Archive for the ‘Uncategorized’ category

Who will win the cable race?

November 15, 2010

My job offers me a wonderful opportunity to test a variety of leading-edge products and technologies. In the past month, I have had the privilege of testing three different categories of cables.

In the copper cable world, it has become routine to perform cable impedance characterization. As signals get even faster, s-parameters like insertion loss and return loss are also called out. At even higher rates, near-end and far-end cross talk s-parameters are required to characterize their transfer functions even for short, 1m cables.

To read the rest of this article, please visit my Scope Guru on Signal Integrity Blog, on EDN’s site

Proper Practice for Preventing Probing Pitfalls

May 13, 2010

When visiting engineering labs world-wide, I am often asked to characterize or debug a system. Before I do any of that, I always take a very close look at the connection to the device under test. In about 25% of my characterization sessions, the gotcha I run into is improper connection to the device under test. Here are some of the most common, along with some tips for avoiding them.

To read the rest of this article please visit my Scope Guru on Signal Integrity Blog, on EDN’s site.

Pulse Aberration Verification

May 11, 2010

Question: I wrote you concerning the use of automatic measurements a couple of months ago and you were able to answer my question quite handedly.  I have another question, this time concerning the verification of pulse aberrations on fast Risetime pulse sources.

We are calibrating a Tunnel Diode Pulser that has a Risetime of 125 ps and pulse aberrations of <1%.  In the past we have used an Tektronix 11801 Sampling Oscilloscope with an SD-26 Sampling Head.   The Tunnel Diode Pulser is feed through a D-11 Delay Line so the Risetime of the initial pulse can be viewed.  The Delay Line will contribute to a decrease in the Risetime.  Can the pulse aberrations be accurately quantified and certified to a known value using the Tektronix 11801 Sampling Oscilloscope with SD-26 Sampling Head and D-11 Delay Line?  Would it be better to feed the Tunnel Diode Pulse directly to the input of the Sampling Head?

Answer:  The Tunnel Diode Pulser specifications described in your question match those of the Tektronix 067-0681-01, most recently offered by Tegam with an aberration specification of <1 % (typical) in a 1 GHz (350 ps rise time) system.  TD Pulsers were originally developed at least 40 years ago, before there were NIST traceable aberration standards.  A TD Pulser was assumed to have by design an inherently flat pulse response, suitable as a reference waveform for fairly low bandwidth oscilloscopes and pulse generators with relatively slow rise times such as the Tektronix PG506 (<1 ns rise time). 

While modern sampling oscilloscopes have better response characteristics than the models of 40 years ago, the aberration specifications, at least for Tektronix models, are typical (not guaranteed).  The aberration specifications may be higher than 1 %, depending on the defined region (time interval after the transition) for the aberration specification.   As a result, unless a sampling oscilloscope response is specially calibrated, the oscilloscope is not suitable as a traceable aberration standard.  For example, SD-26 aberration specifications are typically +/-3 % from 300 ps to 5 ns after the step, and +/-1 % from 5 ns to 100 ns after the step.  These specifications, even without a delay line, are not adequate for verifying a 1 % aberration specification.

Traceability for aberrations and settling parameters is now possible.  NIST offers the service, “Repetitive Pulse Waveform Measurements, Including Settling Parameters (65250S)”.  Learn more here.  

Returning to the question about the delay line, the effect of the DL-11 may be measured by comparing waveforms with and without the DL-11, using a system that doesn’t need a trigger pick-off.  One such system is a sampling mainframe and TDR module such as a Tektronix 11801 and SD-24, or a Tektronix DSA8200 and 80E04.   To help answer this question, measurements were made on a DL-11 using a DSA8200, 80E04, and SMA cables with combined length of about 1 m.  The DSA8200 was set to display a Math waveform having a 350 ps rise time filter, for a 1 GHz system bandwidth.  A horizontal setting of  2 ns/div was used, since that is the defined setting when a TD Pulser is used in PG506 calibration.  The DL-11 Trigger output was terminated in 50 ohm.   The DL-11 showed about 1.2 % overshoot at 4 ns after the edge, when compared with the reference waveform.  This is likely due to compensation for delay cable loss in the DL-11.  With different horizontal settings, the aberrations would measure differently relative to the level at the right edge of the screen, since it takes at least 500 ns for the DL-11 pulse response to settle to a constant level.

In summary, the DL-11 does affect the pulse response enough to be significant for a 1 % aberration measurement.  It is possible to measure and account for the DL-11 pulse response under defined conditions.  However the 11801/SD-26 sampling system pulse response may also have significant aberrations, according to the sampling system specifications.  The sampling system specifications are not guaranteed or traceable without a special calibration.  It is necessary to have a traceable aberration and settling parameter standard if traceable aberration measurements are desired.

Why Are Most Scopes Grounded?

