1 LED - 5 functions

Sometimes need of an indicator for a magnitude of which is not essential to know the value, but if it is inside of a certain range or below a given limite.

In these cases, a character display (which must be read) would be less effective than some LED (which is immediately appreciable). 

In the past We had been made a circuit with classic comparators typically LM311 or LM339, and a quadruple operational, typically LM324, which provided a two-color LED on the following guidelines: 

• steady green LED - green low value
• green LED flashing  - rising value
• flashing green / red - mean value 
• flashing red  - high value
• solid red - over limit

In this way, with only one point on which to focus the attention, you can see 5 different situations of tension that can be a divider or a shunt or any sensor or circuit point. 

In practice, this kind of indicator is always effective result, despite its simplicity, because it is easy to read and allows the user a quick evaluation. Situations of flashing, then, attract the attention much more than a numerical indication to read and interpret. 

Since we have a voltage input, we can discriminate also the state of charge of batteries, the levels of voltages coming from the pressure sensors or current, but also as an indicator of speed limits, on the output voltage of converters F/V and any other application where it is necessary to identify at first glance the status of a segnale.

With the actual microcontrollers, with AD converter, a revision of the circuit could only pass from the all-analog to all-digital, with the advantage of a much smaller number of components and the possibility to vary the switching points without need to calculate dividers and resistors have uncommon values. 

The Analogic-Digital Microcontroller.

A practical case can occur with a heat sink for semiconductors, which it is necessary to know if the temperature exceeds a given value, in order to, for example, act on the adjustments or insert a ventilation system.

Given the need to evaluate an analog voltage, use a microcontroller with ADC module, the module that is also available on the tiny PIC 10F series.
In particular, We find it in 10F220 / 22, similar chip that differ only by the available memory.

We can, of course, use any other microcontroller with ADC, but, since they serve only two pins for the control of the LED and one for the 'input of the voltage to be monitored, the PIC10F 6-pin (4 I / O) are more than adequate , besides having a minimum cost and to occupy a space as never reduced (in the version SOT23-6).

The electrical circuit is, of course, extra simple:

IC1 is a common 78L05 for supply a stable voltage to the PIC: this is indispensable, because the ADC module use the Vdd as Vref for the conversion. In this way the circuit can be supplied starting from 7V to over 24V. If you already have a voltage of 5V with a few tens of milliamps, IC1 and C3 may be omitted (but no C2). C1    100nF C2    2.2uF-10uF C3    10uF-100uF R1/2 220ohm

IC1 (or a well stabilized Vdd) is essential: even if the PIC is plug-in itself as low as 2V and does not require a particular stability of the Vdd, however, serves a supply voltage accurate and of good quality, since it is the Vref for reading AD, otherwise we will have difficulty in interpreting the results of the conversion.

The ADC module of PIC10F22x is 8 bit and, in a simplified application such as this, the accuracy of the reference voltage provided by a linear three-terminal is sufficient for the conversion. However, if the need to do a conversion to 10 bits 0 12 bits, it would be indispensable to a different microcontroller, provided with input for an external Vref, supplied by a precision voltage reference (LM4040 and the like), in addition to a Vdd low noise.

The LED is a two-color three-pin, with one red and one green; in dependence on the type used, it is necessary to vary the resistances R1 and R2 to achieve the desired brightness, of course not exceeding the limit 25mA of current supplied from the pins of the microcontroller.

The LED indicates to the user the status of the temperature reached, going from green to yellow to red with the increase of temperature itself, whose situation is of immediate reading. They were laid down the following conditions:

The "yellow" is those that appears when you switch to green and red with a duty cycle of about 50%.
"yellow / red flashing" is obtained by keeping the red LED and toggling the green; so is the "green / yellow flashing" is obtained by keeping the lights green and switching the red.
The "yellow flashing" is obtained by switching both red and green.

It will be able to easily obtain other effects, by combining the times of switching on and off of the two LEDs.
The operation is simple: the microcontroller reads the voltage to be evaluated and, according to the limits set, controls the switching of the LEDs.

We can also employ an output for a relay or use a microcontroller with a greater number of I / O for controlling a LED bar and so on.

Software.

