Kirk Gray

Loop Gains Still Rule Feedback Yesterday, I watched two webinars making the same claim: That loop gain measurements don't work for nonlinear, time-varying systems. Since that includes every switching regulator ever built, it says that the Bode plot method never works for converters. Did Middlebrook and Cuk, followed by Venable and the entire industry get it wrong all those years? No, of course not. It's total nonsense. The loop gain measurement, as Middlebrook showed long ago, is the most rugged method for verifying the feedback loop and for seeing how to improve designs. Why does this continue to be misunderstood? Your guess is as good as mine. Measure your loops. It's just not that hard to be rigorous in your design.

161

22 Comments

How to turn the voltage analog input into a current one? Sometimes you suddenly need to connect another 4–20 mA sensor to your system, but you don’t have this type of analog input available. But you have a 0–10 V one. For example, the S7‑1200 CPU has built‑in analog inputs that can only be configured as voltage inputs. In fact, it is quite easy to convert to mA input. See my circuit simulation. You just add a 500 Ohm resistor in parallel with the input, and that is it! #otomakeit #efficiency #industrialautomation #controlsystems #controlpanel #Siemens #PLC

54

13 Comments

This @shows a PWM-based DC motor speed control using a 555 timer IC configured in astable mode. The 555 timer generates a continuous pulse width modulated (PWM) signal, where the duty cycle can be adjusted using the variable resistor VR1. Diodes D1 and D2 control the charging and discharging paths of the capacitor, allowing independent adjustment of ON and OFF times. A narrow pulse gives low average voltage and slower motor speed, while a wider pulse provides higher average voltage and faster speed. The output from the 555 timer drives a power transistor (2N3055), which delivers the required current to the DC motor. This design is efficient and ensures smooth control without significant power loss, commonly used in robotics and small motor control systems. #ElectronicsEducation #PWM #DCMotorControl #555Timer #ElectronicsProject #TransistorSwitching #PulseWidthModulation #MotorSpeedControl #DIYElectronics #EngineeringBasics #ElectronicsRD

28

Do you make the same mistake as I often do? :) How often this happens to me: - The hardware configuration is completed - Code is completed and compiled without errors and warnings - Downloaded to the PLC and started testing And for some reason, the analog inputs begin behaving crazily: showing open circuit, or value jumps from almost zero to almost max with minimum change on the input After countless mistakes like this, I know where to look immediately. But it wasn’t always that clear for me. Go to hardware configuration, open your Analog Input module properties and check what type of input is configured: voltage/current and the range. Most probably it is wrong :). Sounds stupidly simple? Yes. But how many times does it get forgotten! Planning a new project? Message me to see how I can help. #otomakeit #efficiency #industrialautomation #controlsystems #controlpanel #PLC #Siemens

120

12 Comments

Precision circuits are failing EMI compliance tests. High-speed interfaces are experiencing signal degradation. Mission-critical systems are showing premature field failures due to thermal stress. These aren't isolated incidents; they represent a fundamental gap in how we approach electromagnetic interference management in modern electronics. The technical specifications tell the story. While conventional solutions deliver 38-42 dB suppression, precision applications demand up to 53 dB to protect sensitive analog stages. Standard 3.5-4.1 kΩ common mode impedance leaves residual noise that degrades performance, whereas 5.2 kΩ at 100 kHz provides efficient EMI blocking that meets stringent requirements. Current handling separates theoretical performance from operational reality. Supporting 5.1 A continuous load without saturation ensures consistent filtering, while typical 3.5-4 A capacity creates heat buildup that undermines reliability. Premium ferrite cores with 98% magnetic retention operate stably to 165°C, outperforming standard materials with 88-90% retention that degrade beyond 140-150°C. Signal integrity depends on precision. Leakage inductance control below 2% variation preserves high-speed protocols like USB, CAN, and Ethernet, while 5-7% variations create intermittent failures. Broadband coverage from 10 kHz to 5 MHz addresses modern switching harmonics more effectively than narrow 20 kHz-2 MHz solutions. The strategic advantage lies in IEC and UL compliant components tested to 105,000 hours MTBF with sub-1.9% thermal losses. These specifications define reliability standards for SMPS, automotive ECUs, medical imaging, industrial drives, and telecom infrastructure where performance margins determine market leadership.

