We show you how easy it is to make your very own dummy load and how to use it in some common hobbyist applications.
THE BROAD OVERVIEW
A circuit under load reacts very differently to an unloaded circuit. The best way to test the correct operation of a circuit is to test it whilst it is under the expected load range. For purely resistive circuits with minimal voltage fluctuations, this can easily be done by placing a suitable valued resistor across the voltage potential.
However, what if you have a dynamic load or a voltage supply that changes over time such as a battery? For these dynamic situations, there is no substitute for a constant current dummy load.
This isn’t, of course, the only use for a simulated constant current load. For example, in Issue 20, we used a constant current load to verify the output of our “Fire powered” mobile phone charger. Using a constant current load while testing the device enabled us to take voltage readings of the device under constant current load to calculate the power output.
You can also use a constant current load to test the capacity of battery cells and even the thermal output of transistors, regulators, and MOSFETs, for example.
HOW IT WORKS
Electronic loads come in a whole variety of sizes and styles as can be seen by our collection below. They all follow a very similar design; they all use a combination of load resistors/current shunts to detect the current, a MOSFET to control the current, and a way to pull the heat away from the MOSFET as it dissipates the power. Of course, they also have a way for the user to select the desired current.
Our constant current load design is an iteration of a very popular design that you can find all over the internet, possibly either designed or made popular by Dave Jones from the EEVblog. Our design is built using easily obtained components, and we have increased the input voltage from 5V to 9V, which will enable us to fully saturate the gate of the MOSFET for higher performance. This extra headroom allows the op-amp output to reach the 5V required to drive the MOSFET. This translates to a much higher current capability of over 3A, provided you can keep the MOSFET cool enough.
The circuit is essentially little more than an operational amplifier in the voltage comparator configuration, connected to a MOSFET and Shunt resistor. The input to the op-amp is connected to a potentiometer which provides a reference value. The op-amp will do everything it possibly can to keep both of its inputs the same voltage. Therefore, when the non-inverting pin of the op-amp (attached to the potentiometer) is greater than the voltage on the inverting pin (attached to the 1Ω Shunt resistor) the op-amp will attempt to rectify this difference by increasing the voltage at the output. This, in turn, saturates the gate of the MOSFET, which allows current to flow from the MOSFETs Drain to the source.
This current then increases the voltage seen by the inverting input. If this voltage is greater than the reference voltage from the potentiometer the op-amp will shutoff voltage to the gate of the MOSFET, which will also stop the current flowing to the shunt resistor.
This reduced current will reduce the voltage seen by the inverting input and the process repeats. This results in the high-frequency switching you see here.
Note: We can see the operational amplifier has a certain lag before it can detect the differential voltage across its input pins, resulting in signal above. This shot was taken whilst dissipating a 1A current from a 12V supply. i.e. 12W.
Whilst this isn’t a perfect system, due to inefficiencies in both the MOSFET and the operational amplifier, it is certainly a very effective and inexpensive solution that any hobbyist could make using parts they have salvaged or purchased for small change at their preferred distributor.
The Fundamental Build: Breadboard Prototype
| Parts Required: | Jaycar | Altronics | Core Electronics |
|---|---|---|---|
| 1 × LM358 Operational Amplifier | ZL3358 | Z2540 | COM-09456 |
| 1 × 10KB Linear Potentiometer^ | RP8510 | R2225 | COM-09939 |
| 1 × 50KB Linear Potentiometer^ | RP8516 | R2227 | CE05181 |
| 1 × LM7809 Voltage Regulator | ZV1509 | Z0509 | 002-512-KA7809ETU |
| 1 × 1N4004 Diode* | ZR1004 | Z0109 | COM-14884 |
| 1 × IRF540N N-Channel MOSFET | ZT2466 | Z1537 | COM-10213 |
| 3 × 10k 1/4W Resistors* | RR0596 | R7058 | PRT-14491 |
| 1 × 1Ω 10W Resistor* | RR3340 | R0411 | KIT-13053 |
| 1 × 10μF Electrolytic Capacitor | RE6070 | R5065 | CE05274 |
| 2 × 100nF Ceramic Capacitors | RC5360 | R2865 | COM-08375 |
* Quantity shown, may be sold in packs. Breadboard and prototyping hardware is also required. ^We found 16mm potentiometers more suitable for prototyping.
