EDN Network

Three chip antennas from Taoglas—the ILA.257, ILA.68, and ILA.89—provide Wi-Fi 6/7, ultra-wideband (UWB), and ISM connectivity. Manufactured using a low-temperature co-fired ceramic (LTCC) process, the antennas deliver high radiation efficiency and frequency stability in ultra-compact packages. According to Taoglas, they also require a smaller keep-out area than competing antennas.

The ILA.257 is a 3.2×1.6×0.5-mm antenna for Wi-Fi 6/7, providing tri-band coverage across 2.4 GHz, 5.8 GHz, and 7.125 GHz with strong radiation efficiency and stable signal integrity. Its small footprint and minimal keep-out area make it well-suited for wearables, portable electronics, and industrial IoT devices.

Engineered for UWB operation from 6 GHz to 8.5 GHz, the ILA.68 3.2×1.6×1.1-mm antenna delivers a stable omnidirectional radiation pattern with consistent repeatability and low insertion loss. It supports applications such as indoor positioning, access control, and short-range radar in space-constrained IoT and automotive systems.

Designed for the 868-MHz and 915-MHz ISM bands, the ILA.89 supports global LPWAN and LoRa deployments with up to 47.9% radiation efficiency and 0.56 dBi peak gain. Its 4.0×12.0×1.6-mm footprint, simple layout, and regional variants help reduce design complexity and speed time-to-market for small IoT devices.

The ILA.257, ILA.68, and ILA.89 antennas are now available from Taoglas and its authorized distributors.

Taoglas

The post Chip antennas boost Wi-Fi and UWB signal integrity appeared first on EDN.

OKW now offers CONNECT fast-assembly handheld plastic enclosures with optional integrated cable glands, making it easier to install power and data cables.

Cost-effective CONNECT is ideal for network technology, building services, safety engineering, IoT/IIoT, medical devices, analytical instruments, data loggers, detectors, sensors, test and measurement.

(Source: OKW Enclosures Inc.)

CONNECT’s two case shells snap together for fast and easy assembly: no screws are required. This offers the choice of two ‘fronts’: one shell is convex – perfect for LEDs – while the other is flat and recessed for a compact display or membrane keypad. Inside the flat shell there are mounting pillars for PCBs and components.

CONNECT enclosures feature open apertures at each end. For these, design engineers can specify a combination of ASA+PC blank end panels and soft-touch TPE cable glands with integrated strain relief. Cable diameters from 0.134“ to 0.232“ are accommodated. The two long sides provide ample space for USB connectors.

These UV-stable ASA+PC (UL 94 V-0) enclosures are available in six sizes from 2.36″ x 1.65″ x 0.87″ to 6.14″ x 2.13″ x 0.87″. The standard colors are off-white (RAL 9002) and black (RAL 9005). Custom colors are also available.

The cable glands come in volcano (gray) and black (RAL 9005). The end parts are off-white (RAL 9002) and black (RAL 9005). Other accessories include wall holders, rail holding clamps for round tubes up to ø 1.26″, and self-tapping screws.

OKW can supply CONNECT fully customized. Services include machining, lacquering, printing, laser marking, decor foils, RFI/EMI shielding, and installation and assembly of accessories.

For more information, view the OKW website: https://www.okwenclosures.com/en/Plastic-enclosures/Connect.htm

The post Handheld enclosures add integrated cable glands appeared first on EDN.

Manufacturing of a modern component-laded printed circuit board (PCB) is an amazing fusion and coordination of diverse technologies. There’s the board as substrate itself, the stencils and masks that enable precise placement of solder paster, and the pick-and-place mechanical system that places components (both ICs and passive ones) on the appropriate lands with pinpoint precision and repeatability, all culminating in most cases in a sophisticated reflow-soldering process.

Most of the loaded components use surface mount technology (SMT) and tiny contacts to their respective lands on the PCB. However, it wasn’t always an SMT world. In the early days of PCBs, the situation was somewhat different. Most of the components were dual inline package (DIP) ICs and passives with tangible wire leads, where their connections went through holes in the board (Figure 1).