Obviously, given the type of chip used, which is a Baseline with minimal resources, the program is in Assembly and is also simple roll.

The flow chart of the program, even in the case of a simple program like this, is an important element for the programmer, since its drafting allows to have a clear picture of what you want to attract and allows to transform the logical path into instructions in so much easier than if it had been drawn. The "active" is essentially a loop recycling at a rate of about 1/4 of a second the following: reading of 'ADC evaluation of the results in the table view before control of LED timing The time delay determines the flashing speed. Changing it, they will get different cadences. Since it is a subroutine, it can be replaced with ease.

In our case, the voltage to be measured arrive from a cheap sensor active, Microchip MCP9701:

	In particular, the 'output proportional to temperature consists of a voltage that varies from 19.5 mV / ° C. The component is provided in the classic container TO-92, but also in SOT23 and SC70 for SMD. It can assess from -40 ° C to + 150 ° C, with an accuracy of ± 2 ° C from 0 to 70 ° C.
	It is provided for a direct connection with the 'analog input of the microcontroller and requires no conditioning circuit or linearizations particular in the field of readings expected. Of course you can use other similar sensors, as MCP9700 or LM35, adjusting the values to which you switch LED status.

We observe that the resolution of one step of the output value of the AD converter 8-bit, with 5V reference voltage is:

Vstep = Vref / step = 5 / 256 = 0.019531V

which is precisely the 19.5mV provided by the manufacturer. This greatly simplify the software, since it does not need any tuning or complicated calculations.

MCP9701 has an offset at 0 ° C equal to 400mV, which allows it to measure even at temperatures below 0 ° C.
Then, the voltage at a given temperature above 0 ° C is equal to:

Vt = .400 + (0.0195 * T)

For example, at 50°C the output will be:

Vout50 = (19.5 x 50) + 400 = 975 + 400 = 1375 mV

We can easily draw up a table that indicates the relationship between the temperature and the output voltage from the sensor:

But, in practice, as said before, we are not interested even know the value of the voltage, in that we can directly carry out the evaluation according to the 8-bit value resulting from AD conversion:

The values are rounded, as in 'application does not matter higher accuracy is required, however, with a little complicated correction algorithms (bitwise) and described in the application listed at the bottom of these pages.

Now let's see how to turn the flowchart in instructions.

All 'beginning place is the usual head introductory-explanatory, which is helpful to understand the purpose of the program, as well as outline the connections between the chip and the rest of hardware.

Now we select the processor.
Since PIC10F220 and 10F222 differ only by the amount of available memory, the most practical thing is to create a source suitable for both. The selection of the processor in the control bar of MPLAB guarantee operation of automatism, managed with #ifdef that selects the appropriate includes.

Basically, if you chose the MPLAB project creation or through the menu Configure/Select Device the PIC10F220 :

the statement #ifdef will let the compiler executes only the highlighted lines:

In the case in which it was selected 10F222:

Note that the definition of the processor appears in the list of symbols (.lst file resulting from the compilation) preceded by a double underscore __, that, in the Convention of Microchip, indicates a private label of 'build environment.

In this simple way you can write a single source automatically adaptable to two different processors.

Note that establish the general reference for the numerical values to the root decimal, r = DEC, which is more "natural" than the default hexadecimal.

Now for the Configuration Word, which is simple, given the limited resources available in the chip. In particular:

•  _MCPU_ON  enable an internal pull up on MCLR pin, avoiding to add an external component
•  _MCLRE_ON  enable GP3 as MCLR
•  _IOFSCS_4MHZ enable the internal clock to 4MHz
•  _WDT_OFF and _CP_OFF  disabile Watchdog and code protection

We observe that this chip allows you to:

• have a 4MHz internal clock, which is essential given the very low number of pins: if they were given over 2 to an external oscillator would remain only two more available as I / O. Among other clock it can be accessible from outside through one of the pin; here this option is not enabled because they serve all the available I / O.
• place an integrated pull-up on the MCLR pin, for which not even need to add an external resistor pull-up: low count of components in  the circuit 

We assign now the limit values of the steps relating to the various modes of operation:

If We use other sensors (MCP9700, LM35, etc) or other voltage limits and step, need to change the table.

