|Product Performed to Expectations:||9|
|Specifications were sufficient to design with:||10|
|Demo Software was of good quality:||10|
|Demo was easy to use:||10|
|Support materials were available:||10|
|The price to performance ratio was good:||10|
|TotalScore:||59 / 60|
I have organized this road test in three parts shown below. Part 1 covers product description, unpacking and exploring the contents, power-up and performing basic measurements.
In part 2 I used the Keysight DSOX1102G oscilloscope to measure various characteristics of electric signals. My focus was to evaluate how easy would be for a student to learn how to use the control panel of this oscilloscope and how to use the built-in measurement functions of the DSOX1102G oscilloscope. In part 3 I evaluated the additional functions of the DSOX1102G oscilloscope: serial protocol analyzer, digital voltmeter, frequency generator, and frequency analysis function. Here are the details for each part:
What I am planning to cover in part 1: product description, unpacking and exploring the contents, power-up and performing basic measurements.
First I want to thank Element14 and Keysight for selecting me as one of the roadtesters for the InfiniiVision DSOX1102G Oscilloscope. I knew when the package comes from a delivery confirmation email and I was happy to find it on my porch. I took it inside and I took this picture:
then I opened it,
and I explored the contents. Here is what I found inside: The oscilloscope, power cable, two probes packed together in a bag, a certificate of calibration, and various documents with useful information about this product.
One of the document showed me where to go on-line and download the manuals for this oscilloscope. I took the document to the other room and I went to the Keysight web site address listed there. Here I found multiple documents available for download:
I then came back to further explore the DSOX1102G oscilloscope:
(my dog was also excited of exploring this oscilloscope)
I connected the power cord and I press the power button. The oscilloscope started quite fast compared to other oscilloscopes that I worked with in the past. Here is a short video:
Next I opened the bag with probes:
and I attached the color rings to each probe:
Next I connected the probes to the oscilloscope and I calibrated the probe compensation using the small screwdriver included in the probes kit. The “yellow” probe was a bit undercompensated ; the “green” probe was just right. Here is a short video:
The probes may be undercompensated, like in this picture:
or overcompensated like in this other picture:
or they can be “just right” compensated like in this picture:
With this setup I then performed a few basic measurements of this waveform. First I set the trigger on rising edge of channel 1 (yellow):
and I setup the built-in measurement function to measure the peak-to-peak value:
The knob labeled “Entry” highlights the measurement when rotated and selects it when pressed. Similarly I have selected to measure “Amplitude”, “Rise time”, and “Duty Cycle”. The four measurements are displayed at the bottom of the screen:
Notice that the duty cycle measurement displays “No edges”. This is because the displayed waveform represents only part of a period of the signal, so I had to adjust the time base to display multiple periods and thus to provide the needed information for the built-in duty cycle measurement:
Since this is an “entry level” type oscilloscope, which means that a lot of users may be just beginners in the electronics field, I will focus my review on “entry level” audience. For this reason, I will take a break here and explain briefly the meaning of the waveform that we see on the oscilloscope.
In a very simplistic way, an oscilloscope is an instrument that can be used to visualize how signals vary in time. The following figure shows the graphical representation of a square wave signal and the image of this signal as shown on the display of an oscilloscope:
We can view the oscilloscope as an instrument that captures successive snapshots of the signal waveform and display them one after another on the screen. The trigger function of the oscilloscope ensures that each snapshot starts at the same location within the signal period so the image on the display stays stable. The time base function of the oscilloscope ensures that each snapshot has the same length in time, thus when sequentially displaying these snapshots the image is clear and stable.
Keysight DSOX1102G oscilloscope has a crystal-clear and colored screen and offers very useful built-in functions besides the displaying signal waveform. Some of these functions, like the automatic measurement of signal amplitude, frequency, rise and fall time, RMS value, mean value, duty cycle, pulse width, overshoot and undershoot, performed digitally rather than visually approximating them on the screen help me significantly in design and troubleshooting of electronic circuits.
