If you’ve ever worked with power electronics, you probably know integrated transformers aren’t just your basic step-up or step-down parts. As someone who’s been supplying these for over a decade—talking to engineers at motor drives, EV chargers, and renewable energy inverters every day—I’ve seen how one tiny spec can make or break a project: frequency response. Too many folks sleep on it, and that leads to headaches down the line. Let’s break this down straight, no jargon fluff, and tie it to why it matters when you’re picking an integrated transformer for your build. Integrated Transformer
First, let’s get what frequency response actually means for these transformers, not just the textbook definition. Integrated transformers (we’ll call them ITs from here on, since that’s what everyone in the field shortens it to) are way smaller than traditional ones— they’re built right onto a PCB, usually combining the primary, secondary, and even auxiliary windings all in one compact core. Their frequency response is just how their electrical characteristics (like impedance, gain, and phase shift) change when the input signal’s frequency goes up or down. Unlike big, beefy transformers for grid use, ITs operate across a much wider range of frequencies—think 10 kHz to 1 MHz, depending on the application. That wide range is where the problem starts: if the frequency response isn’t tailored to your use case, the IT won’t perform how you need it to, no matter how nice its specs look on paper.
Let’s start with a super common pain point: signal distortion. I had a customer last year building a 50 kW EV on-board charger. They went with a cheap IT from a company that didn’t publish proper frequency response curves, just the rated voltage and power. When they cranked up the switching frequency to 150 kHz (standard for modern EV chargers to shrink the overall size), their output waveform had massive harmonic distortion—over 8% total harmonic distortion, way above the 3% limit required for automotive compliance. When they swapped to our ITs, which are designed with a frequency response optimized for 100–250 kHz, the distortion dropped to 2.1%. Why? Because our IT’s frequency response curve is flat across that whole range, meaning the gain stays consistent, so every frequency component of the signal is amplified or attenuated the same amount. The cheap one they used had a steep rolloff above 120 kHz— it was acting like a low-pass filter, eating up the higher harmonics they needed for clean power. That’s the kind of issue that doesn’t show up in a datasheet’s basic specs, but kills a project’s timeline.
Next up: efficiency, which is the name of the game for almost every power electronics build. Let’s tie frequency response to core and copper losses, the two big drags on efficiency. Core loss comes from the magnetic core in the IT, and it spikes when the frequency drifts outside the transformer’s resonant frequency. Wait, resonant frequency is a huge part of frequency response— every integrated transformer has a self-resonant frequency (SRF), the point where its inductive reactance equals capacitive reactance. If your operating frequency is too close to the SRF, you get wild impedance swings, and core loss skyrockets. I had another customer a few months back working on a solar micro-inverter that ran at 60 kHz. They picked an IT with an SRF of 70 kHz— way too close. When they tested, their efficiency was 91%, 3% lower than their target. That extra loss meant they had to add a bigger heatsink, which ruined their compact design. We sent them an IT with an SRF of 180 kHz, so 2x their operating frequency, and their efficiency jumped to 94%. Why? Our IT’s frequency response is engineered to keep impedance stable away from the SRF, so core and copper losses don’t jump at their working frequency. Too many suppliers just slap a random core on an IT and don’t tune the SRF for different use cases— that’s where we stand out, because we test every unit’s frequency response at every stage of production.
Don’t sleep on phase shift either, especially for high-frequency or parallel operation. Phase shift is the delay between the input signal and output signal, and it changes with frequency. For things like interleaved power supplies, where multiple ITs work in parallel, that phase shift mismatch can cause current imbalances between units. If one IT has a 5-degree phase shift at 100 kHz and another has a 12-degree shift, the current in the second unit is way higher, leading to overheating and early failure. Last year, a team doing a 2-parallel 100 kW motor drive came to us with exactly this problem— their original ITs had a 10-degree spread in phase shift at 80 kHz, so one unit was carrying 60% more current than it was rated for. We supplied ITs with a matched frequency response, so phase shift was within 1 degree across all units, and the current balance fixed right up. That’s not a spec you’ll always see in a datasheet, but it’s critical for parallel builds, which are super common in industrial and automotive setups.
