RoadTest: InfiniiVision 1000 X-Series Oscilloscope DSOX1102G
Author: hlipka
Creation date:
Evaluation Type: Independent Products
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?: Tektronix TBS1202B-EDU
What were the biggest problems encountered?: There is no real manual - the one available is basically only a quick-start guide which explains how to turn on all features. But an explanation of how to use them properly is missing.
Detailed Review:
My thanks go to Keysight and Element14 for making me part of this road test. I feel very honored for being selected.
I applied to this road test because I really wanted to compare the DSOX1192G against my Tektronix TBD1202B-EDU which was I used until now (and which I actually also got from E14 for participating in a design challenge). Keysight list the Tek as one of their main competitors, so I was curious to see how they compete with each other.
While both scopes have similar specifications, there are some significant differentiators:
Both scope proved two channels, each with a sample rate of 2GSamples/sec (which don't downgrade when activating the second channel). They also both come with a 7-inch LCD with 800x480 pixels (although the DSOX manual says its a 8.5 inch display...).
For the first part of the evaluation, I did the same power supply measurements as for the last road test, but this time with both scopes. I will use them side-by-side so you can have a look how they compare to each other.
Upon unboxing, the scope feels quite heavy. The data sheets says about 3kg, which is not that much - but the TBS only brings 2kg to the table. Both are about the same size - but lets have a look at how they compare to each other:
The Tek scope feels less crowded, there is much more space between all the buttons and knobs. Partly this is because it just needs less of them (due to Keysight having more functionality), but it also makes more use of the available space and has a larger button area.
What you can also see is that the USB port on the Keysight is recessed, so you cannot fit any USB drive in there. Most current ones should be fine, but I have several that won't fit.
So, time to unstack the scopes and to turn them on:
Its not really noticeably in the video, but the fan on the Keysight is quite load and annoying (the Tek is completely fan-less). So I will replace the fan (but probably not for that road test)
When I have a specific USB drive connected the DSOX will hang during startup until I remove it. This seems like a firmware bug.
Regarding the usable screen area, I actually measured them:
In addition, Tektronix has added the feature to completely remove the menu from the screen. Even when you do that on the DSOX, the are will show additional information (even if its not used currently, e.g. the current Math setup). OTOH, the Keysight scope uses much finer traces and fonts, so it makes the most of the available screen:
Here its pretty good to see that the Keysight uses much finer lines and fonts. So even if its divisions are (physically) smaller on the screen, it allows to see more detail - at least when you use the high-resolution mode for acquisition.
Due to the high-resolution mode the DSOX has lower noise, which is a big advantage when doing measurement like noise and ripple as I did for the power supply.
When looking at the layout of the controls, I don't have areal preference between the two. The Tektronix was my first scope, so I got used to it quite easily (its designed for education, so you expect it to be easy to learn). it still happens that I use the trigger controls on the Keysight when I want to adjust the horizontal, because I'm used to that. The knows on the DSOX feels nice, they are rubberized which gives you the feeling of having more control. But whats really nice is that you can press them all, either for selection or to reset they corresponding function to zero (e.g. move the signal to the middle of the screen).
Its also nice that in the menu you can press the corresponding button for a selection and it switches to the next setting. This makes it faster to select a setting. On the TBS, you need to use the rotary selection knob. This one doesn't have indents so it sometimes happens that you select the wrong item.
On the TBS, there is also no acceleration function on the rotary knobs so it takes a while when you need to change something by a larger value (e.g. the horizontal offset). All in all both scopes are responsive, but the Keysight feels faster over all, and typically all settings and changes are one click faster to reach.
Due to the high-resolution mode the DSOX has lower noise, which is a big advantage when doing measurement like noise and ripple as I did for the power supply. In this mode, the sample rate is set to a higher value, and several measurements are averaged into one data point. This is different to the normal averaging mode which does this calculation over multiple waveforms - high resolution does this within one capture.
What I noticed is that, when using AC coupling, the DSOX doesn't show the trigger level marker on the left side, as its doing in DC mode (it does when changing the level, but it disappears afterwards). This can be confusing. The bottom of the screen shows the voltage offsets for both channels - but this gets removed when you add measurements (even though there is plenty of space to show both). And with measurements you always get the cursors on the screen (the TBS does without, so are not confused as easily).
Proper probing and good probes are essential, so I have a look at that:
I do like the Tektronix probes - in comparison they are easier to handle, are smaller, the cable is thinner and the ground clasps feel much better. And the tip is fine enough to fit a pin header or a via, so you can leave the probe there. With the Keysight probes you always need to hold them down which occupies one hand (or two if you need to probe two signals). I'm seriously thinking about using the Tek probes with a Keysight scope (which is perfectly fine, both scopes use standard BNC connectors without any special additions)
The last measurement I did was the transient response to load changes, and this is where the Keysight with its large sample memory really shines:
Being able to capture both the on- and off-transient with one shot, and then just zoom in into each of the flank and see them in all detail is a real time-saver. It allows you first get an overview, look for the interesting details and then just zoom in without redoing the capture. Just for that its worth getting the DSOX over the TBS scope. If you are a beginner this might not sound so important, but it really changes the game.
