Hands-On Survey of PC Scope Modules--Test Tools or Tech Toys?

Up & Running in Two Minutes

First and foremost, all three modules are easy to connect up and get running. I was apprehensive because loading new software often produces some level of aggravation. Either software doesn’t run first time or you mess up something that’s working fine. (My PC is a rather old 150-MHz model that is packed with stuff, although it does have a 128 Mbyte RAM.) I was even more concerned on this occasion because of possible hardware conflicts. The Pico and TiePie units link to the PC’s parallel printer port (and I only have one). The Cell unit links to the RS-232C serial port, and my PC has two, but they are already committed to the mouse and my Psion organiser. In the event, and for each case, I simply replace either my printer or Psion cable for a cable to the DSO. No conflicts occur, everything works fine, and I even see a sensible message in all combinations if I try to use the right software to operate the wrong hardware.

Because my PC has just one parallel port, as do many PCs, and having it occupied by the cable to the DSO module, means you have to swap cables from the Pico and TiePie units each time you want a printout. Using a second serial port, which many PCs have as standard, makes Cell’s module more convenient for printouts.

Despite that inconvenience, I swapped cables indiscriminately throughout the course of the checks—without powering down—and throughout my testing nothing went phut.

It’s Intuitive, of Course

There can’t be many PC user-interfaces that vendors don’t claim to need only intuitive operation. If a user interface is really intuitive, then I expect to be able to operate basic functions without much figuring out, and certainly without a users’ manual. Nobody has time today, especially using a scope, for consulting handbooks or help pages. Several days, weeks, or months may pass until you need the DSO again and you don’t want to relearn basic operation. (I’m still in total command of my HP1740A without manuals.)

At this stage, I found some divergence between the three models under examination. Both Cell’s ScopeManager software and Pico’s PicoScope software are straightforward to use and the important controls for input channels and timebase settings are obvious for anyone familiar with scopes. I found the TiePie module’s operation somewhat cryptic, although at no time did I have real problems displaying waveforms or making measurements. In fact, with all modules, I soon delighted in being able to drag-and-drop trigger levels, slide around cursors, and print and save results without any reference to manuals or help files.

Being a real scope user, though, I like to set amplitude and time scales in units/division of the graticule. Only the Cell unit helps you to do this. Both Pico and TiePie units steer you to select a full-scale setting (like a DMM). Similarly, the Cell unit’s front panel is the most “virtual” by providing push buttons rather than pull-down menus for setting up measurements.

Performance & Printouts

With a range of stable repetitive inputs up to 1 MHz, there is nothing to choose between the measurement results on all the models. On a 1 V, 50-ns edge, the Cell unit produces a smooth rise with the least number of distinguishable digitising levels. Although all units specify 8-bit amplitude resolution, Pico’s and TiePie’s units produce noticeably more waveform steps. Pico’s waveform is even a bit “bumpy”. Printouts of waveforms themselves are very similar. One noticeable exception is that the TiePie unit doesn’t print the graticule, nor does it print markers along either the X or Y scales to help you assess waveform dimensions from a printout. (See printouts in the product review boxes on pages 12, 14, and 16.)

One aspect of the Cell and TiePie displays that initially had me foxed was that overload signals don’t run off the scale (as with analogue scopes), but appear instead as horizontal lines that run along the upper and lower graticule lines. Until I realised what was happening, I thought I was observing pulses with incredibly flat peaks. Cell’s unit helps here by at least changing the colour of the display to red when waveforms hit end-stops. Pico’s trace continues off-scale as on a real scope.

But What’s Its Spec?

Call me a T&M “old-timer”, but when I receive a new product I invariably dive into the users’ manual to study the “technical specifications” section. I still expect this to be the place to find the most complete and honest statement of a product’s performance. Glossy brochures don’t always display similar openness, and often flash banner specs for optimal performance in a special set of conditions.

In the case of the users’ manuals that come with these three scope modules my reaction ranged from frustration at not finding anything that you can call a technical specification to being somewhat satisfied.

Cell’s user’s manual is virtually devoid of any technical specification and its spec page simply tells you the module’s size, weight, and power consumption. Okay, Cell’s separate data sheet is most complete and gives me everything I want to know, but this information would be a useful addition to the manual. Astonishingly, Cell’s 30-page handbook has no contents pages.

Pico’s user’s manual has two contents pages but they precede only seven more pages of information. (The same manual repeats the seven pages in Français, Deutsch, and Italiano.) The Pico manual has a specification page, although there is no mention of the unit’s analogue bandwidth.