April 23, 2010

Gina BoniniQuestion: I am curious why A/C powered oscilloscopes do not have the ground lead isolated from the A/C chassis ground. Digital multimeters that plug in the wall do. Aside from hand held or portable scopes, shouldn’t that be the norm now by now, or are they, and my scope is just old??

Comments?

Answer: Interesting observation.  There’s actually a few different reasons:

  1. Digital Multimeters (DMMs) don’t draw a lot of power, so it’s pretty easy to power them with an AC/DC power supply that has low-power isolation transformers.  These transformer isolate the product from ground. The power requirements of many scopes are beyond what can be served by this design. That said, Tektronix does offer a unique oscilloscope with isolated channels, the Tektronix TPS2000 series.  This basic oscilloscope is specially designed with isolation transformers.
  2. DMMs and their leads are carefully designed so it’s very hard for you to touch any metal on them, so it’s hard to get shocked if they’re not grounded. In general, that’s not at all true of oscilloscopes. There are a few exceptions, like the Tektronix TPS2000 series and the Fluke handheld scopes. In these cases, the individual inputs are floating.
  3. DMMs are only capable of measuring low frequencies, so the stray capacitance between the chassis and earth ground doesn’t affect the measurements much. In addition to the safety issue (in #2 above), the large capacitance of a floating scope would serious affect the quality of most oscilloscope measurements.
  4. The dielectric in the transformers in most power transformers are not designed to have large voltages applied across them. Over time, high voltage differences across the transformer’s isolation barriers will cause them to fail, creating a reliability and a safety issue.

 For all of these practical and safety reasons, most oscilloscopes will continue to be grounded.

Analyzing Frequency Content

April 8, 2010

Question: When I look into a analog signal in frequency domain (FFT), I see harmonics with certain dB. What I also see (which I shouldn’t) is some noise (!!) in between the harmonics with certain dB. What is this noise ?

Answer: Thank you for the question on analyzing frequency content with an oscilloscope. Depending upon your signal, instrument setup and other factors, there are a number of possibilities as to why you are seeing noise between the signal. As a general rule of thumb, to minimize noise coupling into the scope I first try to ensure proper grounding with short lead lengths. Depending on the signal type, you can also perform filtering or post-processing through some averaging techniques to reduce noise. Again, without knowing anything about the signal or your setup, these are some general considerations that may or may not be relevant.

I would, however, recommend a very nice application note on using Tektronix oscilloscopes for FFT analysis. You can download the application note here.

I hope this helps and thanks again for the question.

Oscilloscopes and Audio Testing, Plus How to Win a Scope!

March 16, 2010

Question: What is needed on the scope to show that a tube or valve preamp stage is used in a pedal direct into the sound board:

The tube sound is often subjectively described by uncritical listeners (that is, not audio professionals) as having a “warmth” and “richness”, but the source of this is by no means agreed on. It may be due to the non-linear clipping that occurs with tube amps, or due to the higher levels of second-order harmonic distortion, common in single-ended designs resulting from the characteristics of the tube interacting with the inductance of the output transformer.

http://en.wikipedia.org/wiki/Tube_sound

I wish to show that a digital audio signal run through a tube post gain stage produces this “warmth”.   I can detail the exact 2nd harmonic characteristics of this sound at least in part mathematically.    ‘With that information, how would I go about illustrating the difference as scope traces?  The tube post gain vs straight digital.

Question 2:  What do I need to do to win a scope?  Used, cosmetic defect, I  don’t care, I’ve used them since 2nd year engineering school and lately have been wanting to work with them again.    

3rd and last question:  In the procedure taking a line level stereo audio signal and modulating it for FM transmission, how can a scope be used to optimize design for the highest quality audio reproduction possible?

Answer: Thanks for the excellent questions. I’ve divided the answer into three parts.

1. Vacuum Tube versus Solid State Sound

The discussion here will be about consumer audio products, but there are many similarities to professional audio products, too.

As a class, not all vacuum tube amplifiers sound the same.  And as a class, not all solid state amplifiers sound the same.  However, it has been generally found that there are differences in ‘tube’ sound versus ‘solid state’ sound.  There are differences in amplifier topologies; and differences in interfaces between preamp, power amp, and loudspeakers (i.e. load impedance and resonances).

 The following amplifier characteristics are thought to be some of the causes of audible differences:

Clipping characteristics, the proportions of low order harmonic distortion products and high order harmonic distortion products, frequency response, and damping factor (output impedance).

Clipping: The effects of clipping can be eliminated, as long as the amplifier, loudspeaker, and playback sound levels are such that the amplifier never clips.  Any amplifier that is driven into hard clipping will cause most music to sound terrible (although the intent of some electric guitar playing intentionally makes use of this highly distorted sound). 

If the intent is to play music very loudly at very low distortion, then either a very efficient loudspeaker is required, or a very powerful amplifier is required. 