11

Why do we rely on magnetics for stepping down our current measurements when there are inherent issues related to magnetic coupling? What are these issues you ask? Mainly, issues related to non-linearity in the difference between the primary and secondary currents due to low currents that are too weak for excitation or too high or DC offset that the generated flux is no longer contained to the core after saturation.  Slight saturation isn't even that uncommon.  You have a poor measuring device under certain conditions but poor is always relative the need. What options are there instead of magnetic CTs to measure the primary currents? Faraday optical current transformers that pass a fiber around a conductor and measure the current based on the magnetic field generated current's effects on the polarization of the light in the fiber. Rogowski's coils use the magnetic field to generate a voltage with an air-core transformer.  Air is not saturable and you get away from an iron core due to not needing to drive a current. DC can't be measured because mutual coupling is needed. Hall sensors measure the small voltage that develops when a conductor's electrons are pushed to one region due to a magnetic field.  The benefit of this method is that it can measure DC. The last one is a resistive shunt. This method steps down the current to a small voltage by precisely known resistance of a segment and measuring its small voltage drop.  Why are magnetic CTs still used if some of these are either cheaper or able to handle DC better? Magnetic CTs have an incredible track record due to their simple design. How many other transducers can you put into the field and expect 50-75 years of services? These other methods have electronics, which might not last that long, and often need an external power supply.  Probably an odd driver would be that you have an entire industry that has built itself around measuring 100 > amps.  Switching to these other transducers requires something to read the small integrated current or voltage signal or polarization angle. This can be done but there are fixes for this obviously, but large utilities like standardizing on procedures, test equipment, and equipment specifications. That is just a lot of hassle to avoid non-linear saturating effects, which are so minor that relays often have algorithms to detect and compensate for slight CT saturation. If you enjoy these articles, consider signing up for a two-part tutorial on "Sequence Components and Faults 1 & 2".  Each part is 90 minutes and $100 each.  Contact me for more information. I will be very flexible if the time slots don't work for you as long as you stay flexible in lining up with a few others.  I am expecting it to be good and if you go to my profile toolkit link, there is a beta browser Fault Analysis program that I am putting together for the tutorial to provide a lot of bang per your buck. #utilities #renewables #energystorage #electricalengineering

99

6 Comments

Understanding PCB Thickness—Why It’s More Than Just a Number! Ever wonder why a board might be 1.6 mm thick instead of 0.8 mm? In our latest blog, we break down how board thickness impacts mechanical strength, signal integrity, heat dissipation, and manufacturability. Whether you’re slimming down a wearable or ruggedizing an industrial device, thickness matters. Read more: https://lnkd.in/gjW2Cn7X Send your RFQ, we add value! Learn more: ktts.us Request for quote: https://lnkd.in/gAWMHAqh Book a call: calendly.com/meetings-ktts Subscribe for the latest updates: https://lnkd.in/gua6gy9k Shayan Khan David Dillman Biofel Carpio Amera A Catherine (Fergie) Ferguson wendy folise Will Tyson, CPIM Basak Sezgin Harshad Dholakiya Eli L. Smith, ETA CPP™ Benjamin Williams Scott Yeakel Cheryl F. Heidi Derrick Andrea Tibbitts Kyle Thompson Michael Engler Matt Casey Justin Nerdrum Nikita Bordbar Claire Caires Dan Arbuckle Ken Thompson Suma Akter Holland Jackson Shahmeer Khan 🇺🇸 Garrett Jackson Carla Andreina Fuentes Lopez Isaiah Watson Tejas T. #PCBDesign #BoardThickness #ElectronicsEngineering #TapRen #KTTS #KTTechnicalSolutions

9

I present to you Sierra Circuits' last design guide of 2025: The Trace & Space Handbook. This handbook contains practical guidelines for setting line widths and clearances. Following these rules will ensure manufacturability and compliance with industry standards. My team broke down how factors like current requirements, target impedance, and copper weight impact your choices for line width and clearance. Anything else we should add? Download here and share your feedback: https://lnkd.in/gdmVWsyy

