Embedded Developer

This is probably the most time consuming part of the project requiring a lot of patience, testing and a 30cm x 30cm piece of wood. Seriously, you’ll need a piece of wood. Here we are going to convert this:

Into this:

Working out the Dimensions of the Cube

The exact dimensions of the cube will depend upon the length of the legs on the LEDs. The legs will be soldered together with the cathodes of the LEDs forming a horizontal plane and the anodes connecting the layers vertically.

Take one of the LEDs and bend the cathode at a point as close to the body of the LED as possible. The cathode should be at 90 degrees to the anode and parallel to the flat base of the LED. Now bend the top 3mm of the leg of the anode through 90 degrees. You are aiming for something which looks like this:

In the final cube we want all of our LEDs to be above each other. If we don’t then each layer will be slightly offset from the on beneath it. We won’t have a cube more of a leaning tower. So the final act of manipulation for the LED is to bend the anode of the LED slightly out from the body of the LED and then to straighten the anode again. The effect should be that the 3mm at the end of the anode should be enough to place the LEDS above each other. When they are connected in the cube they should look like this:

Layer Template

Measure the distance from the centre of the now horizontal cathode to the end of the leg and subtract about 2mm. This will tell you how far apart the LEDs are in the horizontal plane. The 2mm will be used to overlap with the neighbouring LED and will be used to connect the cathodes. Now this is where the wood comes in. You need to drill a grid of holes 8 x 8 which are each this measured distance apart. You will end up with a template in which you can place the LEDs in each layer making soldering easier.

The holes in the template must be small enough to hold the LED firmly but not too firm. The LEDs will need to be removed later without placing too much strain on the soldered joints.

Building the Layers

So now let’s get the soldering iron warmed up and on with the first layer. Take 8 of the LEDs and place them along the top row with the cathodes all pointing to the right (or left if you prefer). The cathodes of the 7 LEDs to the left will overlap slightly with the LED to the right. The rightmost cathode will go off into space. Solder the cathodes together. Congratulations, you now have a line of LEDs.

Now let’s add the remaining LEDs in the layer. I started on the left as I hold the soldering iron in my right hand. Take another 7 LEDs and place these under the top row down the far left column. Again the cathodes should be overlapping. Solder these together. Repeat with the remaining columns. At this point you should have a horizontal row of LEDs connected together with 8 strings of 7 LEDs hanging from it. Eventually you should have something looking like this:

Now test the layer. Simple enough, you just need a power supply and a current limiting resistor. For the LEDs I had chosen a 5V supply fed through a 68 Ohm resistor is adequate. Connect ground to the cathode of the LED which is flying off into space (remember, it’s on the top row, to the right). Now touch each leg of the LEDs in turn with the positive output of the supply (through the current limiting resistor of course). The LEDs should light up and as you touch the anode.

One final bit of soldering is to add a stiffening wire to the layer. Cut and strip a piece of wire. The wire should be long enough to cross the entire layer. Now solder the stiffening wire to the cathodes along the bottom of the layer. See the image above. One thing to note here is that although one piece of wire is sufficient, adding one or two more would not harm.

At this point you will have one complete layer. Time to remove it from the template. This should be done carefully so that you do not put too much stress on the joints. Lifting it up gently with a screwdriver should help. Do not be in a hurry, gently freeing the LEDs first may help. Put this layer to one side and repeat another 7 times.

Connecting the Layers

Now that we have all of the layers built I suggest that they are tested once again. This repeated test may save you pain later. Just imagine how difficult it will be to fix a bad joint in the middle of the cube. We are just checking that the LEDs are still connected and none of the joints have been broken removing the layers from the template.

Now drop one of the layers back into the template. What we now need to do is to place a second layer on top of this one so that the anodes of the layer in the template touch the anodes of the LEDs in this layer. Once in place we need to solder the anodes together – good game, good game. I had the advantage of having a few clamps which could be adjusted vertically as well as horizontally. I would imagine that cardboard (or wood) shaped to the correct height should help support the layers. Remember that whatever you use you will need to be able to remove it without damaging the cube.

Now you have the layer supported, solder the anode of the LEDs in the top layer to the anode of the LED in the layer directly beneath it. Once this has been done – test. Connect the cathode of the bottom layer to ground and touch each of the legs on the top layer in turn with the +ve supply (going through the current limiting resistor). The LED on the bottom layer should light. Move the cathode to the top layer and repeat the test, this time the LED on the top layer should light.