Figure 1 Dual-inline package (DIP) was dominant in the early days of ICs and is still favored by makers and DIY enthusiasts; but most devices are no longer offered this way, nor can they be. Source: Wikipedia

Not only did this require costly drilling of hundreds and thousands of space-consuming holes, but component installation was a challenge. The loaded board—with these through-hole components mounted on one side only—went through a wave-soldering process which soldered the leads to the tracks on the bottom of the board.

The advent of SMT

The use of surface-mount technology began in the 1960s, when it was originally called “planar mounting”. However, surface mount technology didn’t become popular until the mid-1980s, and even as recently as 1986; surface-mount components represented only around 10% of the total market. The technique took off in the late 1980s, and most high-tech electronic PCBs were using surface mount devices by the late 1990s.

SMT enables smaller components, higher board densities, use of top and bottom sides of the board for components, and a reflow soldering process. Today, active and passive components are offered in SMT packages whenever possible, with through-hole packages being the exception. SMT devices can be placed using an automated arrangement, while many larger through-hole ones require manual insertion and soldering. Obviously, this is costly and disruptive to the high-volume production process.

The demand for SMT versions is so overwhelming that many products are available only in that package type. SMT makes possible many super-tiny components we now count on; some are just a millimeter square or smaller.

Due to the popularity of SMT, vendors often announce when they have managed to make a former through-hole component into a SMT one. Doing so is not easy in many cases for ICs, as there are die-layout, thermal, packaging, and reliability issues.

There are also transitions for passives. For example, Vishay Intertechnology recently announced that it has transformed one of its families of axial-leaded safety resistors into surface-mount versions using a clever twisting to the leads in conjunction with a T-shaped PCB land pattern (Figure 2). This is not a trivial twist because these resistors must also meet various safety and regulatory mandates for performance under normal and fault conditions while being compatible with automated handling.

Figure 2 Transforming this leaded safety resistor from a through-hole to SMT device involved much more than a clever design as the SMT version must meet a long list of stringent safety-related requirements and tests. Source: Vishay

In other cases, vendors of leaded discrete devices such as mid-power MOSFETs have announced with fanfare that they have managed to engineer a version with the same ratings in an SMT package. No question about it; it’s a big deal in terms of attractiveness to the customer.

What about the SMT holdouts?

Despite the prevalence of, and desire for, SMT devices, some components are not easily transformed into SMT-friendly packaging that is also compatible with reflow soldering. Larger connecters for attaching discrete terminated wires to wiring blocks are a good example. If they were SMT devices, the stress they endure would flex the board and weaken their soldered connections as well as affect the integrity of the adjacent components. Their relatively large size also makes SMT handling a challenge.

But that dilemma is seeing some resolution. Connector vendor Weidmüller Group has developed what it calls through-hole reflow (THR) technology. These are terminal-block connectors for discrete wires that do require PCB holes and through-hole mounting for mechanical integrity. Yet, it can then be soldered using the standard reflow process along with other SMT devices on the board.

One of the vendor’s families with this capability was developed for Profinet applications and supports Ethernet-compliant data transmission up to 100 Mbps (Figure 3).

Figure 3 One of the available families of THR connector blocks is for Profibus installations. Source: Weidmüller

These connector blocks use glass-fiber-reinforced liquid crystal polymer (LCP) bodies to guarantee a high level of shape stability. The favorable temperature properties of the material (melting point of over 300°C) and the in-built pitch space (stand-off) of 0.3 mm (minimum) are well-suited for the solder-paste process. They come in choice of two pin lengths of 1.5 mm and 3.2 mm to precisely match board thickness, all with very tight tolerance on dimensional stability and pin centering (Figure 4).

Figure 4 The connector pin must have the right length and precise centering for reliable contact. Source: Weidmüller

The reflow wondering profile is like the ones required for other SMT components, so the entire board can be soldered in one pass (Figure 5).

