LPC2103 Hardware Design Overview
An overview of some of the hardware design decisions made for our LPC2103 Reference Design.
This page aims to give you a basic overview of some of the main decisions that were made while designing the LPC2103 Reference Design base board. It isn't exhaustive, but rather it focuses on some of the main problems that we faced when putting the board together.
You may also want to consult the LPC2148 Hardware Design Overview, since these two chips share a number of common features and certain hardware was added on one board that isn't present on the other. For example, we've added a CR2032 battery to the RTC on the 2103 board to keep track of the current date and time even when power is not available ... on the 2148 Reference Board, we simply power the RTC from the main voltage supply meaning the date and time will be lost when power is removed from the device.
Battery Input/Step-Up Converter
In order to be able to run the LPC2103 with standard AA batteries (which typically have an upper rating of 1.5V for standard alkaline cells or 1.2V for rechargeable NIMHs) a step-up converter was added. One of the main problems with step-up converters, though, is that they usually have a lot of 'ripple' in their output (meaning that there is variation in the output voltage). This isn't a huge problem for the MCU itself, but it can be a very significant problem for analog sensors since the voltage the sensors are using as a reference point isn't fixed or dependable, rendering the conversion results inconsistent and inaccurate. To work around this problem, the choice was made to use a step-up converter with 5.0V output, and then pass that into a standard voltage regulator that would regulate the voltage 'down' to the required level. The advantage of this approach is that you have a much more stable power supply, rendering your sensor conversion results more accurate, but the disadvantage is that you will also consume more power since you're using two different voltage regulators and powering any additional components required by them. The choice to go with a more stable power supply or lower power consumption will obviously depend on your requirements, though more expensive step-up converters with less ripple may be a valid compromise as well. In this case we chose to err on the side of stability and affordability.
We chose to use the NCP1402 step-up converter since it's relatively inexpensive (0.45 €/100 on Digikey), it requires few external components, and the 200mA it's capable of supplying is likely enough for most situations where we would chose to use an LPC2103 (reading values from a few basic analog or digital sensors, etc.). Following the suggestion in the datasheet, a 47µH inductor was used, but identifying an appropriate diode posed more problem (we ended up trying four different diodes before settling on one that worked well for us!). The main problem was finding an affordable diode with a very low foreward voltage drop (often written as Vf), which usually means using a Schottky diode, as well as a very fast reverse recovery time (typically written as trr). If the trr is too high (or rather, too slow), the power output will be unstable. After a bit of trial and error, we settled on the Zetex ZLLS410TA since it has a low 0.38V Vf, a very fast 3ns reverse recovery time (trr), and is capable of handling up to 10V and 570mA, which is much higher than we will be using in this case. It also has the advantage of coming in a small, space-saving SOD323 package, but this wasn't really a deciding factor for us. The choice of input and output capacitors (which serve as mini power reservoirs to help further stabilise the power supply) followed the recommendations in the datasheet.
In order to make it as easy as possible to test the voltage level on the battery (since it will drop over time as the battery or batteries discharge), two small test points were also added to the board (TP4 and TP3). This allows someone to easily measure the batteries current voltage using a standard multi-meter or any sort of clamp-on test probe, rather than trying to connect a probe to the fine-pitch JST connector or some other tiny part or pad on the board (potentially causing a short or some other form of nastiness). These test points can also be used to supply the battery power during testing and development if you are unable to use the JST PH-series connector for one reason or another.
In order to allow people to also power the board using a standard connector, a 2.0/2.1mm DC barrel was also added to the board (you can choose which input source to use by setting JP2/PWRSEL to the appropriate position). Using the DC barrel has one main advantage ... it will consume less power than using a battery since you avoid the step-up converter altogether!
Caveat Emptor (Buyer Beware!): One thing worth noting is that you may not being buying what you think you're buying if you just go by the specs published in large letters in the title of a datasheet! These are best case scenario figures, and seldom have any relationship to the real-world. In the case of the NCP1402, it's advertised as having a very low 0.8V startup voltage, which may be true in certain very limited situations, but we found that on a board that draws around 100mA the 5.0V version of the step-up converter actually needed around 1.5V to get a stable supply. No one is lying to you ... those figures are in the charts in the suppied datasheet, but you need to carefully read all of this information or test things thoroughly before settling on any one component. In this case, it was a bit of a disappointment, since it meant the board couldn't be run from a single AA NIMH cell (which is typically around 1.2V fully charged), or even a single half-empty alkaline cell (which starts around 1.5V fully charged). Adding a second battery of either type largely exceeds to 1.5V minimum, but it also adds extra weight and means the battery supply will require more physical space. While there are step-ups that would be able to provide 100mA at 1.0V, they tend to be more expensive, and we decided once more to lean more on the side of providing an economical solution, rather than pushing the boundaries on the power supply and increasing the total production cost.
3.3V/1.8V Voltage Regulation
The LPC2103 is different than some other members of the LPC2100 family in that it runs off two different voltages: it has a lower-power 1.8V Core and uses 3.3V to power the peripherals. This results in a lower power design, but also necessitates a slightly more complicated approach to regulating the power supply since you will need to use two different voltage regulators (one for 3.3V and one for 1.8V) or a regulator with two different outputs, which is what we've chosen to do here using a 3.3V/1.8V version of Micrel's MIC5320. VOUT1 is the 3.3V supply and VOUT2 is the 1.8V supply.
In order to reduce power consumption as much as possible, a jumper was also added to the power indicator LED (LED3/3.3VRED) so that it can be disabled if it isn't required. This won't cut your power consumption in half, but a number of little power-saving steps like this can add up to a reasonable reduction in the total energy footprint of your design.
Parts Used
- U2: Micrel MIC5320-SGYD6 TR 3.3V/1.8V dual 150mA voltage regulator (Datasheet, Digikey)
Real-Time Clock (RTC) Battery Supply
Unlike the LPC2148 Reference Design, on which we decided not to include a CR2032 battery to save surface space, we've added a battery here to power the RTC (real-time clock) when the board itself is not powered. This means that you can set the date and time with the LPC2103's internal RTC, and it will remember and continue to adjust the date/time as long as the CR2032 battery is connected and has enough of a charge left in it. This can be important for remote sensors that are put into a 'sleep' mode most of the time and wake up every x number of minutes to read some sensors, log the data, and then go back to sleep. Having the RTC correctly recording the date and time allows you to also include this information in the logged data.
In order to extend the battery life (a CR2032 normally has ~200mA), a BAT54C dual diode was added to the design. This diode, which has two 'inputs' and one common 'output', allows the RTC's power supply (VBAT) to draw power from the main 3.3V supply when its available (saving battery power), but will automatically switch to the 3.0V CR2032 lithium battery when the main 3.3V supply is no longer present. The reason that this functions this way is because the diode will naturally draw energy from the highest voltage source available to it, which is 3.3V when the board is normally powered, or 3.0V when nothing is available on the diode's other 'input'.
Two small test points were also added to the PCB to allow users to easily measure the voltage level on the CR2032 without removing the battery from the enclosure (which would cause them to lose the current date/time on the RTC). This allows you to judge whether or not the CR2032 needs to be replaced in a non-intrusive manner.
JTAG Connector
The JTAG connector follows a similar design to the one found on the LPC2148 Reference Design, with the exception that on the LPC2103 the RTCK pin needs to be pulled UP, whereas the LPC2148 requires this pin to be pulled down. It pays to carefully read the datasheet and user manual, since even in the same family there can be small variations that can significantly affect your overall design!