🕵️ In this week Newsletter we’ll have a look at the 130W LLC Transformer sample that I ordered from our Lab facilities in Madrid. Remember my last Newsletter on a 130W LLC Transformer from ST Devices App Note? Yes, we're going to analyse that sample!
It’s already a good thing to be able to compare simulations, but it’s even better to have the sample in your hands and being able to test it!
⚙️ Specs reminder
Take a look at Figure 1, just as a reminder of the specs and a 2D cross section of the bobbin of the simulated Transformer.
☝️The important point is that the Transformer is a center tap design with grouped secondary windings, as shown above.
🧑🏭 Building samples
Figure 2. Sample construction process
In Figure 2 you can see the sample wound process, starting from step 1 to 5.
✅ Steps 1-3:
The primary winding approximately extends to the middle of the bobbin. A single served 105x0.05mm litz wire was used and, in the middle of the bobbin, several layers of margin tape (5mm width) are applied to create the barrier between the primary and secondaries.
💰 That’s an inexpensive way to create chambers using regular bobbins. The other way to do it is by 3D printing a custom bobbin, again at the prototype stage.
✅ Step 4-6:
The secondary windings are wound grouped together as shown in Figure 1. A 135x0.071mm litz wire was used for the 2 secondaries and, finally, 3 turns of triple insulated 0.25mm wire was used for the auxiliary winding.
📐 Measurements
Table 1. Transformer measurements
💻 In Table 1 we can check all important parameters of the Transformer. I’m particularly interested in the leakage inductance value of 174.2uΗ @100kHz which, by the way, was predicted at 175uH by Frenetic Online!
That 174.2uH is the secondary (one of the two grouped windings) leakage inductance as measured from the primary side with the respective secondary shorted. That is the inductance value that will dictate the resonant frequency along with the series capacitor.
📈 In the test board, the operating frequency measured at 88 to 106kHz. In fact, due to the natural low frequency ripple of the PFC stage in front of the LLC stage, a constantly changing frequency will be observed as jitter. That is perfectly normal because the LLC converter is just trying to compensate for the 100Hz (x2 line frequency) ripple of the PFC. That can be easily verified using a 400V PSU, that presents a couple hundred mV of voltage ripple, instead of volts that the PFC has, which results in a minimized frequency jitter.
💭 Leakage inductances Q&A
🤌 Why not use the value of leakage inductance with ALL secondaries shorted (as measured in Table 1 – 1st row) to calculate the resonant frequency?
🧑🏫 In a center tap topology only 1 out of the 2 secondaries is conducting every time. The leakage inductance of one of the secondaries is reflected to the primary during each half cycle not both! And that is an important detail!
🤌 What if the measured leakage inductance shorting secondary 1 is different than that of secondary 2?
🧑🏫 Then the resonant frequency will not be the same during each half of the period. That can be easily observed if we look at the current of the primary winding as shown in Figure 3. That is the reason why we group the secondary windings in center tap topologies. That way we can ensure the best inductance/leakage inductance/current sharing match between windings.
Figure 3. Slight differences in resonance between half cycles
🔎 Hi-Pot Testing
The Transformer was designed with reinforced isolation and complies with IEC60950 safety standard, therefore it handles 4kV between the primary and secondary with a sinusoidal 50Hz AC test voltage for 60s. Our sample passed the test with 60uA of leakage current, but the test was repeated again at 2kV between primary/secondary, and the core observing 20uA of leakage current. Both values are much lower than the 10mA mark we set for failure. In addition, no arcing was observed during tests leading to a successful Hi-Pot test.
📊 Simulation VS Reality Performance
Figure 4. Natural convection on top and forced air cooling at the bottom.
🔥 In Figure 4, we can check the simulation and real thermal images of the sample Transformer. Bear in mind that the camera error is +-2°C or +-2%, whichever is greater.The temperature differences are minor and within the measurement error range!
🤌 How do I know the velocity of air (m/s) blowing though the fan to specify it in Frenetic Online simulation software?
🧑🏫 Here is the setup:
Figure 5. Natural convection and 50x50x10mm fan cooling setup
🤓 The fan chosen was a 50x50x10mm 12V located 10mm away from the Transformer, as shown above. To be honest, this Transformer doesn’t need any cooling, but it’s another easy test to do, so why not?
Here is a quick way to find how many m/s does a fan support from datasheet info alone. In particular this one (Mfr part No: EE50101S1-999-A), from its datasheet, has 12CFM at 12V, and that’s enough information. Using the equation below we calculate the maximum air speed:
In ft2 we measure the area of the fan. In our case using the fan dimensions: 50x50mm=2500mm2->0.0269ft2
✨ One cool observation is that fan cooling caused the transformer losses to increase from ~2W to 2.8W. That is due to the difference in the core losses as clearly shown in Figure 4. This is kind of counterintuitive, but the truth is the core losses decrease with a higher temperature, so the Transformer performs better at 65°C than at 35°C. The copper resistance increases as temperature rises, but in this particular design this doesn’t make a significant difference in the winding losses, so the core losses dictate the increased total power losses in the two cooling scenarios.
🦸 Book the last seats available for the next Frenetic's Training Program!
Achieve Magnetics Superpowers with unique modules and lectures exploring High-Frequency Magnetics design. Next session is starting on May 23rd 2023.
🚀 Check out all the information and pre-book your seat here.
😎 I hope you’ve enjoyed this Newsletter! Stay tuned for the next one.
References
ST devices, May 2016, “48 V - 130 W high-efficiency converter with PFC for LED street lighting applications”, AN3106