This circuit is fairly simple to construct. Follow the Fritzing diagram we show here, and be sure that you insert the diode and electrolytic capacitor with the correct polarity. If you intend to test the breadboard prototype with high current, use thicker and shorter connections to improve the current handling capabilities. With that said, you will still need to keep the power draw on this device to under 3W or so because the components can overheat without adequate cooling, and the breadboard itself isn't designed to handle high currents for too long without causing damage.
We found that 9mm potentiometers were difficult to use in the breadboard because they have short and thin legs. If you have 16mm potentiometers on hand, use those instead for a reliable connection. The 9mm potentiometer is fine for the final build, however.
Note: You may notice that the potentiometers are wired differently in the Fritzing diagram compared to the circuit diagram. The potentiometer will work fine in either orientation. The direction of travel is different however.
TESTING THE PROTOTYPE
To test the prototype we attached our Powertech MP3084 bench power supply to the circuit set to 12V with a 3A current limit. We then set the potentiometers on the breadboard to find the maximum power dissipation obtainable, which we measured on the Unit-T UT804 bench multimeter.
In this setup, our circuit was only able to dissipate about 15W drawing a maximum current of 1.25A at 12V. We suspect this was largely a result of all the added resistance from the breadboard connections and thin gauge wires. Moving to a PCB option was the only way to test the proper performance of the circuit.
The Main Build:
Dummy Load
| Parts Required: | Jaycar | Altronics | Core Electronics |
|---|---|---|---|
| 1 × LM358 Operational Amplifier | ZL3358 | Z2540 | COM-09456 |
| 1 × 10K Linear Potentiometer^ | RP8510 | R2225 | COM-09939 |
| 1 × 50K Linear Potentiometer^ | RP8516 | R2227 | CE05181 |
| 1 × LM7809 Voltage Regulator | ZV1509 | Z0509 | 002-512-KA7809ETU |
| 1 × 1N4004 Diode* | ZR1004 | Z0109 | COM-14884 |
| 1 × IRF540N N-Channel MOSFET | ZT2466 | Z1537 | COM-10213 |
| 3 × 10k 1/4W Resistors* | RR0596 | R7058 | PRT-14491 |
| 1 × 1Ω 10W Resistor* | RR3340 | R0411 | R0411KIT-13053 |
| 1 × 10μF Electrolytic Capacitor | RE6070 | R5065 | CE05274 |
| 2 × 100nF Ceramic Capacitors | RC5360 | R2865 | COM-08375 |
| 1 × SPST Rocker Switch | SK0984 | S3210 | POLOLU-1406 |
| 1 × Mini SPDT Toggle Switch | ST0300 | S1310 | CE05253 |
| 1 × Black Banana Socket | PS0408 | P9262 | CE05221 |
| 1 × Red Banana Socket | PS0406 | P9261 | CE05165 |
| 1 × 2.1mm DC Jack | PS0522 | P0622 | PRT-10785 |
| 1 × 5mm Red LED* | ZD1690 | Z0800 | CE05103 |
| 1 × LED Panel Mount Clip* | HP1102 | H1553 | - |
| 1 × 12V 40mm Cooling Fan | YX2503 | F0010 | - |
| 1 × Male Pin Header Strip | HM3212 | P5430 | FIT0084 |
| 1 × Female Header Strip | HM3230 | P5390 | ADA598 |
| 2 × Potentiometer Knobs To Suit | HK7730 | H6109 | ADA2046 |
| 1 × TO220 Heatshink | HH8522 | H0665 | PRT-09576 |
| OPTIONAL: | Jaycar | Altronics | Core Electronics |
|---|---|---|---|
| 1 × LM7805 Voltage Regulator (For Testing Purposes) | ZV1509 | Z0509 | 002-512-KA7809ETU |
| 1 × Black Banana Plug | PP0391 | P9282 | 003-BNNCROSSB |
| 1 × Red Banana Plug | PP0390 | P9281 | 003-BNNCROSSR |
| 1 × USB Plug | PP0790 | - | ADA1387 |
* Quantity shown, may be sold only in packs. Hook up wire, cable ties and 3mm mounting hardware is also required.
^ PCB is designed for 16mm potentiometers. 9mm potentiometers can be used but need to be wired to the PCB.
Note: Our build was created using components sourced from Jaycar. Some modification may be required to the enclosure if you purchase your components from an alternate supplier.
