Friday, January 28, 2011

Today's Power Supplies More Than a Battery with a Knob

In the past the majority of variable power supplies used for testing were viewed as little more than a battery with knob. They were large, heavy, and had little in the way of measurement capability. The measurement capability typically consisted of an analog display or a digital display with only three digits of resolution. The power supply user often did not trust the power supplies measurement capability so the supplies voltage was set and monitored using a DMM connected across the supply's output. If they needed to measure the supplies output current a shunt with a DMM or a scope with a current probe was used.
In the last decade the test and measurement industry has really witnessed the rise of high performance supplies and application specific supplies. High performance supplies and application specific supplies on the market today come with a whole host of advanced features here a quick summary of some of them:

When it comes to power supplies smaller is better
There are two dominate power supply designs used throughout electronics world, linear and switching. In the past in testing linear supplies were dominate because of their clean output power. Switchers on the other hand had a lot of noise on their output power. The downside of linear supplies was they had a huge bulky transformer which made them large and heavy. Although linear supply designs can still be found, the test and measurement industry has really switched to switcher supplies. Why? Because switchers are smaller and advances in electronic filtering has made switcher outputs just as clean as linear outputs. For instance, Agilent's N6700 modular power supply family can deliver up to 1200 W per mainframe in a 1U full rack size (see figure below) with outstanding output noise specs.

Throw away the current shunt and current probe
Today's high performance supplies have built-in high accuracy measurement digitizers in them for capturing voltage and current in parallel with measurement resolutions up to 18 bit. The digitized points can be integrated together for increased accuracy or the points can be used to capture sharp transients which are common occurrence in today's high speed digital electronics. This beats the current measurement methods of the past. For instance you no longer have to deal with the complexity of setting up a measurement shunt with a DMM or external digitizer and you do not have deal with unwanted series resistance that results from shunts. Using a current probe you do not have to worry about adding series resistance to the circuit, but current probes are not very accurate (typically >1% error) and they typically can't measure current below 10 mA. Below is a screen shot from Agilent's N6705B showing digitized current pulses on its scope like display.

Modern power supplies are really making waves
Modern power supplies have high speed output control loops for dealing with sudden output transients. Supply designers have learned how to manipulate these fast output control loops to create high power voltage and current waveform capabilities in high performance supplies. This has given test engineers the ability to simulate engine crank profiles, to simulate a handheld battery powered device being dropped, or to simulate power line noise on a DC level with just a power supply. 

You brake it, you buy it
Prototype cellular base stations and satellite modules are very expensive. If something goes wrong with one of these expensive devices during the test process and they suck too much current from the supply it can lead to some costly damage and design delays. In the past test engineers had to build in expensive protection circuitry between the supply and the device being tested. Today's high performance supplies have made test engineer's job easier by adding a long list of safety features. As an example, Agilent's N6700 modular power supply family has a built-in watchdog timer. The the watchdog timer can be activated and set by the user to start timing after each command received by the system software. If the timer value that the user sets runs down before the next command from the system controller the power supply will shut off its outputs. This protects the device being tested in the event that the system controller or software freezes up or crashes.

There is a power supply for that app
Some of todays advanced applications have unique power requirements. To meet these unique power challenges test and measurement vendors have developed advanced application specific supplies. An example is Agilent's E4360A Solar Array Simulator. This supply is designed to be a high speed and high power current source. Its output I-V characteristics simulate the I-V output curve of a solar panel or an array of solar panels.This type of supply is used to test satellite power systems and terrestrial solar panel max power point tracking devices (for more info check out this post Another example is Agilent's N6781A and N6782A. These advanced supplies were designed for battery drain analysis of handheld devices and low power optimization of handheld devices and the components that go into them. These supplies have many advanced features but the most impressive is there seamless current ranging capability. Meaning they can measure transition from one current range to another without any discontinuities in the output power. This means they can capture a sudden burst of current that goes from uAmps to Amps with 18 bits of measurement resolution throughout the whole pulse (for more info check out this post

The supply features and the application specific supplies I just covered are just examples of the wide range of features that can be found in modern power supplies. Recently power optimization has become a major factor in the design process in the electronics industry. This emphasis on power optimization will continue to drive more advanced power supply features from the test and measurement industry well into the future. One thing is for sure the high performance power supplies of today and tomorrow are really revamping the power supplies image from a battery with a knob to a sophisticated piece of instrumentation.