Now, let’s talk about noise. Integrated transformers are used in everything from audio gear to medical devices, where noise is a non-starter. The frequency response directly impacts electromagnetic interference (EMI), which is that annoying electrical noise that messes with nearby components. If an IT has a peak in its frequency response at a frequency that’s a common EMI band (like 150 kHz to 30 MHz, the FCC’s required limit range), that peak will radiate noise through the PCB or wires. A customer building a portable medical ultrasound machine came to us after their product failed EMI testing because their off-the-shelf IT had a frequency response peak at 250 kHz, exactly where FCC requires emissions to be under 10 dBuV/m. We designed a custom IT with a flattened frequency response that had no peaks in the 150 kHz to 1 MHz range, and they passed first time. No reworks, no delays to their launch— that’s the kind of impact frequency response can have on regulatory compliance, which is make-or-break for medical and automotive products.
Wait, let’s clear up a common myth here: some people think a “good” frequency response is just a flat curve across all frequencies, but that’s not true. It depends on your application. For a low-frequency transformer (say, under 10 kHz) for a grid-connected UPS, a flat response is great, but if you’re working at 500 kHz for a high-density charger, a flat response is actually less important than keeping the SRF way above your operating frequency. That’s why our team doesn’t just give out generic ITs— we work with you to test what frequency range your project runs in, then tune the IT’s core material, winding count, and core geometry to shape the frequency response exactly for that use case. For example, we use nanocrystalline cores for ITs running above 100 kHz, because their frequency response stays stable and core loss is way lower than ferrite at high frequencies, whereas ferrite works fine for lower frequencies under 50 kHz.
I’ve seen so many engineers cut corners on frequency response because it’s not as flashy as power rating or size. But here’s the thing: power rating is a static number. Frequency response is dynamic, it changes with your operating conditions, and it’s what makes the IT work for your specific build. Last month, a young engineer reached out to me frustrated because their project was failing even though they’d spent twice the budget on a top-brand IT. We pulled up the frequency response curve, and sure enough, their operating frequency was right at the SRF of that brand’s IT, causing 5x the normal loss. We swapped them to our IT, tuned for their exact frequency, and they got their prototype working in a week. That’s the value of knowing your stuff when it comes to integrated transformers.
So what should you actually look for when you’re checking frequency response for your IT? First, don’t just trust the datasheet’s single curve— ask the supplier for the full sweep, not just a simplified one. The full sweep will show you where the SRF is, where the impedance rolls off, and if there are any sharp peaks in the EMI bands. Second, match the SRF to your operating frequency: aim for at least 1.5x your max operating frequency, to avoid that impedance spike. Third, if you’re running multiple ITs in parallel, ask for phase shift specs at your exact frequency— make sure they’re matched within 2 degrees, to avoid current imbalance.
At the end of the day, integrated transformers are specialized parts, and frequency response isn’t an afterthought. It’s the backbone of how they perform, whether you’re building an EV charger, a solar inverter, a motor drive, or a medical device. If you’re currently testing ITs and seeing distortion, low efficiency, or EMI issues, odds are it’s a frequency response mismatch. We’ve been solving these exact problems for engineers for years, and we can help you pick or design an IT tailored to your frequency needs. If you want to chat through your project, bounce ideas off our team, or ask for a custom frequency response test sample, just reach out to our team to start the conversation.
Pad Mounted Transformer References
- Mohan, N. (2012). Power Electronics: Converters, Applications, and Design. John Wiley & Sons.
- El-Hamamsy, A. F., et al. (2019). “Frequency Response Characterization of Integrated Power Transformers for High-Frequency DC-DC Converters.” IEEE Transactions on Power Electronics.
- IEEE Std. C57.12.90-2015, “Standard Test Code for Dry-Type General-Purpose and Distribution Transformers.”
Nantong Yawei New Energy Technology Co., Ltd.
As one of the most professional integrated transformer manufacturers and suppliers in China, we’re featured by quality products and good service. Please rest assured to wholesale durable integrated transformer made in China here from our factory. Customized orders are welcome.
Address: Room 28-101, Building 27 and 28, No.333 Kaiyuan Avenue, Sunzhuang Subdistrict, Hai’an City, Nantong City, Jiangsu Province, China
E-mail: admin@nantongyawei.com
WebSite: https://www.nantongyawei.com/