PCB DESIGN
The PCB was designed in the new Eagle version 9.3.0 using the standard libraries. It is a single layer design, so it can be easily milled on a PCB mill or etched using the toner transfer method. We will include the Eagle Board and schematic files on the website for anyone wanting to make their own PCB.
The PCB has been designed to house all of the components, including the potentiometers. If you are building this into the enclosure, the potentiometers and LED will not be soldered directly to the PCB, but rather need to connect via wires that can be soldered to the pads. Our build used a PCB milled using our Bantam Desktop PCB milling machine.
CONSTRUCTION
Constructing the device is fairly straightforward. Use the overlay image shown here to guide you in placing and soldering the components to the PCB. Take care though, as the lack of solder mask can make soldering some components a little tricky. For all the wire connections, excluding the potentiometers and input from the power supply under test, we used male and female pin header connectors to allow for easy placement. The LED was placed in the front panel and the wiring was extended from the PCB.
3D PRINTED ENCLOSURE
We have designed an enclosure for you to print yourself. The files are available on the resources section of our website. You could consider other enclosure alternatives if you do not have access to a 3D printer, such as a large Jiffy box, for example, providing you give it ample ventilation.
ENCLOSURE
We printed our enclosure using Flashforge white filament on a Flashforge Creator Pro at 200-micron layer height. It was printed in two pieces both being positioned flat on the build surface. Each side will take about 5 hours to print at these settings.
FRONT PANEL
We printed the front panel at 200-micron layer height using Flashforge White and Black filament. The Bi-coloured front panel was printed using a single extruder. To get the dual colour we simply embedded the lettering into the 3mm panel by a specific depth of 0.5mm. We then set the printer to print from 0mm to 2.5mm in white. This produced the base of the panel as shown here.
It was then a simple matter of printing the top portion after instructing the slicer to start printing at 2.5mm, and changing the filament to black.
The panel image shown above was the first attempt at the front panel. This suffered from warping, which you can see on the top right of the panel. In an attempt to disguise this warping we tried to use a black paint to cover the visible white protruding through the black surface. This, of course, was not effective leaving us no option but to reprint. We highly recommend you use either a raft or brim on the bottom half to reduce the warping.
FEET
We chose TPU/rubber because we were using a fan, which would constantly be running. The rubber feet should help reduce noise and vibration being transferred to the desk or bench. If you’re not so worried about noise then normal PLA should suffice. We printed the feet at 200-microns using TPU/rubber filament with a very slow speed of 25mm/s. The slow speed means it takes about 25 minutes to print a single foot.
ASSEMBLY
To assemble the project you first need to attach the LED, potentiometers, switches and banana jacks to the front panel. Then attach the DC jack and fan to the rear enclosure using 3mm bolts and nuts.
Using heavy gauge wiring (eg. 22AWG), connect the negative (black) banana jack to one side of a 2 pin female pin header. This connects to the ground pin on the PCB, which is labelled PSU_IN on the PCB. The other pin is the positive input from the load. Connect this to the positive (red) banana jack, via the toggle switch. This switch will allow us to quickly remove or apply the load from a circuit, which can be very useful to analyse overshoot as we will show later.
To connect the LED from the front panel to the PCB we made a short lead with two 2-pin header sockets. We trimmed the legs of the LED and used heatshrink for a neater appearance. Pay attention that you connect the LED to the PCB using the correct polarity.
We used rainbow cable to wire the potentiometers from the front panel to the PCB. Solder the wires directly to the PCB noting that the first pin of the fine adjustment potentiometer (10K) is not used.
Connect the negative pin on the DC jack to one side of a 2 pin header socket. The other side of this header needs to be wired to the power switch.
The positive connection of the DC jack will connect to the remaining switch contact. This switch will allow us to remove power to the device itself.
Once done connect the female pin header to the PCB in the INPUT male header making sure the polarity is correct.
Solder the fan wires to a 2 pin female pin header. This plugs into the header on the PCB.
You will notice in the photo shown here that we used heatshrink on the connections and cable ties to avoid the wiring from touching the heatsink or fan.
TESTING
Our tests will not only test the circuit, but also give you an idea on how the load can be used.
The three most common uses for our USB load are:
First things first, let’s see how the device performs and what sized loads it can dissipate.
MAXIMUM AND MINIMUM LOAD OUTPUT
To figure out the range of the device, the first thing we need to do is discover the minimum resolution of the device. That is, what is the smallest load the device can simulate.
