InfiniiVision 3000T X-Series Oscilloscope MSOX3034 - Review

Table of contents

RoadTest: InfiniiVision 3000T X-Series Oscilloscope MSOX3034

Author: ipv1

Creation date:

Evaluation Type: Test Equipment

Did you receive all parts the manufacturer stated would be included in the package?: True

What other parts do you consider comparable to this product?: Rohde & Schwarz RTB2004, Tek MSOX3034

What were the biggest problems encountered?: Menu and user experience could have been better.

Detailed Review:

Introduction

Keysight is well known for industry standard test equipment and the MSO3034T is no exception. In this roadtest, I take the scope through its paces, looking at the inbuilt demos, features etc and then compare it with a similar(ish) scope offering from a competitor. I dive into the user experience from a  day-to-day operations point of view and then look at a few common measurements before performing a small experiment for an application. A big part of my review revolves around the triggering features of the scope as it distinguishes this instrument from others. That is a lot to cover so please use the TOC to navigate the writeup. Lets jump in.


image

Basic features and first impressions

An oscilloscope is traditionally an instrument designed for reading analog signals and then displaying them in a 2D plane. The scopes of today go beyond this definition and the MSO3034T is an exceptionally well built instrument. It features multiple instruments in one box which includes the oscilloscope, Logic analyser, DMM, Function generator and more. The scope under test here is a 350MHz device with 4 analog and 16 digital channels which is above the hobby grade user and yet below the RF lab set of devices. The offering boasts of a number of features and the list is long. My test unit came with four fixed 10x scope probes and a digital channel cable, the standard add-on knick-knacks which included extra tips, coloured rings and even a BNC adapter.


image

The display on the module is a beautiful 8.5 inch capacitive touch screen with a matte finish, right under which are the multi-purpose buttons. One of the things that I appreciate are the amplitude knobs for each analog Channel. There's also the option to turn off touch for those who do not wish to make use of the additional feature but I highly recommend experimenting with it as it can shave off a lot of time during testing.

The unit I'm going to compare this with is a Rhode n Schwartz oscilloscope and it features a much larger display albeit a glossy one which we will talk about later. The keysight is heavier, feels well-built and my unit comes fully loaded with all the bells and whistles.

The boot up time for the oscilloscope was a disappointment, though not too much as doesn't necessarily impact the way you use it. Another gripe that I have with the instrument is that it feels cramped in the display as there are menu items and text squeezed into the little abyss of the corners of the display.

On a positive note the front panel feels well-organized, has all the necessary buttons that a slightly seasoned test engineer would  appreciate. The menu complexity is justified by the sheer number of features in the system and doesn't take long to get used to.

Every time you turn on the device the oscilloscope prompts you to go through the demos and the demos themselves are quite educational. The demos include a capture of a glitch, a Serial bus error capture, identifying a noise source, debug a mixed signal design with analog and digital channels, validating the frequency hop of an FSK signal using gated fft etc.  I would have liked to see these converted into a lab exercise of sorts where the system would ask the user to set up the scope for a capture by guiding through the steps. I assume that this would be opinionated hence my not making for the best of features.

 

Header 1Header 2
imageimage

 

 

 

If you happen to click the left Arrow or the back arrow under the screen while in the demo menu you will see the built-in training signals menu( similar to my preceding ask albeit without guidance) as well which was a fun thing to go through. The same can be accessed using the dedicated help button.

The generated training signals range from CAN to LIN and even Flexray and a few others I didn’t even know existed. For the most part it's an entertaining experience and will leave you playing with this very expensive toy for a very long time.

I will add a video of all the signals being used later. The space will be used as a placeholder for the same.

Spec comparison with Rhode n Schwarz RTB2k

I compared the keysight with the RnS RTB2K which is a 300MHz scope for less of a  cost.