Repeat adding the remaining layers. Just for safety, I would test every layer in the cube as it is built up. This repeated testing sounds like a big overhead but trust me it’s worth it.

So at this point you should now have a cube and if you are like me, lost two days of your life in pursuit of happiness and all that is good in life.

Coding4Fun have just published my article on how to build a 512 LED cube controlled by a Netduino Mini.

A while ago I created a circuit and some code to control the TLC5940 LED driver using a Netduino Mini. This chip allowed the programmer to control 16 LEDs using a serial interface. In the conclusions I noted that it should be possible to control two or more of these chips by cascading the serial output from one chip into another. This article discusses the additional hardware required along with the software to control up to 32 LEDs using two of these chips linked together.

It is assumed that you have read and understood the previous article before you proceed further.

Hardware

Linking two of these chips is a relatively simple affair and requires the LEDs you wish to control and only 1 additional component, namely a 2K2 ohm resistor, and some additional wire (assuming you are using breadboard).

Starting with the simplest part, the 2K2 ohm resistor connects the IREF pin on the second TLC5940 to ground. This sets the reference voltage for the LEDs connected to the second TLC5940.

The next part of the problem is to connect the two chips together. This is achieved by connecting the following pins:

DCPRG on TLC5940 is connected to V_cc as it is on the first driver chip and the LED controller pins connected to the appropriate LEDs.

This PDF (PWMLEDControlv0.03.pdf) contains the full circuit diagram showing the connections required.

Software

Now we have the additional LEDs in the circuit we will need to be able tell the software driver for the chips how many LEDs we want to control and then be able to control the brightness of the LEDs. In order to do this an additional parameter was added to the constructor numberOfLEDs. This parameter allows the system to work out how many TLC5940’s should be connected together. Once we have the number of chips calculated we can then work out the size of the buffers for the Dot Correction and the Grey Scale data.

The class uses two buffers for both the Dot Correction and the Grey Scale data. This may seem superfluous but doing this allows two different modes of operation:

• Basic – The LEDs can be controlled using an interface which is simpler to use but is slower.
• Fast – This mode is faster but requires the programmer to perform some of the tasks in the driver. The assumption here is that the programmer will have more understanding of the data which is changing and can therefore write more optimised code.

For the moment we will be discussing the basic mode. In this mode, the software driver hides much of the implementation and the user only really needs to perform three tasks:

• Construct a new instance of the software driver.
• Set the brightness of the LEDs (only the lower 12 bits are used).
• Tell the driver to output the data to the TLC5940 chips.

Constructor

Tlc5940 leds = new Tlc5940(numberOfLEDs: 32);

You can of course use the additional parameters to define the pins used for the various control signals or just use the defaults.

Setting a LED

The class contains a method which overloads the array indexing operator. Setting the brightness for a LED is simply a case of using this operator as follows:

leds[10] = 767;

This will set the brightness of LED 10 to the value 767.

An important note here is that the LEDs are numbered from 0 to 15 on the TLC5940 connected to the Netduino Mini, 16 to 31 on the next LED in sequence and so on.

Lighting the LEDs

The final piece of the puzzle is to output the data to the series of TLC5940s connected to the Netduino Mini. This is achieved by calling the Display method in the Tlc5940 class.

Source Code

The source code for the controller class and a simple main program can be found here (CascadedTLC5940.zip).

Conclusion

The work with this chip has not yet finished as there are still some points which need addressing in order to round off this class:

• Dot Correction – this current outputs all 1’s and so there is no fine control for any of the outputs.
• Error Detection and Status Information – It is possible to detect two types of errors and retrieve the status information from the TLC5940. This needs implementing.

One final item which needs addressing is to round off the code once these final feature have been implemented. A topic for another day.

I have recently been working with the TLC5940 LED Driver chip. This provides 16 PWM outputs which can be used to drive common anode LEDs and I suspect motors etc. For the moment I will be using it to drive 16 LEDs to prove the principle and work out how to interface the chip to the Netduino Mini.

The article and code contained here should allow the creation of a circuit capable of doing the following (and a whole lot more):

Objective

The objective of this exercise is to use the TLC5940 to control 16 LEDs under the control of the Netduino Mini. This will require the following to be developed:

• Timer circuit to control the grey scale output.
• Control signals and reference voltage to control the output of the TLC5940
• Control logic and software for the Netduino Mini

A description of PWM and how it works is assumed. If you require more information then you can check out the Wikipedia article on PWM.