Thursday, January 20, 2011

Photovoltaic Capacitance and Time Domain Measurements

I noticed that my PV test blog posts have been getting a lot of hits so I wanted to capture an area of PV test that I have not touched on yet and is out of the normal GPETE realm. That area is photovoltaic capacitance and time domain testing. This type of testing provides deep dive into the physics of PV cell for improving improving the cell's efficiency at the material fabrication level..
We are all familiar with I-V measurements made throughout the PV product design cycle. With I-V measurements we can calculate PV parameters like Isc, Voc, FF, Pmax, etc.Capacitance measurements and time domain measurements are required to completely characterize solar cells. Because traps in the bulk material directly affect carrier recombination at the interface and in the bulk, it is essential to characterize these traps so as to minimize their impact on solar cell performance. Capacitance measurements are the main method to evaluate traps in the bulk. Understanding trap behavior is also important when studying multi-junction PV cells and for controlling the PV cell band gap. To optimize PV cell performance it is also important to know the carrier diffusion length, because it is one of the key parameters impacting PV cell efficiency. Time domain measurement is the principal method used to measure carrier diffusion length. The following figure lists PV parameters that can be obtained from capacitance and time domain measurements.
In the next couple paragraphs I will briefly explain the parameters in capacitance and time domain measurements. I will warn you the math gets a little heavy so if you are just interested in a solution for making these measurements refer to the link at the end of the post.

CV measurements, which are the most common capacitance measurements, can be used to estimate the carrier density (Nc) using the following equation.
Here q is the electron charge, Ks is the semiconductor dielectric constant, ε0 is the permittivity of free space, A is the surface area of a PV cell and Vbi is the built-in potential. A 1/C2 - V plot is called a Mott-Schottky plot, and the Nc distribution over the depletion width (W) is obtained from the slope of Mott-Schottky plot as shown below (click to enlarge).
An AC voltage capacitance measurement (CVac) provides the information about the defect density (Nd). This technique is known as drive-level capacitance profiling (DLCP), and it is used to determine deep defect densities by studying the non-linear response of the capacitor
as a function of the peak-to-peak voltage dV (=Vpp) of the applied oscillating signal. The density that 
can be obtained using DLCP is also called the drive level density (Ndl), and it is defined as shown below. (Note: In the previous equation the subscripted symbols C1, C2, etc. have the units of capacitance per volt, capacitance per volt squared, etc.) 
A capacitance versus frequency (Cf) measurement is helpful to understand the dynamic behavior of PV cells as well. The results of a Cf measurement are often plotted as complex numbers in the impedance plane where this information is known by many names, such as Nyquist plots, Cole-Cole plots, complex impedance plots, etc. 
A variety of time domain measurement methods are being developed to evaluate the recombination parameters of solar cells, such as minority carrier lifetime (τ), surface recombination velocity (S) and minority carrier diffusion length (Ld). One of the most popular techniques is open circuit voltage decay (OCVD) where the excitation is supplied either electrically or optically (see figure below). In the electrical case a constant current equal to Isc is forced into the solar cell and the voltage decay across the solar cell is observed after abruptly terminating the current. In the optical case a light pulse is used to stimulate the solar cell instead of a current. For the short circuit condition the current flow across the solar cell is measured after removing the light stimulus, and this is called the short circuit current decay (SCCD). 

Each of the various measurement parameters just discussed could be measured with a complex setup of multiple GPETE products and some software to post process the results. A much better way to do it is to use  Agilent's B1500A semiconductor device analyzer. The B1500A provides a one box solution with software built-in to make high accuracy and high precision capacitance, time domain, and I-V measurements for PV material testing.

