Skip navigation

In a recent blog, I highlighted three ways high-speed converter innovations are changing our world.


High-speed converters are playing roles in other areas as well. As more and more devices move into the digital domain, our newest converters are simplifying the RF signal chain. Let’s take a journey into the evolution of the humble car radio as an illustration of this change.



In the classic version of the FM radio in your car, frequency translation, filtering, and amplification were done by analog RF blocks to tune into the station of choice, eliminating all other radio stations except the station you wanted before going to the analog-to-digital converter (ADC).





In the digital radio, we push the ADC up close to the antenna. The station select functionality is pushed into the digital domain – the converter is now required to capture the entire FM radio band (100× the bandwidth).


Better for automobile owners.

For you and your friends in the car, the potential functionality of the radio is greatly increased. If some additional digital circuits are added, we can make it possible to tune into multiple radio stations simultaneously (e.g. get weather or traffic info in “sidebands".) Or, by making the digital circuits programmable, you could access a software download that would allow your radio to adapt to new standards, such as HD radio.


Better for automobile manufacturers.

For Analog Devices customers, the RF signal chain is greatly simplified, making hardware design easier and allowing for a radio that would be capable of providing new services in the future.


That’s just one example of how this kind of RF signal simplification can provide real benefits for manufacturers and consumers. This level of simplification, enabled by the unprecedented bandwidth of Analog Devices high-speed converters can extend to many more applications such as cable, 5G, and radar; to deliver new services and capabilities.


Keep this in mind when you cue up that favorite playlist on your next road trip.

You can call me biased but when I hear the term “IoT platforms” and the exclusivity of conversations around software offerings under this term, I cringe. Does this mean hardware is a given?


Yes I’ve woken up and smelled the coffee, I get the concept of IoT where it’s the data that counts and the infinite possibilities of what you can do with this data to improve uptime efficiency of e.g. cities, factories or our bodies (boy do I need that one…)


So how important is the hardware? Well in IoT applications that require very rudimentary sensing, such as a coarse measurement of whether somebody is in the room, or if a garbage container is full; you could argue that hardware is less relevant. Cloud computing and its availability to the masses means that insight can be gained from a huge number of these rudimentary sensor readings and value can be gained from these insights.Tying these sensor readings with time, location and other sensors is a big part of what is envisioned in the explosion of sensors in the coming years and the ability to monetize with software and services. I get it.


There’s a caveat however. I believe that many of the most challenging IoT applications can only be realized with hardware that actually does not exist today.There are still significant challenges in hardware design, and will be for a long time. It’s fairly logical but let me share some reasons why. 


Firstly, what’s the advantage of IoT? Well it’s the remote monitoring and control. This goes for product solutions as well. It makes sense to “future proof” your hardware making it a scale better than you need. After all what’s easier? Running around swapping out units to provide the next big feature that your product requires or a simple software OTA (and isn’t that the whole point of IoT?)


Secondly, I’ve said it before and I’ll say it again: Send garbage up to the clouds and that is what you'll get back - don’t expect insight out of the cloud. There is a base line of how good your data needs to be to make your analytics work effectively. Good data requires careful hardware.


And lastly, and yes I am biased. (well I am the semiconductor dude don’t you know). But actually, this is because I’m also passionate about the piece we play in this. For Analog Devices, IoT is simply extending a signal chain up to the cloud. But we have seen how signal chains have developed over more than 50 years (and personally I’ve been on that journey for 30 of them – back to the uptime efficiency of bodies...). No one has ever asked me could they do less, they have always asked me what can be measured faster, easier, at lower power, at higher WHATEVER to give their customers something better than anyone else. Trust me. Pay attention to your hardware.


From Rant to Reality

And if you’d like to check out one of ADI’s IoT Arduino based Wireless Development Platform for Internet of Things applications feel free.

Full disclosure: I’m in the high-speed converter business helping Analog Devices to write the next chapters on Signal Conversion.


A new generation of high speed converters engineered to exploit deep sub-micron CMOS technology and proprietary architectures is poised to provide industry-leading performance in critical high-dynamic range parameters.


Here are three ways they’ll be driving the next wave of gigahertz bandwidth, software-defined systems.


1. Making 5G Come Alive

From car-to-car communications to watching the big game on your handheld device, you’ll want the speed and immediate response time promised by the 5G network. A key to putting the power of 5G into the hands of people everywhere will be using high-speed 28nm analog-to-digital converters (ADCs) and digital-to-analog converters (DACs) in both base station applications and 5G connectivity equipment.


Analog Devices has a long history of developing high speed converters—including the new 12-Gsps 16-bit RF DAC, and 3-Gsps 14-bit RF ADC.


  1. Keeping Us Safe

With growing unrest in hotspots around the globe, threat monitoring and detection are more important than ever. All-digital phased array radars are dramatically enhancing our capability to identify and mitigate risks to our security. High-speed RF ADCs from Analog Devices are pushing the state-of-the art in analog input bandwidth and dynamic range for pipeline converters, while also delivering significant reductions in size, weight, and power (SWaP). These innovative products are enabling all-digital phased array radars to be added to existing aircraft and ships, making our world safer.


  1. Delivering a Better Television Experience

Game of Thrones. Stranger Things. Atlanta. There’s a lot of great TV out there to enjoy. And to do that, people are demanding that their broadband and wireless service operations provide higher quality, always-on data, and video streaming. The AD9162 DAC delivers unprecedented bandwidth and dynamic range, enabling ultra-high-definition (UHD) and 4K television across more channels at extraordinary streaming and download speeds— bringing the future of TV into homes for all of us to enjoy.


