RoadTest: ADP3450 Analog Discovery Pro USB Oscilloscope
Author: Gough Lui
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?: PicoScope 3304D, 3305D, 3306D.
What were the biggest problems encountered?: Limited memory buffer size, long blind times, limited connection bandwidth, instabilities in Record Mode.
Digilent Analog Discovery Pro ADP3450 USB/Ethernet Mixed Signal Oscilloscope RoadTest Review
By Gough Lui
November 2021 – January 2022
Happy New Year fellow element14 Community members and welcome to the first RoadTest I’ve delivered under the new Verint-based platform. Please bear with me in case of any formatting or presentation issues – I’m still working my way around the new layout. This comes delayed as I've hit a bug with the RoadTest form submission that needed to be worked out ... or worked around. Nevertheless, the review is now live!
Towards the end of last year, I was fortunate enough to be selected for the Digilent Analog Discovery Pro ADP3450 USB/Ethernet Mixed Signal Oscilloscope RoadTest. Those who may be familiar with the original Analog Discovery and Analog Discovery 2 would know these devices to be an educator’s favourite. The ethos behind those devices was to offer basic but usable instruments for the cost of no more than an average textbook, enabling students to do laboratory exercises from the comfort of their own home and offering them an instrument they may find useful into the future. The Analog Discovery Pro ADP3450 seems to be the “grown-up” brother to the Analog Discovery 2, offering higher specifications, improved connectivity and autonomy with an increased price-tag, while retaining the use of the Digilent WaveForms software and its multi-instrument integrations that made the original so useful. As an educator and researcher myself, I felt it to be worth investigating to see how the ADP3450 works, how it performs compared to contemporary devices and its specifications, and how useful its Linux Mode can be.
Thanks to element14 and Digilent for choosing me as one of the RoadTesters for the ADP3450. I hope you find this review informative, useful and interesting. I would appreciate if you could leave a “Like” if you enjoyed this review, take the time to explore the detailed blogs if you would like more information on a specific area or leave a comment if you have a question. I will endeavour to answer any questions as soon as I can.
The Digilent Analog Discovery Pro ADP3450 is a four-channel mixed-signal (or mixed-domain) oscilloscope capable of analog and digital measurements in both time and frequency domains. At the time of publication, the ADP3450 is listed (on element14 Australia) for a price of AU$2150.83 excl. GST for the base kit or AU$2414.91 excl. GST for a kit including probes and test clips.
As a test instrument, it is somewhat unique as a multi-instrument device, consisting of four-channels of oscilloscope input capable of 125MSPS at 14-bits with 55MHz bandwidth, two-channels of arbitrary waveform generator output capable of 125MSPS at 14-bits up to +/-5V swing, 16-channels of digital I/O (not just input) with two trigger I/O ports (again, not just input or output). However, with these ports, it claims to offer much more, including the capability of being a network analyser, spectrum analyser, voltmeter, impedance analyser, data logger, logic analyser, pattern generator, protocol analyser and more through the Digilent WaveForms software with SDK. Furthermore, the instrument is connected by USB 2.0 and Gigabit Ethernet, with four USB 2.0 device ports to support peripherals such as supported Wi-Fi adapters for wireless connectivity and mass storage devices, plus the capability of “Linux Mode” operation independent of a connected host.
Compared to the Analog Discovery 2, the Analog Discovery Pro ADP3450 is more expensive, not as compact, and requires an external power supply. In return, it offers proper BNC ports for probes and wider bandwidth. Perhaps it may appeal to students who may be doing field experiments, require a flexible test device or do not have the budget for big-brand big-box units but may benefit from the tight integration of the multiple instruments. A key flexibility is the digital I/O being truly I/O, as it allows the unit not only to sniff, but also to participate in bus transactions. Another is the Linux Mode capabilities with Ethernet connectivity, allowing the instrument to essentially operate independently of the host – something very few instruments can do.
For true professionals, however, the specifications seem much less appealing – for example, the 55MHz bandwidth is nothing particularly exciting, nor is the 15MHz waveform generator bandwidth. The supported protocol decode is somewhat lacking compared to some other options on the market. The buffer memory (e.g. oscilloscope – 32kS per channel ordinarily, 64kS per channel in extended, 128MS across all channels in record mode) is particularly and the input has only two “real” voltage ranges, although the 14-bit resolution is quite an improvement over the 8-bits offered by others. The ADP3450’s specifications also lack some important values – for example, input cross-talk figures, input noise-floor, effective number of bits, channel-to-channel skew, etc.