For clipping testing, a function generator, scope, and load resistor are required. The shape of the clipped signal will vary from amp to amp, and according to the amount of overdrive to the amp.

Harmonic Distortion at moderate levels: For most amplifiers that are played at levels well below clipping, the distortion is very low. 

For harmonic testing, a low distortion function generator, scope that has an FFT function, and a load resistor are required.  The distortion may be low enough so that a scope cannot see any harmonics other than the 2nd and 3rd (higher order harmonics may not be visible).

In order to see low amplitude high order harmonics of amplifiers, a spectrum analyzer may be required.

Check the level of the harmonics, the number of harmonics, and the rate that their amplitude is reduced as the harmonic number increases.

Frequency Response: Most music fundamentals and harmonics are within the frequency response of the amplifier, but the amplifier may attenuate the signal at the frequency extremes.

For frequency response testing, a function generator, scope, and load resistor are required. 

The gain of the amplifier should be tested over its rated frequency range, both at low power, and at rated power.

Damping Factor: Different damping factors react with loudspeakers to cause different frequency responses, and different transient responses.

A function generator, scope, and multiple load resistors are required to test damping factor.

The output voltage is tested, as the load resistance is changed.  The damping factor (output impedance) is calculated from the output voltage changes, versus the load resistance changes.

2. Winning a Free Scope

Winning a free scope can be difficult but the more entries you have the better the chance. Tektronix participates in many events and shows throughout the year where we give out scopes as well as other prizes. I’d encourage you to check the ‘events’ page on the Tektronix site regularly for update-to-date details (www.tektronix.com/events) . In fact we will be giving away 5 Tektronix MSO2024 Digital Oscilloscopes during Embedded Systems Conference (ESC) Silicon Valley next month. And, we give away a scope each month at our Scope Central online community (www.tektronix.com/scopecentral). You can enter daily–thus increasing your odds of winning!

3. FM Stereo Transmission Quality

The quality of sound from an FM transmission depends on many factors.

The music source material (i.e. CD), playback source (i.e. CD player), sound boards, compressor (presence or lack thereof), and FM Transmitter are all responsible for the quality of sound.  Also, the correct setting of signal amplitudes at each point along the way is paramount to ensure good sound quality.

If the desire is to retain the original dynamic range of the music as it is recorded on the CD, then there should be no compressor in the station audio chain.  Some classical and Jazz stations follow this model.

The maximum input of each stage, including the transmitter modulator input, should be rated by the manufacturer.  These ratings will be used in setting the proper levels in the audio chain. A test CD with full amplitude sine wave (0 dB) should be put in the CD player, and the output sent down the audio chain. 

Starting from the CD player output, to the first input stage, a scope should be used to measure the signal level.  Set the gain levels of the audio at each point along the path so that the signal amplitude is near to, but not at the maximum input rating of each stage along the way.  Too much signal amplitude will cause distortion (or if the signal is way to large it will clip); and too small of a signal amplitude will have a poor signal to noise ratio.

Music CDs have wide variations in their recorded level output.  And without pre-playing and testing each CD for the loudest portion of a recording, it will be an unknown. 

A station philosophy decision must be made as to whether or not to change the gain on the sound board for each CD to maximize the signal to noise ratio.  The danger of adjusting the level individually for each CD, is that a loud portion will come along and cause clipping.

If it is decided to do gain adjustments, then for low level CDs, the gain of the sound board will have to be raised.  Again, the scope can help to set the level.  Then, it will have to be set again according to the next CD recording level.

A compressor can assist in reducing the severity of incorrect level setting, but the tradeoff is a reduction of dynamic range of the music. If there is a compressor in the audio path, it will be a compromise between dynamic range, versus the level you set into the compressor.  If the level is set so that the full scale CD does not cause the compressor to compress at all (might as well not use a compressor).  If the level is set too high, the compressor will be working all the time (the dynamic range will be severely limited).  The setting between those two levels is a compromise, and the best setting is up to the taste and goal of the station.

Bandwidth and Rise Time

March 12, 2010

Question:  How is bandwidth related to rise time for oscilloscopes?

Answer:  Historically, oscilloscope frequency response tended to approximately follow the rule: Bandwidth x risetime = 0.35. This corresponds to a 1- or 2-pole filter roll-off in the frequency domain. Today, at the high end, most real-time digital oscilloscopes more closely follow this rule:

Bandwidth x rise time = 0.45.

This corresponds to a much steeper frequency roll-off above the specified bandwidth. The steeper roll-off is more desirable in digital oscilloscopes that oversample by 4x, 3x, or even less because it prevents aliasing by eliminating any signal above the Nyquist frequency (1/2 the sample rate – the minimum sample rate required for accurate signal representation).

For more information on the topic of bandwidth vs. rise time check out this white paper.