this sense, the op-amp not only amplifies but also buffers the signal coming from pins 2 & 6.
​
The schematic of this building block circuit is shown below in figure 6.
​
This circuit has two individual op-amps (although they are both inside the one chip). U2.1 is used purely to create a low-power voltage rail at exactly half of the supply voltage. This voltage is needed since the AC component of the signal revolves around a DC component which is exactly halfway between 1/3 VCC and 2/3 VCC - this equates to 1/2 VCC. Thinking about it, this is exactly the same as in a 'normal' non-inverting configuration using a dual-rail supply, the AC component of the signal is superimposed on a DC component, this time of 0V. Since 0V is often circuit ground, the feedback network can simply be connected to ground as well.
​
Op-amp U2.2 is then configured as a non-inverting amplifier with a gain calculated as follows:
​
This means that the signal is theoretically 2/3 of the supply voltage (in this case, about 3.3V in magnitude). This is still a far shout from the desired 5V peak-to-peak, so why not simply just increase the gain of U2.2? Again, it is due to this concept of saturation at the op-amp output. The output of many op-amps does not go all the way to the supply voltage, in the case of the LM358 (and at this very low current) one could probably expect a maximum output voltage of 1V - 1.5V below the supply voltage, in this case, that would be a maximum output voltage of about 3.5V since it is being run from a 5V supply... Suddenly that theoretical 3.3V does not sound so bad.
​
One way to get around this issue is to use a 'rail-to-rail' op-amp which, as the name implies, can drive its output from (very close to) one supply rail to (again, very close to) the other supply rail. The drawback to this far more ideal type of op-amp? They are harder to get hold of, are often more expensive, and can be misleading in how they are labeled - for example, they can have a rail-to-rail input, but not necessarily a rail-to-rail output, and still be called a rail-to-rail op-amp. Using one of these rail-to-rail op-amps, however, would allow for a greater signal to be present on the grid and a pin-compatible replacement for the LM358 could be something like the MCP6242 (I believe this is also one of the cheaper ones out there).
​
Again, a simulation of this circuit can be found here.
​
The Final Schematic
Combining these two building blocks, along with the tube and the power supply module, the schematic shown in figure 7 is obtained.
​
The only additional components (apart from the aforementioned tube and power supply module) are resistors R6, R7, and R8.
R6 is simply a pull-down resistor and prevents any charge from building up on the grid and affecting the voltage, this could be optional since the grid is being actively controlled directly from the op-amp output (as opposed to a triode-based audio amplifier where an AC coupling capacitor prevents any DC path to ground); however, it does not adversely affect the circuit's operation hence it is still in the schematic.
​
R7 is used as a form of isolation allowing the anode and deflector voltage to fluctuate with the grid voltage - and hence amplify the input signal as a normal triode would - whilst allowing the target plate voltage to remain constant, thus allowing the shadow to 'open' and 'close'.
​
R8 is used to 'cathode bias' the tube, another technique also used with standard triodes. Effectively, this resistor creates a voltage drop across it proportional to the current flowing through the tube. If the average current increases for any reason, the voltage drop across this resistor also increases placing the cathode at a higher potential. This, in turn, means that the control grid is now repelling electrons more strongly than before which leads to a reduction in the average current flowing through the tube. Now that the average current has reduced, the voltage drop across the resistor reduces, and so the cycle repeats until a 'happy medium' is found.
​
R9 & R10 ARE PARAMOUNT as they provide a minimum load for the high-voltage power supply module. Without this minimum load, the output voltage can rise up to OVER 500V which is incredibly dangerous and risks damaging the module's output capacitor, the magic eye tube, and you. They must also be two 22kΩ resistors in series in order to create the appropriate resistance AND to ensure that their voltage rating is not even close to being exceeded. Since these are essentially a safety-related component, they must be derated appropriately so as to ensure maximum reliability.
​
If you want to know more about triodes and how to use them, there is a fantastic document here.
​
In Operation
Once the circuit was in operation, I connected an oscilloscope to the control grid signal at both int input and the output of the op-amp. The waveform was captured and is shown in figure 9 below.
​
The peak-to-peak voltage can be seen at the bottom of the capture. The oscilloscope shows us that the op-amp is indeed doubling the peak-to-peak voltage of the waveform. Also observed is a tiny amount of saturation at the peak of the blue waveform (where it appears to flatten out) which rather accurately shows the maximum output voltage of the LM358 in this application (also known as the 'voltage swing'). The final observation to make is the noise present on the amplified waveform. this stems from switching noise injected by the power supply module onto the USB positive rail that the op-amp is unable to filter out by itself (also known as Power Supply Ripple Rejection). The fix to this (although it really doesn't matter in this application) is to add capacitance between the USB's positive and negative rails.
​
A short video of the tube pulsing is shown below, this is using the exact circuit from figure 9.
A NOTE ON HEALTH AND SAFETY
The high potential that this tube operates at is POTENTIALLY LETHAL, most likely you'd get a tingle and nothing more; however, if you have a heart condition or similar, this circuit becomes very dangerous indeed. If you were to remove the load from the power supply, the open-circuit voltage can reach (and has reached) UPWARDS OF 500V. DO NOT touch the circuit whilst it is powered and do not modify anything with the power on, it can damage the components, or worse, it can damage you. It should also be noted that the back of the tube does get hot due to the heated filament, so again, do not touch it whilst it is powered or shortly after having been powered.
​
​
​