Over the Christmas break I was thinking about my blogs for this year and decided as well as following the builds for my various lab equipment projects (they've unfortunately been on hold while other work took priority!), I would do a series of experimental blogs too. These will be a mix of evaluations of off the shelf boards and experiments with various design concepts. This is the first of these and I will look at an LM2596S board by a company called eBoot which I purchased from Amazon a few weeks back. It came as a pack of 6 for £8.99 and on Amazon Prime came free on next day delivery which makes the boards just £1.50 each. So for this price, how good can these boards be? The boards can be found on Amazon here: https://www.amazon.co.uk/gp/product/B01GJ0SC2C/ref=oh_aui_detailpage_o05_s00?ie=UTF8&psc=1
The specification from the Amazon listing is as follows:
- Input: DC 3 V to 40 V (input voltage must be 1.5 V higher than the output voltage, no boost)
- Output: DC 1.5 V to 35 V voltage is continuously adjustable, maximum output current is 3 A
- Features: all use SANYO solid-state capacitors, 36 u thickening circuit boards, high-Q inductance with LED indicator output
- Size: 45 * 20 * 14 mm (with potentiometer)
Things that could negatively affect the performance of a board like this are the quality and appropriateness of the passive components used and the attention to detail in the layout. At first glance the layout looks reasonable so I will move on to the components. They quote they use SANYO solid-state capacitors but looking at the markings on the parts I am not sure they actually are SANYO parts. The only parts I can find marked RVT are actually by ELNA. The data sheet can be found here: http://http://www.elna.co.jp/en/capacitor/pdf/catalog_17-18_e.pdf
The part on the board is 220uF 35V rated which for me doesn't sit well with the maximum output rating of 35V as you need some headroom above that. With the 35V rating and the output set to 35V then any ripples or spikes will take the voltage over the spec for the capacitor and over time it's likely to cause it to fail prematurely. The ESR of this part at 180mOhm is also a little high for a 3A output. Peak current is likely to be higher than this so I am expecting to see relatively large fluctuations in the output voltage at the switching frequency when the current demand is highest on the output capacitor. The ripple current spec for this part is also only 500mA which seems quite low for this application. I'd have chosen a 50V Aluminium Polymer cap for this application as it would have the appropriate voltage rating for the output, an ESR in the low 10's of mOhm and a ripple current rating of several amps. A good example would be something like this range from Nichicon: http://nichicon-us.com/english/products/pdfs/e-pcr.pdf
They specify a high-Q inductor but it's impossible to tell just from looking at the board exactly what it is. For the most efficient power supply I would try to ensure the the DC resistance (DCR) of the inductor was minimized.
I don't have an LCR meter currently but I am looking to obtain one, when I do I will probably take the parts off one of these boards and accurately measure them to see what their specs actually are. That'll be covered in a short follow up blog if/when I get the required equipment.
But for £1.50 a board, I guess I can't complain too much......
Anyway, the data sheet for the LM2596S can be found here: http://www.ti.com/lit/ds/symlink/lm2596.pdf
The specification from the data sheet is as follows:
- Input Voltage Range Up to 40 V
- 3.3-V, 5-V, 12-V, and Adjustable Output Versions
- Adjustable Version Output Voltage Range: 1.2V to 37V ± 4% Maximum Over Line and Load Conditions
- 3-A Output Load Current
- Requires Only 4 External Components
- Excellent Line and Load Regulation Specifications
- 150-kHz Fixed-Frequency Internal Oscillator
- TTL Shutdown Capability
- Low Power Standby Mode, IQ, Typically 80 μA
- High Efficiency
- Uses Readily Available Standard Inductors
- Thermal Shutdown and Current-Limit Protection
- Available in TO-220 and TO-263 Packages
If you are interested in using this part in your own designs it can be purchased from Farnell here: http://uk.farnell.com/texas-instruments/lm2596s-adj-nopb/ic-dc-dc-converter-3a/dp/1469194
Q. What do I actually want to test?
A. For this blog I want to look at the following:
- Efficiency for various input and output voltages and load currents.
- Output regulation at various constant load currents.
- Output regulation under transient load conditions.
- Temperature for various input and output voltages and load currents.
- Look briefly at the EMI performance of the module.
Q. What equipment will I need for performing these tests?
A. For the efficiency tests I will need to be able to set and monitor the voltage and current at the input and output and provide a either a constant or transient load. I'll also need to be able to view the voltage and current waveforms on the output to check the regulation under various constant loads and transient loads. For the EMI testing I am limited by not having access to a spectrum analyzer so I will have to improvise. The test equipment I plan to use for this are as follows:
- TTi Bench Power Supply
- TTi LD300 DC Electronic Load.
- Keysight DMM
- Keithley 2015 THD Bench DMM.