 

MSOX3034T

RTB2K

Tektronix MDO3034

Display

8.5”

10.1”

9”

Touchscreen

yes

yes

no

Memory

4M Points

10M Points

10M Points

ADC Res

8-bit

10-bit

8-Bit

Sample Rate

5GS/s

1.25GS/s

2.5GS/s

Refresh Rate

1,000,000 wfms/s

50,000 Wfs/s

280,000Wfs/s

 

Let's take a look at both instruments one by one and compare some of their most notable features.

Display

The RnS display is bigger and hence better. The experience feels a bit more polished, due to the larger real estate. The Keysight on its own is great and the matte finish is just what you need. The touch screen feels cramped at times but the physical buttons make things easier.

The need for touch becomes much more clear once you start to use trigger zones. There are very few other features that cannot do without touch but it’s something you get used to and then it’s hard to go back. I tried using the scope with the touch turned off and quickly realized that there were no menu buttons on the right side of the display for the menu. The menu itself is quite dense and the little knob/dial is the only way to cycle through it. I am now using a touch stylus like the one sold at the dollar store instead of poking the screen.

 

Sample Rate vs Refresh Rate vs Memory Width

I did a video on the subject in 2017 and won't repeat the content. Here is a short summary.

The Sample Rate is the interval between consecutive samples that are converted by the ADC. The higher the Sample Rate, the higher the bandwidth of the instrument but the sample rate is usually adjusted in software as it can be lower for slower signals.

Memory is where the samples go and it can fill up pretty fast if you are sampling at a high rate. Nyquist criteria notwithstanding, the sampling rate is adjusted so as to fill up the memory with enough samples based on the time-base setting of the scope. Also processing more data, especially for operations like FFT will slow things down.

Refresh Rate is the rate at which the instrument completes the cycle of ‘Trigger-sample/capture-process/display’. The concept of dead time effects the refresh rate which creates windows in time when the instrument does not capture data.

https://youtu.be/gl172iN5wcU please have a look at this video.

Memory

Memory is useful for storing the data samples or data points and the keysight features 4Million points vs the 10Million on the RnS. The lower memory is not an absolute disadvantage since the higher refresh rate makes up for it quite a bit. Lets try to visualize this a bit.

Imagine having an analog signal that pops up every now and then due to a sporadic event. It may be a simple glitch and I use a microcontroller to generate the glitch in a demo later which is similar to the demo included in the Keysight.

Now this glitch may be:

  1. A pulse with a larger amplitude(Amplitude glitch)
  2. A pulse with a smaller amplitude(runt)
  3. A pulse with just the wrong width(logical glitch)/ Timing Glitch
  4. A pulse that is misformed and does not conform to expected output

Now imagine that if you understand what you are looking for then you set up a trigger and then wait. The advantage of the larger memory is that you have a longer trace captured after the trigger event but the caveat is that it takes longer time for the process to get ready for the next round of capture hence the slow refresh rate. Some will argue that this is not the case and you can use a faster processor but the takeaway here is that the larger memory means you capture more in a single shot.The keysight’s faster refresh rate comes into the picture when you don’t have a clear picture of the type of glitch. This means that it will appear on the keysight even when the RnS misses it. This allows a user to make a better decision on how to capture it and then consequently setup the scope for debugging.

Sample Rate

The Keysight features 5GS/s which leave the RnS in the dust and in my experiments, the capturing experience was hugely different. Not a lot of notes about this.

Zone Triggering

The zone triggering was one thing that the touch screen just cannot do without (unless you connect a mouse). It allows you to set a rectangular area of intersection or clearance but obviously works in conjunction with a basic trigger. So zone triggering is more like a triggered output filter more than anything. Here is a video demonstrating the same with a signal that I emulated.[Zone Triggering]

Display and controls

The display feels cramped and I prefer the user experience on the RnS over the Keysight. The menu on the Keysight feels obstructive to me and I struggle with finding certain items. The controls for the individual channel are better than the RnS since each channel has an amplitude and position control.The horizontal controls work for the most part but the layout of the rest of the controls feels mushed together. Due to the absence of a clear demarcation between control groups, one has to initially look for specific buttons and a lot of the functionality is implemented in software such as the trigger options etc. This is similar to the RnS and nothing new but I would have appreciated some additional controls.It just was not as pleasant of an experience as I had hoped for.