Figure 5 The recommended reflow soldering profile for these THR connectors matches the profile of other SMT devices. Source: Weidmüller

Another connector family supports various USB connections (Figure 6).

Figure 6 A range of THR USB connectors is also available. Source: Weidmüller

With these THR connectors, you get the mechanical integrity of through-hole devices alongside the manufacturing benefit of automatic insertion (Figure 7) and reflow soldering. There is no need for a separate step to manually insert the connector and have a separate soldering step. You can also use them for through-hole wave-soldering as well, if you prefer.

Figure 7 Even the larger-block THR connectors can be automatically inserted using SMT pick-and-place systems. Source: Weidmüller

Connectors such as these will undoubtedly lower manufacturing costs while not compromising performance. Once again, it’s a reminder of the vital role and impact of mechanical know-how and material-science expertise to less-visible, low-glamour yet important advances in our “electronics” industry.

Bill Schweber is a degreed senior EE who has written three textbooks, hundreds of technical articles, opinion columns, and product features. Prior to becoming an author and editor, he spent his entire hands-on career on the analog side by working on power supplies, sensors, signal conditioning, and wired and wireless communication links. His work experience includes many years at Analog Devices in applications and marketing.

Related Content

• Consumer Connectors Get Ruggedized
• Be aware of connector mating-cycle limits
• Give Me Back My External Wi-Fi Antenna Connector, Please

The post Through-hole connector resolves surface-mount dilemma appeared first on EDN.

The most surprising thing to me about the Oura Ring 4, compared to its Gen3 predecessor, is how similar the two products are in terms of elemental usage perception. Granted, the precursor’s three internal finger-orientation bumps:

are now effectively gone:

and there are also multiple internal implementation differences between the two generations, some of which I’ll touch on in the paragraphs that follow. But they both use the same Android and iOS apps, generate the same data, and run for roughly the same ~1 week between charges.

One key qualifier on that last point: I bought them both used on eBay. The Ring 4, which claims 8 days of operating life when new, may have already accumulated more cycles from prior-owner usage than was the case with the Gen3 forebear, which touts 7 days’ operating life when new.

Smart ring “kissing cousins”

They look similar, too: the Gen3 in “Brushed Titanium” is the lower of the two rings on my left index finger in the following photos, with the Ring 4 in “Brushed Silver” above it:

And here’s the Ring 4 standalone, alongside my wedding band:

A smart ring enthusiast’s detailed analysis of the two product generations, complete with an abundance of comparative captured-data results, is below for those of you interested in more of an on-finger relative appraisal than I was able (and, admittedly, willing) to muster:

Sensing enhancements

Perhaps the biggest claimed innovation with the newer Ring 4 is Smart Sensing:

Smart Sensing is powered by an algorithm that works alongside the research-grade sensors within Oura Ring 4 to respond to each member’s unique finger physiology, including the structure and distinct features of your finger (i.e. skin tone, BMI, and age).

 The multiple sensors form an 18-path multi-wavelength photoplethysmography (PPG) subsystem, which adjusts dynamically to your lifestyle throughout the day and night.

As the functional representation in this conceptual video suggests:

there are two multi-LED clusters, each supporting three separate light wavelengths (red, green and infrared), with corresponding reception photodiodes in the rectangular structures to either side of each cluster (three structures total):

To complete the picture, here’s the inner top half of my Ring 4:

Six total LEDs, outputting to three total photodiodes, translates to 18 total possible light path options (which is presumably how Oura came up with the number I quoted earlier), with the optimal paths initially determined as part of the first-time ring setup:

and further fine-tuning is dynamically done while the ring is being worn, including compensating for non-optimum repositioning on the finger per the earlier-mentioned lack of distinct orientation bumps in this latest product generation.

What are the various-wavelength LEDs used for? Generally speaking, the infrared ones are capable of penetrating further into the finger tissue than are their visible-light counterparts, at some presumed tradeoff (accuracy, perhaps?). And specifically:

• Red and infrared LEDs measure blood oxygen levels (SpO2) while you sleep.
• Green and infrared LEDs track heart rate (HR) and heart rate variability (HRV) 24/7, as well as respiration rate during sleep.

