UnintegratedCircuit
Part 3 - Generating The High Voltage
Choosing A Topology
As mentioned in the previous section, a step-up converter needs to be used in order to provide the high potential that the anode requires. Stepping a voltage up can be achieved with different topologies: boost, flyback and SEPIC (Single-Ended Primary-Inductor Converter) are some commonly implemented examples.
​
The boost topology is good when one only intends to step up a voltage from a lower level to a higher level, working particularly well when the step-up ratio (output voltage divided by input voltage) is relatively small. Obvious limitations are that the maximum output voltage is restricted to a modest level and that the input voltage must always be lower than the output voltage.
​
The flyback topology is best used when one desires multiple output voltages (i.e. 5V to +/-12V) or when one requires electrical isolation between the input and the output since a flyback converter utilises a coupled inductor (also called a flyback transformer) which means there is no electrical connection (possible exception of the common rail, which is usually circuit ground anyway). Due to the use of this flyback transformer, multiple transformer secondary windings can be wound onto a single core, with each winding has a polarity and turns ratio with respect to the primary winding hence the ability to produce multiple, positive or negative output voltages.
​
The SEPIC topology can act as a step-up or step-down converter without any circuitry changes (i.e. will maintain a constant output voltage regardless of whether the input is greater than, equal to or less than the output voltage). This means it's especially suited to battery-powered applications where the input voltage may start, after being freshly charged, at a voltage higher than the output voltage, but may end, once depleted, at a voltage lower than the output. An example of this scenario would be generating a 3.3V rail from a single Li-ion cell which has a maximum voltage of 4.2V and a minimum voltage of about 2.5V (as with common protection ICs like the DW01).
​
Based upon the above descriptions, I opted for a simple boost converter since the anode (output) voltage will always be greater than the USB (input) voltage and there is no need for multiple output voltages since the logic rail, grid and filament voltages are generated from the USB either directly or via the microcontroller. Additionally, the input voltage is fixed so it will never rise above the output voltage.
Choosing A Boost Converter IC
The specification for the boost converter IC was fairly harsh:
​
-
It should be as cheap as possible;
-
It should be as small as possible (but still solderable by hand);
-
It should be as simple to implement as possible;
-
It must have high light-load efficiency;
-
It must be able to produce a modest output voltage;
-
It must be able to run from a USB input supply.
​​
Now, when considering all of these factors, it seems like one would require a small miracle to find a chip that has all of this functionality... Enter the TPS6104x boost converter IC by Texas Instruments. Stumbling across this chip was the result of a brief parametric search on the LCSC website. The advantages of this distribution company are substantially cheaper prices on a lot of products (both from big brand names, although this is less common, and from offering alternatives by more niche Asian brands) and offering a large selection of Asian manufacturers' products which can help fulfil more niche requirements or, as mentioned above, cheaper alternatives.
​
This TI chip, however, was perfect for this project as explained below and as will become apparent in the following couple of sections. Firstly, it's cheap - about £0.24 at the time of writing - even for just a single IC. Secondly, it's offered in a SOT23-6 package (about 2mm * 3mm). It also has a relatively simple schematic as shown below in Figure 1 (the typical application schematic on the first page of the datasheet).
​
As if all of this wasn't enough already, the datasheet shows examples of ways to implement full load shutdown (i.e. presenting 0V at the output when the part is in shutdown instead of a diode drop below the input voltage which would be the case with the schematic shown in Figure 1) and various methods of controlling the brightness via PWM and true analogue signals.
​
Without a shadow of a doubt, I'd be hard-pressed to find a chip that offers this level of flexibility, with this performance, in this footprint & simplicity and at this price point. Consequently, the next part will go through the design of a boost converter module which will be optimised to this project.
​
​
​