To do this, we simply need to connect the constant current load to a power supply and dial in the smallest load value we can. For our device, this turned out to be 1.125mA, however, at these very low currents it’s difficult to dial in a specific current.
To test the maximum output of the load, we used a Powertech MP3084 Lab Bench Power Supply, set to 12V with a 3A current limit. We set the load to dissipate 250mA, which equates to a 3W load.
We monitored the temperature of the heatsink using a Digitech QM1571 Multimeter and took a recording after 10 mins. After this, we increase the current and repeat until the 3A/36W mark is reached. This test allows us to ascertain how hot the MOSFET will get while dissipating higher loads and, therefore, give us an idea of the maximum performance.
Whilst the load seems to handle dissipating 36W the temperature rise at this level of dissipation is less than optimal.
The datasheet for the IRF540N suggests the MOSFET has an operating temperature of up to 175°C, as seen here.
However, this temperature will be referring to the temperature of the internal PN Junction, which we can’t measure, and is likely to be much hotter than the external temperature we are measuring.
This will also be referring to the absolute maximum temperature before failure, not a sustained long-term temperature.
For the longevity of the device, it is best practice to keep the temperature increase to less than 80°C above ambient. If you want to dissipate higher loads more regularly consider a larger heatsink or other cooling methods.
However, this device as it stands is perfectly fine for long term use with loads under 20W.
Now we have a baseline for what the load is capable of dissipating, let’s demonstrate its use by using it to do a variety of different tasks we often find ourselves doing.
BATTERY BANK CAPACITY CHECKER
Our number one use for our electronic dummy load has been to test the capacity of USB battery banks. Whilst most battery banks have their claimed capacity written on them we have found this is very rarely in line with actual real-world use. Therefore, we often test our battery bank's capacity under real-world conditions.
To do this using our electronic load, you need to connect some banana plugs and wires to a USB male connector. This will allow you to easily attach the USB battery bank to the load. When you’re done it should look something like this:
Note: Use thicker wiring to reduce the resistance. We want the load, not the wiring to dissipate the power. If you feel your wiring getting warm under load, use thicker or better-quality wire.
We then want to measure the current and voltage. You can use a multimeter for this, but the job is made much easier with a USB tester. This measures the voltage and current at the same time, and even calculates the mAh for you. We are using a Uni-T UT658 USB tester.
Connect your fully charged powerbank and USB tester similar to how you see our setup here. You then need to dial in the desired current using the current display on the multimeter or the USB tester. The load will apply a constant drain of your chosen range to the battery bank, until the bank cutoff voltage is reached and the battery bank shuts down.
If you’re just using a USB tester all you need to do is disconnect the tester and place it into a USB socket to find the maximum capacity which they record.
However, if you’re using the much more accurate multimeter approach, you need to record the voltage every few minutes. Using the results you can graph and calculate the power and capacity of the battery. In our test, the battery bank reported a value of 1404mAh, which is close to the claimed capacity of 1500mAh.
BATTERY CAPACITY TESTER
Another use for this device is to test the capacity of your batteries. Let’s say, for example, you have a 12V lead acid battery that you want to use for a project but have no idea how of its actual capacity. This is the perfect device because no matter what the input voltage changes to, the device will still keep a constant current draw on the battery.
Therefore, if let’s say we dial in a 500mA draw on the fully charged 12V battery, the voltage of the battery would start at about 13.8V (unloaded) and be considered “flat” at about 10.9V.
Since the current is constant, all we need to do is monitor and record the voltage with respect to time. To do this, simply attach the fully charged battery to the device with a multimeter in series to measure the current. Dial in the desired current, in our case 500mA.
We then attach the load to the battery with the multimeter in parallel to measure the voltage. We then need to record the voltage reading every 15 minutes. To get an approximate idea of how long the battery should last, we can use the following formula.
Discharge Time = Ah rating / current
Since our lead-acid battery has a rating of 1.3Ah we can expect a test time of around:
1.3 / 0.5 = 2.6 hours
This means, if we record the voltage every 15 mins for two and a half hours, we should capture the full range of the battery from fully charged to fully discharged. We can display that data in an easy to read graph and, therefore, be able to easily calculate the actual capacity of the battery.
