In the last few blogs I have been spending a fair amount of time looking at various performance measurements for LDOs (see Measuring LDO Noise Spectral Density for a Novice, Measuring LDO Power Supply Rejection Ratio for a Novice, and Measuring LDO Line Regulation for a Novice). In this installment of this series I’ll explain how to collect load regulation for an LDO. It is quite similar to the topic of my last blog which was on measuring the line regulation of an LDO. Recall for the line regulation that the input voltage is varied and the output voltage is monitored. The change in output voltage divided by the change in input voltage is the line regulation. Taking the logarithm and multiplying by 20 results in the line regulation expressed in dB as the Typical Line Regulation (in dB) Plot of an LDO Under Test figure showed in my last blog. Now we will look at the change in output voltage in relation to the change in the output current.
Once again, the reason I have been spending some extended time with these measurements is due to the stringent space application requirements from customers. It is important to rigorously test a device going into space to make sure that it will perform to expectations in the harsh environment of space. This is not to say that commercial testing is not stringent, but it is merely an indication of how much more stringent testing is required for space applications. It is one thing to have some type of unexpected performance or failure in a commercial application, it is quite another to have an event in a space application. Once the device is up in space on a satellite there is no opportunity to do an onsite repair. While working on space product developments I see products like these where it is necessary to do additional testing to provide additional assurance that the device will be space worthy.
Let’s now take a look at the actual measurement and what is required. Much like the line regulation measurement, a DC power supply and a digital multimeter are required. Once again for this measurement the Keithley 2230-30-1 DC source meter and the Keysight 34461A digital multimeter are used. Any equivalent equipment may be used, but keep in mind that good source and measurement accuracy is required. As I mentioned in my previous blog the Keithley 2230-30-1 has 0.03% output voltage accuracy and 0.1% output current accuracy. The Keysight 34461A digital multimeter has 6 ½ digit output capability. These are both shown below in the following two figures.
Keithley 2230 DC Source Meter
Keysight (Agilent) 34461A Digital Multimeter
The measurement setup for this test is fairly straightforward. The Keithley 2230 DC power supply connects to the LDO input and the Keysight 34461A connects to the LDO output. The first step is to use the multimeter to accurately measure the two resistors used for the output load resistance RL. At least two different load resistors are required because the goal of the measurement is to find the change in output voltage versus the change in output load current. In order to change the output load current the output load resistance must be varied. To find the maximum change an open circuit (RL ≥ 1MΩ) is used first and then a resistance value that results in the maximum load current is used. This procedure provides the maximum and minimum output load currents. The input voltage in this case is held constant and the output voltage response is recorded for the two load currents. The output load resistance measurement is important because during the measurement the output voltage is measured on the digital multimeter and the resultant load current is calculated from output voltage divided by the load resistance. The resultant measurement provides ΔVOUT /ΔIOUT which is the load regulation for the LDO. Ideally there would be no change in the output voltage between the no load and maximum load conditions, however, every LDO will have some change in its output voltage when driven to it maximum load current.
Load Regulation Measurement Setup
Ideally one would have a switch on the output of the LDO to select between output load resistors. However, in this case I performed the measurement on an LDO evaluation board that did not have such provisions so I soldered down the 1MΩ resistor and performed the measurements and then soldered down the load resistor for maximum load current and performed the same measurements. I was able to use a modified version of the Python script from the line regulation measurement to at least automate the measurement process. It required some minor modifications to perform the load regulation measurement, but the script needed very little change since the output voltage is monitored in both case.
In this case I used the same LDO under test which has a default output voltage of 5.0V with an input voltage range up to 5.5V. The maximum load current is 400mA. For this device the maximum load current can be achieved with a 12.5Ω load resistance. In order to handle the required power as well as make the load resistance more accurate I used four 49.9Ω 1% tolerance output load resistors in parallel to achieve the required load resistance. These resistors result in a resistance of 12.475Ω. Not only did they yield the appropriate resistance but had the proper amount of power dissipation required. The 400mA of current from a 5V output results in an output power dissipation of 400mA × 5V = 2W. Using four 0.6 watt 49.9Ω axial lead resistors meets this power requirement with margin. Once the data is collected the results are tabulated in Excel where I can perform some calculations to derive the load regulation. These results are given in Table 1.
Load Regulation Measurement Results
The first four columns are self-explanatory. These are the output voltage measurements and the calculated load currents based on these output voltage measurements divided by the measured output load resistor values. The change in output voltage divided by the change in output current is then calculate din the fifth column. This is simply the change in output voltage between no load and maximum load currents divided by the corresponding change in no load and maximum load currents. The sixth column is the percent of voltage change and it is simply the difference in voltages between no load and maximum load currents divided by the voltage under the no load condition. In the last column is the calculation of interest. We now take the %V Change and divide by the difference in no load and maximum load current and multiply by 1000 to get the resulting %/mA units that we are looking for in this measurement. In this case the resulting load regulation comes in at 0.002441826%/mA which is very good load regulation.
This measurement is a little more involved than the line regulation but is still not a very complex measurement and is relatively straightforward to perform. The calculations are a little more involved but still not overly complicated. Once again automating the measurement with Python was very handy. I had to do a bit more work to change the load resistors, but the measurement is pretty easy to perform. Any time we can automate measurements it helps to speed up the process and make more efficient use of time. I hope this series of blogs has proven useful and provided some good insight into measurements for LDO performance. It was good for this novice to spend some time on these measurements after having spent so much time working with ADCs over the last several years. I admit that I thought of an LDO as a pretty simple device, but performing all these different measurements has given me a new appreciation for these regulators.