Chun Soong Wong spoke about the testing and measurement challenges in the power management sector and how they are helping users address such issues.
Amid the latest trends and developments in the electronics industry, how is the test and measurement industry keeping up? This month’s In Focus looks into the factors driving the innovations and developments in T&M, the new challenges, and opportunities.
Rohde & Schwarz is one of the leading providers of test and measurement (T&M) instruments worldwide. Founded in 1933 by Dr. Lothar Rohde and Hermann Schwarz, the company now has around 12,000 employees, with 77 offices around the world.
Chun Soong Wong is an Oscilloscope Product Manager for Asia at Rohde & Schwarz Regional Headquarters Singapore Pte Ltd. In an interview with EE Times Asia, he spoke about the testing and measurement challenges in the power management sector and how they are helping users address such issues; and provided T&M best practices that users should consider.
EE Times Asia: From a power management perspective, what are the key challenges that you always hear from your customers?
Chun Soong Wong: In terms of power customers, there are two issues: one is higher energy level—they are looking for higher and higher voltage and current; and on the other hand, there is also the demand for faster transistors—switching has to be fast, with higher bandwidth. All of these are coming into the power design field.
These demands are actually inducing other challenges. For example, very high-powered switching up and down, will create an electromagnetic interference (EMI) not just on the product but also the measurement device and to the environment around it. This is something that we are looking into—how to improve the measurement capability and how to reduce the noise that our customers are going to face in their designs.
EETA: What are the main factors causing these challenges?
Wong: It really comes from the demand of the market. For example, we talk about the electrification of cars. We are also looking at green energy—there is a lot of demand to convert solar power and feed them back into the grid. Of course, there’s IoT. These requirements involve batteries, and conversion of energy. One of the methods to do this conversion is to do the switching fast enough, and to have higher power, so that it can convert the energy in a shorter amount of time. For example, if you want to use an electric car and you wanted to charge within 10 minutes, this is something that is driving the industry to go into this high power and fast switching. That, in turn, is demanding a change in the technology used in the power industry.
Say, maybe in the past 20 years, our power electronics has been mainly focused on silicon IGBT—these are what we have been using for quite a while. With the high-power and faster switching requirements, this technology is moving towards another material per se, for example, gallium nitride (GaN) or silicon carbide (SiC).
Another area that is interesting, is the need to measure very small voltages—and this is also another power concern. The key reason is that, in for IoT devices to have a long battery life, you must have as small consumption as possible. That, in turn, drives the engineers to look into how they can further reduce energy—and to do very small signal measurements, which can be coupled by the environmental noise and the terminal noise. To go below the sub-nano amps or sub-pico amps is really challenging in the electronics world.
EETA: How then are you helping customers address these issues?
Wong: We look into how we can offer a T&M solution. The power industry is still a very conservative industry. What have been working for them—with the silicon IGBT, over the past decades, they will expect a very similar incremental performance in the future. But for them, they are not really going to jump into very fast switching or very high power in just a matter of one or two years; it will take quite a long time to do that.
For us, it is really to look into the immediate requirements of the customers, and to roll out a solution that will match what they need in terms of design as well as manufacturing in the future. We look into what we can offer in terms of better specifications. For example, if you are going for higher voltage, we’ll design a probe that can withstand higher voltage.
But these are also very difficult challenges that we face because we can’t just have a solution overnight that can fulfil all these requirements.
There are not a lot of players in the T&M industry that work on their own probe design—usually they OEM it out from a third party and get them to work on these developments. At R&S, we design our own high-voltage and high-frequency probes—and this is a commitment that we have.
EETA: What makes these solutions unique?
Wong: There can be two major areas: the probe is what connect the T&M instrument to the devices they are measuring. The probe design is very critical because whatever that is being translated by these probes to the measurement instrument will be what is being measured. Therefore, the probe design must have low noise and support very high voltage and frequency.
At R&S, our commitment is to quality. When customers switch to R&S products, this is something that they can expect: a high-quality probe that can fulfil the signal integrity that they want. At the same time, we have a very good specification in terms of bandwidth. Both goes hand-in-hand in making sure that the measurement that we measure will be repeatable and the precision they get will be as accurate as possible.
We are also working on a power-rail probe—there is quite a market there, not only in the IoT environment. If you talk about high precision, for example, quantum computing, that is trying to measure very small changes in power amplitude—those are very important.
For these sectors, the quality of the power will be critical because any small noise appearing in the power rail will propagate to the rest of the circuit. That, in turn, will make measurements difficult to determine whether they are real signals or just noise from the environment or the power supply. This is the demand we are seeing—and we want to measure this power as clean as possible.
To address this, we’ve developed a power rail probe technology that provides 1:1 attenuation. We don’t attenuate the signal, and importantly, do not amplify. Any attenuation or amplification will introduce noise into the signal path. At the same time, we provide a bandwidth range of up to 4GHz. This is also critical because if you think of the RF trend, we are going from 2.5GHz to 5GHz in the future. All these high-frequency components will exist in the power rail, and this is something that a lot of power rail designers do not pay attention to, because usually power engineers are being offered the lowest end of scope that is available in their company.