All three LED types were also present with the Gen3 ring, albeit in a different multi-location configuration than the Ring 4 (albeit common to both the Heritage and Horizon Gen3 styles):

The labeling in the following Ring 4 “stock” image, by the way, isn’t locationally or otherwise accurate, as far as I can tell; the area labeled “accelerometer” is actually a multi-LED cluster, for example, and in contrast to the distinct “Red And Infrared…” and “Green And Infrared…” labels in the stock image, both of the clusters actually contain both green and red (plus infrared) LEDs:

Also embedded within the ring is a 3D accelerometer, which I’ve just learned, thanks to a Texas Instruments technical article I came across while researching this writeup, is useful not only for counting steps (along with, alas, keystrokes and other finger motions mimicking steps) but also “used in combination with the light signals as inputs into PPG algorithms.”

And there’s also a digital temperature sensor, although it doesn’t leverage direct skin contact for measurement purposes. Instead, it’s a negative temperature coefficient (NTC) thermistor whose (quoting from Wikipedia) “resistance decreases as temperature rises; usually because electrons are bumped up by thermal agitation from the valence band to the conduction band”.

Battery life optimizations

As noted in the public summary of a recent Ring 4 teardown by TechInsights, the newer smart ring has a higher capacity battery (26 mAh) than its Gen3 predecessor, which is likely a key factor in its day-longer specified operation between recharges. Additionally, the Ring 4’s Smart Sensing algorithms further optimize battery life as follows:

In order to optimize signal quality and power efficiency, Oura Ring 4 selects the optimal LED for each situation, instead of burning several LEDs simultaneously.

and

Smart Sensing also helps maximize the battery life of Oura Ring 4 by dynamically adjusting the brightness of the LEDs, using the dimmest possible setting to achieve the desired signal quality. This allows the battery life of Oura Ring 4 to extend up to eight days.

Here, for example, is a dim-light photo of both green LEDs in action, one in each cluster:

Generally speaking, the LEDs are active only briefly (when they’re illuminated at all, that is) and I haven’t yet succeeded in grabbing my smartphone and activating its camera in time to capture photos of any of the other combinations I’ve observed and note below. They include:

• Single green LED (either cluster)
• Concurrent single green and single red LEDs (one from each cluster), and
• Both single (either cluster) and dual concurrent (both clusters) red LED(s)

I’ve also witnessed transitions from bright to dim output illumination, prior to turnoff, for both one and two concurrent green LEDs, but not (yet, at least) for either one or both red LED(s). And perhaps obviously, the narrow-spectrum eyes-and-brain visual sensing and processing subsystem in my noggin isn’t capable of discerning infrared (or even near-IR) emissions, so…

Third-party functional insights

Operating life between integrated battery recharges, which I’ve already covered, is key to wearer satisfaction with the product, of course, as is recharge speed to “full” for the next multi-day (hopefully) wearing period.

But for long-term satisfaction, a sufficiently high number of supported recharge cycles prior to effective battery expiration (and subsequent landfill donation) is also necessary. To wit, I’ll close with some interesting (at least to me) information that I indirectly (and surprisingly, happily) stumbled across.

First off, here’s what the Ring 4 looks like in the process of charging on its inductive dock:

In last month’s Oura Gen3 write-up, I shared a photo of the portable charging case (including an integrated battery) that I’d acquired from Doohoeek via Amazon, with the dock mounted inside. Behind it was the Doohoeek charging case for the Oura Ring 4. They look the same, don’t they?

That’s because, it turns out, they are the same, at least from a hardware standpoint. Requoting what I first mentioned last month, the “development story (which I got straight from the manufacturer) was not only fascinating in its own right but also gave me insider insight into how Oura has evolved its smart ring charging scheme for the smart ring over time. More about that soon, likely next month.