Function generator

This is one of the more amazing addons since it saves bench space. The experience was straightforward and Keysight could have added a cable but I appreciate the added instrument.

The user experience

Before I get into the trigger experiments, I want to share my user experience with some basic operations.

Setting up the instrument

The first step in using an oscilloscope is to understand your DUT. I start with a microcontroller device that generates pulses. This could be a communication line with no explicit clock signal. Usually with modern scopes, one would connect the scope and then press the autoset button and wait. There are other things to consider and things you can/should set manually like

  1. Trigger type - Auto/Man: I usually set this is normal
  2. Acquisition: Envelope detect/persistence for signals or roll mode and even tweaking the signal record length/segments/sampling rate.
  3. Automatic Measurements: Peak-Peak voltage, frequency and in some cases, rise or fall time.

With the Keysight, it became easier after doing it a couple of times and I have come to appreciate the instrument.

Mess on a desk

The footprint of the scope is exactly as expected and does take up some space. My RnS is lighter and feels stable sitting atop a cardboard box though I would advise getting a dedicated shelf.


image

The logic analyzer cable is all 16-bits packed into one and I appreciate the quality. Strangely, I prefer it over the RnS cable setup which came with two cable sets of 8-bits each.

The touch screen

The touch experience is something to get used and frankly it's clunky. The ‘Main Menu’ in the case of the RnS is a large button on the lower right corner of the screen whereas for the Keysight it is a little icon in the top left. The menu is ugly and feels like windows 3.11.(yes, I am that old)

The buttons on the front panel can be used to open certain function menus directly and it works out most of the time but it is not a substitute for a good menu system.


Trigger types

Header 1Header 2
imageimage

 


Measurements - Analog

Let's get real and for this section, I have a few experiments. Glitches and pulses to be exact.

Measure from DAC of SAMD21

The DAC in question is a 10-bit dac and nothing fancy. The output was set to ramp and both scopes display similar output. Neither has an upper hand over the other and so the 10-bit ADC on the RnS does add value over the keysight.

image

image

image

Note that the 10-bit ADC vs 8-bit ADC does not necessarily make a huge difference in this particular case.

Glitch

I tried using the in-built demo in the Keysight and sure enough, the RnS cannot isolate the timing glitch generated by the system. I setup the RnS to Envelope detect and I was able to capture some of the glitches but not all for sure.


image

 

I then decided to create a circuit of my own. For my experiment, I use a Teensy 4.0 which has a Cortex M7 to generate a clock like signal. I then introduce anomalies in the signal which is essentially a calculated timing glitch.

In the code, I make it such that the clock signal goes to LOW immediately after going high after every n cycles. I tested with setting this to every 4th cycle being a glitch and then set it to 100 and 4000 etc.

The result was that the RnS did manage to see the signal when it was more frequent, but lost track of it in longer gaps until I set up the trigger to happen at a particular state.


It is a rare occurrence but in situation where the glitch is very infrequent and there is no alternative trigger source, the Keysight and it’s 1Mwf/s refresh rate beats everyone. But it took me a while to figure out a setup to demo this feature.

For all other scenarios, it really does not matter since I would set up an alternative trigger system. The high refresh rate gives you superpowers but they are not essential for everyone.

 

There are some elements of this review that I am going to redo and I would consider a work in progress. I am going to stick to this one for a bit longer but here is a quick video for now.

 

Update: I have been using the scope on and off for the last few months and it turns out that our units came with trial licenses of the features. After the trial expires the scope becomes a different animal with serial decode and a part of the capture memory disabled. To make matters worse, the waveform generator no longer functions. This is a big thing because the perpetual license for everything is more than $3000 Canadian and even though you still have the hardware, the features are disabled, staring at you.

The scope is still an amazing analog device and the digital capture works well but with the reduced memory and decode features, the value for embedded systems designers goes down substantially. I wish keysight would include basic I2C SPI, UART and LIN decode as well as the waveform generator as a default offering.

Anonymous