Can a USB Oscilloscope Replace a Benchtop Scope? A Hands-On Review of the PicoScope 6824E
イチケン / ICHIKEN
This review looks at the PicoScope, an oscilloscope from Pico Technology of the UK. The unit has no display of its own. It connects to a computer over USB, and everything happens in software on the PC. The review was provided courtesy of Pico Technology. The reviewer, Ichiken, tries to answer two questions: how a PC-based oscilloscope differs from a conventional benchtop instrument in hardware and daily use, and where its strengths and weaknesses lie. Their overall view is that PicoScope is a software-centric product. How useful it is depends largely on how well you use that software.
A Company Built Around PC-Based Instruments
Pico Technology specializes in PC-based instrumentation. Besides oscilloscopes, it makes data loggers, network analyzers, and diagnostic tools for automobiles, and all of them are designed to connect to a computer. The model reviewed here is an 8-channel oscilloscope from the PicoScope 6000 series, the 6824E.
Hardware: Smaller Without a Screen, With Some Trade-Offs
The most obvious difference from a benchtop scope is size. With no display, the unit only needs to capture waveforms and talk to a computer, and Ichiken suggests that is why it can be so compact. Housings vary by model and are mostly plastic. The 6000 series puts more emphasis on cooling, so part of its enclosure is an aluminum heat sink.
The front panel has eight analog channels, each selectable between 1 MΩ and 50 Ω input. Ichiken says this is typical of mid-range to high-end scopes. One figure stands out to them as low. Other manufacturers often rate the 1 MΩ input for a maximum of 300 V RMS, but the PicoScope's maximum is ±20 V. Channels A, B, G and H are ordinary BNC connectors. The middle four have BNC connectors that can also power and communicate with active probes.
Looking at Pico's probe lineup, Ichiken first thought it focused on relatively low-voltage analog signals and digital signals. On a closer look they also found clamp-on current probes and high-voltage differential probes, so power-electronics measurements appear possible. The front also has digital inputs. Optional 8-channel digital probes can be connected in pairs for 16 digital channels. The unit ships with four passive probes rated at 500 MHz.
The rear panel has trigger input/output, a 10 MHz reference input for synchronization, the function generator output, a power input, and the USB connector. Ichiken calls this a fairly standard set of rear connectors.
Power comes from an external AC adapter. Ichiken notes that USB scopes and modern compact scopes often use external supplies. They find the adapter awkward to deal with and would prefer a built-in supply, though they accept that the design may make that difficult. The lower-end PicoScope 2000, 3000 and 4000 series run on USB bus power alone, so those need nothing but a computer.
Startup is quick. After the USB cable is connected and the dedicated software is launched, the scope is usable within about 10 seconds.
The Software Model: Free, Fully Featured, Frequently Updated
Ichiken stresses that the software is the most important part of this product. It is a free download, and every function is available from the start with no additional licenses. As a result, all PicoScope models offer essentially the same features. What differs between models is the hardware: number of channels, sample rate, memory depth, and so on. The software is updated fairly often. Ichiken finds that reassuring, noting that some manufacturers rarely update theirs.
Everyday Operation With a Mouse
The interface is built for mouse use. Channel settings sit on the left side of the screen. Channel 1 is on by default, and other channels can be enabled and their traces dragged into position. Each channel has manual and automatic ranging. In automatic mode the software adjusts the amplitude scale per channel. Ichiken says they haven't seen many oscilloscopes that do this.
The timebase control is at the top left. It can be set by clicking or by scrolling the mouse wheel over it, which Ichiken compares to turning the physical knob on a benchtop scope. Waveform updates felt responsive to them. Next to the timebase is a setting for whether the capture should prioritize memory length or sample rate. The vertical range can be adjusted with buttons or the wheel in the same way. Coupling options are AC, DC and DC 50 Ω, again typical of mid- to high-end instruments. Probe type and attenuation can also be set.
Waveforms can be zoomed on both axes with the mouse wheel. Ichiken compares this favorably with touchscreen scopes. On many of those, pinch-to-zoom is sluggish and doesn't always do what you intend, so for this kind of waveform manipulation the mouse felt easier to them. Still, they admit that without dedicated knobs and buttons some operations are slightly frustrating, which they put down to getting used to it.
The display settings let you rename waveforms and change trace colors. Ichiken says they use renaming quite often.
DSP Filtering and Resolution Enhancement
Under the channel settings is a DSP section with a low-pass filter that can be switched on and off. Ichiken reports that it works well. There is also a resolution enhancement feature. It uses signal processing to add up to 4 bits of apparent resolution on top of the A/D converter's native resolution, so 8-bit hardware can be taken to 9, 10 or up to 12 bits. Ichiken zooms in on the top of a waveform to show the difference and warns that this is only a processing-based improvement. The result is not the same as the original signal.
