Wednesday, April 27, 2011

Optimizing PXI Modular Test System Throughput Webcast

A super smart colleague of mine just did a great modular test webcast entitled "Optimizing PXI Modular Functional Test System Throughput." Below you will find an overview of the webcast and at the bottom a link to the webcast recording. 

Webcast Overview:
Reducing overall test time in high volume electronic manufacturing sites is critical to meeting manufacturing capacity and cost targets. Optimizing throughput, while not sacrificing test coverage or production yield, is essential to retaining a competitive edge.

Best-in-class manufacturers use statistical methods during test system development to achieve high throughput without compromising test coverage or test yields. During test system development, statistically based Gauge Reliability and Repeatability studies are used to verify the stability of the test methods and test system. After establishing a reliable and repeatable test, methods to optimize throughput can be explored. The statistical methods can be used again to verify test throughput optimization will not impact test yields.

This Webcast will explore these statistical methods used during electronic functional test development and deployment to help the test engineer achieve accurate, reliable results.

Who should view this webcast:
  • Test engineers responsible for automated test
  • Aerospace and Defense
  • Electronic test
  • Communications & Semi-conductor testing
Webcast Link (Hosted by EE Times):

Friday, April 22, 2011

Overview of Agilent's New Modular PXI and AXIe Instruments

This video is a good overview of Agilent's new PXI and AXIe product line. The only thing that is missing is Agilent new high performance AXIe arb the M8190A. Click here to read about the M8190A

Monday, April 18, 2011

53131A, 53132A, and 53181A Universal Counter Discontinuance Announcement

Today (April 18th 2011) Agilent is announcing the discontinuance of the popular 53131A, 53132A, and 53181A universal counters. A custom component used in this counter family is no longer available and is forcing this unplanned product discontinuance. The last order date for this counter family is November 1, 2011, subject to availability.

Compatible replacements exist. The new 53200 Series of Counters are functional equivalents to the previous generation counter family with vastly improved performance, features, and usability. They include “53100 Emulation Mode” enabling 53100 Series Standard Commands for Programmable Instruments (SCPI).

The 53200 Series Counters provide more speed, resolution and accuracy at the same price or better. Features of the 53200 series include:
  • 350 MHz standard inputs, with timing resolution down to 20 ps (up to 12 digits of resolution on 1 second gate time)
  • Up to 75,000 frequency measurements per second 
  • Large color graphical display for data logging, trending, histograms and more 
  •  LXI/LAN, USB and GPIB connectivity
  • Gap-free measurements for modulation domain analysis and true Allan Variance / Deviation calculations (53230A only)
  • Optional pulsed RF measurement capability (53230A only)
You can also read the Agilent RF & Universal Frequency Counter/Timers Programming Comparison Guide for a list of those areas where 531xxA Series users might find differences in operation when using a 53200 Series counter. These differences are few, but documented in order to make it easier to verify programs.

When 531xxA Series compatibility mode is selected, all programming is performed through one of the 53200 Series’ remote interface (LAN, USB, GPIB). The counter display responds according to the remote commands received. Pressing any front panel key while in 531xxA Series compatibility mode returns the counter to 53200 Series mode as prompted.

Tuesday, April 12, 2011

Understanding Resistance Measurement Errors

This post covers resistance measurement error external to the DMM and how to prevent it. Quality bench-top DMMs offer two methods for measuring resistance: 2-wire and 4-wire ohms. For both methods, the test current flows from the input HI terminal and then through the resistance being measured and finally into the LO terminal. For 2-wire ohms, the voltage drop across the resistor being measured is sensed internal to the DMM. Therefore, test lead resistance is also measured. For 4-wire ohms, separate “sense” connections are required. Since no current (or very little) flows in the sense leads, the resistance in these leads does not give a measurement error. Errors for dc voltage measurements (see Agilent Application Note 1389-1) also apply to resistance measurements. However, there are additional error sources that are unique to resistance measurements.

Power Dissipation Effects — When measuring resistors designed for temperature measurements, such as thermistors and RTDs, be aware that the DMM’s stimulus current will dissipate some power in the DUT. This dissipated power will raise the temperature of the DUT and therefore change its resistance such that its resistance value no longer represents the ambient temperature of the environment it is in. If power dissipation is a problem, select the DMM’s next higher measurement range to reduce the amount of stimulus current used to make the measurement and therefore reduce the temperature error to an acceptable level. The figure below shows the stimulus current for Agilent’s 34401A at each resistance measurement range.

Settling Time Effects — Capacitive and inductive elements are a part of every measurement setup. The amount of setting time that is needed to make an accurate measurement depends on the values of the reactive elements and the time constants associated with their current paths. Quality DMMs like Agilent’s 34401A, have the ability to insert automatic measurement settling delays. These delays are adequate for resistance measurements with less than 200 pF of combined cable and device capacitance, which is particularly important when measuring resistances above 100 k. Settling time errors are most pronounced after a connection change, such as a switch closure in a test system, or after a range change in the DMM. In these types of scenarios allow a settle time that matches your accuracy needs. To check for setting time errors, make resistance measurements (on a known resistance) immediately after and at set time intervals after a range or connection change. Repeat this a number times and calculate the average and standard deviation for each set of measurements in a particular time interval. Once the average and standard deviation measurements of each time interval begin to match you can determine what your minimum setting time is.