133

7 Comments

Technical Note: Why do most Low-Voltage VFDs use electrolytic capacitors? If you’ve ever looked inside a low-voltage drive, you’ve seen them: large DC-bus electrolytic capacitors. They’re not there by accident. - High energy storage in a compact size Electrolytics provide very high capacitance in a small footprint, critical for wall-mounted and OEM-friendly LV drives. - Cost-effective DC bus stabilization They smooth rectified AC into stable DC, keeping inverter switching clean and torque output consistent, without driving system cost up. Short ride-through capability They store enough energy to help the drive ride through brief voltage sags and line disturbances. Industry-proven technology Decades of field use, well-understood design rules, and predictable performance make them the go-to choice across most vendors. ⚠️ The trade-off? Electrolytic capacitors age over time. Heat, ripple current, and long periods without energization can accelerate degradation — often silently. ▶️ This is why DC-bus capacitors are typically the life-limiting component in LV drives, and why storage, preventive maintenance, and operating conditions matter. Understanding this helps avoid surprises in the field and improves reliability conversations with customers. ✅ Preventive Maintenance Tips for DC-Bus Electrolytic Capacitors (LV VFDs) 1️⃣ Control temperature — always • Heat is the #1 life reducer of electrolytic capacitors. • Keep ambient temperature within drive specs. • Ensure clean airflow, working fans, and check air vents. Rule of thumb: every 10 °C increase cuts capacitor life roughly in half. 2️⃣ Periodically energize stored or backup drives • Long periods without power accelerate electrolyte degradation. • Energize drives every 3–6 months to reform the dielectric layer. • Follow manufacturer storage and reforming guidelines. 3️⃣ Minimize DC-bus ripple current • Excessive ripple increases internal heating and ESR rise. • Verify: • proper line impedance, • correct reactor / choke sizing, • harmonic mitigation where required. 4️⃣ Monitor capacitor health indicators • Many LV drives estimate capacitor life based on: • operating hours, • temperature history, • ripple current models. • Use these as trends, not absolute guarantees. 5️⃣ Inspect cooling components during PM cycles • Fans, filters, and heat sinks directly impact capacitor life. • Replace failed fans immediately — capacitor damage often follows silently. 6️⃣ Avoid repeated power cycling and harsh starts • Frequent energization/de-energization stresses DC-bus components. • Where possible: • avoid nuisance power cycling, • use controlled startup practices. 7️⃣ Plan capacitor replacement proactively • Treat DC-bus capacitors as consumable components, not lifetime parts. • In critical applications: • plan replacement during scheduled outages, • don’t wait for failure. #VFD #IndustrialAutomation #PowerElectronics #Reliability #Drives #HVAC #WaterWastewater

11

1 Comment

Happy #CostModelMonday. I was at #IMAPS #CHIPcon last week talking about #chiplets, so I thought I'd share another bit of analysis I did for my presentation there. I'm always saying that the cost trade-off between breaking up a monolithic die into chiplets is simple: silicon cost goes down, packaging cost goes up. The fun part comes when you have to answer about a billion questions about what nodes you're breaking up, what's the IP reuse situation, what type of packaging are you using, and so forth and so on, to actually get to a point where you know how much you saved on silicon and what you paid for the packaging. Changing the answer to one of those questions--or, to put it another way, changing one single variable in your whole chip and package design--changes the cost. That's why trade-offs are so important. The baseline in the first row of the table below shows the results of breaking a large monolithic SoC die (flip chip packaging) into multiple chiplets on an interposer. If it was a 10nm die that was broken up, the results showed that the chiplet design was about 10% more expensive, which means it was a case where the silicon savings are NOT enough to offset the increase in packaging costs. Then I made a couple changes to see what might happen. What does the table below tell us? Well, the second row shows that yield matters a lot, assuming some perfect yields would save us enough that the SoC flip chip package and the chiplet package would become the same cost. The third row, on the other hand, tells us that shifting some routing from the top of the interposer to the bottom doesn't save all that much.

8

It’s Not Just a Probe Station... It’s Also a Shield Most researchers know our SATs probe station for its precision contacts, temperature control, and XYZ movement. But here's something few realize: It silently doubles as a Faraday cage. That’s right while you’re running sensitive low-frequency EIS measurements, the chamber helps block out ambient lab noise and electrical interference. No extra enclosures. No aluminum foil hacks. Just clean, stable, low-noise data by design. Because real measurements aren't just about the signal They're about removing everything that isn't. Follow SATs for instruments made by researchers, for researchers. #TechHighlight #EIS #FaradayCage #ProbeStation #MaterialScience #ScientificInstruments #LowNoiseMeasurements #PhDLife #STEMCommunity #SATsInstruments #AcademicResearch #Electrochemistry #Nanotech

14