They then compare it with changing the actual hardware resolution. Going from 8-bit to 10-bit hardware removes most of the visible quantization error, and at 12 bits it is hard to see any at all. With software enhancement added on top, the display can go up to 16 bits.
Flexible Multi-Window Display
Ichiken considers the display flexibility a strength that comes from doing everything in PC software. Additional scope views can be added from the View menu, for example channels A and B in one scope view and C, D and E in another. The same waveform can appear in several views at once. Ichiken says few oscilloscopes can do that, apart from rough workarounds with math functions that might slow the instrument down. Views can be tiled, split into tabs, or dragged out as separate windows and moved to another monitor.
They push this to see what happens. Up to eight scope views fit in the original window. A ninth opens a new window, and more windows appear as views are added, again at 17 and beyond. Adding many views makes the display noticeably slower. Ichiken's explanation is that the oscilloscope only acquires data and sends it to the computer, and all processing and drawing happen on the PC. Display speed therefore depends heavily on the computer's specifications.
Reusable Setups and Multiple Scopes on One PC
Because every PicoScope uses the same software, the same measurement environment is easy to reproduce even when the hardware changes. Setups can be saved as PicoScope configuration files and loaded on other units with the same settings.
One computer can also run several PicoScope units at once. Ichiken demonstrates four connected and running together. They suggest a production-line example: several USB scopes monitor signals from equipment, and parts that fail a mask test are rejected.
Mask Testing and Actions
To demonstrate, Ichiken generates a mask automatically from the current signal using the Tools menu. Waveforms that stay inside the mask pass. The Actions feature can then respond to a failure: produce an external trigger, launch an application on the computer, output decoded data, and so on. For clarity they choose to play a sound, an explosion-like effect, which fires once for each failure. The counter shows nine failures. The software also keeps a history of the failing waveforms so you can check what each one looked like. Ichiken found this easy to set up and useful.
Where the Software-Centric Approach Pays Off
Ichiken notes that most oscilloscopes they have reviewed were hardware-heavy products, while PicoScope feels software-oriented. How easy it is to use, and how far you can take it, depends on the software. They see its strengths in combination with a computer: integration into R&D setups and automated measurement in particular. SDKs and sample programs for this kind of embedded or automated use are also free.
The software also works without any hardware. It includes a demo emulator that generates various test signals, and all features, including everything in the Tools menu, are available. Anyone can try out what the software can do before buying.
Touch operation is possible if a touchscreen monitor is connected, but Ichiken was not impressed. Response time is fine. However, the software seems optimized for mouse use. Compared with touchscreen scopes from other makers, it felt less comfortable to them: some actions take more movement than with a mouse, and some can't be done by touch at all. Their recommendation is to use a mouse and keyboard.
Specifications: Bandwidth, Resolution, and Sample Rate Trade-Offs
The 6824E has 500 MHz bandwidth, a maximum sample rate of 5 GSa/s, and hardware resolution selectable between 8, 10 and 12 bits depending on how many channels are in use. Software enhancement can add up to 4 more bits.
8-bit and 10-bit can be used on all channels at once, but 12-bit is limited to two channels at a time. Ichiken says frankly that this is disappointing. If 12-bit is available, they would like to be able to use it on every channel. In practice, they say, this is essentially a 10-bit oscilloscope. They mention that another model offers 12 bits on all eight channels, though with a reduced bandwidth, which they give as 20 MHz.
Sample rate also falls as more channels and bits are used. Ichiken walks through the combinations on screen:
- 12-bit: 1.25 GSa/s with one or two channels.
- 10-bit: 5 GSa/s with one channel, 2.5 GSa/s with two, 1.25 GSa/s with three or four, and 625 MSa/s with five or more.
- 8-bit: 5 GSa/s with up to two channels, 2.5 GSa/s with three or four, and 1.25 GSa/s with five to eight.
Their reading is that the total rate is shared across channels, with memory also playing a role. They note that manufacturers handle this trade-off very differently depending on design philosophy. Ideally the maximum sample rate would hold no matter how many channels or bits are used, but they point out that this would make the instrument considerably more expensive. Depending on how you use the scope, you could run short of effective bandwidth, so they advise keeping these limits in mind.
Noise Floor
Ichiken measured the noise floor with a 10 ms timebase, using the automatic measurements for peak-to-peak and ripple RMS. With 1 MΩ input they read about 2.1 mV peak-to-peak and about 200 µV RMS. The averaged values came out at about 2.2 mV and 200 µV. They also say that changing the hardware resolution left the noise level almost unchanged. With 50 Ω termination they measured about 1.8 mV peak-to-peak and about 170 µV RMS. Ichiken judges these to be fairly average values.
Built-In Signal Generator and AWG
The rear-panel generator output was connected back to an input and controlled from the "Gen" section of the software. Frequency, amplitude and waveform type can be changed, with many waveform types available, including white noise. The arbitrary waveform generator (AWG) can import waveforms or let you draw one by hand, and a drawn waveform appeared on screen as expected. You can also type a bit pattern such as "0101..." to create a digital waveform. Ichiken suggests this is useful for testing serial decoding.