Here is an example of a sinusoidal signal displayed on the Keysight DSOX1102G:
Here I used the save function of the DSOX1102G oscilloscope, which saves the measured waveform either as a picture (like in the case above) but also it can save it as .csv file very useful for post processing and analysis of data in Microsoft Excel.
The measurements at the bottom of the screen show the signal amplitude, the peak-to-peak magnitude of about 2V, the RMS voltage of 680mV, and the cycle average (which here is around 0V since the sinusoidal signal is centered on 0V) . The rms value for a sinusoidal signal of 1V amplitude (2V peak-to-peak) should be 0.707*1V = 707mV. The measured value on the DSOX1102G screen is 680mV, so I think the difference may come from the sinusoidal signal not being a perfect sinusoidal signal. The RMS (root mean square) voltage is what we measure with multimeters when we probe time varying signals. Not all oscilloscopes have built-in measurement for RMS voltage, so in many cases we need to use a digital multimeter (DMM) to measure the RMS value; however, DMMs are usually limited in bandwidth so they do not measure accurately the RMS value of high frequency signals. I like this DSOX1102G oscilloscope since all the built-in measurements work up to the full bandwidth of the oscilloscope.
So what is the physical meaning of the RMS value of a time varying signal?
A DMM measures the rms value of a sinusoidal signal, which is equal to the peak voltage multiplied by 0.707.
Here is a clarification for the "rms" (Vpeak*0.707) and also the "average" (Vpeak * 0.636) values of a sinusoidal waveform:
The factors 0.707 and 0.636 result from the following analysis:
0.707 comes from the rms voltage definition. V_rms, is a constant (DC) voltage that produces the same average power dissipation on a resistor as the sinusoidal voltage V_max*sin(ω*t). The power dissipation P = V*I = V^2 / R . So for V_rms, which remember is a DC voltage, it’s easy: P_avg = V_rms^2 / R; however, for a sinusoidal voltage P_avg = ( V_max*sin(ω*t) )^2 / R = [V_max^2 * (sin(ω*t))^2] / R . Since these two average powers have to be equal, it results that: V_rms^2 = V_max^2 * (sin(ω*t))^2. The sinusoidal term is an average over many periods of a sin square function, which from trigonometry or from an intuitive graphic (similar to a sinusoid but shifted up and varying between 0 and 1) has the value of 0.5. Inserting 0.5 in the above equation V_rms^2 = V_max^2 * 0.5, and taking square roots on both sides V_rms = V_max * squareroot(0.5) = V_max * 0.707
0.636 comes form calculating the time average of voltage (not power like in the rms case above) for one half cycle. By integrating the sinusoidal voltage over half of period we obtain V_max * (2/pi) = V_max * (2/3.14) =V_max * 0.636
In part 2 I have used the Keysight DSOX1102G oscilloscope to measure various characteristics of electric signals. My focus was to evaluate how easy would be for a student to learn how to use the control panel of this oscilloscope and how to use the built-in measurement functions of the DSOX1102G oscilloscope.
First I would like to summarize how I accessed the measurement panel of the DSOX1102G oscilloscope:
I started by pressing the measure button on the front panel:
Then I pressed the measurement type key on the right side of the display screen:
Pressing the type key opened the measurement selection window on the screen, as shown in the above picture. I then selected the measurement by rotating the “Entry” knob on the right side next to the screen:
To activate a measurement I pressed on the “Entry” knob, and the corresponding measurement was displayed at the bottom of the screen. Here is an example of rise time measurement on a screen captured using the “save to USB” function of the DSOX1102G oscilloscope:
I like the cursors that are automatically displayed when selecting the rise time measurement because they show how rise time is measured on the waveform. This is a good feature for learning and I think it is beneficial to entry level students.