Whilst this isn’t a great graph due to there being too few data points, it does show that the battery capacity is much less than we calculated it should be:
(current × time) = (0.500 × 60 mins) = ~500mAh
This didn't make sense initially, until we remembered that it is industry standard for lead-acid batteries to have discharge and charge ratings based on a 20 hour period.
consumption = capacity / hours = 1.3 / 20 = 0.065A (65mA)
Therefore, the 1.3Ah rating only exists when the current draw is 65mA. Which is shown in the datasheet here.
Therefore, as disappointing as the test was, we were able to prove the battery will have a capacity of around 500mAh when used in conjunction with a load of 500mA. This test can easily be conducted at various levels and with it, you can reproduce and verify the claims made in datasheets. For example, after conducting this experiment I was able to ascertain that my battery was not performing to specifications. After opening the battery, I learned that one of the cells was quite low on electrolyte.
VOLTAGE REGULATOR THERMAL CHARACTERISTICS
Another potential use for this device is to test the performance of a power supply under specific situations. For example, let’s say you have a requirement for a specific voltage which requires the use of a linear voltage regulator. We can use this device to test the circuit in a real-world application.
Let’s assume you have a 12V power supply and need to step the voltage down to 5V using a voltage regulator. This regulator circuit would be housed in an enclosure. You can use this dummy load to simulate the load on the power supply circuit under different load conditions, to monitor the temperature curve in relation to load.
To demonstrate this, we will use a 7805 linear voltage regulator to step the 12V supply down to 5V, and draw an increasing load from 50mA to 400mA. We will record the temperature of the regulator after 20 minutes at each load level.
This test can be used to discover if your circuit needs additional cooling, or whether your enclosure needs modification to assist in heat dissipation etc.
This experiment allows us to analyse the temperature increase of an 7805 linear voltage regulator as we increased the current demands on the circuit. Whilst we did this experiment in the open air we have in the past used this same process to analyse the increase of temperature in enclosed spaces. This helps when designing power supplies for projects which need to be fully enclosed to protect them from the elements, etc. It’s worth mentioning that in an enclosed space the temperature is not linear. Small increases in current can create significantly higher temperatures when the heat inside the enclosure has no way to be removed.
With that said, it shows quite clearly that this brand of 7805 in the open air, without any heatsinking, becomes unstable at currents exceeding 240mA.
POWER SUPPLY TESTING
Another interesting use for an electronic load is to test and compare the performance of power supplies. We often use a load to verify a power supply is capable of maintaining the claimed specifications of a purchased unit, and to also determine the specifications for a power supply we have designed and built.
Let’s say we have a USB phone charger that is rated to provide 1A of continuous charging current. We can use our constant current load to verify that the device is capable of delivering the claimed current, and we can use an oscilloscope to monitor the voltage ripple at this load.
We can then use this to compare against other similarly rated supplies. Take, for example, the following experiment. Here we compare a USB charger from Huawei with an Apple branded charger that has similar specifications.
Both were able to provide the claimed 1A charging current and both show tolerable ripple across the load.
OVERSHOOT
Both supplies reacted nearly identically when a 1A load was quickly applied to them by using the load switch. Both overshot the 5V output very briefly with the Apple ringing for a slightly shorter duration.
SWITCHING NOISE
In this situation, the two supplies were both on par. The Apple charger slightly ahead, due to its slightly lower ripple and slightly faster-switching speed.
We were quite surprised by the results of this test. The two devices were chosen as we perceived there to be a significant difference in quality equal to the difference in cost. In reality, the two devices behaved near identically. Both were able to easily provide their rated current and neither showed any significant issues with ripple, noise or overshoot.
WHERE TO FROM HERE?
There are many ways you could improve the functionality of this electronic constant current load, including a way to display the current on the device. An analogue meter, like the MU45 from Jaycar, would be suitable. This would help avoid the need to first attach the device with a multimeter in series to dial in the desired current.
You could add an LCD readout to display the voltage and current, and calculate the power/capacity of the device under test. This could be done using the Adafruit INA219 current sensor breakout board and a generic LCD and Microcontroller combination. However, this also means the cost increases significantly.
It would also be a simple upgrade to add temperature control to the fan. This would allow the device to be silent unless it was operating under heavy loads. This could easily be done using another operational amplifier and an NTC thermistor and a transistor. Something similar to the setup below should work fine. Note the potentiometer can be used to set the threshold at which temperature the fan turns on.