The market for USB-connected host-driven instruments is relatively small and fragmented, with many attempts but very few successes, especially in the mixed-signal/mixed-domain area. Perhaps the closest contender is the Picoscope 3404D MSO (AU$1948.95 excl. GST) which offers four channels of analog input at 8-bit 1GSPS with 70MHz bandwidth, 16-channels of digital input, a single 20MSPS 12-bit arbitrary waveform generator with >1MHz bandwidth and 128MS of buffer memory. It even comes with all probes included and can operate from USB bus power. For a little more, one could step up to the 3405D MSO (AU$2516.25 excl. GST) and have 100MHz analog bandwidth and 256MS of buffer memory or the 3406D MSO (AU$3303.15 excl. GST) for 200MHz of analog bandwidth and 512MS of buffer memory. The Picoscope is known for its wide protocol decoding support and has a faster USB3.0 interface, but compared to the ADP3450, it doesn’t have Ethernet, Wi-Fi, Linux Mode, pattern generator, digital I/O capabilities and does not integrate the use of the AWG and oscilloscope channels to perform impedance analysis.
For educators looking at all-in-one benchtop solutions, the latest NI ELVIS III starts at AU$5,588 without any accessories, which is quite a noticeable jump over the ADP3450. Another option is the Digilent Analog Discovery Pro ADP5250 which doesn’t have the USB-host ports, Ethernet or Linux mode, and has only half the oscilloscope, arbitrary waveform generator, trigger channels and digital I/O channels. In return, it adds on a digital logic analyser offering 32 channels, provides a “real” multi-rail power supply and a digital multimeter. Accordingly, it lists for about AU$4515 excl. GST which is also significantly more than the ADP3450.
The ADP3450 is thus quite a unique multi-instrument combination device. While its price is not as inexpensive as the Analog Discovery 2, it is still priced competitively once the number of instruments and specifications are taken into account. As purely a measurement instrument, the paper specifications may not appeal to the professionals, but it hides the fact that the ADP3450 is much more tightly integrated than other instruments thus offering features such as impedance analysis and is simultaneously more flexible in being able to operate independently of a host in Linux Mode with its SDK. It is also not only capable of digital bus sniffing, but can also perform true I/O and participate in digital buses, making it even more versatile than most digital-capable instruments. This makes it especially well-suited for educational, experimentation and hacking purposes. It seems to have an upper edge when it comes to ADC resolution, although this comes at the cost of sample rate and bandwidth and the WaveForms software may be an asset where users are already familiar with it from previous experience with the Analog Discovery 2. In the education space, it is cheaper than the other bench-top options while also being small enough to be portable, thus perhaps could make an ideal instrument to loan out to students for them to catch-up on labs at home. Likewise, final-year or research students without the budget for a plethora of big-box instruments could consider this as a piece of test equipment that may still be sufficient even after they graduate.
For more information, see Digilent ADP3450 In-Depth – Ch1: Market Survey & Feature Introduction.
The Digilent Analog Discovery Pro ADP3450 arrived in a colour-print cardboard box about the size of a mATX motherboard box. The packaging seems well thought out, clearly labelled and secure, ensuring a straightforward unboxing experience with no surprises.
The main unit itself feels quite light and clusters most of its functionality on the front and rear panels. The front is home to a power LED, digital I/O including digital I/O voltage output, four oscilloscope channels and two arbitrary waveform generator channels. The rear panel is home to the two trigger I/Os, four USB host ports, Gigabit Ethernet, USB device port, reset button and barrel power input. A power switch resides on the left-side panel, with passive cooling occurring using the vents on the bottom and side panels. The top panel is solid with recesses for feet, suggesting it is intended to have other equipment potentially stacked on top. The bottom panel also features flip-out feet which prop up the front of the unit for easier access.