- Keysight Oscilloscope
- PerfectPrime Quad Thermocouple Thermometer.
- Ramsey Electronics Shielded Test Enclosure.
Q. How do I perform the Efficiency Tests?
A The percentage efficiency of the converter is simply(100 Power In Power Out so to calculate these we need to know the current and voltage on the input and the current and voltage on the output We could do this by measuring current and voltage with a single multimeter on both the input and the output for each and every test condition but that would be a slow and tedious method with lots of continual reconnecting of the meter Another way would be to have four multimeters this would enable all the measurements to be read at once for each test condition The problem there is I don't have 4 multimeters So we could say well we know what voltage and current are going in from the readout of the power supply and this is true to an extent except it doesn't take into consideration the voltage drops along the wires used to connect the power supply to the board Fortunately the has a remote sense option so I can feed back the voltage right at the input to the board and the power supply will compensate for the voltage drop in the cables I'll load it up and measure with a DMM initially to ensure all is working correctly but after that I will leave it to the power supply and take the voltage at the board to be as set on the supply In addition power supply may not give an accurate current measurement under some conditions(as I found out doing testing for another blog so when I need to know the current accurately I still need to use a good DMM In this case I will put my Keysight in the circuit to measure the current at the input At the output the DC electronic load will give an accurate current measurement at the output so I just need to measure the output voltage directly at the board For this I will use my Keithley 2015 THD bench multimeter For the purposes of this blog I shall test at the following input and output voltages and output current
- Input Voltage - 3V, 12V, 24V, 40V
- Output Voltage (Limited by 1.5V below input voltage) - 1.5V, 3.3V, 5V, 12V, 15V, 24V
- Output Current - 10mA, 30mA, 100mA, 300mA, 1A, 3A.
Q. What do I mean by output regulation?
A. There are two aspects to this, firstly as the current increases does the output stay at the selected voltage or does it droop, secondly how clean is the output, does it get noisy and have significant spikes at the switching frequency?
Q. What are constant and transient loads?
A. A constant load means set a constant current and measure the output. A transient load varies the current are differing rates. This will show how quickly the regulator can react to keep the regulation correct and whether there are any significant over or under voltage errors in regulation as a result.
Q. How do I measure the output regulation?
A. For constant current this can simply be measured using a multimeter. Set the output voltage to be correct at low/no load and then measure at increasing load currents. If the measured output voltage remains constant the device has good regulation, if it starts to sag then this indicated poor regulation. For transient loads an oscilloscope should be used to monitor the output voltage and current waveforms. The LD300 electronic load has a current monitor output which is a voltage scaled at +/-50mV per Amp. The current waveform will show how quickly the electronic load is changing the current whereas the output voltage waveform will show if there are any under or over voltage conditions while the regulator tries to adjust to the changing output load. Going from a high current load to a low current load is likely to induce a temporary over voltage condition whereas going from a low current load to a high current load is likely to induce a temporary under voltage condition.
Q. Why should I worry about the temperature of the regulator and how do the input voltage, output voltage and load current factor in?
A. It's all related to the efficiency! Any power that goes in but doesn't end up in the load must be coming out as heat! Heat in an electronic device is usually wasted energy (exceptions would be for example an ovenized oscillator or precision voltage reference).
Q. What is EMI and why do I care with this regulator?
A. EMI stands for Electro-Magnetic Interference. Whenever current flows through a conductor a magnetic field is created and the strength of the field is proportional to the current. Depending on whether it's DC, AC, switching on/off, etc will affect the electric field generated. Extending this basic principle to electronic devices, some devices will be very benign and not radiate much at any frequency whereas others will generate a lot of EMI at various frequencies. A switching regulator such as the LM2596 is a type of device that requires careful design to minimise its EM emissions. If a device radiates a high level at frequencies which can interfere with other devices around them them these other devices may start to malfunction. Because of this governments around the world mandate strict requirements for compliance with EMI regulations within certain frequency bands to ensure they do not cause any issues. EMI testing is a specialist field and way beyond the scope of this blog and as I don't have a spectrum analyzer I am quite restricted in what tests I can perform but I will try and create some meaningful examples to try and illustrate the phenomenon.
Q. Why are we doing a Q&A when you should be showing the test setup?!
A. Good question! Ok on with the details of the setup then!
The input of the LM2596S board will be from the TTi Power Supply The KeySight will be wired in series with the positive input voltage to measure the input current. The remote sense will connect directly at the input of the board such that it will compensate for any losses in the cable and also the burden voltage of the ammeter As stated previously with this compensated supply voltage the input voltage will be read directly from the set voltage on the power supply
The output of the LM2596S board will connect directly to the TTi LD300 DC Electronic Load. The Keithley 2015 THD will be connected directly to the output of the board to measure the voltage at the board output therefore eliminating any losses in the output cables. The current can be read directly from the electronic load.