Another feature which can be quite handy is the serial bus decode feature of the DSOX. This is something the Tektronix doesn't have, so the object of comparison are the typical low-cost logic analyzers (starting from the Saleae clones you can get at ebay for 10 bucks, to be used with SigRok, up to a real Saleae for maybe 200$). The serial decode feature will set you back 140EUR (in Germany), so it competes more with the Saleae than its clones. Lets take for example the DSLogic: about100 EUR, for 16 channels with 100MHz sample rate and 16M sample memory. This is much more than the DSOX can provide, so you can capture longer data streams with the DSLogic. But using a oscilloscope has some advantages:
Each of these features provides a unique value, so lets see how this looks like. First, I show how to use the bus decoding and triggering functionality to analyze an I2C signal:
When you need to use higher sample rates, or the times between the interesting parts of the capture are too long, you need to use the segmented memory function:
This feature is not only interesting for analyzing digital signals like serial buses, but also when searching for rare signals. You can then trigger on one of the conditions where the problem might happen, do up to 50 captures and analyze them later on for the anomaly you are searching. This comes at a cost though - adding memory segments mean that each segment gets shorter. So with a constant sample rate the time base gets short (less information on the screen), or, with the same time base, the sample rate gets lower so you can zoom in to see details. For this you would need a scope with higher memory depth.
One thing what is not shown here is that the DSOX, although its only a two-channel scope, can show the external trigger as third channel. Although it shows only digital values (based on the current trigger level) seeing it can be areal value (e.g. when its connected to the chip select signal of a SPI bus. Unfortunately it seems that the external trigger view can only be activated when using the 'normal' acquisition mode, but the same seems to be true for the bus decoding.
So it depends on your actual use-case whether you should purchase that option or not. If you mainly work with serial buses, its probably worth the money. Being able to check signal integrity is really important and can help a lot. Triggering is also more difficult here so being able to e.g. trigger on a I2C address is very handy. And thats something all the not-so-expensive logic analyzer won't do (not even the Saleae - the only exception seems to be the new Digilent Digital Discovery).
But when you need to look at more signals, ore need really long sample times a dedicated logic analyzer is the way to go. As a comparison, I did look at the same signal with my LogicPirate, which is a MSP430 LaunchPad together with serial sRAM. It can sample 8 channels at up to 40MHz, and has some basic triggering capabilities. Here I did sample at 2MHz with 128k memory:
The OLS software is quite sophisticated when analyzing known signals like I2C buses. It shows many details about the transaction directly in the signal view, and the protocol analyzer shows all the details of the I2C transfers. But when compared with what the DSOX is showing
one can see that this comes at a price. While the scope doesn't show as much of the I2C signal as the logic analyzer, it shows much more details about the actual signal. From the logic analyzer I would never have guessed that the signal quality is as bad.
When applying for this test I was curious to see whether having a semi-third channel is enough, or whether you actually want a four-channel scope. The answer is a clear "maybe". For many features two or three channels are enough. But with SPI you probably will run into situations where you need to see both data lines, and this won't work with the DSOX. In these cases you either need a four-channel scope, a good logic analyzer, or you need to combine both (use the DSOX to trigger, and use the trigger-out to trigger on your logic analyzer).
Last but not least, the DSOX (in its 'G' version) comes with a built-in function generator. While its not a full-fledged arbitrary wave generator, it provides quite some functionality:
So it provides quite a lot more functionality than a simple XR2206 generator, and I think it will be sufficient for my lab for the next few years. And it also means I don't need the space for an additional generator... But about 200 bucks are a significant amount of money, and can buy you already a decent function generated. So as long as you have the bench space available, for a hobbyist having two separate devices might be the better option. But if you are low on desk space, go for it - it packs quite a bunch of functionality (for example, I forgot to mention in the video that it not only does AM modulation, but FM and FSK too).
There are quite some features, and some quirks too, which are not captured by the videos. For example - any scope screen shot saved on USB has the serial number of the scope added to the top of it. Why? Model number and time stamp are fine, but why add the serial number? Not everyone is fine with making it public so you are forced to crop the image
But I only have two real complaints:
Apart from these gripes, I can only recommend the DSOX1102G. Its a pleasure to work with, its responsive in all situations, quite easy and intuitive to use in most situations. I think it delivers some more functions than most competitors, even if its actual specs (memory depth, number of channels, bandwidth) are not the best.
This scope is probably a little bit overwhelming as a first scope for a hobbyist, so something cheaper and easier might seem to be better. But on the other hand, it will grow with your needs when you have more complex problems. Features such as the serial decode and triggering, or the segmented, can be a tremendous help. You won't need them every day, but when you need them you really need them. And there is still a ton of things I did not touch in this review (and some I even did not try out, such as the Bode-plot function). But they never were in my way, so you can start easy and learn over time how to use it properly.
Top Comments
Hendrik, your review is something two CRO about!
Great Stuff!