To learn more about the B1500A click here

Monday, January 17, 2011

Connecting and Controlling an LXI Instrument Using Its Web Interface

From speaking with LXI instrument owners I have noticed that a large majority of them do not even realize that their instruments have a built in web interface hosted by a web server inside the instrument. Often these web interfaces provide a way to control the instrument remotely with only a web browser. In this blog post video I show you how to connect to and access an LXI instrument's web interface. 

Tuesday, January 11, 2011

A Low Cost Way to Capture Agile PRFs and PWs in Radar Test

In military radar applications the pulse repetition frequency (PRF) and sometimes pulse width (PW) parameters of a radar's pulsed RF/microwave signal may vary during operation. There is typically two reasons why PRF and PW may be varied:

  1. PRF and PW are tied to a measured target's range and resolution as well as the max range of the radar. A sophisticated radar system that can change modes from searching to tracking or from one search range to another must change its PRF and PW to match the radar's current mode. This is often referred to as mode changes.
  2. Military radars often employ PRF and PW that are constantly changing to help prevent an enemy from "spooking" the radar (creating a false target). These radar's employ complex highly secret algorithms for constantly varying the PRF and PW. This is often referred to as an agile pulse.
In this post I wanted to cover cover a fairly low cost solution for continuously capturing these parameters for verifying an agile PRF/PW algorithm, analyzing pulse noise from one mode to the next, or verifying tolerances from one pulse signal change to another. The solution is made up of a simple RF power detector and one or two gap-free sampling universal counters like the 53230A. The power detector is used to strip off the RF carrier and just output the pulse signal. The gap-free sampling counter or counters are then used to make the needed continuous timing measurements on the pulse signal. Why gap-free sampling counter?
  1. Counters provide high accuracy and high resolution timing measurements.
  2. Gap-free means no pulses will be skipped so you can get a complete and continuous picture of a group of pulses.
  3. Since counters are just making timing measurements with the edge event of a signal, you can capture a large amount of continuous pulse data with less memory and simpler post processing compared to an instrument that is digitizing the entire pulse signal. For instance the 53230A can store up to 1 million readings. There is not many scopes out there that can store 1 million digitized pulses in memory.
  4. Finally counters are low cost compared to RF / high speed digitizing instruments
As I mentioned above you may want to use one or two counters. You only need one if you are just interested in capturing one of the pulse parameters at once. The one counter based solution can be seen below.
The counter in the above figure is set for timestamped measurements which returns the time from positive signal edge to positive signal edge or negative signal edge to negative signal edge. Using the counter for timestamped measurements in this application returns the pulse repetition interval (PRF) from pulse to pulse. The PRF can them be obtained by inverting each reading. Below is an example measurement of an agile PRF signal using the setup above.
The below setup uses two counters to capture both PRF and PW simultaneously. One counter is used in timestamp mode for capturing the PRI and the other is making PW measurements. 
In the above setup, after the pulse parameter measurements are made on the desired set of pulses some simple software can be used to post process and combine the timing measurement data to give you a PRF, PRI, PW, and the duty cycle of each pulse in the set. 
Depending on the gap-free sampling counter you use, it may not measure the first couple of leading pulses from the set. This is because the counter may need to set the right input conditioning and edge leveling before the timing measurements can start. Also if you are using two different measurement modes in the counter, such as timestamps and PW, each mode may skip a different amount of leading pulses which you will need to know when using two counters in parallel to properly align your measurement. Finally, you want to be sure you use a high quality RF power detector for this type of test. Using a low quality detector can lead to carrier noise on the output pulse signal, which can really lower your measurement accuracy. 

Monday, January 3, 2011

Analyzing Close In Noise with a Gap-Free Sampling Counter

I wanted to expand a bit on the MDA capture portion of the video in case it was not clear. Like I mentioned in the video the Y axis was frequency and the X axis was time. 10 MHz is located at the center of the Y axis. The span of the Y axis was approximately 8 Hz. Please add a comment or email me if you have any questions.

Click here for more info on the 53230A Universal Counter