There’s a lot more to look forward to when it comes to high-speed converters. Follow the EngineerZone Spotlight for more blog posts in the future.

The amount of data that goes up to the cloud is creating a data lake that continues to get bigger and bigger. I’m not adverse to this but what’s the point? How much data do you need or use? Which neatly brings me onto useless IoT applications. (I had to find a reason to get onto this…)


I recently read a review of a wireless kettle (and believe me before I instigate a wave of responses – “annoyed from Andover” anyone? I’m sure there are many perfect wireless kettles out there but honestly why can’t you just get off your chair and flip the switch?)…there are some applications that just don’t make sense to me.



I’ll share some feedback on getting this kettle going “On the first attempt to boil the kettle, it was hampered by a forced-debugging, which consequently caused its base station to reset. The kettle's base station appeared unable to connect with the kettle itself. On re-calibration it seems that it had never connected at all. Then the flaky WiFi connectivity required constant checking and resetting.” – all for a cup of hot water anyone? Now if that kettle had been monitoring the level of water and looping back into my system to pass up a local value of how much power it used to boil back into my home efficiency system then I get it. This is what I call sending insight versus just data.


Smart partitioning and embedding of algorithms in the sensors allows the most critical data to be interpreted in real time at its source. Algorithms embedded into intelligent sensors and in the cloud allow for interpretation beyond what can be done with silicon alone. In fact, this leads to the possibility for prediction and anticipation of future system behavior. Accelerating adoption of IoT solutions in mission critical applications is dependent on the ability to build secure systems – intelligent partitioning enables this.


In Industrial automation, active machine monitoring can transform factories by radically improving uptime efficiency. For example, accurate MEMS vibration sensors and algorithms to predict machine performance and reliability both locally for real-time optimization and intervention and in the cloud where information from multiple systems across multiple factories can be aggregated, analyzed and acted upon to improve productivity.


In mobile healthcare, clinical grade monitoring of vital signs and other wellness indicators are becoming increasingly important. For example, energy efficient optical sensors and ultra-low power MEMS sensors that reliably assess patient health in real-time using advanced algorithms. Another example is motion rejection algorithms which more accurately interpret heart rate from optical measurements at a patient’s wrist. Resultant information can then be transferred to the cloud for remote processing that may include more intensive data analysis used to derive further bio-markers and trends.


And that’s why you should use the cloud effectively (and why you should not have to wait hours for your kettle to boil…)


From Rant to Reality:

For intelligence at the edge you need low power processing. More information on the latest ultra low power MCU that enables 10 times system-level power savings

In short because we’re at the battlefront of every IoT application. We hardened semiconductor warriors are on the front line taking any and every kind of analog signal and transferring it to the digital world.


Does light, pressure, vibration temperature movement come in a nice neat digital format? You bet not. So Analog Devices (ADI) is one of the translators of real information into usable data. We focus on very difficult modalities because that’s our history, but simply said the first stage of the IoT battle is to sense (we do that too!), measure and then then transform into data. And frankly, how well you do this sets the stage for success further down the line. If you put garbage in then, sorry to disappoint you, but you’re going to get garbage out (big secret unveiled!). Need a location that’s within a mm, or a temperature that ensures you understand different growing conditions for unique crops, or a vibration that tells you exactly what ball bearing will fail next in your rotor? You need precision, and reliability. Don’t like changing batteries? You need measurement that is always on but also consuming ultra-low power (or self powered) to ensure even the most demanding IoT applications are empowered with the right data at the right time.


(If you really like changing batteries or love just writing analytics code for the sake of it don’t read any further and sorry to have wasted your time….)


So that’s why ADI is talking to you about IoT…. We’re one of the most important army commando units at the front line of the IoT battle enabling customers to create secure, reliable and robust IoT systems.


And silicon innovations enable interpretation of the world around us by bridging the physical and digital domains with technologies that sense, measure, interpret and connect.


From Rant to Reality

If you’d like to check out the latest low noise, high performance IoT MEMS sensor for Condition Based Monitoring Applications, click the MEMS sensor link for information.


EcoAnalog: Hybrid Vehicles

Posted by tnelson654 Employee Jun 15, 2017

On my bike ride downtown for coffee this morning, I waited in a double left turn lane surrounded by cars. I used to dread those moments because of the exhaust fumes. It's much better now thanks to hybrid vehicles and the new start/stop feature in traditional vehicles. Hybrid vehicles and stop/start gasoline-powered vehicles eliminate tons of carbon from polluting the atmosphere. Literally, tons!

According to the International Energy Agency in 2016, the transportation sector was responsible for approximately 23% of the world’s total carbon dioxide (CO2) emissions from fossil fuel consumption. On the occasion of their 9 millionth hybrid vehicle in April 2016, Toyota calculated that their hybrid vehicles produced approximately 67 million less tons of CO2 compared to conventional vehicles of similar size and driving performance. Today, almost every major automaker offers hybrid vehicles. So, certainly, that number is many times larger for the industry as a whole.


12V Stop Start

My first hybrid had the stop/start feature; it automatically shut off the engine at the first stop light and then started it back up when I wanted to go. It took some getting used to! And now that system has been applied to regular (non-hybrid) vehicles. For city driving, shutting off the engine at stop lights saves fuel and therefore CO2 emissions compared to idling at the stop light. Estimates range but many suggest 10% better fuel economy compared to the same model without stop/start. This has the potential to reduce much more CO2 than hybrid vehicles because there are so many more traditional vehicles on the road. I'm not going to attempt the math, but it's probably on the order of hundreds of millions of tons. Between the two technologies, that's a huge impact.


