Keysight: expert to expert
Keysight: expert to expert RSS FeedRSS

Keysight logo-129 px

A collective blog from the experts in measurement and design, discussing the latest tools for circuit-level modeling through system verification for general RF/microwave, 4G and 5G communications and aerospace/defense applications. Learn about these applications and the EDA simulation software, test and measurement equipment and techniques behind state-of-the-art RF, microwave and high speed design.

Racehorses or workhorses? Bandwidth alone won’t cut it in the scope world.

May 14, 2012

The race to high bandwidth for real time oscilloscopes has now reached above 60 GHz.  Agilent Technologies recently introduced its 63 GHz 90000 Q-Series, which is nearly double its previous generation oscilloscope of 33 GHz.  The milestone was achieved in less than 18 months of design and manufacturing time; the quickest bandwidth turnaround in Agilent scope’s history.  The 63 GHz is achieved with Agilent’s RealEdge technology, an architecture that takes advantage of Agilent’s proven proprietary Indium Phosphide (InP) transistor technology and also shows Agilent’s long experience in RF design which is unique in the T&M industry.

Agilent Scope

While the bandwidth is a nice banner specification, oscilloscope users need more than just the bandwidth to get their jobs done.  Recently while visiting an oscilloscope user we ran into a number of issues that the 90000 Q-Series was ideal for solving.  The 90000 Q-Series allows for +/- 5 V input voltage into its front end, even on its high bandwidth channels.  This is different than other competitive offerings that are available in the market.  The +/-5 V, made it possible for this user to be able to debug an issue where the user’s TDR was not outputting the correct signal.  Typically the signal would be 800mV worth of input swing, but the equipment was not terminated correctly, causing it to output 2.5V of signal swing.  The TDR also featured a very fast rise time (approximately 12 ps (10/90)) and the user was able to simply input the signal into the 90000 Q-Series, hit the auto scale button, and the waveform was properly depicted. The other oscilloscope in the lab being evaluated was limited on its input range and was saturated by the 2.5 V signal, making the scope unusable for this signal.

 

Scope trace

 

Another interesting situation occurred when using the 90000 Q-Series for general purpose debug.  In this case the user wanted to recover a clock signal on a long pattern.  Unfortunately the customer didn’t know the pattern length of the signal that was being evaluated.  This is problematic for some of the clock recovery software that is available on the market.  The user was trying to guess the pattern length of the signal on a competitive oscilloscope to try to get it to recover the clock.  This was not working as the pattern length was significantly different than what the user originally thought.  Fortunately we had the 90000 Q-Series with us and using the 90000 Q-Series’ clock recovery software, we not only were able to recover the clock, but the 90000 Q-Series identified the pattern length to use on the other oscilloscope to help it recover the clock.

The 90000 Q-Series also features the world’s lowest intrinsic jitter with 75 fs, but possibly more important it maintains an extremely stable time base, which means whether you are making a jitter measurement at 1Mpt or 200 Mpts, you get the same answer.  One other oscilloscope user was extremely frustrated by what the user thought was a problem on his device with his clock caused by running long patterns through the device.  Using another oscilloscope, the user was seeing his jitter vary from a few ps of total jitter at shallow memory to over 20ps with a PRBS 15 pattern.  The issue had caused weeks of frustration for the designer as he was unable to identify the root cause in his circuit.  After explaining the situation, we had a 90000 Q-Series available and decided to measure the jitter on the device.  Interestingly enough the measurements with a PRBS7 and a PRBS15 looked nearly identical (with the exception of increased ISI, which you would expect).  The problem was not with the device, but with the competitive scope that was being use.  The time-base was not stable and would give worse answers at deep memory.  Weeks of work was now solved.

Eye Diagram

Users tend to focus on bandwidth as the key specification when evaluating an oscilloscope, but the reality is that the oscilloscope (even at high bandwidths) must be able to do general purpose debug work.  The 90000 Q-Series provides both high bandwidth and the ability to do general purpose work.