Timer (Clock) Circuit

The TLC5940 uses an external timing signal to control the 16 grey scale outputs to the LEDs. This signal can be as high as 30 MHz. The design presented here uses an 8 MHz crystal / resonator. When I was developing the circuit I have little knowledge of how to build the timer and so used a NE555 configured as a 40 kHz astable oscillator. This was crude but was good enough to start the project. The final design uses an oscillator circuit from a document recommended by Mario on the Netduino forums. You can read the original post here. Mario recommended circuit 1d which is the one I finally used:

The final inverter is not really necessary but the chip contains 6 inverters and I was only using two so I fed the output from the timer through an final inverter. This had the effect or producing a squarer signal – although the final signal is still slightly rounded.

In the final circuit I actually used a resonator. I simply dropped this in place of the crystal connecting the outer two pins to the logic gates and the centre pin to ground.

Control Signals

As well as the clock signal, the TLC5940 requires several control signals to determine how the chip should treat the grey scale data. This section looks at each signal and describes its function.

BLANK (Pin 23)

This pin determines the output state of the TLC5940. When this pin is high the outputs are turned off. Setting the pin low turns the outputs back on. More importantly, it changes the internal counter for the PWM control and resets the counter to 0. It is to this counter to which we now turn out attention.

The TLC5940 uses 12 bits to determine the output level for the LEDs connected. Each LED is controlled independently and the full array of 12 bits requires a total of 192 bits (12 bits x 16 LEDs) of data to determine the output levels for the LEDs. The 12 bits for each LED gives a total of 4096 (remember this number, we’ll be using it later) levels for each LED (0 to 4095 inclusive). The TLC5940 uses this data and the current counter to determine the output state of each OUT pin.

Setting BLANK to high and then low causes the counter to be set to zero. An incoming clock pulse causes the counter to be incremented by 1. The chip then turns on all of the outputs which have a greyscale value which is non-zero. Each subsequent pulse cause the chip to do the following:

1. Increment the counter
2. For each output compare the counter with the greyscale value for the output pin. If the value is greater than the counter then turn the output on, else turn the output off.

This process repeats until the internal counter reaches 4095 at which point the process stops and the outputs are effectively turned off (the counter value is greater than or equal the output value for the LEDs). This point is important as providing only one pulse to BLANK and a series of 4096 (or more) clock pulses makes this a one time only output. The LED on each output lights for the specified duty cycle and then turns off.

The trick here is to ensure that a BLANK pulse is generated for every 4096 greyscale clock pulses. This is achieved by attaching the output from the timer circuit to both the greyscale clock on the TLC5940 and a divider circuit. The divider counts 4096 clock pulses and then outputs a new BLANK signal to the TLC5940 effectively restarting the counting process. So the modified clock circuit looks something like this:

The clock signal is split and one connection to the GSCLK (grey scale clock) pin on the TLC5940, the other connection is to the clock inputs of a 14-bit counter/divider (74HC4060). The divider counts the clock pulses and sets the output pins of the chip to represent the current value of the counter. The counter is reset once it reaches the maximum value and the whole process restarts. The counter does not output the lower three bits but this is not an issue for us as we are interested in counting to 4096. We can use the output from bit 12 (2 ^ 12 = 4096) as a BLANK signal for the TLC5940. So for the first 4095 pulses the TLC is setting the outputs according to the values held for each pin. When the counter reaches 4096, the TLC stops displaying data but then a BLANK pulse is received from the divider. This then restarts the whole process. The result, we have the outputs continuously cycling the the LEDs on for the required duty cycle.

One final point about the BLANK signal. This could be used by the microcontroller which is controlling the circuit to turn off the outputs. To allow for this, the output from the divider is ORed with the BLANK signal coming from the microcontroller. The output from the OR gate is passed on to the TLC5940.

VPRG (pin 27)

This signal determines the type of data being transmitted to the TLC5940.

When VPRG is high the chip expected 96 bits of data for the Dot Correction (DC) registers. The DC registers allow the controller to scale the output current supplied at each output pin. This is covered later in this article.

When VPRG is low, the TLC5940 expects 192 bits of data for the grey scale registers.

XLAT (pin 24)

A low-high pulse on this pin latches the data received from the microcontroller into the DC or the GS registers. The data is latched on the rising edge (i.e. low to high) of the XLAT signal.

