Archive for the ‘High Speed Buses’ category

New Advances in Oscilloscope Technology

March 25, 2010

Question: What are some new advances in oscilloscope technology?

Answer: Besides making scopes faster, wider, deeper, there are many new capabilities in scope performance and time-saving features. For example digital storage oscilloscopes have traditionally been used for physical layer testing of AC/DC parametric measurements such as rise time, voltage swing, high/low values, etc. However with the wide adoption of mainstream buses like I2C and USB many scope now have built-in decode capability. In fact the new Tektronix MSO70000 mixed signal scopes have 20 GHz bandwidth across 4 channels as well as an integrated 16 high speed digital acquisition system. This allows for isolation of faults that occur only during specific bus cycles like a read or write cycle (DDR3 memory) or for monitoring of a wideband RF system that incorporates digital subsystems and controls like SPI or I2C.

Isolating faults during Read or Write cycles on a Mixed Signal Oscilloscope

Triggering on a Read Burst

Often, while monitoring signals in the digital domain, it’s important to be able to view the same signals in the analog domain as well. Maybe there is a glitch or some crosstalk that is causing an error on the bus or other digital problem. The next step would be to get an analog probe and probe the same point (double probe). While most new high bandwidth probes have lower loading than before this is inefficient and can still add some non-negligble loading to the bus. Some recent advances in scope architecture now allow for simultaneous signal acquisition in both the digital and analog domain, with a single probe. This is possible through an internal multiplexer technology called iCapture. Below are some examples of how this is done. For an interesting read check out the article “Twelve Things You Didn’t Know a Scope Could Do”  (please be patient–the PDF might take a few moments to download).

Advertisements

Clock Recovery Settings to Generate An Open Eye

March 22, 2010

Question:  I am evaluating a system that implements SSC to reduce EMI.  However, I cannot generate an open eye or get meaningful results, yet the system is working fine.  It is a 6Gb/s system and I am using the settings of 1MHz loop BW and Type I PLL.

Answer:  These are default CDR settings and may not match your device.  Ideally, you should ask your CDR vendor what specific loop BW and PLL type they are using.  If that is not possible, you should use the minimal settings that are described in the standards document for the specific technology and speed you are implementing.  Here are some rules of thumb if neither of the options above are available:  There are numerous implementations where the loop BW is set to the “golden PLL” of bitrate/1667.  Another rule of thumb is that when deploying SSC, a 2nd order PLL is often used in the CDR system due to the more demanding tracking requirements.  Finally, try increasing the loop BW beyond the “golden PLL” to determine correlation to the specific CDR implementation.

Also (and on a completely separate topic), I’m proud of my new Guest Blog at EDN, just introduced today. Come visit Scope Guru on Signal Integrity!

Pros & Cons of Adding a Cap Filter to High-Speed Link

March 19, 2010

Question:  I added a 2pF cap to the driver side of my high-speed link to reduce HF effects.  However, I am seeing effects that I do not understand in my analysis.  What effects will this cap cause?

Answer: A cap will filter out higher-frequency content.  However, it can also slow down the rising edge of the signal reducing the eye-height.  This filtering can also look like ISI/DDDj/PDj and reduce your eye width.  An alternative technique for removing high frequency content from your analysis is to use the bandwidth limit filters in the scope to determine what the affect of rise-time filtering/adding caps is.

DesignCon 2010

February 3, 2010

Question:  What’s Tektronix up to at DesignCon 2010?

Answer:  At Tektronix, the future is 3D.  And as the world’s standard in oscilloscopes we decided to have a little fun with that message this year. 

As the exhibit hall doors opened at 12:30 yesterday, our booth was immediately inundated by hundreds of conference attendees anxious to see what’s new this year.  To create even more excitement about our new MSO70000 series oscilloscope, the industry’s first mixed-signal high performance scope, we brought in an Avatar-like character to engage curious attendees as they walked by.  Our own Na’vi handed out retro 3-D glasses and invited people to gaze into the future of DPX technology.  After getting their photo taken with her, many turned toward the nearby demo of DDR validation and debug featuring the new MSO70000.  Even Chiphead stopped by to check it out!  If you’re one of the few in the Silicon Valley who haven’t made it to our booth yet, please stop by today to “see into the future” and meet our Scope Gurus in person.

Displaying an Eye Pattern on My Oscilloscope

January 21, 2010

Question:  Can you please tell me how to use a scope to measure the eye-opening of digital signal?  And, is it true that only certain scopes have this capability?

Answer:  Any real-time scope can display an eye with varying degrees of clarity. The faster the oscilloscope is able to re-trigger, the clearer the eye.

For complete (and fully illustrated)  instructions on displaying an eye pattern on your scope, please visit this this page on our website.

For even more advanced jitter and eye analysis the DPOJET software can be used. For quick eye measurements the easiest method is to use the “One-touch wizard”.

DPOJET: Jitter and Eye-diagram Analysis Tools

What is Dynamic Range?

January 15, 2010

Question:  What is Dynamic Range, and why is it important?

Answer:  Dynamic Range is the maximum ratio of the levels of two signals simultaneously present at the input of an instrument. In other words what is the smallest and largest signal levels that can be measured at the same time without errors? Different instruments provide different dynamic range performance due to architectural advantages. Oscilloscopes typically have wide bandwidth but may have less dynamic range relative to other instrumentation like spectrum analyzers.

Spectrum analyzers look at a selected band of frequency in a high dynamic range.  These instruments are very good as resolving close-in frequencies relative to a carrier.  Example:  a sine wave oscillator with low level AM coupling may look like noise or jitter in the time-domain on an oscilloscope.  The spectrum analyzer will help resolve the low level discrete frequency components to enable diagnosis of the coupling agents.

An analog to digital converter (ADC) range is based on the resolution of the ADC. Most real-time oscilloscopes use an ADC with 8 bits of resolution and have a theoretical dynamic range of about 48 dB (6 dB per bit). In order to effectively optimize an oscilloscope for best use of the dynamic range careful control of the volts-per-division control is recommended. The volts-per-division setting (usually written as volts/div) is a scaling factor that varies the size of the waveform on the screen. If the volts/div setting is 5 volts, then each of the ten vertical divisions represents 5 volts and the entire screen can display 50 volts from bottom to top, assuming a reticule with ten major divisions. If the setting is 0.5 volts/div, the screen can display 5 volts from bottom to top, and so on.

Below are two examples of voltage measurements that can be compared with good and poor dynamic range utilization. Note the measurement differences from both setups despite measuring the same signal.

Measurements    Mean    Std Deviation

High (Good)        311mV    0.4mV

High (Poor)          299mV    4mV  

Pk-Pk (Good)      428mV    2.4mV

Pk-Pk (Poor)        528mV    12.5mV

High Dynamic Range

Low Dynamic Range

Explanation of Jitter

December 29, 2009

Question: What is jitter?

Answer: This simple and intuitive definition is provided by the SONET specification:

“Jitter is defined as the short-term variations of a digital signal’s significant instants from their ideal positions in time.”

This captures the essence of jitter, but some of the individual terms (short-term, significant instants, ideal positions) need to be more specific before this definition can be unambiguously used. In all real applications, jitter has a random component, so it must be specified using statistical terms. Metrics such as mean value and standard deviation, and qualifiers such as confidence interval, must be used to establish meaningful, repeatable measurements.

Additional information about jitter, including application notes and white papers, can be found at www.tektronix.com/jitter.