A waveform created with the math channels can also be sent out through the generator. Ichiken demonstrates this and suggests it could be used for things like examining amplifier waveforms or measuring op-amp characteristics.
Spectrum View and Spectrum Mode
A spectrum view can be added alongside a scope view to show frequency analysis. With a 10 kHz triangle wave, the main peak sits at 10 kHz with other peaks elsewhere. Ichiken says PC processing opens up many possibilities and would find features like a waterfall display interesting to see from Pico.
There is also a separate spectrum mode, selected from the device settings, which prioritizes spectrum-analyzer operation. The timebase controls turn into frequency span controls, and the number of FFT bins can be set up to about one million. Fewer bins run faster at lower frequency resolution. Very large bin counts make the FFT take a long time, and small ones feel stress-free to use, in Ichiken's experience.
Adding a scope view in spectrum mode lets you see which time-domain waveform corresponds to the current spectrum. Clicking between the two works in both directions, which Ichiken found useful. Spectrum results can be mask-tested too. Ichiken set a mask so that capture stops when a spectral component appears in an unmasked region, which effectively triggers on the spectrum. They then examined the corresponding waveform and called the feature quite convenient.
Triggers and Rapid (High-Speed) Triggering
Trigger types range from standard edge to runt and logic triggers. PicoScope can do serial decoding, but at the time of recording Ichiken found that triggering on decoded results did not appear to be supported.
The feature that most interested them was the high-speed trigger. In normal triggering, each captured waveform is sent to the computer. The oscilloscope also keeps recent captures in an internal ring buffer, which you can browse through the waveform history button; in their setup it held 64 waveforms. Ichiken explains that sending each capture to the PC uses a lot of the scope's internal resources, so rare events can be missed. The high-speed trigger instead fills the scope's internal memory with captures and then sends the whole batch to the PC at once.
To compare, they enabled capture time and trigger timestamps, used auto trigger, and set the timebase to 50 µs per division. With normal triggering they collected 632 captures. The second trigger came 19 ms after the first. The third came 56 ms after the first, 37 ms after the previous one. That leaves 20 to 30 ms of dead time between captures, and anything that happens in that window is missed.
With the high-speed trigger, 1,000 captures were buffered almost instantly and then moved to the computer. Under the same conditions, each trigger followed the previous one by about 60 µs. With a 50 µs capture window, that means only about 10 µs of dead time. Ichiken concludes that infrequent waveforms are far less likely to be missed, and calls this the main strength of the feature.
Measurements, Rulers, Decoding, and Math Channels
Automatic measurements report individual waveform parameters on both the amplitude and time axes, across multiple channels. Measured values can be given pass/fail thresholds and linked to Actions, for example to send an external trigger, output decoded data, or save the waveform. A similar feature called DeepMeasure displays all waveform parameters at once, but it cannot be linked to external actions. Ichiken found it slowed things down somewhat, presumably because of the extra load on the computer.
The ruler function is what other manufacturers call cursors. Ichiken notes that on touchscreen scopes they often touch the wrong axis or the waveform when trying to grab a cursor. The mouse gives fine control, and they liked the feel.
Serial decoding is also free from the start and covers a full range of protocols, including CAN, LIN and I²C. When digital probes are connected, the digital inputs appear as selectable decode sources.
Math channels combine waveforms and apply functions: filters, buffers, trigonometric and scientific functions, and more. They can also plot quantities such as frequency or duty cycle as graphs rather than single numbers. In one demo, the measured blue trace was a signal that rises in frequency over time. A math channel applied a low-pass filter twice, and the computed result shrank in amplitude as the frequency rose. Ichiken points out again that this heavy processing runs on the computer, not the oscilloscope. If operation becomes sluggish, the PC is the cause.
A Software Update Mid-Shoot
A software update arrived during filming, announced by a red circle at the top right of the screen. After installing it, Ichiken found several additions. A new "Default Setup" button appeared at the top. isoSPI was added to the serial decoders. A new overlay display stacks all captured waveforms on top of each other to make glitches easier to spot. Actions were also improved. Previously you couldn't choose event details, such as which automatic measurement should fire an action. After the update you can choose which measurement triggers it. Ichiken considers this a significant improvement in usability.
Verdict: Best Suited to Low-Voltage and Embedded Work
Ichiken admits they were initially unsure about an oscilloscope that has to be connected to a computer, but found it was not a problem once they got used to it. Their overall judgment is that this is not a product for power systems. It fits embedded development and other relatively low-voltage work.
The biggest positive for them is that the fully featured software is free. They note that on other oscilloscopes, decoding the protocols commonly used in embedded development sometimes costs extra. They understand manufacturers are commercial businesses but still find that a bit much. Frequent software updates add a sense of security, and because the scope already works with a computer, moving waveform data around is fairly seamless. That compares well with the usual routine of saving waveforms to a USB stick and carrying them to a PC.
The main drawback is unavoidable: you need a computer. Ichiken says they would like a large workbench, since boards and other equipment are usually there too and a computer makes the bench more cluttered.
This video is brought to you courtesy of Pico Technology. Hello. Today we have an oscilloscope called PicoScope from Pico Technology in the UK. I'll review it. This oscilloscope has no display on the main unit. It's the type that connects to a computer with a USB cable. Today we'll look at the hardware differences, actual use, and differences from other oscilloscopes.
Pico Technology is a British manufacturer specializing in PC-based instrumentation. In addition to oscilloscopes, they develop data loggers, network analyzers, and diagnostic tools for automobiles. As you can see from the other products, they're all designed to be connected to a computer. The one we'll be using this time is an 8-channel oscilloscope. For more information about the products we use today and other oscilloscope products, the links are in the description. If you're interested, please check them out. Let's start using it.
Let's start with the exterior. With no display, it's a lot smaller. It only needs the basic waveform acquisition function and communication with a computer. Does that mean it can be made this small? The exterior depends on the model. It's basically made of plastic, but the PicoScope 6000 series we'll be using emphasizes cooling, with part of the housing serving as an aluminum heat sink. Let's look at this in a little more detail.
The front has eight analog channels. There are two input resistances: 1 MΩ and 50 Ω. This is a common configuration for mid-range to high-end oscilloscopes. On other manufacturers' oscilloscopes, the maximum input voltage for the 1 MΩ input is often 300 V RMS. For PicoScope, the maximum is ±20 V. Compared with other manufacturers' oscilloscopes, that value is quite low.
Regarding the analog channels, A, B, G, and H here have normal BNC terminals. The middle four have BNC terminals that can supply power to active probes and exchange signals. If you look at Pico Technology's website, there are active probes that can be connected. Looking at the lineup, I thought there were a lot of products for viewing relatively low-voltage analog signals and digital signals. But if you look closely, they also have clamp-on current probes and high-voltage differential probes. Power electronics measurements could also be made.
Here's the digital input. Digital inputs are available in addition to analog inputs. These digital probes are optional. Two 8-channel digital probes can be connected, making 8 × 2, or 16 channels. Four channels' worth of passive probes are included, with a frequency bandwidth of 500 MHz.
Let's look at the back of the unit. There are trigger input/output terminals, a 10 MHz input for synchronization, and a function generator output terminal. These are fairly common specifications. The rest are the power input and a USB connector for connecting to a computer.
The power supply is an AC adapter. Many USB oscilloscopes and modern small oscilloscopes use external power supplies. With an AC adapter, connecting this area is troublesome. I'd really like to see a built-in power supply, though I think that's difficult because of the design and so on. By the way, lower models such as the PicoScope 2000, 3000, and 4000 don't require an AC adapter. They operate on USB bus power alone. In other words, all you need is a computer.
Now we'll turn on the power. In addition to the oscilloscope itself, a computer is required to run the software. Launch the dedicated software and connect the USB cable to use it. Let's do that. Connect the cable. It's displayed. Press here. Once the oscilloscope is turned on and started, it's ready to use. Startup is quite fast, too. Within about 10 seconds of connecting it, you can use it.
The software is important when using this oscilloscope. It's free to download, and all functions are available free of charge from the start. No additional licenses are required. For this reason, the available functions are basically the same across all PicoScope oscilloscopes. What changes is the hardware performance: the number of available channels, sampling rate, memory capacity, and so on. The software itself is updated fairly frequently, which makes me feel more reassured as a user. Some companies don't update their software very often.
From here, we'll look at specific operations and other details. You can see the screen now. The UI seems to be designed primarily for mouse operation. The channel section is here on the left. Clicking it opens various settings. By default, channel 1 appears to be on. If you also turn on channel 2, it appears like this. If you grab an axis point like this, it looks like you can move it.
Here in the channel settings, there are manual and automatic options. Apparently, these select automatic or manual range adjustment. I haven't seen many oscilloscopes that automatically adjust the amplitude range for each channel. PicoScope seems to do this automatically.
Let's add more waveforms. When you add them, it looks like this. All channel waveforms are currently displayed on one screen. That's a little hard to see, so we'll add a waveform window. From Views at the bottom left of the screen, add a scope. It can be displayed like this. Under management, you can show only A and B in Scope 1, for example, and C, D, and E in Scope 2. You can do that, too.
The ability to duplicate waveforms and display them in separate windows isn't very common on other oscilloscopes. Is this possible because everything is handled by software on the computer? It's quite user-friendly, or perhaps expandable. I think it's very flexible. Right now there are four waveforms arranged vertically. You can also tile them like this. There's quite a lot of flexibility here.
As I mentioned earlier, you can duplicate all the waveforms and display them like this. There actually aren't many oscilloscopes that can do this. Some might be able to imitate it with the MATH function, but that might slow the oscilloscope down. You can also move the waveform into a separate window like this. With multiple displays, you can move that window to another display.
I have three scopes right now. What happens if we keep adding more? This is a little slow. I think that comes down to the computer's specifications. I now have eight scope windows open on the original screen. At nine or more, a new window appears to open. As you add more scopes, windows pop up like this. As before, you can rearrange the waveforms and tile them.
Right now, this scope only has one tab. You can put this into one scope, for example, as shown here, or add multiple tabs. You can view different waveforms in separate tabs. With the waveforms in windows like this, what happens if we increase the number of scopes further? At 17 or more scopes, another window opens. That's how it works.
However, displaying a lot of waveforms like this makes it slightly sluggish. For this operation, the oscilloscope only acquires data and sends it to the computer. All the waveform data processing is done on the computer. If the waveform display is slow, for example, I think that depends heavily on the computer's specifications.
Let's look at the basic usage. In the upper left of the screen, there's an area where you can adjust the time axis, sample rate, and so on. Click there to adjust the time axis. You can change it with a click here. I just clicked to open it, but you can also hover the mouse over it and turn the wheel to change the timebase. On a benchtop oscilloscope, you'd turn a physical knob to adjust the timebase and amplitude. This feels quite similar. The responsiveness of the waveform display seems fine, too. It's running quite smoothly.
Next to the timebase is an indication labeled "sample." Let's look here. You can choose whether to prioritize memory length or sample rate when measuring.
Next we'll adjust the input range. I touched on this earlier. Start on the left side of the screen. You can adjust the voltage range using these buttons, or turn the wheel here. It looks much more sluggish. The voltage range can be adjusted this way. Below that is the coupling mode: AC, DC, and DC with a 50 Ω input. As we discussed earlier, these are fairly common coupling options on mid-range to high-end products. Then you can choose the probe type and adjust the attenuation ratio.
Next is the DSP section. You can turn the low-pass filter on and off. It's working quite well. Below that is something called resolution enhancement. In addition to the resolution of the unit's built-in A/D converter, software signal processing provides an effective increase in resolution. Let's see how well this resolution enhancement works.
I'm at 8 bits right now, and only the top of the waveform shown here is magnified. As you can see, the colors of the waveforms and so on are quite different. It's relatively easy to do. Using a mouse feels much easier. As we increase the resolution from the current 8 bits, signal processing can take it to 9 or 10 bits, and apparently up to 12 bits. Up to four additional bits are available. Just a word of caution: it's not the same as the original signal. This is only a resolution improvement through signal processing. Keep in mind that the resulting waveform differs from the original.
Let's also look at improving the resolution of the hardware A/D converter. Just above this section, it's currently set to 8 bits. If we set it to 10 bits, the quantization errors largely disappear. At 12-bit resolution, the quantization error is barely noticeable. You can also enhance the resolution through software here. This is how it works. It seems to go up to 16 bits.
There's also this Display section. You can rename waveforms and change their colors. That's about it. I imagine some people will use these features. Changing the name of a waveform is not a good idea. I use this one quite a bit. You can edit it from here.
We've looked at the operation so far. Normal oscilloscopes have dedicated adjustment knobs and physical buttons. With the USB oscilloscope I'm using today, everything is controlled with a mouse or keyboard. Some parts are slightly frustrating, but I suppose it's a matter of getting used to it.
You can also zoom waveforms in and out without adjusting the input amplitude range or timebase. Turn the wheel here to zoom in and out. The same is true for the time axis. You can zoom in and out like this. Quite a few oscilloscope touch panels aren't very responsive. Sometimes you spread your fingers on the screen, but the operation doesn't work as intended. For this kind of waveform manipulation, a mouse feels easier to use. The other features are all in this Tools section. We'll check those later.
With PicoScope, the software is the same even when the hardware is different. The advantage is that it's easy to build the same environment even if the oscilloscope hardware changes. From this Save section, you can store various data. There's something called a PS configuration file, which saves the unit's settings. That's quite convenient if you want to use the same measurement settings on other devices.
You can also use multiple PicoScope units with one computer. I have four connected now, as you can see, and everything is working fine. For example, you could put several USB oscilloscopes on a production line and look at the signals coming from the equipment. If a signal fails the mask test, reject it; if it passes, mark it as passed.
Let's try it. Select a mask from the Tools section. You can automatically create a mask from the signal currently being acquired. Doing that creates a mask like this. Right now, all the waveforms pass through the white part of the mask, so they're passing.
How do you use this mask to trigger some external action? Under Tools, there's an option called Actions. Press it to produce an event. You can do various things, such as triggering or launching an application on your computer to run something. You can also output decoded data. This time, to make it easy to understand, let's play a sound. Some sound like explosions. Let's use this one. If you pull the signal with this, it sounds like this. I think it plays the sound once for each failure being acquired. There you go.
It now shows nine failed attempts. A history is retained so you can see what kind of signal failed. Doing this is pretty easy. I think it's quite good. What do you think?
Having used PicoScope so far, I think quite a few of the oscilloscopes I've reviewed as Ichiken have been hardware-focused products. PicoScope feels more software-oriented. How easy the software is to use and how fully you can use it seem to be the keys to making the most of PicoScope.
It can be used for embedded applications in research and development, for example. Since it's used with a computer, I think it could be particularly strong in applications that take advantage of that, such as automatic measurement. SDKs and sample programs for embedded applications and automatic measurements are all available free of charge. If you're interested, please check them out.
By the way, even if you don't have a PicoScope unit, the software installed on your computer has a demonstration emulator. You can check how it operates in advance without an oscilloscope. It starts up like this. You can generate various demo signals, for example. All the features are available here, too. You can read waveforms like this even without an oscilloscope. Open the Tools section to try all the features available in PicoScope. This lets you check what it can do without having an oscilloscope on hand. You can download it from the link in the description. If you're interested, give it a try.
By the way, if you connect a touchscreen LCD, touch operation is also available. However, PicoScope seems to be optimized for mouse operation. Compared with other manufacturers' touchscreen oscilloscopes, the feel of the controls, or ease of use, isn't very good.
I have no problem with things like response time. Compared to a mouse, there is more movement, and there will be operations that cannot be done in the first place. For this part, I'm going to be quiet. It would be better to operate it with a mouse and keyboard.
From here we will also look at oscilloscope performance. PicoScope itself has various lineups. We will be using the 6824E. Frequency bandwidth is 500 MHz, and sample rate is up to 5 GSa/s. Resolution depends on the number of channels used: 8, 10, and 12 bits. We saw a little bit about resolution earlier. By processing on the software side, an additional 4 bits are available.
I'll actually try to do it while looking at the screen. Here at the top of the screen, it's called hardware resolution. There are three options to choose from: 8, 10, or 12. 8-bit and 10-bit can be used for all channels at the same time. In the case of 12-bit, only two channels are available, as shown here. They cannot all be used at the same time.
I'll be honest about this. I think it's a disappointing specification. Since we can use 12 bits, I think it would be good if all channels could be 12-bit at the same time. Therefore, the amplitude resolution of this oscilloscope is basically 10-bit. By the way, another model has a frequency bandwidth reduced to 20 MHz, but an 8-channel 12-bit model is also available.
The rest is about sample rate. This is up to 5 GSa/s, but as the number of channels and bits used increases, the sampling rate decreases. For example, if the resolution is now 12 bits, in this case the sample rate is 1.25 GSa/s. If the number of channels is increased to two, it's unchanged at 1.25 GSa/s.
When the resolution is set to 10 bits, the sample rate is now up to 5 GSa/s. This is when only one channel is running. If you increase the number of channels to two, it goes down to 2.5 GSa/s. Maybe that means up to 5 GSa/s for all channels combined. And depending on how much data you have in memory and so on, I guess that means it's going to change.
2.5 GSa/s if you increase the number of channels. I'd increase the number of channels you're using to two: 2.5 GSa/s. With three it goes down to 1.25 GSa/s. It's the same up to four. Add more channels, and at five channels or more it drops to 625 MSa/s.
Shall we set the next resolution to the lowest, 8 bits? Maybe with no more than two channels, I think 5 GSa/s would work. That's right. I'm running two channels simultaneously, and 5 GSa/s is working well. If the number of channels is increased to three, it's reduced to 2.5 GSa/s. Even with four, it's 2.5 GSa/s. If you use more than five channels, the sampling rate goes down to 1.25 GSa/s. If you run all eight channels of this, it's 1.25 GSa/s.
For more information on this area, I'm just wondering if there is an oscilloscope for the split. When a manufacturer increases the number of channels or the number of bits, the increase or decrease of the sample rate varies quite a bit depending on the design philosophy. If you change the number of channels and the number of bits, I hope they always maintain the maximum sampling rate. But if you actually do that, it will be considerably more expensive. I think that's something to consider as well. Depending on how you use it, there is a possibility that we may run out of bandwidth. Be careful.
Next we will look at the noise floor. The time axis should have a range of 10 ms. How is it measured in terms of amplitude? There was automatic measurement. If we choose the peak-to-peak and ripple RMS values here, I see it comes out this way. Now coupling is 1 MΩ input. Peak-to-peak would be about 2.1 mV. The RMS value of the ripple is about 200 μV. I can see averages, etc. I'd say 2.2 mV. Also, the ripple RMS value is about 200 μV.
I just checked. With this hardware resolution, the noise level is almost exactly the same. It would be good. The next step is to set the input impedance of the oscilloscope to 50 Ω. With terminating resistors, we take measurements. For peak-to-peak, I would say about 1.8 mV. The ripple RMS value is about 170 μV. For this noise floor, the value is relatively average.
From here we will also look at other features in oscilloscopes. First is the signal generator. Shall I try to connect the output of the signal generator on the back side to this input? From the oscilloscope operation screen, at Gen, you can do it this way. The waveform is displayed. Now this is being done manually. If you set it to automatic, it can automatically match it.
Here, the frequency and amplitude can be changed, along with the waveform type. Various waveforms can be made here. White noise is also available. Good. There are approximate waveforms. The other thing is this AWG. Press this to create your own waveform. There's also waveform import and a waveform drawing function. For example, like this. Now press Apply. The waveform is coming out as I just drew it.
In addition to this, you can even create something like a digital waveform. For example, if you type 0101 like this and apply it, you see the digital waveform you just created here. You can make some kind of digital signal, for example for testing decoding. It can be used in such places.
I'll come back to this later. Using the math channels here, you can make some waveforms, and the waveform can also be output. For example, I just created this red waveform. This can also be output from the signal generator. Choose it this way. This is output from the signal generator, and the waveform you are actually seeing is on the analog channel of the oscilloscope. Looking at amp waveforms or measuring the characteristics of operational amplifiers—it looks like you can do a lot of things.
Earlier in this View section, I did the scope display. We haven't done a spectrum display yet. This is the one that does the frequency analysis of the waveform. As you can see, the top is the input waveform. Below are the results of the frequency analysis. If we take a longer time range, frequency analysis results are available. It's a 10 kHz triangular wave right now. The peak appears just at 10 kHz. There are peaks elsewhere.
I'm sure the computer can add a lot of features. For the PicoScope people, for example, you could display waveforms in a waterfall. I think it would be interesting to do something like that.
By the way, it was running in oscilloscope mode. From Instrument here, you can switch to Spectrum mode. What's the difference between the spectrum we saw in the previous view and Spectrum selected under Instrument? You can choose the mode from this side of the instrument. I'll make it Spectrum. This mode gives priority to spectrum analyzer operation. The top part here was the part about adjusting time earlier. In Spectrum mode, it adjusts the spectrum. The frequency display range can be changed in this way.
And then there's the number of bins. This can change the number of points to be measured. The fewer these bins are, the faster the operation is, but frequency resolution will be reduced. Conversely, the number of bins can be increased up to about one million points. This way, the time for the FFT will be very long. So if you set a small number of bins, it works rather crisply. I have the impression that it will be stress-free to use.
Still from View, adding a scope lets you understand which waveform's frequency you're looking at right now. I think this is a pretty useful feature. In this waveform part, the waveform and spectrum analyzer results can be viewed like this. I'll explain the detailed operation of this part later. From the waveform, you can also look at the results of frequency analysis. Conversely, if you click on this spectrum, from the results of frequency analysis you can see what waveform it is. You can go back and forth in both directions.
In addition, the spectrum analyzer results can be masked. Here's the situation. For example, here and now, there are some areas that are not masked. If the spectrum rises here, we can stop capturing and see that waveform. Now you can actually move it around. Since the spectrum has just risen at the mask, capturing stops there, and then you can see the waveform. This can also be triggered from the actual spectrum. I think it's pretty convenient.
Next we will look at triggers. I had touched on this lightly earlier, so we're taking a closer look. In this trigger type section, there are standard edge triggers, runt, and logic. Is this what logic looks like? I see. PicoScope itself can do serial decoding. Triggering from the results of that decoding doesn't seem to be supported at this time.
The other thing I'm curious about is this high-speed trigger. First of all, with a normal trigger, when the waveform coming into the oscilloscope triggers it, the waveform data is sent to the computer as it comes in. If you press the button next to the trigger called Waveform, you can see the history of the waveforms sent so far. You can see it this way. The history of this waveform is stored in something like a ring buffer inside the oscilloscope, and that's what we're seeing right now. In this state, we now have 64 waveforms stored in the ring buffer.
With a normal trigger, sending the acquired waveforms to the computer eats up a lot of resources inside the oscilloscope. You may miss some waveforms that appear infrequently. If you want to trigger on such infrequently appearing waveforms, cue the high-speed trigger. The fast trigger is pressed like this. Now the waveform is being buffered in the memory inside this oscilloscope. When that memory is full, the waveform data is sent to the PC in a batch. In this way, even waveforms that appear infrequently can trigger it. We'll see how much this reduces the number of missed waveforms.
In the Waveform section here, turn on capture time and trigger timestamp. The trigger setting is automatic. The time axis is set to 50 μs per division. We will now take the waveform. If you do, you start getting waveforms like this. When they accumulate, press Waveform. Then the sequentially stored data is displayed on the computer like this. I've done it 632 times.
Let's see the first one. The first trigger is indicated as 1, and 2 is the second trigger. Then 3, 4, 5, 6, and so on. If we look at this waveform, we are now taking waveforms for a period of 50 μs. But 19 ms have elapsed since the previous trigger. The third trigger is 56 ms after the very first trigger, and 37 ms have passed since the previous trigger. Therefore, between triggers, a period of inactivity of 20 ms to 30 ms occurs. During those tens of milliseconds, if the waveform you wanted to take was coming, that means you missed it.
Here's what would happen with the fast trigger I mentioned earlier. We're going fast here. And again. That was over in an instant. I'm going to do it one more time. Watch this waveform closely. Now a buffer for 1,000 waveforms is saved all at once in the oscilloscope. When that measurement is done, the data is moved to the computer.
The measurement conditions other than the trigger are the same as before. The time axis is looking at a period of 50 μs. If you look at this second trigger section, the second trigger occurs after 60 μs has elapsed since the previous trigger. How about the third trigger? Again, it triggers after 60 μs has elapsed since the previous trigger. Therefore, the period when the waveform is not being taken is 10 μs. For this reason, compared to a normal trigger, you're less likely to miss waveforms that appear infrequently. This is the power of the high-speed trigger.
We will also look at other tools. You can see many things from the tools on the left. For example, do you see the automatic measurement here? I touched on this briefly earlier. The idea is that you can look individually at the information the waveform has: amplitude, time axis, and measurements when multiple channels are used.
Incidentally, if the value measured by this automatic measurement exceeds a certain threshold, you can make a pass/fail decision. If you link it to the action we just tried to use, you can output a trigger externally, output decode data, or save the waveform.
A feature similar to the automatic measurements we have just seen is called DeepMeasure. This has no function for linking to external triggers, etc. It displays all the information the waveform has. Using this will slow things down a bit. Maybe the workload on the computer side is getting heavier.
Then there's the ruler function. This is what other companies call the cursor function. On the touch panel, when choosing these details, I often touch the axis or waveforms by mistake. The mouse can be used to control quite a lot of details. I think the feel of the operation is rather good.
Serial decoding is next. This is all available free of charge from the start! CAN, LIN, and I2C—we have a complete line of such protocols. When a digital probe like this is connected, digital inputs are available for selection as the decoding input channel.
The last one is the math channel. This can perform calculations on waveforms together, use a waveform as a source, and multiply it by some function or something. You can do a lot of things like this. Filters, buffers, functions, trigonometric functions, and scientific functions—there are quite a variety of types. You can see the frequency, the duty ratio, and so on. This is not a numerical display. You can graph these frequencies. That's the way I see it.
Here's what we're doing now: a low-pass-filtered signal. And I'm not talking about applying a low-pass filter. I'm working on it. This blue waveform is the measured waveform, with the frequency getting higher and higher. The acquired waveform is low-pass-filtered twice. Then the computer performs the calculation. As the frequency increases, the amplitude gets smaller and smaller. We are able to compute such waveforms.
Such operations run on a computer. It's not like the oscilloscope itself has a heavier workload. But the computer being operated just became too slow. There are some things that make operation sluggish.
Regarding software updates, a software update actually arrived when I was shooting this video. It's this red circle in the upper right corner of the screen. The software update is now complete! Various functions have been added. We will pick a few and look at them.
For example, at the top of the screen, earlier there was no button called Initial Settings. It's newly added. This is the default setup. For another example, if you look at the serial decoding section, isoSPI has been added; it was not there earlier. To make it easier to find glitch waveforms, an overlay has been added to this waveform area. You can overlap all the waveforms you take like this.
Furthermore, this action part has improved quite a bit. When executing the action just now, you didn't have a choice of event details. When the action event was set to the result of an automatic measurement, you couldn't actually choose the automatic measurement item. After the update, you have a choice of which one to trigger on. You can choose an automatic measurement as the trigger in this way. This is pretty good. It greatly improves usability. I think it's a pretty good update.
So today I reviewed the PicoScope, an oscilloscope from Pico Technology. Oscilloscopes that are connected to a computer?
At first I wasn't so sure. Once you get used to it, it's not a problem.
As for my overall impression of this USB oscilloscope, it's not the kind of product you would use in a power system. I thought it might be a product suitable for embedded development and other areas where the voltage is relatively low.
The good news, after all, is that the software is fully functional and free of charge! With oscilloscopes, analysis of protocols commonly used in embedded development sometimes costs extra money. Oscilloscope manufacturers are also commercial companies, so I don't blame them. I sometimes think that it's a little bit like that, isn't it?
And also, the computer software is updated frequently. There is a sense of security in this area for the user.
It will be used with the computer, so I think it's nice to be able to exchange waveform data rather seamlessly. Often, waveform data is stored on a USB memory stick and taken to the computer.
And then there are the disadvantages. I need a computer for this. There's nothing we can do about that; it can't be helped. I wish I could secure a large workbench like this. In most cases, there are other boards with it, and it tends to be messy.
Last but not least, if you're interested in PicoScope and want to use it, please consider it!
If you found this video helpful, please give us a high rating and subscribe to our channel! If you have any concerns, please leave a comment as well! Thank you for watching the video until the end.
Article published · Updated