DCPRG (pin 19)

This pin is not used in this circuit and to connected to V_cc

IREF (pin 20)

This pin can be used to set the maximum current output on each of the output pins. This is done by connecting the pin to ground through a resistor. The value of the resistor is determined by the following calculation:

R = (1.24 / I_max) * 31.5

Where I_max is the maximum current to be supplied. The value of 1.24 is the reference voltage on the IREF pin.

For the LEDs I am using I want a maximum current of about 26mA. This gives a resistor value of 1502 ohms, so a 1k5 resistor should do here.

XERR (pin 16)

This pin is an output from the TLC5940 and can be used to report when an error is detected. This pin is not used in this project and is connected through a 10K pull-up resistor to V_cc.

Grey Scale and Dot Correction

The TLC5940 uses a value in the range 0-4095 to represent the duty cycle for each of the outputs. This gives an effective brightness for each LED given by the following calculation:

Percentage brightness = (Grey Scale value for output N / 4095) * 100

The dot correction reduces the current flowing through the output pins. This can have two effects, it can be used to adjust the current to bring it in range for a specific LED or it can be used to fine tune the output current for an output pin. This can effectively reduce the brightness of a LED. The change in maximum current for a pin is given by the following calculation:

I_out = I_max * (Dot correction value for pin N / 63)

Where I_max is set using the resistor connected to the IREF pin (discussed above).

Final Circuit

The final circuit looks like this:

Controlling the Chip

The whole circuit is controlled using a Netduino Mini. The four pink control lines in the above circuit are used by the Netduino Mini to control the TLC5940.

The application encapsulates the TLC5940 functionality in a single class Tlc5940. This project only deals with one chip controlling 16 outputs (even though several can be linked together). The class only exposes 2 methods:

• Constructor
• Display

Constructor

The constructor sets up the system to use the SPI interface and takes parameters to determine which additional control pins should be used.

Display

The display method takes two parameters, one for each of the dot correction data and one for the grey scale data.

The main program contains a small demonstration of the functionality of this class and the TLC5940 chip. It slowly fades in 16 LEDs from off to full brightness and then fades them out to the off state again. This is repeated in a loop forever. For an example of the output see the video at the top of this article.

The code is pretty well commented and so further discussion is not provided here.

You can find the source code here (TLC5940.zip).

Conclusion

This article is really a starter for this chip as there is a whole lot more which is possible. There are a number of areas where I would like to expand both the code and the circuit, namely:

• Resetting the counter when it reaches 4096. At the moment this runs on until it reaches the maximum value.
• Use the dot correction values to allow different LEDs to be used on each pin. Different LEDs have different forward currents and this can be fine tuned using the dot correction data to allow the maximum brightness to be achieved by each individual LED.
• Look at using the CS pin on the SPI interface for the BLANK signal from the Netduino Mini to the TLC5940.
• Add some error checking code to check if the TLC5940 has detected any errors.
• Link two or more TLC5940’s together.

Adding these features will make this class more rounded and fully encapsulate the feature set offered by this circuit.

The first cube was completed in July and the initial posting generated some interest. I agreed to make a second and this one has just completed it’s first transatlantic journey and has made an appearance at MakerFare in New York. The following image appeared on Time Out New York’s web site:

Building the second cube was fun as I was able to implement all of the lessons learned from making the first cube.

Whilst the blog may have been a little quiet recently I have never the less been busy writing about the cube. There is a Wiki article on the Netduino web site showing how the cube was constructed.

I have also released the source code to the project on Codeplex.

I’m just about to start to write this up but I thought I’d share the cube working before I do. Here is a video showing a few animations:

Watch this space for more information over the coming days.

In a previous post I mentioned that replacing a single LED in the centre of a matric would be murder and so it will be. The matrix I have in mind is an 8 x 8 x 8 LED cube controlled by a Netduino board. Now it should be obvious that there are not enough pins to connect 512 LEDs to the Netduino without some magic being involved. This is where the Persistence of Vision post comes into the picture. In theory, it should be possible to control an array using a group of shift registers and a multiplexer circuit. So for the next few weeks this is what I’ll be looking at.

Multi-Threading

The previous post on Persistence of Vision had a simple while loop which allowed 16 LEDs to be controlled from the main program loop. In the final program this will be too cumbersome and timing is almost certainly to become an issue. To overcome this we will need to have the display logic separated from the control logic. This will allow the display to be continuously updated whilst the main program loop is working out what should be displayed next. The following tests this concept by turning on on LED at a time in a bank of 8. If this project has a chance of working then the 8 LEDs should all appear to be switched on permanently.