TiePie’s 116-page users’ manual provides a much better level of information, and includes five pages of comprehensive technical specifications. The manual caters for novices and experts alike, and even tells you about aliasing and what a scope is for. One glaring flaw, though, is that the manual I received states the unit’s bandwidth and sampling rate to be DC to 20 MHz and 50 MHz, respectively, while the data sheet that accompanied the unit states 50-MHz bandwidth and 100-MHz sample rate.

In the end, I downloaded technical specifications from each vendor’s web site on the basis that, if a vendor doesn’t maintain web data, then what hope is there for ever finding what the spec is. In the future, though, where will you go to remind yourself of today’s spec when, by then, web site data will probably refer to a later model, or not be there at all? Table 1 (below) summarizes current web data.

Measurement Ability Is Basic

One of my concerns about all three units is that the PC user-interfaces look so attractive—Cell and TiePie in particular—they tend to disguise the rather basic measurement ability of these DSO front ends. Cell’s persistence display mode option, for example, is so glamorous that it reminds me of the sort of display you associate with a top-end LeCroy.

To qualify as general-purpose in today’s terms, any DSO needs a 100-MHz analogue bandwidth and a real-time or transient sampling-rate of at least 100 Msamples/s. Although DSO stands for “digital storage oscilloscope”, for most low-cost DSOs, it could equally well mean “dozens of signals omitted”. Even with a 100-MHz/100-Msample/s combination, you’ll only see 10 points on a 10-MHz signal (100 ns period). In other words, if you want to view a 10 ns edge, then the DSO has to rely on equivalent-time sampling, which works only with stable and repetitive waveforms. In this case, a DSO samples just one point on the edge and then ignores maybe the next 100 pulses before it’s ready to sample the second point on another edge, and so on. In this way, your view of a complete edge on the display is actually the result of a single sample taken of thousands of individual and widely separated edges in a pulse train. In consequence, the chances of a low-cost DSO missing a random event in a pulse train are extremely high. Equivalent-time sampling does have the advantage of easily displaying pre-trigger events though.

Even if your DSO can use real-time sampling (for example, at signal frequencies below 1 MHz), then its sampling rate still sets the horizontal resolution and its memory depth sets the length of the record. For the Pico and TiePie modules, with 50 Msamples/s and 32k memories, that works out to only 640 ms of signal at maximum horizontal resolution. By comparison, the Cell module’s 20 Msamples/s and 1M memory gives you a 50-ms record, and this is optionally extendable to 2M or 4M, to provide 100 ms and 200 ms, respectively.

Remember also that an input signal will be 3 dB down (30% error) at a scope’s specified bandwidth. Even a 25-MHz input on a 100-MHz scope will be 3% down, and this error roughly equates to the overall amplitude accuracy specification on most scopes anyway.

With these thoughts in mind, although all three vendors make much of their unit’s extensive signal display and processing features, don’t forget that all this hinges upon quite limited analogue measurements. Cell provides an attractive 150-MHz bandwidth but this is coupled with a rather low 20-Msamples/s sampling rate. Both Pico and TiePie provide 50 Msamples/s (100 Msamples/s if you select single channel operation) but then the analogue bandwidths are only 50 MHz. At least the sampling rates quoted here are for real-time sampling and not the misleading much higher equivalent-time sampling rates that some mainstream scope vendors use in banner specs.

Match the Application

Despite my few reservations, there’s no reason to treat these three models as anything less than serious test tools, providing you match them to suitable applications. In general, these applications will be in production test, service centre, or educational establishments. For example, if you have a known and existing waveform that you simply need to record and archive in production or service, then these units provide an unbeatable simple and low-cost solution. Equally, where you need to perform go/no-go tests of a device (for example, electromechanical switch-bounce), then again these boxes will do a good job.

For general-purpose probing and debugging of unknown signals I would rather soldier on with my hot-and-heavy analogue scope. I have more confidence that I’ll see events such as noise bursts, spikes, and HF oscillations. My ideal—ignoring for the moment the chances of ever owning one of today’s top-end DSOs—would be to use a general-purpose analogue scope for searching the unknown and then bring in a DSO for documenting results when I know roughly what to expect.

In terms of value for money, these DSO modules are hard to beat, especially considering all the units include other virtual instruments. TiePie’s HS801 and Pico’s ADC-220 includes a scope, voltmeter, transient recorder, and spectrum analyser. A later model, HS801 AWG, adds an arbitrary waveform generator. Prices range from 896 euros for Pico’s ADC-200/100 to 2640 euros for Cell’s VDS 2152B. (Cell has recently introduced a fixed configuration Slimline series with the same analogue performance and 256k/channel memory at 910 euros single channel, 1090 euros dual channel.) Table 2 (below) compares essential performance parameters of these three scope modules with low-cost bench DSOs from Agilent and Tektronix that you may consider as competitive units. T&ME

Table 1