High-Resistance Measurement Errors — When you measure large resistances, significant errors can occur due to insulation resistance and surface cleanliness. You should take the necessary precautions to maintain a "clean" high-resistance system. Test leads and fixtures are susceptible to leakage due to moisture absorption in insulating materials and "dirty" surface films. Nylon and PVC are relatively poor insulators (10 ^9 ohms) compared to PTFE Teflon insulators (10^13 ohms). Leakage from nylon or PVC insulators can easily contribute a 0.1% error when measuring a 1 M resistance in humid conditions.

Tuesday, April 5, 2011

Ground Loops and Other Spurious Coupling Mechanisms and How to Prevent Them Part 2

This is a two part blog post is based off a great tutorial I read titled “Tutorial on Ground Loops” by P. M. Bellan. A link to the tutorial can be found at the end of the post. The subject of the two part post is something you won’t find in textbooks from your EE courses, spurious coupling mechanisms in circuits and test systems and how to prevent them. A spurious coupling mechanism is any outside unwanted voltage or current source entering a signal path. There are five main spurious coupling mechanisms: direct conduction, capacitive coupling, inductive coupling, radiated electromagnetic field pickup, and ground loops. The part one post focused on the first four types of spurious coupling (click here for part one). In this post we will focus strictly on ground loops.

Ground Loop Theory
Kirchoff's law tells us the total DC voltage in a complete circuit (Vloop) is zero. When considering AC components in a circuit, Kirchoff's law is superseded by Faraday's law. With Faraday's law Vloop = - dF/dt. F is the magnetic flux in the loop. The magnetic flux is given by F = BA. B is the magnetic field linking in the loop and the area of the loop is A. Faraday's law is what makes the basis for Transformers, inductors, and also ground loops. The make up of a ground loop signal depends on linked spurious time-dependent magnetic flux produced by nearby external circuits (fluorescent lighting, switching power supply, motor, computer) as well as the resistance and inductance in the loop.

Ground Loop Examples
The figure below shows and example of a measurement setup where a ground loop is created. The ground loop is formed by the the measurement instrument being connected to ground, the sensor or device under test (DUT) is connected to ground, and both instrument low sides are connected together via the shielding in the coax cable.This creates a loop (shown in green) for spurious time-dependent magnetic flux to enter (shown in blue) and creates unwanted current flow (shown with green arrows).
Used by permission from P. M. Bellan
Eliminating the ground loop in the above example is just a matter of locating and eliminating the inadvertent short (assuming you know it exists). What if we are making a measurement on a DUT that is suppose to be grounded? In this case we can look to our measurement or test instrument to brake the ground loop. For instance most high quality DMMs and power supplies have their low side connector isolated from ground with the option to short it to a ground potential. In this case you would want to ensure the low side connection of the instrument remains isolated from ground to prevent a ground loop from being created.Of course the isolation does have some finite impedance for DC and AC that you may want to check in the instrument's specifications when performing sensitive measurements. Some test and measurement instruments, such as scopes, typically have the low side connector tied to ground. In this case you would need a differential circuit (differential probe for scopes), isolation transformer, or optical isolation to break the ground loop. These types of ground isolation methods introduce their on limitations on the measurement.
The next example involves a multi-channel measurement setup and each channel shares a common low side connection. This setup may be a instrument, such as a digitizer, with multiple channels or a complex test system with switching. This time a loop is created between the low sides of two measurement channels as shown in the figure below. 
Used by permission from P. M. Bellan
One common way ground loops of this type are broken is by disconnecting the shielding or low side return from either the instrument or the DUT. This method will brake the ground loop but it will also degrade the signal since it brakes the signal path return. 
The third example ground loop involves two test and measurement instruments connected together. As shown in the figure below, each instrument is connected to ground and the loop is completed by each instrument's low side being connected together.
Used by permission from P. M. Bellan

This type of ground loop can be broken if one of the instrument's low side is isolated from ground.

Methods for Eliminating and Reducing Ground Loop Error
  • Breaking the loop: As mentioned throughout the post, there are multiple ways to break the loop to eliminate ground loop error including remove an unintentional short, use an instrument with an isolated low side, use a differential circuit / probe, use an isolation transformer, or use optical isolators depending on your application.
  • Increasing inductance of loop path to decrease loop current: By adding inductance, and therefore impedance, into the ground loop you can lower the ground loop current without affecting the desired signal. This can be accomplished by wrapping cabling, whether it be twisted pair or coax cable, around shield beads, a ferrite core, or an iron-core transformer. The more turns the more inductance is added to the loop. For a real world example of this look at the video cord on your computer monitor it will most likely have an iron-core around it. This method is more effective for higher frequency or sudden transient ground loop signals. 
  • Eliminating the area of the ground loop: Make the area of the loop smaller by using shorter signal cabling, shorter power cables, twisting power cables together, and using AC power outlets that are adjacent to each other
  • Removing linked magnetic flux: If possible remove noisy devices out of the test measurement area. This could mean moving a test away from a motor or replacing a switching supply with a linear supply.

Click here to check out the "Tutorial on Ground Loops"