The algorithms differ by automaker. While traditional vehicles utilize a second lead-acid battery for restarting the engine, battery chemistry advances allowed hybrid vehicles to move from nickel-metal hydride to lithium-ion battery stacks. Accurately measuring and controlling the batteries is one of the keys to maximizing the performance of these systems. The stop/start system needs to be sure the battery is capable of restarting the engine before it stops it. The hybrid battery stack needs to keep each cell in the stack balanced with all the others, sharing equally and remaining fully charged. Isolating and level-shifting precision measurements of tiny voltages on a stack of hundreds of volts in an abusive environment is a challenge. It has to be robust, reliable and stable across temperature variations for the life of the battery system. And more systems use ADI than any other. Each generation is more integrated than the previous one, more precise, more reliable; and as they become more complex, the firmware support makes it easier to implement.


Schematic for battery monitoring


It's just what we do. So when I'm waiting at a red light on my bike, I don't have to breath all that exhaust, the planet doesn't have to absorb all that CO2 and I'm happy about that.

Like nearly everyone around me here, I am an engineer. And like you, since you are reading this "green" post, the future of the environment is important to me. I recycle, I commute via bike, and I have solar panels on my home. Many times in my career, I wondered if I should be working or volunteering somewhere that is directly involved with the cause. Should I be inventing a car that doesn't pollute, should I be building solar energy farms, should I be doing something more meaningful? Should I be somewhere else? And then I realized that I am in exactly the right place. I am doing extremely impactful and meaningful things with my engineering skills to help the environment! I just didn't realize it.

In future posts, my like-minded colleagues and I will link some recognizable (and a few obscure) "green" movements with the innovative technologies being developed here at Analog Devices. It's really quite inspiring! I believe we are really making a difference in the world while making the best use of our skills as engineers.

Chances are, if there's an emerging technology that might really help the planet, then we are already working on it. What green initiatives are you passionate about? Want to know if we are involved a particular environmental effort? Then ask in the comments. We will respond and may even address it in a future post!

The LTC2185 is a 125Msps 16-bit ADC with excellent noise and linearity performance while only consuming 185mW per channel. It is ideal for demanding low power applications that require excellent AC performance. A high performance ADC like the LTC2185 requires a high performance amplifier driving it to maintain the excellent performance. The ADA4927-1 delivers the linearity performance required by the LTC2185 while only consuming 215mW. The well designed package of the ADA4927-1 allows for a simple layout that reduces parasitic capacitance in the feedback path that can erode the phase margin of the amplifier. This combination of ADC and driver allows excellent performance from 62.5-125MHz a region where other high speed amplifiers are lacking. 


The LTC2185 is a two-channel simultaneous sampling parallel ADC which offers a choice of full-rate CMOS, or double data rate (DDR) CMOS/LVDS digital outputs. Pin-compatible speed grade options include 25Msps, 40Msps, 65Msps, 80Msps and 105Msps with approximate power dissipation of just 1.5mW/Msps per channel. It includes popular features such as the digital output randomizer and alternate bit polarity (ABP) mode that minimize digital feedback when using parallel CMOS outputs.  Analog full power bandwidth of 550MHz and ultralow jitter of 0.07psRMS allows under-sampling of IF frequencies with excellent noise performance. To maintain this level of performance the LTC2185 needs to be driven with an appropriate amplifier like the ADA4927-1.


The ADA4927 is a high speed differential current feedback amplifier. Fabricated on Analog Devices’ silicon-germanium process, the ADA4927-1 has excellent distortion and an input voltage noise of only 1.3nV/rtHz. This allows it to drive high speed ADCs like the LTC2185. The gain of the ADA4927-1 is set with external feedback resistors located next to the input pins.  By keeping the feedback pins and input pins close on the package, the ADA4927-1 provides a clean layout and minimizing the parasitic capacitance in the feedback network. This make the ADA4927-1 an ideal choice for driving high performance ADCs, like the LTC2185, from DC to 125 MHz.


Figure 1 shows a schematic of the ADA4927-1 driving the LTC2185. The corresponding layout is shown in figure 2.  The feedback pins on the ADA4927-1 are adjacent to the input pins which minimizes the parasitic capacitance of the feedback node and improves the phase margin of the amplifier. It also Simplifier the layout by making it possible to place feedback resistors directly across the two pins and not having additional trace length in the feedback path. There is a simple filter between the amplifier and ADC that reduces the wideband noise of the amplifier and improves the SNR of the system. This filter also attenuates the sampling glitches from the ADC before they reach the amplifier. This helps keep the output network of the ADA4927 from oscillating in response to these glitches. This filter network can be modified to accommodate a wide range of input bandwidth requirements.

Figure 1:  Schematic showing an ADA4927-1 driving one channel of the LTC2185


Figure 2:  Layout showing an ADA4927-1 driving once channel of the LTC2185


Figure 3 and figure 4 show the SNR and SFDR of the LTC2185 and ADA4927-1 combination. The SFDR stays above 67dB out to 125MHz while the SNR is better than 63dB to the same frequency. This combination only consumes 250mW. With a sample rate of 125Msps, this combination provides good performance through the entire 2nd Nyquist zone where other amplifiers begin to have poor linearity. 