The misconception is that they think power shouldn’t have a very high frequency because it is just mostly DC, so designers are looking at power at 20MHz, maybe the most is at 1GHz. This is harming them right now because with RF, that is going to suck power in terms of the GHz range, and you won’t be able to see these power noises if you are just using a very low-bandwidth scope. Our probes not only offer very good sensitivity but also having the bandwidth requirement that will allow users to see much more into their signal.
On the oscilloscope itself, there is also quite a lot of development. We just launched our newest scope, the RTO6, which features the capability to process the signal into spectrum domain. When you collect a scope waveform, it is usually amplitude versus time. Our scope can convert that into FFT, and you can look into all the channels in their respective FFT domains at the same time. You can trigger on time, frequency, or zone. All these features are available, making the analysis very easy and allowing users to do a much more complex setup if they suspect something is wrong with their circuit—both in power as well as in RF.
With a scope like this, you have a lot of tools in your hands to find out all the problem.
EETA: What do you think will be the next level of innovation for the T&M industry?
Wong: From a T&M perspective, there’s a lot direction. From a power standpoint, we are seeing trends such as higher ADC performance, as well as having more and more channels.
Most vendors will take off-the-shelf ADCs to use in their designs. There’s a good and bad here: the good thing is they can have a fast turnaround time to offer their customers. On the other hand, a lot of these off-the-shelf ADCs usually do not have a very good noise performance.
One of the things that we look into a good ADC is the Effective Number of Bits (ENOB). If you translate into RF, this is more like a signal-to-noise ratio. When you offer high precision, you want to have a high-precision result, not a high-precision noise. This is something that the industry is working towards. In the future, this is something that we will see, is how to achieve a high precision in terms of your products, which also translate to how to achieve higher accuracy in terms of the measurements that you are trying to make.
Overall, I think what we can see here is more on the direction of cost reduction. The major players in this power field have actually offered 6-channel scopes, 8-channel scopes, but those are really expensive because if you think about a scope that has 4 channels, and you want to offer 8 channels, you are actually doubling the cost of the hardware. The next innovation is how to reduce the cost of these 8-channel scopes. Where it comes to a price point, where power engineers are able to accept the cost of ownership—because right now, it is still highly expensive.
Another potential innovation is a stronger push to include RF analysis. For now, our scopes have FFT. The next direction we are looking into is probably a built-in spectrum analysis.
We will also see developments on the probe side. Nowadays, when we talk about the high-voltage and high-frequency requirements in the power industry, one of the things that it induces is the noise. To circumvent that, some major players have introduced an optical probe. Instead of using the traditional electrical cables to connect the probes to the scope to the DUT, they use optical cable in between. The advantage is that it suppresses a lot of the common-mode noise that is in between the DUT and the instruments. That is working quite well; but that, of course, is still a very expensive approach. Imagine, a scope class, like the 3000-class series, and you want to introduce a probe that is almost the same price as the instrument. This is something that a lot of design engineers are having difficulty with—the cost of ownership.
I think the more vendors that are offering these optical probes, that more chance that prices will go down across all these vendors.
EETA: Can you suggest some T&M best practices in power delivery and power rail measurements?
Wong: There are a few things to look into. One is precision, which doesn’t mean accuracy—it just means more decimal points into your result. For better accuracy, you need to lower the noise level, so you need to look into the noise performance of your measurement instrument. That will dictate most of the performance.
The other part in terms of noise, it really depends on the sensitivity you are using for your probe as well as your scope, because some probes offer a very huge attenuation ratio; at that range, the noise will be amplified. This is not a good way to do the measurement.
Another good practice is always look into the loading. For example, if you have a probe connected into your DUT, it changes the behaviour. If it is a current probe, it will be like an inductive loading; if it is a voltage probe, there will be losses. All these will affect your measurement. So, look into the probe loading profile, especially at the frequency that you are measuring. As we go to higher and higher frequency measurements, that probe loading profile will be different. Look into the frequency versus impedance profile of your probe, of your scope as well, and make sure that at the frequency of interest is not degrading too much. There will definitely be degradation, but make sure that it is not a worse degradation of the whole probe loading profile. If there is, you have to select a different probe that could operate better in such frequencies.
Power is distributed throughout your circuit. If there is a noise in your power, you can imagine all the noise that are going to be in your circuit. If you try to correct it in a later phase, if you add more capacitor or inductor, or these passive circuits to correct it, you are going to make your circuit so big and that probably will not meet your design requirement in the end. Therefore, EMI another thing you should look into when doing test and measurement on your power design.
Stephen Las Marias is the editor of EETimes Asia. He may be reached at firstname.lastname@example.org.