Here’s the Ring 4 and dock inside the second-generation Doohoeek case (which, by the way, is also backwards-compatible with the Gen3 ring and dock):

And as promised, here’s the full back-and-forth between myself (in bold) and the manufacturer (in italics) over Amazon’s messaging system:

As I believe you already realize, while Doohoeek’s first-generation battery case that I’d bought from you through Amazon works fine with the Oura Gen3, it doesn’t (any longer, at least) work with the Ring 4. For that, one of Doohoeek’s second-generation battery cases is necessary. Can you comment on what the incompatibility was that precluded ongoing reliable operation of the original battery case with the Ring 4 charging dock (although it still works fine for the Gen3)? A USB-PD handshaking issue between your battery and the charging dock? Or was it something specific to the ring itself?

Hi Brian,

thank you for your question! Here’s a brief technical explanation of the Ring 4 compatibility issue with our original charging case:

Our first-gen charging case used a smart current-detection algorithm to determine charging status. Under normal conditions, when the ring reached full charge, the current would drop and remain consistently low—triggering our case to stop charging. This worked flawlessly with Oura Gen3 and initially with the Ring 4.

However, after a recent Oura firmware update, the Ring 4 began exhibiting unstable current draw patterns during charging—specifically, prolonged periods of low current followed by unexpected current spikes, even when the ring was not fully charged. This behavior caused our case to misinterpret the ring as “fully charged” and prematurely terminate charging.

To resolve this, we redesigned our charging logic in the updated version to implement a more robust timing-based backup protocol.

We appreciate your interest and hope this clarifies the engineering challenge we addressed!

Best,

Doohoeek Support Team

This is perfect! It was obvious to me that whatever it was, it was something that a firmware update couldn’t resolve, and I’d wondered if ring-generated current draw variances were to blame. I suspect the Ring 4 is doing this to maximize battery life over extended charge cycle counts. Thanks again!

p.s…I also wonder why you didn’t change the product naming, box labeling, etc. so potential buyers could have reassurance as to which version they’d be getting?

Hi Brian,

Thank you for your insightful feedback — you’ve clearly thought deeply about how these systems interact, and we really appreciate that.

Yes, the current behavior on the Ring 4 appears optimized for long-term battery longevity

Regarding your question about naming and packaging:

We actually had already mass-produced the outer shells and packaging for old version when Oura pushed the update that changed the charging behavior. Rather than discard those components (and create unnecessary waste), we decided to prioritize a firmware-level fix and use the same exterior.

That’s why the outside looks identical, but the internal charging behavior is now completely updated.

If you’d like to confirm whether your unit is the latest version, you can check the FNSKU barcode on the package:

Old version (no longer in production) ONLY used: X004HYCA09

New version (may change in future production) currently used: X004Q62DV9

Customers can also contact us with a photo of the label, and we’d be happy to verify it for them personally.

Thanks again for your support and sharp eyes.

Best,

Doohoeek Support Team

Very interesting! So it IS possible to firmware-retrofit existing units. Would that require a unit shipment back to the factory for the update, or did you consider developing a Windows-based (for example) update utility for customer upgrade purposes (by tethering the battery case’s USB-C input to a computer)?

Hi Brian,

Great question.

Unfortunately, a firmware update is not possible for units that have already been shipped. The hardware design does not support customer-side or even a cost-effective return-to-factory update process.

The only practical solution we could implement was to correct the firmware in all newly produced units moving forward, which is what you have received.

We appreciate your understanding!

Best,

Doohoeek Support Team

And with that, having recently passed through 2,000 words, I’ll wrap up for today. Stay tuned for the aforementioned teardown-to-come (on a different Ring 4; I plan to keep using this one!), and until then, I as-always welcome your thoughts in the comments!

—Brian Dipert is the Principal at Sierra Media and a former technical editor at EDN Magazine, where he still regularly contributes as a freelancer.