Figure 3:  SNR of the LTC2185 driven with the ADA4927-1

Figure 4:  SFDR of the LTC2185 driven with the ADA4927-1


Using the ADA4927-1 to drive the LTC2185 provides excellent linearity while keeping the power consumption low. The fact that the ADA4927-1 stays very linear out to 125MHz allows this ADC amplifier combination to be used in demanding communication and medical applications that require the use of the second Nyquist zone of the LTC2185. The pin out of the ADA4927-1 and filter design minimize the complexity of the layout while maintaining excellent performance on a low power budget.


Follow the EngineerZone Spotlight to know when the next Combo Circuits From the Lab blog is posted.


Unplug and Go Ride a Bike

Posted by tnelson654 Employee May 11, 2017


On our website, it says:

Solving the problems of tomorrow and today relies on intelligently bridging the digital and analog worlds. At Analog Devices, we create unmatched technologies and solutions to solve our customers’ problems.

Forgive the hyperbole, but our customers' problems are the world's problems. And I'm very proud of the fact that the technologies we create have a big impact on solving the world's problems. But sometimes, the intelligent thing to do is to leave technology out. Just unplug. Go off the grid and go ride a bike.

I have a nice little home out in the suburbs with solar panels on the roof and a hybrid in the garage. But the price I pay for it is a 25-mile commute that can sometimes take over an hour. Imagine four lanes of stop-and-go highway traffic, mostly stopped. No amount of navigation assistance, multi-zone air conditioning or fuel-mapping optimization is going to make the situation any better. So, five years ago on "Bike to Work Day" I decided to try commuting all the way to work and back on my bike. Navigation was by sight and memory, air conditioning was the wind in my face and fuel mapping was a granola bar. And with that, I was hooked!

Erik Soule and Todd Nelson in Milpitas
I ride almost every day now. But "Bike to Work Day" is always special because it is a time to help my co-workers who think it might be possible for them to maybe just ride once. May 11, 2017, is Bike to Work Day in the Bay area and I'm hoping to add a few more people to our commuter group. The community sprinkles "energizer stations" on all the popular bike-commuting routes and gives out snacks, energy bars, coupons and some even serve coffee and warm breakfast. It's a happening! We aren't really talking about how we aren't burning fossil fuels or emitting CO2. We aren't talking about how we are freeing up traffic and parking spots for those who drive. We aren't talking about the health benefits. We are simply talking about how much fun it is.


Bill Skirkey in Chelmsford
We are just happy to ride.

(It's nice to know it's part of the solution, too.)


May is National Bike Month and cities around the United States are designating a Bike to Work day. If you bike to work, tell us about it on Twitter. #BikeToWorkDay @ADI_News


High Precision Voltage Source

Posted by clon Employee May 9, 2017

Today in our Combo Circuit from the Lab series, we are showcasing an ultra-high precision programmable voltage using ADI/LTC products together. The AD5971 with the LTZ1000, ADA4077 and AD8675/6 can be used to provide a programmable voltage source that achieves 1PPM resolution with 1PPM INL and better than 1PPM FSR long-term drift. This powerful combination helps provide radiologists with the superior image clarity, resolution and contrast they need, enabling them to see smaller anatomical structures. Just think of what this means when applied to an MRI (magnetic resonance imaging). Enhanced imagery of organs and soft tissues will allow medical professionals to more accurately detect heart problems, tumors, cysts and abnormalities in various parts of the body. This is just one of many applications for this programmable voltage source. 



Other applications that require 1ppm of accuracy are:

  • Scientific Medical and Aerospace Instrumentation
    • Medical Imaging Systems
    • Laser Beam Positioners
    • Vibration Systems.
  • Test and Measurement
    • ATE
    • Mass Spectrometry
    • Source-Measure Units (SMU)
    • Data Acquisition / Analysers
  • Industrial Automation
    • Semiconductor Manufacturing
    • Process Automation
    • Power Supply Control
    • Advanced Robotics


For Test and Measurements Systems, the 1ppm resolution and accuracy improves overall test equipment accuracy and granularity,  leading to finer control and excitation of external sources and nano-Actuators. In Industrial Automation, the 1ppm resolution and accuracy provides the precision that is required to move, alter or position an actuator on a nano-scale.




The AD5791 is a 20-bit, unbuffered voltage output digital to analog converter with 1 ppm relative accuracy (1 LSB INL) and 1 LSB DNL (guaranteed monotonic). It has an impressive 0.05-ppm/°C temperature drift, 0.1 ppm p-p noise and better than 1-ppm long-term stability. The AD5791 contains a precision R-2R  architecture which exploits state of the art thin-film resistor-matching techniques.  It operates from a bipolar supply up to 33V,  it can be driven by a positive reference in the range of 5V to VDD-2.5V and a negative reference in the range of VSS + 2.5 V to 0 V. The AD5791 uses a versatile 3-wire serial interface that operates at clock rates up to 35 MHz and that is compatible with standard SPI, QSPI™, MICROWIRE™, and DSP interface standards. The AD5791 is offered in a 20 lead TSSOP package.


The LTZ1000 is a ultra-stable temperature controllable reference. It provides a 7V output with an impressive 1.2µVP-P of noise, long-term stability of 2µV/√kHr and temperature drifts of 0.05ppm/°C. The part contains a buried zener reference, a heater resistor for temperature stabilization and a temperature sensing transistor. External components are used to set operating currents and to temperature stabilize the reference - this allows maximum flexibility and best long-term stability and noise.