The Keysight MSO-X 3054A Oscilloscope shall be used to examine the voltage and current waveforms of the LM2596S board. Channel 1 connected directly to the output of the board with a x10 probe. Channel 2 shall connect to the current monitor output of the electronic load by a BNC cable and shall have it's input impedance set to 1MOhm.
The test setup is shown below:
Note: I have drawn the multimeter on the output just connecting in the middle of the output wires. In the actual setup the multimeter is connected directly at the output of the LM2596S board to eliminate voltage drops in the cabling under high current loads.
Now on with the actual testing!
The following table shows the measured input current for each setting of input voltage, (unloaded) output voltage and output current.
There are a few things to note in this measurement data.
Firstly I could get no meaningful output at 3V input even though the spec says it should work. I believe this is down to stability and could be fixed by changing the capacitors and inductors to values and types selected specifically for the lower voltage range.
I was also unable to get the output to hold up with 24V input and 15V output at 3A load. As I increased the load current to about 2.4A it would immediately collapse. It would also collapse at a much lower output current of a little over 1A if I tried to increase the load too quickly. This is a stability issue again.
Finally you will see the 40V input column is blank. As I set up for the measurements and applied the 10mA load the converter board failed an emitted the magic smoke. I believe it came from either the inductor or the voltage setting POT but it's now totally dead. I will take the components off and test them to determine which has failed for this to happen at such a low load.
The following table shows the calculated efficiencies across the range of measurements taken above.
It can be seen from this that the board can have some very good performance with certain setup conditions. It peaks at 93.4% efficient at 24V input and 15V output with a load of 1A. I expect this was roughly where the initial design targeted the initial design parameters. The highest efficiency for a 5V output was again at 1A output but this time with a 12V input but it was much lower peaking at 84.4% which is still reasonable. The efficiency tails off some more at 3.3V output peaking at 78.7% at 1A load, with a 12V input.
Output Regulation - Constant Load
The following table shows the regulation in terms of voltage droop in volts across the range of measurements taken above.
For the most part, just looking at the voltage measured on the output on a multimeter the droop on the output under load is very slight. The worst case is 140mV at 3A load on a 5V output. What this doesn't show is how much ripple there is on the output and this will be measured on the oscilloscope once I have switched out for a new board.
I'm going to go ahead an post this at the stage as I had intended to have it finished today and as I am not sure when I will be able to set everything up to do further testing I would rather put this up for feedback and I will come back and update the sections below once I have done the additional testing.
The following oscilloscope captures show the output waveforms (AC coupled) for the highest current (worst case) output measurements from the range taken above. [AWAITING FIXED TEST SETUP]
The following oscilloscope captures show the no-load output waveform (AC coupled) for a selection of input and output voltages. [AWAITING FIXED TEST SETUP]
Output Regulation - Transient Load [AWAITING FIXED TEST SETUP]
The following oscilloscope traces show the output waveform (AC coupled) triggered on the rising and falling edge of the transitions between high current draw (3A) and low current draw (10mA) for a typical input voltage (12V) and output voltage (5V) combination that you might come across.
Temperature [AWAITING FIXED TEST SETUP]
Temperature probes were connected to the LM2596S and the output inductor and capacitor with a fourth probe measuring the ambient room temperature. The input voltage, output voltage and output current were then set to those of the worst case efficiency found above and the setup was left to "soak" until the temperature measurements settled. The following table shows the temperatures measured on each of the devices as well as the delta with the ambient temperature to give an indication of how much heat can be generated in such a circuit.
EMI Performance [AWAITING FIXED TEST SETUP]
Conclusion [INITIAL CONCLUSION, AWAITING FIXED TEST SETUP FOR FINAL VERDICT]
For £1.50 each I have to say these were worth a punt. Are they great? Absolutely not, they don't work over the full range of the specification. Failing to be stable for low voltage inputs and certain input/output combinations and emitting the magic smoke with the 40V input are disappointing but not unexpected at this price point. With some careful analysis and component selection I suspect it could be made to work a lot better than the current boards. I'll use this as the subject for a subsequent blog post!
So would I recommend these boards? Given the very low price I'd give a very cautious yes at this point if your design parameters fall well within the range of acceptable operation above and the load device isn't anything valuable. Even then I would do some exhaustive testing to determine that it wasn't going to suddenly fail at a later date in the final application. It should be noted that the failure mode was to fail with the output rising to the input voltage so if you had a high input voltage and a load which required a low voltage and it failed in the same way mine failed then your load device would get severely damaged.
Thanks for reading and as always let me know your thoughts so far in the comments!