Related Content

• The Smart Ring: Passing fad, or the next big health-monitoring thing?
• Can a smart ring make me an Ultrahuman being?
• Does (wearing) an Oura (smart ring) a day keep the doctor away?
• RingConn: Smart, svelte, and econ(omical)

The post The Oura Ring 4: Does “one more” deliver much (if any) more? appeared first on EDN.

Editor’s note: In this Design Idea (DI), contributor Bonicatto designs a digital filter system (DFS. This is a benchtop filtering system that can apply various filter types to an incoming signal. Filtering range is up to 120 kHz.

In Part 1 of this DI, the DFS’s function and hardware implementation are discussed.

In Part 2 of this DI, the DFS’s firmware and performance are discussed.

Selectable/adjustable bench filter

Over the years, I have been able to obtain a lot of equipment needed for designing, testing, and diagnosing electronic equipment. I have accumulated power supplies, scopes, digital voltmeters (DVMs), spectrum analyzers, signal generators, vector network analyzers (VNAs), LCR meters, etc., etc.

One piece of equipment I never found is a reasonably priced lab bench filter—something that would take in a signal and filter it with a filter whose parameters could be set on the front panel.

There are some tools that run on a PC’s sound card, but I don’t like to connect my electronic tests on my PC for fear that I’ll damage the PC. The other issue is that I am looking for something that can go up to 100 kHz or so, which is not typical of many soundcards. So, it was time to try to design one.

Wow the engineering world with your unique design: Design Ideas Submission Guide

What I came up with in a small bench-top device with one BNC input for the signal you want filtered and one BNC output for the resulting filtered signal (Figure 1). It has a touchscreen LCD to select a filter type and the cutoff/center frequency. So, what can it do?

Figure 1 The finished digital filter system that allows you to select a low-pass, high-pass, band-pass, or band-stop filter type.

You can select a low-pass, high-pass, band-pass, or band-stop filter type. The filter can also be either a two-pole Butterworth or a four-pole.

For the frequency, you can select anywhere from a few Hz to 120 kHz. The are also three gain controls (an analog input gain knob, an analog output gain, and an internal digital gain.)

The cost to build the filter is around $75, as well as some odds and ends you probably already have around.

I also included a download for a 3D printable enclosure. Let’s take a deeper look at this design.

The circuit

The design is centered around a digital filter executed in a Cortex M4 microcontroller (MCU). The three main blocks of the system are an analog front end (AFE), which is composed of four op-amps providing input gain adjustment and antialiasing filtering.

Next is a single board computer (SBC) powered by a Cortex M4. This provides an input for the ADC, controls the LCD and touchscreen, executes the digital filters, and controls the output DAC.

The last block is the analog back end (ABE), which again consists of four op-amps that make up the analog gain circuit and the analog output reconstruction filter.

Let’s take a look at the schematic to see more detail (Figure 2).

Figure 2 The DFS schematic showing the AFE, the ABE, and SBC that provides an input for the ADC, controls the TFT display, executes the digital filters, and controls the output DAC.

Here you can see the blocks we just talked about and a few other minor pieces. Let’s dive a little deeper.

The AFE

The AFE starts by AC-coupling the external signal you want to filter. Then, the first op-amp, after the protection diodes, provides an adjustable gain for the input. This uses a simple single-supply inverting op-amp circuit. RV1 is a potentiometer on the front panel (see Figure 1 above) that allows for a gain of the input from 1x to 5x.

Again, looking at the schematics, we next see a single-pole low-pass filter, which is tuned to 120 kHz. Next are a pair of 2-pole Sallen-Key low-pass filters with components selected to create a Butterworth filter set to 120 kHz.

So now our input signal has been filtered at a frequency that will allow the MCU’s ADC to sample without aliasing. I designed this filter and the ABE filter using TI’s WEBENCH Circuit Designer.

So, we have a 5-pole low-pass filter frontend that will give us a roll-off of 30 dB per octave, or 100 dB per decade.