The ADA4077 is a high precision, low noise operational amplifier with a combination of extremely low offset voltage and very low input bias currents. Unlike JFET amplifiers, the low bias and offset currents are relatively insensitive to ambient temperatures, even up to 125°C. Outputs are stable with capacitive loads of more than 1000 pF with no external compensation.



The AD8675/AD8676 are precision rail to rail operational amplifier with ultralow offset, drift, and voltage noise combined with very low input bias currents over the full operating temperature range.


Some Circuit Considerations:


Low-frequency noise must be kept to a minimum to avoid impact on the DC performance of the circuit. In the 0.1-Hz to 10-Hz bandwidth the AD5791 generates about 0.6μVp-p noise, each ADA4077 will generate 0.25μVp-p noise, the AD8675 will generate 0.1μVp-p  noise and the LTZ1000 generates 1.2µVp-p noise.  Resistor values were chosen to ensure that their Johnson noise will not significantly add to the total noise level.


AD5791 Reference buffer configuration

The reference buffers used to drive the REFP and REFN pins of the AD5791 must be configured in unity gain. Any extra currents flowing through a gain setting resistor into the reference sense pins will degrade the accuracy of the DAC.


AD5791 INL Sensitivity

The AD5791 INL performance is marginally sensitive to the input bias current of the amplifiers used as reference buffers, for this reason amplifiers with low input bias currents were chosen.



Temperature Drift

To maintain a low temperature drift coefficient for the entire system the individual components chosen must have low temperature drift. The AD5791 has a TC of 0.05ppm FSR/°C, the LTZ1000 offers a TC of 0.05ppm/°C , the ADA4077’s and the AD8675 contribute 0.005ppm FSR/°C and 0.01ppm FSR/°C respectively.


Long Term Drift

Long term drift is another important parameter that can cause significant accuracy limitations in systems. Long-term stability for the AD5791 is typically better than 0.1ppm/1000 hours at 125°C. Long-term stability on the order of 1µV per month can be achieved with the LTZ1000.

Lab Results:

INL Error was measured at ambient temperature in the lab by  varying the input code to the AD5791 from zero-scale to full-scale with a  code steps of 5. The voltage at the output of the output buffer (AD8675) was recorded at each code using a 8.5 digit DVM. The results are well within the ±1LSB specification.



The noise measured at mid-scale was 1.1µV p-p and the noise measured at full scale was 3.7uV p-p. The noise contribution from each voltage reference path is attenuated by the DAC when mid-scale code is selected, hence the lower noise figure for mid-scale code.


Long Term Drift:

The system Long term drift was measured at 25Deg°C. The AD5791 was programmed to +5V (3/4 scale) and the output voltage was measured every 30minutes over a period of 1000 hours. Drift values less than 1ppm FSR were observed.



In addition to ease of use the  AD5791 offers a guaranteed 1ppm accuracy, however selecting the correct components and voltage reference is critical to capitalise on the precision specifications of the AD5791. The low noise, low temperature drift, low long-term drift and high precision of the LTZ1000, ADA4077, AD8676 and the AD8675 improves the system precision, stability and repeatability over temperature and time.


DDS vs. PLL (round 2)

Posted by JLKeip May 4, 2017

20 years ago I used to listen to Paul Harvey on the radio from time to time. He was known for delving deeper into news stories than many other reporters would. I enjoyed that. I hadn't thought of him for a while, but as I was writing this, I could hear him and his signature phrase reverberating in my mind: 'And that... is the rest of the story'. So here is the rest of the story on the nuances of the advantages PLL typically has over DDS.


Power Consumption -

This amount of advantage for a PLL depends greatly on the frequency range you wish to synthesize. DDS has so much digital content that as you push up the sample rate, you significantly increase the power consumption, and you need higher sample rates to hit higher output frequencies. If your frequency set is sub 100 MHz or so, the power comparison gets quite competitive:  There exists a DDS with a top speed of 250 MSPS which consumes on the order of 50 mW (AD9913), so the gap between DDS and PLL here is not necessarily large. 

As process geometries shrink, the power gap will also shrink.


Price -

A price difference certainly exists. This gap can also shrink with geometries, but it can also shrink due to economies of scale.  PLLs are much more broadly used and drive much higher volumes, which helps cost structures.  If you are looking at a high volume application, and a DDS solution yields some advantage, you should expect to find some room for negotiation on the price point.


Broad Spectral Purity -

If you need a fractional N PLL, you'll see the gap shrink some, but it won't disappear entirely. An interesting note (exhibited below) regarding Programmable Modulus DDS's - if you utilize lower integer values in the denominator of your frequency equation, you will see the spurs significantly converge, greatly reducing the number of spurs. At its root, the dilemma is choosing between broad spectral purity and tuning resolution.


300 MHz output from a 2.1 GHz sample rate (Programmable Modulus)~299 MHz output from a 2.1 GHZ sample rate Non-Programmable Modulus DDS


Both of these plots use a DDS (AD9915 to be specific) running at 2.1 GSPS. The image on the left has the output set to 300 MHZ (exactly 1/7th the sample rate), whereas the image on the right is programmed to ~299 MHz. The reduction in the number of spurs is dramatic isn't it?


Ancillary Circuitry -

Not a whole lot to say here.  PLLs which rely on external VCOs (as many higher performance ones do), and those with external loop filters will also require a fair bit of board space beyond their footprint, but there isn't much to be done about the reconstruction filter component needs of the DDS.  I have seen a design or two which was able to use the raw output signal, but they are certainly in the minority.