Accessories included the digital I/O break-out cable, a 65W grounded Adapter-Tech power supply with ample power reserve and US/EU mains leads. For those in other regions, they may have to supply their own mains power cable. Unfortunately, not being bus powered makes this unit a little less convenient when out in the field due to the need for a separate supply of power. The unit usually does not include probes, however, for the purposes of the RoadTest they included four P2150 150MHz passive 1X/10X switchable probes. These were Ypioneer branded and appear to be a Chinese generic product. While serviceable, the probes felt a little rough around the edges, the probe tip was not as fine as some other probes and the test hook is a little chunky.
For more information, see Digilent ADP3450 In-Depth – Ch2: Unboxing.
Initial setup of the ADP3450 involves both software and hardware elements. The software must first be obtained from their website, which requires filling in a form that asks quite a number of questions. I feel this is a minor inconvenience, however, to their credit, the software is available for Windows (32/64-bit), Mac OS X and Linux (32/64-bit, x86/ARM, deb/rpm) making it one of the most widely compatible pieces of test equipment software I have seen. Installation on Windows is straightforward, while Linux requires installing Adept Runtime before WaveForms and then fixing missing dependencies.
Hardware set-up includes plugging in the power and turning on the unit. The probes have to be labelled for each channel and probe compensation should be trimmed. Unfortunately, this is where I noticed that the ADP3450 does not have a probe compensation terminal. While the AWG could potentially generate the necessary signal, it would be unwise to directly probe the BNC socket, so perhaps a fixture or a terminal would be a good addition.
Software configuration then follows, which includes updating the device firmware and then setting up the operating mode and memory buffer split which needs to be done over USB initially (as a security precaution). Standard USB or Ethernet allows for host operation of the instrument, while Linux mode (covered later) allows the instrument to operate independently. Network configuration, reference frequency and authentication configuration should also be completed as well. Unfortunately, I didn’t have as much luck getting authentication working, which apparently only works in Linux mode. Unlike some other devices, the configuration also allows for calibrating the device using a trustworthy DMM in case the factory calibration is not good enough. Various user interface, graphics and device options can be configured, while a speed test and status read-out serve as diagnostics.
The software itself is very powerful, but also can appear cluttered and intimidating at first. There are a number of icons for various features which may not be obvious at first, and layers of sub-menus with the ability to break windows into multiple views and run multiple instruments simultaneously in separate tabs. I appreciated the consistency of the software across platforms, which performed virtually identically. The unit itself was silent as it is passively cooled, with the exception of loud, distracting relay clicking upon initialisation at start-up.
Product documentation consists of the quick-start postcard provided with the device, help within WaveForms which provides an overview of the architecture and instruments in the ADP3450, the Digilent Reference which is akin to a “wiki” style website providing a reference manual, specifications and a growing list of tutorials for the device. I felt that the documentation was a bit lacking with some omissions, ambiguities and occasional errors. Finally, support can be found in the Digilent Forums where the Test and Measurement subforum has a number of very active employees that seem to be highly responsive to questions and also build beta-versions of WaveForms to resolve problems.
For more information, see Digilent ADP3450 In-Depth – Ch3: Setup of Waveforms & Documentation.
The Digilent WaveForms software offers modules including Scope, Wavegen, Supplies, Voltmeter, Logger, Logic, Patterns, StaticIO, Spectrum, Network, Impedance, Traver, Protocol and Script. Each of these modules open in their own tab or window with multiple instruments allowed to run concurrently with the ability to freely resize to the screen and invoke multiple sub-window views. All of this is a rather positive design choice, as it makes for efficient and highly flexible use of screen real-estate. It can, however, be cluttered as a result of the flexibility with options are hidden behind text menus and smaller inset icons which are less intuitive but I think they have done a reasonable job of balancing this. Performance sometimes suffers and I also have had a number of application crashes where the program would just “disappear” and sometimes the instrument will complain that it is “in use” and will need to be restarted to regain access.
When it comes to the modules and the ADP3450, the results are quite a mixed bag. The Scope module for example is highly configurable and offers plenty of features, but the ADP3450 itself proves to be a limiting factor as its 32,768 sample buffer and relatively long blind-time make for a limited zoom capability. The limited sample rate also affects clarity for higher-frequency signals, while oversampling for repetitive signals can take quite a bit of time and computational resources. The Record Mode which offers larger record lengths proved unstable for me with sample loss errors. The module does not offer digital decoding from analog channels, while increasing the number of sample history buffers seems to be a computer-side options, thus cannot get around the blind-time issue. This is probably a limitation of the USB 2.0 and Ethernet interfaces as well, as the throughput only ranges from 40 to 53MB/s which is not sufficient for timely transfer of the samples.