Here is an example of fall time measurement on the same waveform:
Other characteristics that define a square wave signal like this are pulse width and duty cycle. Each can be measured on positive pulse or on the negative pulse, like I am showing in the following screenshot:
Notice that the duty cycle is not 50% as “ideally” expected. In this case the positive pulse is longer than the negative pulse, resulting in 55% duty cycle as referred to the positive pulse. 55% added up with 45% duty cycle referred to the negative pulse results in 100% as expected based on the duty cycle definition. Duty cycle is reflected in pulse width, as measured by DSOX1102G as 284ns for positive pulse and 232ns on negative pulse. The ratio should correlate with the duty cycle: 284/(284+232) = 0.55 or 55%.
DSOX1102G can also measure the delay between two pulses. Here is an example of delay from one rising edge of channel 1 signal (yellow) to the next rising edge of channel 2 signal (green):
Notice at the bottom of the screen the visual representation of what this delay value means (yellow rising edge to green rising edge).
Another measurement related to delay is phase difference. The phase of these two signals is measured as 155 degrees, as shown in the picture below:
The delay and phase are related and part of the relationship is also the period of the signal. In the following picture I show these three quantities measured on the two waveforms:
So the 220ns delay and the 513ns period mean that the delay represents 220ns/513ns = 0.429 of the period, which considering that the period can be viewed as 360 degree represents 0.429*360degrees = 154 degree phase shift. This is the value of the phase shift measured by the DSOX1102G oscilloscope as shown in the picture above.
Another group of measurements relates to signal integrity. In the following picture I am showing the overshoot of a square waveform rising edge measured with the DSOX1102G oscilloscope:
Notice the cursors that are automatically displayed when selecting this measure function. Theses cursors show how the overshoot has been measured, in this case from the top level of the waveform to the peak level. The overshoot value is displayed at the bottom. Similarly, the overshoot of the falling edge can be measured, like I am showing in the picture below:
Another characteristic of this waveform that I measured is the preshoot, as I am showing in the next picture:
The preshoot depends on multiple factors, and usually it has a more dominant negative peak right before the rising edge (we cannot see this in the picture above because the driving circuit did not have that type of preshoot). The peak to peak value of this waveform (with ringing) is shown in the picture below:
This is different that the amplitude measurement, which I am showing in the next picture:
So far I was measuring time domain characteristics of the waveforms. The same waveforms can be characterized in frequency domain. The Keysight DSOX1102G oscilloscope has the possibility of analyzing signals in frequency domain through a built-in Fast Fourier Transform (FFT) function. In the following screenshot I am showing a 2MHz square wave signal displayed on the DSOX1102G oscilloscope in time domain (green waveform) and in frequency domain (white trace).
To further relate the time domain and frequency domain representation of this waveform, I have used the built-in cursor function to measure the characteristics of the FFT waveform. I am showing these characteristics in the following screenshot with my annotations in pink color:
We can see the fundamental component located at 2MHz (this is the frequency of the signal also shown in the time domain displayed sinusoid in the previous picture) and odd harmonics at 6MHz, 10MHz, and 14MHz, … (odd harmonics continue beyond the measurement interval). There are also parasitic spurs superimposed to these expected spectral components. Since rectangular signals are typically used in communication interfaces (like microprocessors, systems-on-chip, peripheral modules, and various interface signals between them) the parasitic spurs translate into timing jitter, which may degrade the performance or generate failures.
In the following picture I am showing the FFT transform of a sinusoidal signal with my annotations in color pink:
We can see the fundamental component located at 2MHz and no harmonics, which ideally is expected from a sinusoidal signal. There are some parasitic spurs caused by non-ideal frequency generator. When we design or troubleshoot a circuit these parasitic spurs distort the signals and degrade the performance of our projects. Since they cannot be visually noticed on the time domain waveform this built-in FFT function in the Keysight DSOX1102G oscilloscope is a great tool that can be used to identify the root cause of signals distortion.
In part 3 I have evaluated the additional functions of the DSOX1102G oscilloscope: serial protocol analyzer, function generator, digital voltmeter, and the frequency response analyzer.