Now we define the use of GPIO and RAM.
In particular, the control of the LED output refers, as a symbol, directly to the shadow used to avoid the problem RMW.
So the RAM you must reserve a location for this shadow.
Use in a case so simple a shadow of the I / O is not strictly necessary: the load being constituted only by LED, the problem RMW does not occur, but it is still appropriate to use the principle for the correct procedure for access to 'I / O PIC without LAT register.

In RAM they are reserved also two locations for counters required by routine timing.

To simplify the use of shadow and avoid the repeat of identical instructions that make heavy reading the list, it may want a small macro to transfer the shadow registers of I / O.

We can add some other macro control of the LED; they are not essential, but are useful to clarify the possible listing:

The flashing is made by changing the status LED or LED concerned. This is achieved with a toggle (reverse state) of the value of the corresponding bit in the register I / O. The toggle is performed with the XOR function; remember that XOR with 1 inverts the value of the bit.

The actual program starts with the classic writing of calibration value of internal oscillator.
To understand well that for these PIC the reset vector does not point to 00h, but to 'last location of the Flash, where the manufacturer, in testing the chip, writes the calibration value of'oscillator, in the form of an opcode: 

movlw   valore_calibrazione

So, the first statement that the PIC runs after reset is not what is 00h, but this movlw.
At the next cycle, the wrap around the program counter brings its value to point to 00h, ie all 'area of instructions inserted by' user.
Since WREG is loaded with the appropriate value for the calibration of 'internal oscillator, the easiest thing is to transfer this value to the OSCCAL register.

The line andlw 0xFE is added with the purpose of dismantling the possible output Fosc / 4 where this is available (if it is not used, as in this case).

This action would not be necessary, given that the activities of micro in this application does not require any particular accuracy in the cycle time, but even here it is advisable not to jump instructions that on other occasions may be necessary. Follow a precise procedure is essential not to find themselves with problems whose causes are sometimes identified with difficulty.

The main program starts with the setup of OPTION_REG and the direction of the pin.
To be noted that the Baseline require special instructions OPTION and TRIS, which is obsolete for PIC families superiors.
A semi-graphic helps to understand the function of various bits and assign the right value. Here 's use of the binary number is very clear that not expressing the value in hexadecimal or decimal.

The evaluation of the results of the AD conversion is carried out with a subtraction from the values specified for the various ranges.
It should be noted that it is necessary to use subwf  because  Baseline don'yt have the opcode  sublw.
skpc (skip next line on carry set) it is one of the pseudo instructions supported by MPASM Assembler and that is of more immediate understanding of' equivalent btfss STATUS, C.

It may be interesting to note that the color of the text, easily accomplished by an editor for programming, allowing an immediate evaluation of the components of each line.
For example, we look at how the 'setting of the color (default typical for listed PIC) stand in the opcodes blue, green commentary and black label and macro.

Let's add a couple of subroutines, one related to AD conversion and 'other generation of the delay wanted.

Simple is always extra management ADC module: no need to capture time between readings AD and later since they are already widely rhythmic routine of late.
The delay is of the usual type waste time, since Baseline does not have interrupt; the structure is derived from the site of Golovchenko and may be varied at will to obtain a flashing with different frequency.
Of course you can also use the Timer0, which in this application is not any other function.

END terminates the compilation.

Note that the return from the subroutine requires retlw, given that there is no return in the instruction set of the Baseline. You might think of introducing a equate:

RETURN = RETLW

to maintain uniformity of writing with other PIC, but this is totally unnecessary: ​​we can also write directly return, since MPASM performs equally compile without errors, operating automatic replacement with the correct opcode (and pointing out the operation with a warning).

However, the advice is to use the instruction set of the chip that you are using; This allows you to have a greater awareness of what you are doing in relation to 'hardware.

As for the LEDs, you might think to handle them with two PWM so as to have a passage of the green-yellow-red uniform, but we have seen that in practice the move with the flashing is much more effective for the' operator that immediately identifies with this a time of transition from some situation to another.

Instead, if desired, a pwm may be applied to the on-state of the LED, adding another timing loop to a frequency greater than 100 Hz, in order to reduce consumption, where necessary, given that the two-color LEDs usually require more than 10mA for have a satisfactory brightness.