The same sample buffer and blind-time limitations also affect the Logic module, which is otherwise well-implemented with SPI, I2C, UART, UART, USART (PS/2), CAN, I2S, 1Wire, HDMI-CEC and Manchester support out of the box. The Logic module works well, assuming you can capture the event of interest in the first place.
The Wavegen module is well implemented as well, with excellent choices of functions, sweeps, modulated signals, custom lists or playback of audio files. This reduces the need to create custom lists of values, however, the 32,768 sample capability and ~15MHz bandwidth do not measure-up to standalone instruments. The Pattern module is also especially well implemented with many pattern options of creating clocks, pulses, constants, random noise, counters and ROM logic functions compared with the pattern generator in my RTM3004.
Perhaps the biggest stars are the Network, Impedance and Tracer modules. All three show clever integration of the inbuilt instruments to achieve measurements which either are added-cost options on other equipment or require the user to program their own solutions. The Network module provides a bode plot/frequency response analysis feature which works well although may suffer from the wavegen’s limitations. The Impedance module provides the capability of an LCR meter and impedance analyser but requires user-provided sense resistors. Absolute accuracy is not always guaranteed as the sense resistor value can be critical, but frequency-dependent component behaviour can be measured which provides very valuable insights. The Tracer module allows for I-V curves to be measured on diodes, BJTs and MOSFETs using up to two user-provided sense resistors to emulate the behaviour of a curve tracer and also provides useful information on low-level behaviour of semiconductors. However, because of the limitations of the wavegen, the current levels and voltages are quite limited. While these instruments definitely don’t reach the capabilities of standalone instruments, I felt the utility they provided to be sufficient for basic needs and their inclusion is quite innovative.
The Spectrum module operates very similar to the FFT in the Scope with a few more friendly options around units and FFT bins. Unfortunately, as it is based on the oscilloscope inputs, it is only capable of about 50MHz thus is not a replacement for a dedicated spectrum analyser. It is still useful for lower-frequency work. The Static I/O module duplicates some functionality which can be achieved through the Logic module, but presents it in a more interactive and intuitive way in the form of buttons, sliders, virtual LEDs and 7-segment displays. This is great for educational purposes for reducing the amount of parts needed for certain lab exercises. The Logger module allows for measurements from the Scope to be recorded over time like a chart recorder and feels like a simple extension of the Scope capabilities.
The Supplies module was a bit of a let-down as the ADP3450 doesn’t have a dedicated power supply peripheral, thus only a single digital I/O voltage can be set from 1.2V to 3.3V sourcing up to 300mA and also affecting the digital input threshold voltage. The Voltmeter module also seemed a bit superfluous as the ADP3450 does not have a dedicated voltmeter, thus it is equivalent to using Scope with configured measurements.
As a result, I feel that the “pro” moniker perhaps may be a bit misleading. Compared to previous iterations of the Analog Discovery or other low-end starter instruments “of last resort”, the ADP3450 definitely is a notable step-up. However, compared to the mid-range entry-level standalone instruments or a similar equivalent from PicoScope, some of the ADP3450’s limitations become clear. While there are a few areas where the WaveForms and ADP3450 shine, the basic features of the ADP3450 are still relatively limited in part due to limited buffer memory, sample rate, analog bandwidth, connectivity bandwidth and hardware peripherals. As a result, I feel the ADP3450 will probably suit educational and relatively undemanding hobby users much more than it will suit professionals.
For more information, see Digilent ADP3450 In-Depth – Ch4: Tour of Digilent WaveForms.
