• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 512 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LMG3422R030: Inquiries Regarding High Operational Switching Frequency & PCB Layout Design

Imad Shaham
Prodigy 40 points
Part Number: LMG3422R030
Other Parts Discussed in Thread: LMG3410R150

Hi,

I am an undergraduate student working on my graduation project titled 'The Design and Implementation of a GaN-based Class D Audio Power Amplifier'. For the switching devices, we have selected the LMG3422R030. However, the Evaluation Kit provided alongside the Datasheet uses a switching frequency less than 200 kHz. Our goal is to operate the GaN at a switching frequency of 500 kHz, and so we have the following questions:

1) Is the device capable of hard switching at the desired frequency? The Datasheet mentions a 2.2 MHz maximum switching capability, but is that figure for hard switching or soft switching?

2) What necessary changes are required in the schematic to achieve the desired switching frequency?

3) In the provided Evaluation Kit the PCB consists of 4 layers, but the layers are not described in detail. What is the purpose of each layer in the PCB precisely? Whether that is PWR, GND, signal, or otherwise.

Best Regards and Thanks,

I. Shaham

  • 0
    TI__Expert 8960 points

    Hello,

    1. LMG3422R030 can electrically handle hard switching at 500kHz. Whether it can thermally handle hard switching at 500kHz is determined by cooling ability, current, and especially input voltage.

    2. Schematic changes won't help.

    3. The EVM doesn't assign layers to just one task such as power, signal, or ground. You can look in the LMG3422R030 datasheet for layout guidelines.

    Thanks,

    Travis

  • 0
    Prodigy 40 points

    Thanks for your reply Mr.Travis!

    Can you please refer to me regarding the first question in the datasheet? Where it clearly says that the device can operate in hard switching mode at higher frequencies.

    With Regards.

  • 0 in reply to Imad Shaham
    TI__Expert 8960 points

    Hi Imad,

    The datasheet does not state anywhere that it will handle hard switching at 500kHz, because it is dependent on application, particularly regarding thermals.

    The "gate driver" section of the datasheet states the maximum gate drive frequency of 2.2MHz, and the gate driver does not care about hard vs soft switching. LMG3422 can hard switch at up to 2.2MHz with very low input voltage & current.

    You would need to do the power loss calculations to determine if your design could handle hard switching at 500kHz.

    Thanks,

    Travis

  • 0 in reply to Travis Williams
    Prodigy 40 points

    Thank you again Mr.Travis!

    Regarding the calculations.. is it easy to be done? how it is calculated in general?

    We are aiming for a power of 500W RMS on 8 ohms loadspeaker.. full bridged with +100VDC supply rail.

    Excuse my lack of knowledge

    Thanks!

  • +1 in reply to Imad Shaham
    TI__Expert 8960 points

    Hi Imad,

    If you want the best estimate of switching losses, there's no substitute for a bench test. Power loss testing is one of the major purposes of our half bridge EVMs. Additionally, doing a high accuracy calculation is not trivial. This document goes into more detail:

    https://gansystems.com/wp-content/uploads/2018/04/APEC18-Parasitic-Capacitance-Eqoss-Loss-Mechanism-Calculation-and-Measurement-in-Hard-Switching-for-GaN-HEMTs.pdf

    With that being said, the main losses for almost all GaN hard switchers are the RDSon losses, overlap losses, & Qoss losses.

    Here is the graph for LMG3422R030 inductive switching losses:

    For this example, I will assume a 50% duty cycle on one of the half bridges & 5A RMS.

    RDSon per FET: 3.5A*3.5A * 0.03Ω = 0.37W

    The switching energy from the graph (which primarily comes from Qoss) is ~70uJ at 5A. The actual loss for you application will be much lower, as this was measured at 480V.

    Even at 480V, the total loss per FET at 500kHz is about 35W, which can be managed with heatsinks & airflow. In reality the loss should be several times lower than 35W due to lower input voltage.

    Please note that this is a rough calculation that is intended to provide a rough estimate of losses, but a bench test should be done if you need accurate information.

    Additionally, the LMG3422R030 seems vastly oversized for this application, which is harming efficiency. You might consider switching to a smaller FET such as LMG3410R150 in order to reduce losses. Also, it is less than half the price. We have EVMs available for both if you want to perform power loss testing.

    LMG3410 switching energy graph for comparison:

    Thanks,

    Travis.