What’s the right channel count for mixed signal oscilloscopes?
The mixed signal oscilloscope was first introduced in 1993 and featured two analog channels paired with 8 or 16 digital channels. In the ensuing years, the channel counts for mainstream MSOs – the go-to debug tool for embedded systems designers – became essentially locked in at 2 or 4 analog channels and 16 digital channels. The MSO was embraced by embedded designers, because it extended their visibility from two or four signals to as many as 20 signals, without having to reach for their tool of last resort – the logic analyzer.
Despite this channel count’s long run and broad market acceptance, is it still right for today’s embedded systems? For oscilloscope manufacturers and embedded systems designers alike, it’s a question worth asking. Manufacturers need to know if they are delivering test capabilities customers actually want and are willing to pay for. Designers need the right tool for the job.
This thinking led to multiple research projects involving embedded systems engineers from around the world to look a little deeper into the topic of scope channel counts. What we learned informed much of the thinking in the new 5 Series MSO that increases the number of available analog channels to 6 or 8, and offers from 8 to 64 digital channels. The digital channels can also be re-configured on the fly.
Given the success of 4-channel MSOs over the years, it’s safe to say that the traditional number of analog and digital channels has served most embedded designers well enough, or perhaps more to the point, they have been able to make it work. But a significant number – 35 percent in our research – said they ideally wanted 8 analog channels.
In the past when these engineers needed more than 4 analog inputs, they would try to use two scopes together. This practice of “cascading” multiple scopes creates several challenges. To synchronize the acquisitions, the scopes must be triggered at the same instant, requiring cabling (or double-probing) and creative trigger setups. Comparing data on two displays is difficult, so many engineers take the raw data from the two scopes and use a PC to merge the waveforms for evaluation. This synchronization is time-consuming with two identical scope models, but it becomes even more problematic when using different scope models.
On the digital side, it turns out that less can be just as important as more. In some cases, engineers were frustrated when they were forced to buy 16 digital channels when they really only wanted eight. In our research, approximately 75% percent of respondents said they wanted something other than 16 digital channels. Some wanted more, some wanted fewer.
For embedded systems designers, the scope characteristic that stands out even more than channel count is flexibility. Our research found that 79 percent of embedded engineers want oscilloscopes that are “future proofed” and multi-purposed to serve the varied needs of overstretched design teams.
When we talked to embedded designers about the design phase at which they encountered the need for more channels and flexibility, the most common answer was during system-level troubleshooting. When multiple subsystems start coming together, with multiple processors, multiple power rails, multiple serial buses, and multiple I/O devices, that’s when system-level visibility can become critical. The traditional approach to troubleshooting with a scope is to take multiple two- or four-channel captures, making one’s way backward against the signal path to get to the root cause of a failure. In today’s systems that process input from multiple sensors to drive multiple actuators while communicating over multiple buses, this approach can be painful. These embedded computing systems, incorporating sensors, actuators, processing power and communications will form the distributed smarts in the growing Internet of Things (IoT).
Another pain point embedded engineers shared during our research stems from the proliferation of power rails in today’s systems. To optimize for both power consumption, performance and speed, even a relatively simple system might have a 12 V bulk supply, a couple of 5 V supplies, a 3.3-volt supply and a 1.8 V supply. Verifying and troubleshooting the turn-on and turn-off sequence of these supplies, especially in relation to other control or status signals on their boards requires more channels or more tests.
Some creative engineers reported that they use the variable thresholds on digital MSO channels to check power sequencing. In this case, they set the thresholds of the digital channels slightly below the nominal voltage level of the supply and used this setup to generate a “timing diagram” of power supplies, reset lines, interrupts, status lines and so on. The obvious shortcoming in this approach is the power supplies are represented as binary waveforms, ignoring the analog characteristics of the signals. Most engineers prefer to perform this testing and troubleshooting using analog channels.
For many applications, the traditional 4 analog/16 digital configuration may be all that’s needed. But if something new comes along – and our bet is that it will – it could be comforting to know that some other options are finally available.
Wilson Lee is Senior Manager, Marketing at Tektronix. Previously to joining Tektronix, Wilson has had over 25 years of technical marketing, technical sales leadership roles – with manufacturers such as CTS Electronic Components, as well technical/value add distributors such as Richardson RFPD, and Premier Farnell. Wilson has heavily focused on design within the RF/Wireless, Industrial Power, and Industrial Automation market segments. Wilson earned his Bachelor’s of Science Degree at Cornell University. While living in NY, Chicago and Asia through his career, he currently resides in Greater Portland, Oregon.