Obviously such a structure can also be used to control small relays, opto-isolators or buzzers, creating alarm outputs or care in relation to the measured value. In this case you need to add a hysteresis sufficient to prevent the bounce of the contacts for small changes around the threshold values.

Hardware.

Given the 'extreme simplicity of the circuit, the construction can be a pleasure or as needed of' application.
Here a couple of examples of printed circuit boards (scale approximately 2: 1):

Obviously, the SMD circuit is smaller, but not so much, having to use a single-sided printed circuit. 

However, passive components are used SMD also for the version with the chip in package DIP8 to have a small circuit surface printing. The three-terminal regulator is still in TO92. In the version in the picture, the base of the chip is replaced with a single pin (not included those relating to the pin is not internally connected to the chip version DIP8). A pin connector 2.54 allows you to connect the cables with the coupon directly to another PCB. Given the very low number of components required, the circuit can be achieved in many other ways, even on a breadboard.

Variations.

We see a possible variation: indicate the status of a voltage, for examples that of a 12V battery.
We send all 'analog input voltage to be evaluated through a divider:

R3 and R4 are the divider input. In this case you could have a value of 1: 4, making it possible input voltages up to 20V. In this way we Vin = 20V -> Vgp0 = 5V. They may be respectively 6K2 and 2k2 or 11k ohms and 3K9 or other combinations in the ratio 1: 4. If it is the circuit is connected to the battery of a vehicle, it will be possible to put a zener diode (5V6-1 / 4W) in parallel to C1 to prevent the input voltage reaches a dangerous limit for the microcontroller. If it is not expected that the input voltage exceeds 20V, the zener can be omitted.

Obviously, the divider may be adjusted to the voltage that you want to measure and the same structure adapted to other range of input voltage.

We establish the following conditions of the LED:

Here, too, the values are completely adaptable to the ranges that you want to evaluate. Just change accordingly the values of the constants related.
Then we assign the limit values of the various modes of operation of the LED:

Modifichiamo l' analisi del risultato della conversione per adeguarla alle nuove condizioni.

A slight complication is inserted to get the flashing green / red.

The problem that can be generated is as follows: in the absence of interrupts, in order to exploit the cycle as cadence of the flashing with the system of the toggle, it is necessary that the two control bits of the LEDs are different since it is necessary that the first entry in the SEL5 selection have different values for the two bits that control the LED red and green.
If you switch to SEL4 SEL5, or with a growing tension, there is no problem. In SEL4 the two control bits LEDs have different value, so their toggle produces a flashing alternately.
A flow chart makes clear it.

If you go directly to SEL3 to SEL5 for a rapid increase in voltage (the acquisition cycle is very "slow"), you come up with the two-bit value of the spread, which would lead to a toggle flashing yellow and not all 'alternating red / green desired. It is therefore necessary that all 'input of fasesel5 the two bits are of different value. If not, the cycle of the toggle does not allow flashing red / green. It does not matter how they setup the first time, because the next steps of the cycle will reverse the state and this guarantees the flashing.

Of course you can use other logical paths to generate the flashing, but the one used is the most banal and simple absolute.

Of course, if you prefer a different reaction of the LED, just change the operations performed at each selection.
Here too we can think about the connection of relays which are inserted / switched off depending on the voltage (always remembering to insert a minimum hysteresis in the switching points).

A further variation would be to use the GP3 as digital input and select with the voltage value applied two different configurations of the set of intervention and / or LED's function, thus being able to use a spread chips for two different applications.

As we see, a very simple programming in Assembly on a PIC among the smallest available (not only in size, but also as resources) allows sensitive results.

The program will serve as a basis to create all kinds of indicators. You can associate different conditions of the LEDs and different thresholds, with different outputs from the two-color LED used.

Documentation.

• MCP9700/9701
• AN679 - Temperature Sensing Technologies
• AN1001 - IC Temperature Sensor Accuracy Compensation with a PIC Microcontroller
• AN981 - Interfacing a MCP9700 Analog Temperature Sensor to a PICmicro Microcontroller

• Source and MPLAB project