I first started with the serial protocol analyzer. I was particularly interested in this function since I am working on a project where I control an FPGA from a computer through a RS232 interface. So I turned on this project and I setup the Keysight DSOX1102G oscilloscope with channel 1 probing the TX line and channel 2 probing the RX line. For the measurement settings I found very convenient the “quick reference” in the user guide document; it helped me get the measurement active in a few minutes. In particular I liked the tables with instructions, like this one for setting up the type of serial protocol:
and this one for setting up the trigger mode:
continued on next page:
Here are my steps in setting up the measurement: first I selected the “Serial Bus”:
and then I continued with selecting the bus mode as UART/RS232:
Next I configured the UART/RS232 bus to match the settings I have in my FPGA project:
and then I setup the oscilloscope channels:
Channel 1 was connected to the TX line and channel 2 to RX. This is 3.3V UART bus so I setup the trigger levels somewhere in the middle of the swing. The trigger function was set to trigger on the TX start bit:
With these settings I started to send date through the serial interface from computer to FPGA and back. Here is a screenshot of the captured data packet:
The yellow trace shows the TX line and the green trace shows the RX line. At the bottom of the screen we see the values of the sent data in HEX format (01 2C B0 14). This is an instruction sent to the FPGA. The FPGA responds immediately in this case with zero values (00 00 00 00) followed by a copy of what it has received from the computer (14 B0 2C 01) which is a “reversed order” version of the TX line data. Here is another measurement, this time the FPGA responds with “non-zero” data:
The FPGA data is shown on the RX serial captured data as 62 17 00 00 followed again by a copy of the command received from the computer 14 00 EA 2C (in reversed order).
This was the expected functionality, but I wanted to see how the Keysight DSOX1102G oscilloscope handles wrong data (this is important because we use this scope for troubleshooting and in troubleshooting faulty signals may not follow the serial protocol standard). Here is an example of measured data with a failure intentionally inserted in this experiment:
Notice a red section which is “undefined” due to some previous activity that altered the logic sequence of start/stop bits in the serial stream. Notice also a pause on the blue region which is due to signals not following the serial protocol specs. I am impressed of the serial protocol measurement feature of the DSOX1102G oscilloscope and I am happy to use this function in troubleshooting my projects.
Next I looked at the built-in function generator. I found this feature very useful especially for entry level engineers and students. The function generator can generate square wave signals, sinusoidal, triangular, pulse, DC, and thermal (random) noise. Here is an example of square wave signal:
The signal is not perfect, but I think part of the waveform integrity is due to the wires I used for probing. Since the oscilloscope canmeasure waveform aberrations, I used this feature to measure the overshoot, preshoot, duty cycle, and rise time of this waveform:
Here we can clearly see the preshoot, right at the beginning of the rising edge. Duty cycle is pretty good (very close to 50%) and rise time is as expected for a square wave signal of the frequency range of this function generator. I then increased the frequency to the maximum value of 10MHz:
I then used the FFT function to look at the frequency spectrum:
The frequency spectrum follows the expected behavior of fundamental and 3-rd harmonics, as I showed in the previous part of this blog series (in part 2).
I then performed similar measurements for sinusoidal waveform:
which did not show any visible parasitic spurs.
I then setup the function generator to output a triangular waveform:
which had the frequency spectrum as shown in the picture below:
The pulse and random noise waveforms are shown in the following two pictures:
Next I wanted to test the built-n digital voltmeter, but I wasn’t able to do this because my oscilloscope does not have a built-in digital voltmeter. Even though I followed the instructions:
When I got to the “Features” menu there was no “Digital Voltmeter” selection available there:
I even checked the “Help” feature of the oscilloscope, and based on that I was supposed to find a selection for Voltmeter:
But it looks like somehow this Voltmeter is not available on my oscilloscope. No problem, because I have a DMM that I use, so not having a built-in Digital Voltmeter in my DSOX1102G is not an issue. I wonder if anyone else has missing voltmeter with the road test oscilloscopes.