Frequency Upconversion -

While upconversion is still not a word (no matter how hard I lobby), I still use it! There are two ways to bridge this gap.  One is to work with a Super Nyquist image frequency from the DDS rather than the fundamental.  Doing so reduces the output power of the signal, and it also makes the filter design more complex - a band pass filter must be designed as opposed to a low pass filter.

The other approach is to use a hybrid DDS/PLL based approach, which has the effect of mixing the advantages I have discussed:

Hopefully this gives you all the knowledge you need to make an informed decision on which route to go the next time you need to synthesize a signal for any need you come across.  I encourage you to ask questions here or in the DDS or RF-PLL forums of the Engineer Zone.   You'll also find sortable lists of our PLL and DDS portfolios at the website.


Wireless Current Sense

Posted by KrisLokere Employee May 2, 2017

Continuing with our series, Combo Circuits from the Lab, join Kris Lokere as he tells us what exactly makes Wireless Current Sense.

Measuring the current that flows through a sense resistor seems easy. Amplify the voltage, read it with an ADC, and now you know what the current is. But it gets more difficult if the sense resistor itself sits at a voltage that is very different from system ground. Typical solutions bridge that voltage difference in either the analog or digital domain. But here is a different approach – wireless.

Analog current sense ICs are compact solutions, but the voltage difference that they can withstand is limited by semiconductor processes. It is difficult to find devices that are rated for more than 100V. And these circuits often lose accuracy if the sense resistor common mode voltage changes quickly or swings both above and below system ground.

Digital isolation techniques (magnetic or optical) are a bit more bulky, but work without loss of accuracy and can typically withstand thousands of volts. These circuits need an isolated power supply but that can sometimes be integrated in the isolator component. If the sense resistor is physically separated from the main system then you also may need to run long wires or cables.

A wireless current sense circuit overcomes many of these limitations. By allowing the entire circuit to float with the common mode of the sense resistor, and transmitting the measured data wirelessly over the air, there are no voltage limitations at all. The sense resistor can be located anywhere, without the need to run cables. If the circuit is very low power, then you don’t even need an isolated power supply and can instead run for many years from a small battery.

Design Overview

Figure 1 shows the block diagram of the design. The current sense circuit is based on the LTC2063 chopper-stabilized opamp to amplify the voltage drop across a sense resistor. The micropower SAR ADC AD7988 digitizes the value and reports the result via SPI interface. The LTP5901-IPM is the radio module which contains not only the radio, but also the networking firmware needed to automatically form an IP-based mesh network. In addition, the LTP5901-IPM has a built-in microprocessor which reads the AD7988 ADC SPI port. The LTC3335 is a low-power DC/DC power supply which converts the battery voltage to a constant output voltage. The LTC3335 also includes a Coulomb counter which reports cumulative charge pulled from the battery.

Figure 1 - Wireless Current Sense Circuit Floats with Sense Resistor

Figure 1: A low-power wireless current sense circuit is formed by a low power chopper opamp to amplify the sense voltage, digitized using a low power ADC and reference, and connected to a SmartMesh IP wireless radio module. A low-power DC/DC converter conditions the battery and also keeps track of charge drawn from the battery.

Signal Chain

The LTC2063 is an ultra-low power, chopper-stabilized opamp. With a maximum supply current of 2µA, it is uniquely suited for use in battery powered applications. Because the offset voltage is less than 10µV, it can measure even very small voltage drops without loss of accuracy. Figure 2 shows the LTC2063 configured to gain-up and level shift the voltage across a 10mΩ sense resistor. The gain is chosen so that ±10mV full-scale at the sense resistor (corresponding to ±1A of current) maps to a near full-scale range at the output, centered around 1.5V. This amplified signal is fed into a 16-bit SAR ADC. The AD7988 was chosen for its very low standby current and good DC accuracy. At low sample rates, the ADC automatically shuts-down in between conversions, resulting in an average current consumption as little as 10µA at 1ksps. The LT6656 biases the amplifier, the level-shift resistors and the ADC’s reference input. The LT6656 voltage reference consumes less than 1µA, and can drive up to 5mA loads with low drop-out, making it easy to output a precise 3V, even when powered from the 3.3V system supply.

There are three roughly equal sources to offset error in this signal chain, together contributing about 0.5% relative to a ±10mV full-scale input. They are the offset voltage of the LTC2063 and AD7988, as well as mismatch in the level-shift resistors (0.1% resistors are recommended). A one-point calibration step could largely eliminate that offset. Gain error is generally dominated by inaccuracies in available sense resistors, which tend to be worse than the 0.05%, 10ppm/°C specifications of the LT6656 voltage reference.

Figure 2 - Wireless Current Sense Circuit Floats with Sense Resistor

Figure 2: The current sense circuitry floats with the sense resistor voltage. The LTC2063 chopper opamp amplifies the sense voltage and biases it mid-rail for the AD7988 ADC. The LT6656-3 provides the precision 3V reference.

Power Management

The LTC3335 is a nano-power Buck-Boost converter with integrated Coulomb counter. It is configured to provide a regulated 3.3V output from an input supply between 1.8V and 5.5V. This allows the circuit to be powered by, for example, two Alkaline primary battery cells. For duty-cycled wireless applications, the load current can easily vary from 1µA to 20mA, depending on whether the radio is in active or sleep mode. The LTC3335 has a quiescent current of just 680nA at no load, which keeps the entire circuit very low power when the radio and signal chain are in sleep mode. Still, the LTC3335 can output as much as 50mA, which easily provides enough power during radio transmit/receive and for a variety of signal chain circuits.