The flywheel RC circuit is next. As explained in a previous article, the capacitor in this RC circuit provides a charge to hold up the voltage level when the ADC samples the input. More on this can be found at: ADC Driver Ref Design Optimizing THD, Noise, and SNR for High Dynamic Range

The ABE

We’ll skip the MCU for now and jump to the right side of the schematic. Here we see a circuit very similar to the AFE, but this is used as a reconstruction filter that removes artifacts created by the discrete steps used in the MCU’s DAC.

So, starting from the DAC output from the SBC, we see an adjustable gain stage which allows the user, via the output potentiometer, to increase the output level, if desired. This output gain can be adjusted from 1x to 5x.

Next in the schematic, you’ll see two stages of two-pole Sallen-Key low-pass filters configured exactly like the pair in the AFE. So again, they are configured as a 120 kHz Butterworth filter. 

The last op-amp circuit in the ABE is a 2x gain stage and buffer. Why a 2x gain stage? I’ll explain more later, but the gist is that the DAC has a limited slew rate compared to the sample rate I used. So, I reduced the value in the DAC by 2 and then compensated for it in this gain stage.

A note about the op-amps used in this design: The design calls for something that can handle 120 kHz passing through a gain of up to 5 and also dealing with the Sallen-Key filters (the TI WEBENCH shows a gain-bandwidth requirement of at least 6 MHz). I also needed a slew rate that could deal with a 120 kHz signal with a level of 3.3 Vpp. The STMicroelectronics TSV782 fit the bill nicely.

The last two components are the resistor and the capacitor before the output BNC connector. The resistor is used to stabilize the op-amp circuit if the output is connected to a large capacitance load. The 1uF capacitor provides AC coupling to the output BNC.

The MCU

The brains used in this design is a Feather M4 Express SBC, which contains a Microchip Technology’s ATSAMD51 that has a Cortex M4 core. This is primarily powered by a USB connection (or a battery we will discuss in Part 2).  

This ATSAMD51 has a few ADCs and DACs, and we use one of each in this design. It also has plenty of memory (512 kB of program memory and 192  kB of SRAM).

It runs at a usable 120 MHz and is enhanced with a floating-point processor. All this works nicely for the digital filtering we will explain in Part 2. Other features I used include a number of digital I/O ports, an SPI port, and a few other ADC inputs.

One feature I found very nice on the SBC was a 3.3 VDC linear regulator that not only powers the MCU, but has sufficient output to power all other devices in the design.

On the schematic (Figure 1), you can see that the AFE connects to an ADC input on the SBC, and an SBC DAC connects to the ABE circuit. Another major component is the TFT LCD and touchscreen, powered by the 3.3 VDC coming from the SBC.

Miscellaneous schematic items

That leaves a few extra items on the schematic.

Voltage reference

There are 2 simple ½ voltage dividers to generate 1.65 VDC from the 3.3 VDC supply. One is used on the AFE to get a mid-voltage reference for the single supply op-amp design. This reference is simply two equal resistors and a capacitor connected to ground, and from the center of the series-connected resistors.

A second reference was created for the ABE circuit. I used two references as I was laying this out on a protoboard, and the circuits were separated by a significant distance (without a ground plane).

LED indicator

There is also an LED used to indicate that the ADC is clipping the signal because the input is too large or too small. Another LED indicates the DAC is clipping for the same reasons. There will be more discussion on this in the firmware section in Part 2.

Floating ground

An interesting feature of the SBC is that it contains the charging circuit for a lithium polymer 3.7-V battery. This is optional in the design, but it does allow you to operate the DFS with a floating ground and a quiet voltage supply, which may help in your testing.

Enable

A somewhat unique feature, which turns out to be helpful, is an enable that is used to turn off the system if you pull it to ground.

If you use a battery, along with the USB, and want to use a typical power on/off switch, you would need to break the incoming USB line and the battery line, which makes it a 2-pole switch.

So, to get the DFS to power down, I pull the enable line to ground using a 3-pole SPDT switch, which I found has the typical “O/I” on/off indications. You can use a SPST switch; this will have to be switched to “I” to shut it down and “O” to turn it on.

USB voltage display

A ½ voltage divider, with a filter capacitor, is connected to the USB input and used as an input to one of the ADCs, so we can display the connected USB voltage.

Optional reset

The last item is an optional reset. I did not provide a hole to mount a pushbutton, but you can drill a hole in the back of the enclosure for a normally-open pushbutton.

More information

This device is a fairly easy to build. I built the circuit on a protoboard with SMT parts (thru-hole would have been easier). Maybe someone would like to lay out a PCB and share the design. I think you’ll find this DFS has a number of uses in your lab/shop.

The schematic, code, 3D print files, links to various parts, and more information and notes on the design and construction can be downloaded at: https://makerworld.com/en/@user_1242957023/upload

Editor’s Note: Stay tuned for Part 2 to learn more about the device’s firmware.

Damian Bonicatto is a consulting engineer with decades of experience in embedded hardware, firmware, and system design. He holds over 30 patents.

Phoenix Bonicatto is a freelance writer.

Related Content

• Non-linear digital filters: Use cases and sample code
• A beginner’s guide to power of IQ data and beauty of negative frequencies – Part 1
• A Sallen-Key low-pass filter design toolkit
• Designing second order Sallen-Key low pass filters with minimal sensitivity to component tolerances
• Toward better behaved Sallen-Key low pass filters
• Design second- and third-order Sallen-Key filters with one op amp

The post A digital filter system (DFS), Part 1 appeared first on EDN.

This Design Idea (DI) offers an alternative solution for an application borrowed from frequent DI contributor R. Jayapal, presented in: “A 0-20mA source current to 4-20mA loop current converter.” 

It converts a 0/20mA current mode input, such as produced by some process control instrumentation, into a standard industrial 4/20mA current loop output.

Wow the engineering world with your unique design: Design Ideas Submission Guide

Figure 1 shows the circuit. It’s based on a (very) old friend—the LM337 three-legged regulator. Here’s how it works.

Figure 1 U1 plus R1 through R5 current steering networks convert 0/20mA input to 4/20mA output.

The fixed resistance of the R1 + R2 + R3 series network, working in parallel with the adjustable R4 + R5 pair, presents a combined load of 312 ohms to the 1.25v output of U1. That causes a zero-input current draw of 1.25/312 = 4 mA, trimmed by R5 (see calibration sequence detailed later).

Summed with this is a 0 to 16 mA current derived from the 0 to 20 mA input, controlled by the 4:1 ratio current split provided by the R1/R2/R3 current divider and fine trimmed by R2 (ditto). 

Note that 4 mA is below the guaranteed minimum regulation current specification for the LM337. In fact, most will work happily with half that much, but you might get a greedy one. So just be aware.

The result is a precision conversion of the 0 to 20mA input to an accurate 4 to 20mA loop current. Conversion precision and stability are insensitive to R2 trimmer wiper resistance due to the somewhat unusual input topology in play.

Calibration proceeds in a four-step linear (iteration-free one-pass) sequence consisting of:

1. Set input = 0.0 mA.
2. Adjust R5 for 4.00 mA loop current.
3. Set input = 20.00 mA.
4. Adjust R2 for 20.00 mA loop current.

Done.

The input voltage burden is a negative 1.0 volt. The output loop voltage drop is 4 volts minimum to 40 volts maximum. The maximum ambient temperature (with no U1 heatsink) is 100oC. Resistors should be precision types, and the trimmer pots should be multiturn cermet or similar.

Stephen Woodward’s relationship with EDN’s DI column goes back quite a long way. Over 100 submissions have been accepted since his first contribution back in 1974.

Related Content

• A 0-20mA source current to 4-20mA loop current converter
• A two-wire temperature transmitter using an RTD sensor
• Two-wire interface has galvanic isolation
• Low-cost NiCd battery charger with charge level indicator
• Single phase mains cycle skipping controller sans harmonics
• Two-wire remote sensor preamp

The post Silly simple precision 0/20mA to 4/20mA converter appeared first on EDN.