Hardware

The hardware is relatively simple. We have one shift register connected to the Netduino. This is in turn connected to 8 LEDs through current limiting resistors. The schematic looks something like this:

Software

The software requires us to take the display logic from previous posts and add this to a new class. This class will need to have a method executing in it’s own thread in order to allow the main program and the display driver to run at the same time. So splitting the display code into it’s own class we get something like this:

class LEDCube
{
    /// <summary>
    /// SPI bus to use to send data to the shift registers.
    /// </summary>
    SPI spi = null;

    /// <summary>
    /// CSPI bus configuration
    /// </summary>
    SPI.Configuration config;

    /// <summary>
    /// Buffer holding the display data.
    /// </summary>
    private byte[] buffer;

    /// <summary>
    /// Constructor for the LEDCube class.
    /// </summary>
    public LEDCube()
    {
        config = new SPI.Configuration(SPI_mod: SPI.SPI_module.SPI1,
                                       ChipSelect_Port: Pins.GPIO_PIN_D9,
                                       ChipSelect_ActiveState: false,
                                       ChipSelect_SetupTime: 0,
                                       ChipSelect_HoldTime: 0,
                                       Clock_IdleState: true,
                                       Clock_Edge: true,
                                       Clock_RateKHz: 400);

        spi = new SPI(config);
        buffer = new byte[1];
        buffer[0] = 0;
    }

    /// <summary>
    /// Main loop which continuously updates the display from the buffer.
    /// </summary>
    public void DisplayBuffer()
    {
        while (true)
        {
            lock (buffer)
            {
                spi.Write(buffer);
            }
        }
    }

    /// <summary>
    /// Change the byte in the display buffer.
    /// </summary>
    /// <param name="b">Byte to put into the buffer.</param>
    public void UpdateBuffer(byte b)
    {
        lock (buffer)
        {
            buffer[0] = b;
        }
    }
}

Much of the code should be familiar, we have the SPI bus and config variables along with a constructor to make a new instance of the SPI bus. The new code is really the DisplayBuffer and the UpdateBuffer methods.

• DisplayBuffer is the method which has replaced the part of the main program loop which outputs the data to the SPI bus. This method is run in it’s own thread.
• UpdateBuffer simply copies new data from the caller into the display buffer.

Note that both of these methods use locking to ensure that the methods are thread safe.

The main program loop now looks something like this:

LEDCube cube = new LEDCube();
Thread display = new Thread(new ThreadStart(cube.DisplayBuffer));
display.Start();
while (true)
{
    byte value = 1;
    for (int index = 0; index < 8; index++)
    {
        cube.UpdateBuffer(value);
        value <<= 1;
    }
}

A new LEDCube is created and at first this is not running in it’s own thread. The next two lines create and start a new thread. The thread is executing the DisplayBuffer method. We then start the main program lopp which simply sets each bit in a byte and the updates the buffer in the cube using UpdateBuffer.

Next step, build the control logic for 512 LEDs.

By chance last week end I came across a post discussing persistence of vision. This set me wondering if I could demonstrate this using the Netduino. First a little background.

Background

Persistence of vision is the phenomena which allows you to watch a movie. The projector shows a single still image on the screen. This is then replaced 1/24th of a second later by a new image. The projector achieves this by covering the image with a shutter whilst the frame is moved into place. We do not see the blank screen, just the previous frame. A fuller description can be found on this Wikipedia page.

We will be using this effect to control 27 LEDs using only 11 control lines. These lines will be the outputs from two 74595 shift registers which in turn will be connected to a Netduino Plus.

The Problem

The 27 LEDs will be split into three banks of 9. The aim will be to write a small program which will turn on one or more of the of the 9 LEDs in each bank in turn. If we do this slowly we will see bank 1 turn on whilst banks 2 and 3 are off. Next we will see bank 1 turn off, bank 2 will turn on and bank 3 will remain off, and so on. In theory, if we do this fast enough, we will see all three banks on at the same time.

Hardware

This project uses the hardware and principles from two previous projects:

1. Netduino Controlling an External LED (using a transistor as a switch)
2. Counting Using 74595 Shift Registers

A picture is worth a thousand words so lets have a look at the schematic:

The two shift registers are connected to the Netduino using the SPI interface. The output from IC1 is fed to the serial input of IC2 creating 16 output lines (8 from each register). These lines will power the LEDs and select the bank to turn on.

The LEDs will be powered using bits 0 to 7 of IC1 and bit 0 of IC2.

The top three bits of IC2 will select which bank to turn on. This is done using a transistor as a switch.

The trick to controlling the banks and LEDs lies in the way the circuit is wired up. The anodes of LED1, LED10 and LED19 are connected together and then connected to QA of IC1. Similarly, the anodes of LED2, LED11 and LED20 are connected to QB of IC1. This continues until all 27 LEDs have been connected to the 9 control lines (QA-QH of IC1 and QA of IC2).

The next part of the trick is to connect the cathodes from the LEDs in bank 1 together. These are then connected to a current limiting resistor which is in turn connected to the collector of a 2N2222 NPN transistor. This is repeated with banks 2 and 3 being connected similarly to their own transistor.

The base of each transistor is connected through a resistor to one of the bank control outputs pins (QF, QG and QH) of IC2. The emitter is connected to ground.

Note that in theory it is possible to turn on all three banks at once by setting the appropriate bits in the shift register but the object of this exercise is to show how to make a still image using each bank in turn.

Software

The software uses SPI to send two bytes to the 74595 shift registers. The values sent will control which LEDs are turned on / off. The code looks like this:

SPI.Configuration config;
config = new SPI.Configuration(SPI_mod: SPI.SPI_module.SPI1, ChipSelect_Port: Pins.GPIO_PIN_D9, ChipSelect_ActiveState: false, ChipSelect_SetupTime: 0, ChipSelect_HoldTime: 0, Clock_IdleState: true, Clock_Edge: false, Clock_RateKHz: 400);
SPI spi = new SPI(config);
byte[] value = new byte[2];
value[1] = 0xff;
while (true)
{
    value[0] = 0x81;
    spi.Write(value);
    value[0] = 0x41;
    spi.Write(value);
    value[0] = 0x21;
    spi.Write(value);
}

If you run this code you will see a still image with all three banks looking as though they are permanently on. Set a breakpoint on the line value[0] = 0x81; and run the program again. Single stepping will show each bank being turned on in turn. The reason we see the still image is because the banks are turned on and off quickly.

Putting it all together and wiring it up on breadboard gives you this:

Persistence Of Vision On Breadboard

Following the post earlier this week regarding the implementation of a ShiftRegister class which allows the Netduino to control a series of 74HC595 shift registers I had a look at what would be needed to make the system count and show the output in binary on a series of LEDs. What you see here is the result. The hardware is the same as the previous post, only the software has changed.

One of the main desires is to allow the programmer to use the natural language features of C# to work with this class. The modifications should therefore support operations such as assignment, logical and etc. The main program loop for a counter should look something like this:

ShiftRegister shiftRegister = new ShiftRegister(16, Pins.GPIO_PIN_D9);
for (ushort index = 0; index < 10000; index++)
{
    shiftRegister = index;
    Debug.Print("Count: " + index + ", " + shiftRegister.ToString());
    shiftRegister.LatchData();
    Thread.Sleep(100);
}

In order to support this we will need to overload the implicit assignment operator for an unsigned short being assigned to a ShiftRegister instance. This results in the following code:

/// <summary>
/// Overload the assignment operator.
/// </summary>
/// <param name="usi">Unsigned short integer to assign to the register.</param>
/// <returns>New ShiftRegister holding the unsigned short value.</returns>
public static implicit operator ShiftRegister(ushort usi)
{
    ShiftRegister result = new ShiftRegister(16);   // ushorts are 16 bits.

    ushort mask = 1;
    for (int index = 0; index < 16; index++)
    {
        result._bits[index] = (usi & mask) > 0;
        mask <<= 1;
    }
    return (result);
}

This generates some problems with the base shift register class from the last post. Most noteably the creation of the SPI instance. The run-time system will generate an error should the programmer try and create two objects wanting to access the SPI bus. Reading the above code you can see that the assignment overload requires a new ShiftRegister instance to be created. A few changes are therefore required in order to allow the system to share the same interface. In order to allow this, the base class moves the SPI object from a shared instance to a static object. This ensures that only one of these can exist at any time. The remaining modifications to the class support this change and add a ToString() method for debugging. The modified code and a sample test project can be found here and the following video shows the application in action.

Further Developments

The base functionality assumed that the class is the only class wanting to use the SPI bus. The number of chips and breakout boards using is large and so it is likely that the programmer will want to communicate with several slave devices using the same bus. This is allowed using the SPI protocol and for the moment this is left as an exercise for the reader.