My last experiment that I planned was to analyze the frequency response of an amplifier using the frequency analysis function of the Keysight DSOX1102G oscilloscope. So I built an amplifier using an operational amplifier in inverting configuration, and I applied a sinusoidal signal at the input (sinusoidal signal generated with the DSOX1102G oscilloscope. Here is a picture of my testbench:
(a bit messy I would categorize it, but it functioned as expected)
Here is a picture showing the input and output signals in time domain:
We can see that the output is inverted (or in other words 180 degrees shifted), and from the measured values shown at the bottom of the screen we can calculate the DC voltage gain as 1.73V/0.149V = 11.6.
I than started to setup the frequency analysis function by pressing “Analyze” and selecting “Frequency Response Analysis”:
I was impressed that the DSOX1102G oscilloscope showed me on the screen how to connect the function generator and the channels to the testbench:
After connecting the two channels the way was shown in the diagram I pressed the “Run Analysis” button on the side of the display and to my surprise the DSOX1102G starting the frequency sweep and displayed in real time the measurements on the screen. I captured this in a short video linked below:
The result was the gain and frequency variation with frequency for this amplifier, as shown in the following picture:
The important part looks good and as expected, with the gain dropping by 20dB per decade after the dominant pole. There is some noise at high frequencies primarily due to the inductive loop coupling in my solder-less breadboard amplifier. To confirm this assumption I have re-run the frequency analysis:
The significant region of the gain and phase remained the same, while the high frequency region of noise showed different behavior. This confirmed to me that the noisy behavior is due to noise and noise coupling in my breadboard circuit.
As the gain trace shows, the output signal amplitude becomes very low in the region where the phase trace has the noisy picks. The signal to noise ratio at the amplifier output degrades so much that the oscilloscope cannot retrieve the signal phase information from noise. The noise is a combination of: amplifier noise, oscilloscope channel noise floor, and noise coupling in my circuit. Adding noise coupling techniques like shielding, minimize loop inductance, improved supply decoupling... can reduce the noise coupling effect (and potentially the phase trace spikes), but the thermal/random noise of the amplifier and noise floor of the oscilloscope still remain. In my opinion the noise spikes on the phase measurement are normal from fundamental physical principles. What Keysight could have done to avoid displaying the noisy spikes on the phase trace would have been to limit the frequency range of the analysis at a certain threshold of the signal-to-noise ratio at the measurement node, and just stop the analysis there without plotting the rest of the trace or only the rest of the phase trace. From an engineering perspective I like to see the entire trace with the noisy spikes as long as I understand that they are caused by fundamental physical limitations of circuits and method used and as long as I understand that they do not represent the circuit functionality. This frequency analysis method is a fundamental concept and it has been covered in many textbooks (and the magnitude and phase traces show ideal behavior for magnitudes going down as low as the graph scale has been drawn) but usually books do not mention the limitations of actually performing this measurement in a lab and how these limitations will affect the measured traces. From my circuit design experience the magnitude can be measured down to lower levels but the phase measurement is limited at higher magnitude levels due to a comparator dead band. This comparator dead band limitation comes at higher magnitude levels than the limitation for measuring the magnitude, which we can also see on the Keysight DSOX1102B display. From these displayed traces we can also determine the phase measurement comparator dead band as about -40dB attenuation from the input signal magnitude.
This frequency analysis feature is very useful for students in analog design classes. I am thinking at some point in the future to expand the operational amplifier design project of my CMOS Analog IC Design course and add a hands-on lab where students can evaluate an operational amplifier. I like the DSOX1102G oscilloscope because it provides almost a “full lab equipment set in one instrument” for students especially on-line students who don’t have access to school labs.
This concludes the third and last part of my road test evaluation of the Keysight InfiniiVision DSOX1102G oscilloscope. I am very impressed of the measurements capabilities that Keysight has implemented in such a small size instrument.
Best Wishes to Everyone,