The inclusion of Linux Mode comes courtesy of intelligently using the Xilinx Zynq SoC which powers the ADP3450. By reconfiguring it such that the unit boots from a 4GiB eMMC with the 512MiB DDR3L RAM buffer as the system RAM, it converts the instrument from a host-driven instrument to one that can host itself and run independently from a host. This can be considered similar to attaching a USB MSO to a Raspberry Pi, however, unlike such a solution, Linux Mode doesn’t mean that the instrument is dedicated to the Linux instance – it can still be used as a host-driven instrument over Ethernet or Wi-Fi, offering maximum flexibility. Key advantages include the ability to perform experiments and operate independently of a host, benefiting from higher throughput rates and lower latency. The main downsides are the loss of record mode sample rate – from 100/125MSPS to about 1MSPS, likely due to the loss of buffer memory, and the fact the instrument is now a computer node on the network so good security practices are necessary to avoid compromise. Careful use of the Linux Mode is necessary to avoid unnecessary wear to the eMMC which is not designed to be replaced – logging data to USB mass storage devices is highly recommended.
The implementation increases the flexibility of the ADP3450 quite dramatically. It is actually possible to even run a desktop environment, with a virtual display and VNC remote access to host a copy of WaveForms on the device itself. While performance is relatively limited, it demonstrates the capability of the solution. Other uses include installing Python for running WaveForms SDK scripts for standalone data acquisition and signal generation, Jupyter Notebook for web-browser based development, samba for Windows file-sharing, pyvisa and pymodbus for interfacing with other instruments that speak SCPI or Modbus-TCP protocols. It can even serve as a USB-TMC to Ethernet bridge, making it extremely versatile.
Linux Mode is easily enabled by changing the boot options and both SSH and a USB-Serial terminal is available for administration. In case of problems or an upgrade, the Linux Mode eMMC can be re-imaged. Linux Mode is Debian-based and runs on a 4.19.0 kernel, which is somewhat dated at this point, with the kernel family in long-term support until December 2024. The kernel appears to be built without support for certain devices (e.g. UVC, FUSE) which limits hardware compatibility, while sources do not appear to be openly available for users to build their own kernels. Perhaps the biggest limitation is the 512MB of system memory which can be limiting when a graphical environment and several servers are installed, so having a swapfile as a workaround is recommended for stability if such services are used. One quirk observed relates to the power switch on the side, which appears to be a hardware power switch that does not gracefully shut-down the Linux operating system which is not ideal for system integrity in the long-run.
The network connectivity by Ethernet is quite impressive for a device of this class, offering Gigabit Ethernet with an iperf3 measured performance of above 500Mbit/s in both directions. Wi-Fi connectivity is more limited in the adapters that are supported, with one of the claimed supported adapters (TL-WN722N v1 based on the AR9002U chipset) not working. However, rummaging through my collection of Wi-Fi adapters, I found adapters based on the RTL8812BU, RTL8188CUS, RTL8188EUS, RTL8192EU, RT2870, RT5370 and RT5372 were functional although your mileage may vary.
For more information, see Digilent ADP3450 In-Depth – Ch5: Linux Mode, LAN & Wi-Fi Connectivity.
The WaveForms SDK makes the ADP3450 a much more useful and versatile instrument. By enabling the possibility to automate the instrument through programs interfacing with the WaveForms SDK, it becomes possible to automate processes which would otherwise take lots of time to manually achieve. The SDK is automatically installed with every installation of WaveForms and supports C, C#, Python and Visual Basic.
It should be noted that their SDK is quite different compared to traditional SCPI commands and VISA automation, being a set of functions which are hardware-specific to WaveForms-compatible hardware. This may have advantages in terms of performance and ease of implementation, but I also feel it to be a missed opportunity to teach users the “standardised” SCPI as used by big-box instruments and may represent an element of vendor lock-in.
My experience with the SDK was aided greatly by the number of Python-based examples which covered many use cases, thus allowing more rapid implementation of programs that combine both WaveForms SDK equipment and SCPI-standard equipment in later experiments. Considering that I have never explored the WaveForms SDK, the learning curve was not too bad. Unfortunately, the documentation did have some minor inconsistencies with the online version of the SDK Reference Manual missing data relevant to the ADP3450 which was present only in the PDF document version of the reference. The sample code did occasionally contain errors, some files are hardware-specific, but all had limited comments and magic numbers scattered in the code, requiring furious reference to the documentation to understand the specific configuration value which is used. At times, these values are not as well documented in the reference as desired, leading to some experimentation being necessary. I also found that the SDK features heavy use of ctypes variables due to the interfacing with a (presumably) C-based library, which makes for unpythonic code clutter. Some examples also require the installation of some hefty libraries such as matplotlib and numpy to run.
