Repository navigation
Home
Welcome to the Wiki for my project, a portable, lithium-ion battery powered speaker amplifier with both solar and USB charging capability. This project was created to satisfy the project requirements for ENGR460, and was inspired by the suggestion to use a solar cell that was provided to us as well as a personal love for music and DIY audio projects. The ultimate intention for this project was to attach a Bluetooth audio receiver to the board, thereby creating my own portable Bluetooth speaker with solar charging.
To realize this project, there were four major components to be designed:
- A Li-ion battery charging circuit capable of taking both USB and solar power input: Being able to take solar power as an input meant that an IC capable of some form of power point tracking would be required.
- Battery protection: While most battery charging ICs can provide overcharge protection, not all are able to provide protection from other things that could damage a lithium battery, such as overdischarge, which can damage a lithium battery permanently. Thankfully, many battery protection ICs exist on the market that will disconnect a battery from the rest of the circuit if a dangerous condition is detected.
- Audio amplifier: Considering this project is a portable speaker amplifier, this part is a given. It was desired to use a speaker amplifier IC with adequate power output, wide voltage input range, and good efficiency.
- Voltage booster: While there are speaker amplifier IC's capable of operating using low voltage input, such as the typical 3.7-4 volts one might see off of a lithium battery, many of the amplifiers with significant power output only achieve those power output values with voltage inputs of about 9-12V, so the voltage from the lithium ion battery would need to be boosted in some way.
Above is a simplified block diagram of the components of the system. The boxes highlighted in green comprise the parts that are on the circuit board that was created for the projects. Everything else is external to the circuit board.
When selecting ICs to use for the project, admittedly, I used a lot of Texas Instrument ICs, largely because I trust them as a semiconductor company, their website is easy to navigate to find the right kind of product you might be looking for, and I find their their datasheets and documentation to be much better than most of their competitors.
-
Battery Charging: TI BQ24210 800mA single-input, single cell Li-ion solar battery charger
This IC was a great fit for this project for a number of reasons. While its current output isn't super impressive at 800mA, its hallmark feature is the ability to take both solar and USB inputs. Using solar power to do work like charging a battery requires power point tracking, often referred to as Maximum Power Point Tracking (MPPT). While the datasheet doesn't say that the BQ24210 uses MPPT specifically, it does say that it has "selectable battery tracking mode to maximize the charge rate from solar panel using DPM feature", so some form of tracking is likely being employed. Other great features of this IC include thermal protection, battery detection, status indication, compact size, and a wide input voltage range. Importantly, the minimum input voltage is 3.5V, which is good considering the solar panels we were supplied had an open-circuit voltage of about 5-6V. Thus, the battery charger should still be able to operate if an overcast day partially obscured the solar panel.
-
Battery Protection: TI BQ2970 Li-Ion/Li Polymer Advanced Single-Cell Battery Protector IC
This IC is part of a family of battery protection ICs from Texas Instruments, and simply monitors for conditions that could harm the battery. If a potentially harmful condition is detected, the IC will deactivate the power MOSFETs that allow the negative terminal of the battery to be connected to the circuit's ground. The main thing this IC needed to protect against was undervoltage on the battery.
-
Audio Amplification: TI TPA3138D2 10-W Stereo 3.5-14V Analog Input Class-D Audio Amplifier
This IC made for an excellent choice for a speaker amplifier for the project, particularly for its wide input voltage range, high efficiency, more than adequate power output, low idle current, and lack of a need for inductors or external heatsink. The suggested applications for the IC even include Bluetooth and wireless speakers.
-
Voltage Booster: Maxim Integrated MAX1771 12V Step-Up DC-DC Controller
While I was originally going to go with the Texas Instruments LM2623, I quickly decided against this since its reference design required a capacitor whose value depended on your application, but could also not be determined empirically due to its non-linear characteristics. Because trial-and-error design was not a feasible option for this project, I switched to using Maxim's MAX1771 step up controller because it is pre-configured for 12V output, has a more than adequate output current at 2A, good efficiency, and a wide input voltage range. Observing the graphs from the amplifier datasheet for both output power vs. input voltage, and output power vs. efficiency, it can be seen that much better power output is achieved with a higher voltage input, and better efficiency is also achieved with higher power output. In addition, the power efficiency measurements were done with an input voltage of close to 12V.
The project used a few items that were intended to be connected to the circuit board, namely a battery, solar panel, and speakers.
- Solar panel: The solar panel used was supplied to the students of this class, and is functionally identical to this ALLPOWERS solar cell.
- Battery: A lithium battery was sourced from a commercially available portable power bank, namely the Anker PowerCore 5000. This was done because the power bank was about the same cost as ordering a lithium battery by itself (especially after shipping), and by ordering from a trustworthy company such as Anker, I knew I was getting a good lithium battery. The portable power bank can then be dismantled to isolate the battery, and in addition, the charge PCB can be repurposed for another project.
- Speakers: For the speakers, I chose the Dayton Audio PC68-4 2-1/2" full range cone drivers, for their reasonable size, price, good reviews, ability to handle 20 watts RMS, and low impedance of 4 Ohms. However, any standard drivers would work fine with this circuit, assuming the power handling is adequate to avoid damage.
Above is the final schematic for the project and circuit, designed in KiCAD. Much of the design for each individual section was either based on reference designs in the datasheet for that specific part, an application note for that part, or the schematic of the evaluation module for that part. Texas Instrument's portable audio amplifier reference design, TIDA-01182, was also helpful in understanding how the various parts should have been hooked together.