The LTC3335 also has a handy built-in Coulomb counter feature. When switching, it keeps track of the total charge that it draws from the battery. This information can be read-out using an I2C interface, and can then be used as a predictor for when it may become time to replace the batteries.

Wireless Networking

The LTP5901-IPM is a complete wireless radio module, which includes the radio transceiver, embedded microprocessor, and SmartMesh IP networking software. The LTP5901-IPM performs two functions in this application: wireless networking and housekeeping microprocessor. When multiple SmartMesh IP motes are powered up in the vicinity of a network manager, the motes automatically recognize each other and form a wireless mesh network. The entire network is automatically time-synchronized, which means that each radio is only powered on during very short, specific time intervals. As a result, each node can function not only as a source of sensor information, but also as a routing node to relay data from other nodes towards the manager. This creates a highly reliable, low power mesh network, where multiple paths are available from each node to the manager, even though all nodes, including the routing nodes, operate on very low power.

The LTP5901-IPM includes an ARM Cortex-M3 microprocessor core which runs the networking software. In addition, users may write application firmware to perform tasks specific to the user application. In this example, the microprocessor inside the LTP5901-IPM reads the SPI port of the current measurement ADC (AD7988) and reads the I2C port of the Coulomb counter (LTC3335). The microprocessor can also put the chopper opamp (LTC2063) in shutdown mode, further reducing its current consumption from 2µA to 200nA. This provides additional power savings in use models with extremely long intervals between measurements.

Overall Power Consumption

The total power consumption of the complete application circuit depends on various factors, including how often the signal chain takes a reading, and how the nodes are configured in the network. Typical power consumption for a mote reporting once per second is less than 5µA for the measurement circuit and can be 40µA for the wireless radio, allowing years of operation on small batteries.

Figure 3 - Wireless Current Sense Circuit Floats with Sense Resistor

Figure 3: A complete wireless current sense circuit is implemented on a small PCB. The only physical connections are the banana-jacks for the current to be measured. The wireless radio module is shown on the right. The circuit is powered from two AAA batteries connected on the back of the board.

Combining Linear Technology and Analog Devices signal chain, power management, and wireless networking products enable the design of a truly wireless current sense circuit. Figure 3 shows an example implementation. The new ultra-low power LTC2063 chopper op amp can accurately read small voltage drops across a sense resistor. The entire circuit, including micro-power ADC and voltage reference, floats with the common mode of the sense resistor. The nano-power LTC3335 switcher can power the circuit for years from a small battery, while reporting cumulative battery usage with its built-in Coulomb counter. The LTP5901-IPM wireless module manages the entire application and automatically connects to a highly reliable SmartMesh IP network.

Stay tuned for more Combo Circuits From the Lab blogs here on the EngineerZone Spotlight.

DDS v. PLL nuances (pt I)

Posted by JLKeip Apr 25, 2017

As promised in my last post, here is [PART 1 of] a review of some subtleties of the table contrasting DDS/PLL strengths. I know I didn’t [initially] say anything about breaking this up into 2 parts, but after I started writing, I realized it was getting too long for a single entry, so I went back, edited the last post and finished writing this!  

I’ll start with the characteristics I identified as being better suited to a DDS.

Frequency Resolution – There are a few nuances here…

There exists a DDS which runs at 400 MSPS with a 48 bit tuning word (AD9956). This yields a tuning resolution of no worse than 1.42 µHz, yes that’s microHertz.  The resolution of a standard PLL is limited by the depth of the dividers in the loop, and they are a few orders of magnitude worse than this.  I would be remiss however if I didn’t point out that there are fractional-N PLLs (N being the division factor in the feedback divider in the PLL) which close the gap significantly.  There is a downside to using a frac-N PLL: more spurs in the output.  The signals are more jittery. 

Another thing to note is that a standard DDS cannot hit many exact ratios. In the standard DDS frequency equation the denominator is a fixed power of 2.  So, for example, if you have a 200 MHz sample clock, you can hit 50 MHZ exactly (/4), but you cannot hit 40 MHz exactly – you will be off by some tiny bit: if you used the AD9956, you could either be 0.142uHz less than 40 MHz, or 0.568uHz more than 40 MHz. 

Standard PLLs hit these precise ratios quite easily. As such if you need exact ratios, avoid the standard DDS.  BUT that doesn't mean you need to avoid DDS as a whole, if you like the other things about DDS, look to Programmable Modulus!

The Programmable Modulus DDS (which I, affectionately, refer to as p-mod) is a relatively recent innovation in DDS.  A P-MOD DDS allows you to alter the DDS equation so that the denominator is no longer restricted to a power of 2.  No PLL can compete with a P-MOD DDS (AD9913, AD9914, AD9915, AD9164 e.g.) for frequency resolution; I believe that is inherent.  I suspect I will write more about P-MOD in a future entry;  you can also read about it here.

Returning to the 40 MHz example above, the industrious engineer could design a system which used frequency hopping to spend 80% of the time just below 40 MHz and 20% just above 40 MHz, and you would not only end up with a signal that averaged 40 MHz, you would also mimicking the internal workings of a P-MOD.

I promise the rest of these will be shorter!

Frequency Agility

There are ways to introduce varying frequencies in PLLs, but they are not as well controlled, nor are they as repeatable as the digital approach available with DDS.