However, it is still undoubtedly a very powerful API to have as it allows for the same code to automate the ADP3450, whether run “on-device” in Linux Mode, on host connected via USB, Ethernet or Wi-Fi without any changes. I found the performance of the API to be stable for the simple experiments I have conducted.
For more information, see Digilent ADP3450 In-Depth – Ch6: WaveForms SDK for Python.
I had the chance to apply the ADP3450 across multiple different application scenarios to learn how it performs. I started by replicating classic undergraduate laboratory experiments including the generation of Lissajous patterns which went relatively well. By choosing to have the instruments as free-floating windows, it was possible to configure the Wavegen outputs while watching the Scope. Unfortunately, the limited buffer and re-arm time means that the frame rate of the moving patterns was somewhat limited.
Running a Bode plot on a simple cascade of R-C elements went as well as one may expect, although did show the limitations of the noise-floor and dynamic range of the input. Curve tracing also went well with ordinary diodes, LEDs and MOSFETs, although was limited to lower voltages and currents. I found it useful as it could resolve the small-signal behaviour of MOSFETs and be potentially used to measure the small-signal gain of BJTs although I did not test the latter. The connections were somewhat laborious so I designed a breakout PCB for this, but unfortunately, it is still in the delivery pipeline. The Impedance module also proved very useful in measuring the self-resonant frequency for inductors and capacitors, which is something that is especially useful when trying to match the right components for modern switching converters.
Some features which would be especially useful in educational contexts include the audio capability which allows you to “hear” the input and record the input to an audio file. This functionality worked although required some careful choice of timebase if unbroken audio is expected with on-screen plotting. Another is the StaticIO module’s capability of emulating a seven-segment display which worked well although had slightly low contrast.
Host-less capabilities in Linux Mode were tested with a mains power cycle-by-cycle monitoring project which unfortunately did not capture any significant waveform anomalies during the monitoring campaign, but did illustrate the power of the WaveForms SDK automation examples in needing minimal modification to record into hour-long .wav files which could then be analysed offline on a desktop computer later. The availability of USB ports on the rear also increased the convenience of using an active high-voltage differential probe as it provided a convenient source of power.
Unfortunately, due to time limitations, I didn’t have as much of a chance to test the digital bus decoding features all that thoroughly. I tested I2C decoding which appeared to work fine, although on longer captures would complain of missing samples. I feel that the limited bandwidth and buffer size of the hardware again becomes a limiting factor, and continuous break-free digital bus decoding is probably better handled by dedicated instruments.
For more information, see Digilent ADP3450 In-Depth – Ch7: Using the ADP3450 in Various Applications.
I put the ADP3450’s specifications to the test in a barrage of checks. Testing the input accuracy of the oscilloscope inputs, I measured the voltage measurement error over 32,768 samples averaged. The results at 1V range show that Channel 2 seems to be a little mismatched with the remainder of the channels, however, all channels were able to fit within about 2.5mV of error which is better than the specifications imply, in part due to the averaging. At the 25V range, the error remained within 115mV which is also better than implied by the specification. Reading noise in the 1V range was about 20mV, likely in part due to the Keithley 2450 SMU’s output noise, while in 25V range this was about 110mV.
Oscilloscope input noise was measured with open-circuit inputs. In the 1V range, the channels had a mean peak-to-peak noise of 4.25mV and RMS noise of 0.53mV. In the 25V range, the channels had a mean peak-to-peak noise of 107mV and an RMS noise level of 13.4mV. In both cases, engaging the 20MHz bandwidth limiter had negligible effect on the noise level and both showed about 9-bits of clean resolution which suggests the full 14-bit resolution headline figure is probably not attainable. This is perhaps not unexpected as many oscilloscopes do have a smaller “effective number of bits” than their ADCs would imply and the result is superior to general-purpose 8-bit oscilloscopes. Given this design seems to prioritise signal quality (i.e. higher bits) over higher speed (i.e. higher sample rates), and with a limited number of true hardware ranges, it seems a little unfortunate that the noise level was not a bit lower. Input skew was immeasurable, likely in part due to a single system clock source and also the relatively “slow” clock compared to other instruments.
Input frequency response was within 1.8dB of each other when tested using its own Wavegen. Wavegen voltage accuracy was also good, besting its own specifications by having a measured error of about 10mV across the full range. Output frequency response was measured to be about 16.5MHz when into a high impedance load, besting the datasheet claim of >15MHz, but diminished when connected to a matched 50Ω input with a bandwidth of about 12.8MHz in the low range at 1V set amplitude and 10.4MHz in the high range at 5V set amplitude.
The digital I/O voltage output when unloaded had an error ranging between 8mV and 17.5mV low. When loaded, however, it dips around 128mV under the full rated 300mA load and continues to output even under 1A load and a 400mV voltage drop indicating there seems to be no over-current protection active. The digital I/O input threshold depends on the I/O voltage. Testing showed the threshold voltage uniformity across all channels seemed to increase as the voltage decreases, however, the size of the forbidden region increases as the threshold voltage decreases. The threshold voltages appear to be compatible with most logic families, and the values are roughly in line with those displayed in the WaveForms app.
The digital input switching rate was checked with an Nexys3 board programmed as a clock generator as per previous reviews. It seems that at the default system clock of 100MHz, the 50MHz signal is not reliably captured but a 25MHz signal can be captured with significant asymmetries. It seems the rule-of-thumb of sample rate should be five times as high as the signal frequency is still recommended, so with the 125MSPS sample rate, a maximum frequency of about 25MHz would be recommended.
Instrument standby power consumption with the power switch turned off was an excellent 176mW. Once the instrument is running in Standard mode, but idle, it consumes about 7.18W which is quite a bit more, but still very little compared to big-box instruments which usually consume three to ten times as much. The input voltage to the unit is supervised and needs to be within 19.3V to 20.3V for it to power up and within 19.0V to 20.4V to remain running, consuming about 0.32A to 0.35A when idle.
For more information, see Digilent ADP3450 In-Depth – Ch8: Instrument Performance Tests.
Looking inside the ADP3450 was an interesting exercise. The low weight implied that there would be quite a bit of empty space inside and this was borne out. Internally, the components predominantly came from Analog Devices which extends upon the tradition with the prior Analog Discovery and Analog Discovery 2 series boards, but this one does not seem to carry the co-branding. Regardless, the board was neatly constructed and well-labelled with the silkscreen which should aid future diagnosis, modification and repair if necessary. There are a number of buttons and jumper headers which are undocumented. Much of the ADP3450’s functionality is orchestrated by a Xilinx Zynq SoC which hides under a heatsink. The design has a surprising number of power rails, all derived by switch-mode conversion, occupying the outer edges of the board. Portions of the input matrix for the oscilloscope channels were shielded under a can, connected to ground via desoldering braid (which is an interesting approach). Pairs of oscilloscope channels share an AD9648 ADC. The wavegen output comes from an AD9717 TxDAC and AD5645R nanoDAC. The main boot device is a Spansion 128MBit Flash Memory, with an ISSI 4Gbit DDR3L memory buffer and a Samsung 4GiB eMMC for the main block device in Linux Mode. Various Microchip USB solutions are used alongside an FTDI FT232RQ which forms the USB device interface and a Texas Instruments DP83867CR Gigabit PHY Transceiver for the Ethernet interface. The design shows some special care taken to manage in-rush currents and power filtering.
For more information, see Digilent ADP3450 In-Depth – Ch9: Teardown.
The Digilent Analog Discovery Pro ADP3450 is an interesting multi-instrument device. Compared with the older Analog Discovery 2, its specifications and connectivity are much improved although it now comes at a price comparable to proper pieces of test equipment. For the price, the conventional choice may be a Picoscope 3404D which features analog channels with slightly more bandwidth (70MHz vs 55MHz), higher sampling rates (but at a lower resolution), equal numbers of digital channels, a single AWG channel (with lower bandwidth) and a similar amount of buffer memory with a faster USB 3.0 bus-powered interface. The PicoScope software is pretty good too, with a generous amount of protocol decoding. Despite this, ADP3450 is unique as it offers Ethernet connectivity and Linux Mode for autonomous operation, emphasising higher bit-depth on the channels rather than sample rate and offers digital I/O and trigger I/O capabilities which not only measures but also drives signals. Furthermore, the WaveForms software integrations enable more “virtual” instruments to be made, especially due to the presence of two wavegen channels. On paper, however, the ADP3450’s specifications seem incomplete and may not appeal so much to professionals, but the light weight, relatively compact footprint and reasonable price may make it suitable for educational circumstances where the unit is loaned out to students or may be used in field-work scenarios.
The WaveForms software is a key asset, bringing a lot of appeal as it runs on Windows, Mac OS X and Linux including ARM architecture. Its interface is customisable but consists of various modules including Scope, Wavegen, Supplies, Voltmeter, Logger, Logic, Patterns, StaticIO, Spectrum, Network, Impedance, Traver, Protocol and Script. The pre-defined ability to use the waveform generator outputs and oscilloscope inputs to perform curve tracing, impedance analysis and bode plots is rarely seen as standard features and can be quite an asset even for more advanced users. The user interface is very flexible with the possibility to configure multiple views, separate tabs into separate windows and resize them freely. Performance and stability sometimes suffer, and the flexibility also comes at a cost of visual clutter and settings behind menus or icons. Unfortunately, the ADP3450’s rather conventional hardware architecture does limit usability to some extent due to the 32,768 sample buffer and limited bandwidth which results in significant blind-time when operating in ordinary mode. This could lead to missing glitches and not being able to decode digital data continuously. Record mode at higher sample rates above 10MSPS in my experience was not stable and is not available above 1MSPS when operating in Linux Mode. The mixed signal capabilities were also hampered by the fact that digital protocol decoding is not available for analog channels, meaning that both an analog and digital channel will have to be connected in parallel if you want to examine the digital and analog domains simultaneously.
Linux Mode and the WaveForms SDK for Python were both impressive, as Linux Mode feels like having a Raspberry Pi “glued” to the instrument. This means that you can run code on the device without the need for a host PC, which can lead to energy savings and enable higher reliability. External USB devices can also be used including memory devices and supported Wi-Fi adapters for wireless connectivity. The unit is just powerful enough to host a desktop environment and run WaveForms on itself, being accessed remotely by VNC, although performance and memory limitations are quickly apparent. This does open up the possibility of working with the ADP3450 from a device like a smartphone. The WaveForms SDK was very useful with many Python-based examples that were adaptable for further experimentation. Unfortunately, the sample code had fewer comments than I would have liked to see, had a number of “magic values” and contained occasional errors. The online documentation was not up-to-date, while the latest PDF file included with the SDK still had some areas where values were not comprehensively documented or contained minor errors. Unfortunately, it doesn’t utilise SCPI, so users will have to learn the SDK which is specific to devices supported by WaveForms to get the most from the device.
Instrument performance tests overall revealed good performance with the factory calibration that often bested datasheet values, although some channel to channel mismatch was identified. The inputs were measured to have around 9-noise-free bits, which is superior to many ordinary 8-bit inputs, but is perhaps a little short of expectations given the 14-bit specification. The digital I/O voltage output also did not appear to be current limited, but had high setting granularity and reasonable accuracy. Waveform generator output bandwidth exceeded specifications when unloaded, but when terminated to 50-ohms, fell below specifications ranging from 10.4 to 12.8MHz. Standby power consumption was excellent overall and the unit shuts down if the input voltage goes out of a set range. A teardown reveals the inside to have a lot of empty space, but the tradition of using Analog Devices and Xilinx components continues. Construction quality was great, although it is perhaps the first time I’ve seen desoldering braid used to connect a shield to ground.
It’s hard to come to an overall verdict on the ADP3450, because as an oscilloscope and logic analyser, the limited buffer memory and bandwidth leading to lengthy blind-time means that it is not as capable as some other dedicated instruments when it comes to catching glitches or performing long decodes. But the ADP3450 is a multi-instrument device and what it does well is to integrate the functionality to create useful functionality out of it especially in the Network, Tracer and Impedance modules. Some of the functionality is executed well, such as the Pattern module which provides more features than most pattern generators and StaticIO which seems deceptively simple but makes I/O more intuitive. Instead of having a digital input only, the I/O capability opens up additional avenues of flexibility and it is also capable of autonomous operation with connectivity over Ethernet which is still relatively uncommon for most host-driven instruments. I think that the instrument will appeal most to educational and relatively undemanding hobbyists, but it seems to be an instrument that can have a myriad of uses even if one does outgrow some of its features.