Above is the final PCB design, as well as the 3D render. For clarity, the ground fill has been hidden so the traces can be more easily seen. The PCB could be laid out better for future revisions, but in its current state, the footprints, pin numbers, and traces all appear to be correct. The PCB was laid out one section at a time, starting with the charger, then the battery protection, then the voltage booster, and finally the amplifier. Below is an image of the final, assembled PCB:
Unfortunately, this PCB is not easy to assemble for the average DIY electronics enthusiast. The battery charging and protection ICs are incredibly small and have no legs, so fixing any solder bridges if they appear is not simple. To reflow the solder paste for the surface mount component, an electric skillet was used, since I did not have access to an industrial reflow oven when I was assembling this PCB. The reason I chose an electric skillet was due to this article from Sparkfun in a series of tutorials about surface mount design and assembly.
In this repo is also a BOM file generated by Digikey that can be used to order the same parts that I used in this project. Alternatively, a BOM can be generated by KiCAD.
Testing on the battery charger was done using the USB input, due to the consistent power delivery nature of USB and lack of reliance on a bright light source if testing with a solar panel is desired. Currently, the charger is not fully functional. Initially, when power was applied to the charger, the LED indicating that power is present would light up, but not the second LED that indicates that charging is occurring. Eventually, it was determined that the problem may be due to the thermistor that provides thermal protection, because once it was removed, the charging light came on. Despite the charging LED coming on though, the IC is likely not charging the battery with any significant amount of current. To test, the battery was left overnight to charge, and was at the same low voltage (~3.3V) when it was checked again in the morning, indicating that no charging had occurred.
Above is the section in the charging circuit that is responsible for informing the BQ24210 of the temperature of the battery. The 21.5k resistor Rth forms a voltage divider with the thermistor, the value of which was determined using Equation 3 from the BQ24210 datasheet, assuming the use of a 10k thermistor. In fact, 21.5k is the value that Texas Instruments themselves recommends for this resistor. Initially, I had thought that the thermistor was either of the wrong value or was faulty, but inserting a standard 10k resistor in place of the thermistor (which should have, in theory, tricked the IC into thinking that the temperature was at 25C, a temperature that should allow for the full charging rate) caused the charging indicator LED to turn off, so there may be a deeper problem at play here that is yet to be determined.
The battery protection circuit works as intended. The amplifier and voltage booster circuit was left on overnight to discharge the battery, and eventually, the battery protection IC had done its job by disconnecting the battery from ground when the battery hit ~3.2V.
The voltage boost circuit also works as intended. From a fully charged battery at 4.12V, the voltage boost circuit produces a very steady 12.07V output.
The amplifier is also functional, and does an excellent job of driving the 4 Ohm Dayton speakers I purchased. It can drive the speakers to extremely loud levels without issue. There is however a bit of noise on the output in the form of a high pitched whine when there is no audio signal input, which may be due to switching noise from the voltage booster, or perhaps the use of a shared ground between components. Usage of a quieter amplifier, better isolation circuitry, or separated ground planes may help this for future revisions.
For output power, the amplifier was tested by feeding it a 1kHz signal on high gain, and then turning up the volume of the input signal to the point just before the speakers would begin to distort. RMS voltage and current draw were both measured using my Fluke 117 True RMS multimeter.
- Battery voltage: 4.1V
- RMS voltage at speaker: 5.55V RMS
- Power delivered to 4 ohm speaker: V^2 / R = (5.55)^2 / 4 = 7.7 Watts RMS
- Current consumed by voltage booster and amplifier: 3.7A
Overall, a very respectable result, and the volume I was testing at was a volume I would never normally listen at. The amplifier was capable of even higher output, but any volume higher than the volume I was testing at would cause the speakers to distort, so 7.7 watts seems like a usable maximum based on the speakers I chose, and is a value I'm quite happy with. The current draw, while rather high, is much lower at normal volumes. Even at max volume, the battery would be able to power the amplifier for a little over an hour, since it is a 1000 mAh battery.
Once the circuit has been assembled, there are about three parts the user ever needs to interact with. The first is the power input selection DPDT switch, SW_POWER, which toggles between the USB and solar power inputs. The second is the switch that turns on the amp and voltage booster, SW_AMP. This switch should normally be off when the speaker is not in use. The final element that the user can interact with is the gain on the amp. This is best implemented using standard header pins like the kind you might find on the GPIO row of a Raspberry Pi, and then using a 2x1 jumper to select which gain you wish to use. The selected gain of the amplifier only takes effect when the amplifier is powered on. The gain the user chooses is up to personal preference. High gain will give much higher volume output, but potentially at the cost of sound quality.
The charger also has two LEDs for status indication. D2 will light up when the IC detects that a voltage input is present, while D3 will light up when the IC is actively charging the battery.
While the project is in a usable state, there are a number of changes I would make for a future revision, or if I were to do this project again.
- Fix the BQ24210's charging issue, either by figuring out a proper value for Rth, or by bypassing the thermal protection altogether.
- Select ICs that are a little more friendly for the average DIY enthusiast.
- Lay out the PCB more neatly. While the current design isn't terrible, there isn't any particular overarching logic to the layout.
- Add more test points for debugging, preferably ones that are drilled out rather than pads, so that pins can be soldered in which would make it easier to attach oscilloscope probes.
- Try to eliminate some of the high frequency background noise in the amplifier, either by using a better ground design, or by choosing a quieter voltage booster.
- Design a 3D printable enclosure.