Frequency hops can be implemented with two parallel PLLs and a switch (They call ‘em ping-pong PLLs – no table, paddles or net required), but you either need one complete PLL for every frequency you might want to hop to, OR you need to allow some settling time if you change the frequency of the bypassed PLL.

You can also sweep frequency with a variable divider in the loop, but it is nowhere near as well-controlled or repeatable as a DDS based sweep

Phase Resolution & Agility

In an analog PLL, any adjustment to phase ends up cycling through the loop, so it’s not a precise, repeatable change like you get with a DDS. Digital PLLs enable some level of phase adjustment capability. 

Amplitude Resolution & Agility

Amplitude is not a parameter that PLLs vary.

Look forward to part II in the not too distant future. Questions may be posed either in the comments, or in the DDS Product Section of ADI’s EngineerZone.

When driving a direct sampling high speed ADC, the most likely place to degrade the performance is the interface between the final amplifier and the ADC.  With any direct sampling ADC there will be nonlinear charge produced in the sampling process. This charge is reflected into the input network each time the sampling switches close. If these are left un-attenuated they can be reflected back to the ADC and re-sampled, degrading the simple distortion or inter modulation distortion of the ADC. The input network of the ADC should be as close to 50 ohms as possible to allow for the maximum absorption of this nonlinear charge. Using a highly absorptive filter can improve the SFDR by suppressing the nonlinear tones created in the sampling process. 


Using the LTC6409 to drive the AD9265

The LTC6409 is a differential amplifier with excellent linearity making it an ideal driver for the AD9265. The AD9265 is a 16-bit 125Msps high performance ADC with better than 77dB SNR and 89dB SFDR at 100MHz. These specifications can quickly degrade by poor design choices when designing the input network. In nearly all cases a filter is required between the ADC and amplifier to reduce the wideband noise of the final amplifier. Both the design and layout of this filter are critical. The filter should be absorptive so that high frequency non-linearities from the sampling process are absorbed into 50ohm termination and not allowed to be reflected back into the ADC input. Figure 1 shows an absorptive filter network that can be used between the LTC6409 and AD9265. Figure 2 shows the filter response. The purpose of this filter is not to be highly selective but simply to attenuate the wideband noise of the amplifier and nonlinearities of sampling process. At high frequency the inductors become opens and the capacitors become shorts which directs the high frequency content of the sampling process to 50ohm termination resistors. If the traces are routed at 50ohms there will not be any return reflections and the SFDR of the ADC will not degrade.


Figure 1: Filter network between the LTC6409 and the AD9265 


Figure 2:  Simulated Filter response for the circuit shown in figure 1


Another potential source of distortion is an asymmetrical layout of the input network. With an ideal layout, the differential nature of the signals allows excellent common mode rejection and very good 2nd harmonic distortion. Any deviation from perfect symmetry will cause a mismatch in the differential signals which will manifest itself as 2nd order harmonic distortion. Even a simple design decision to flood copper closer on one side of the differential pair than the other side can cause a difference in ground current in the adjacent ground planes. This adds distortion to the system. Absolute symmetry is required for maximum performance.  


Figure 3 shows the PCB layout of the LTC6409 driving the AD9265 and the filter network. Care was taken to preserve the symmetry of the network as well as position the absorptive elements to maximize their effectiveness. The first set of absorptive elements are positioned so that any high frequency products are immediately absorbed. The main signal path meanders around the grounded copper until it gets to the second set of absorptive elements and finally to the source termination at the amplifier. This network maximizes the performance possible from the LTC6409 and the AD9265.


Figure 3:  Layout of the LTC6409 and AD9265


To compare the performance of the LTC6409 and AD9265, a board was designed to connect to the PScope software via the DC890. The absorptive filter in figure 1 and a reflective filter in figure 4 were populated and tested over frequency.  The AD9265 was clocked with a 125Msps low jitter clock and the LTC6409 was driven with a filtered sinusoidal signal from 48.1 to 178.1MHz.  The SNR and SFDR were recorded with PScope. A sample data collection can be seeing in figure 5. The comparison of SNR and SFDR with the absorptive and reflective filters are shown in figures 6 and 7. The SFDR is consistently better with the absorptive network, at some points it is up to 10dB better. The SNR is also consistently better until very high input frequencies when the SNR is dominated by other factors.


With an absorptive network, the performance of the system improves over a reflective network. The excellent performance of the LTC6409 and AD9265 was degraded when a reflective network was used. The results are clear with the LTC6409 and AD9265 but the practice of using highly absorptive and symmetrical input networks can be applied to any direct sampling ADC and differential amplifier.  By focusing on the interface between the amplifier and ADC, maximum performance can be realized.  


Figure 4:  Reflective filter used between LTC6409 and AD9265


Figure 5: Sample data collection from PScope.  58.1MHz being sampled at 125Msps


Figure 6: SNR comparison between absorptive and reflective filters


Figure 7: SFDR comparison between absorptive and reflective filters


Look for more Combo Circuits From the Lab blogs here on the EngineerZone Spotlight.


It’s a new beginning for Analog Devices and Linear Technology. Together we become the high-performance analog industry leader with an innovation engine fueled by our comprehensive portfolio of analog and mixed signal products and solutions.



We have been telling you that 1+1 > 2 and now we will show you. Over the next several weeks we will be highlighting the powerful combinations possible with ADI and LTC products in our new blog series: Combo Circuits from the Lab.


Read our first blog: Improving Linearity by Using Absorptive Filters by Clarence.Mayott tomorrow, here on EngineerZone Spotlight.

Filter Blog

By date: By tag: