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.

TPA3116D2 and Transformer Coupled Load

Other Parts Discussed in Thread: TPA3116D2

I'd like to use an integrated device like the TPA3116D2 to drive higher impedance loads (~100 Ohms). I also require an isolated output. A transfomer coupled output seems to be the best fit. Has anbody tried this before or does somebody know the pitfalls of using a device like this to drive a transformer coupled load? Is it better to filter the output signal before or after the transformer? Any input on output filter selection etc. is greatly appreciated along with any other design pitfalls. Thank you.

  • Hi, Randy. We have worked with designs using a TPA31xx audio amplifier to drive a transformer load. I've attached a TI Application Report on the subject.

    4572.Audio_Power_Amplifier_Operation_with_Transformer_Load_sloa133.pdf

    Additional filtering (beyond the LC output filter) is generally not required, unless there are special EMC requirements.  There are potential issues with transformer saturation and DC offset.

    - Transformer saturation can permit very high primary currents that put an APA (audio power amplifier) into shutdown.  The attached Application Report covers this topic.

    - Primary resistances are generally low, so DC offset can generate some current in them.  It's just necessary to make sure offset current is insignificant compared to the saturation point of the transformer.  The saturation current information can be found from the transformer vendor.

    As for driving higher impedance loads, the main concern is the LC filter design to avoid peaking. This Application Report on LC filter design may also be useful to you. On the TPA3116D2 EVM, for example, we use 10 µH inductors and 0.68 µF capacitors on the output filter, for driving a 4 Ω load. A critically damped output filter is ideal, with a Q point being at 0.707, but an underdamped filter with Q ≈ 1 is acceptable. Q = Rload× √(C/L), so for the EVM filter Q = 1.04.  With a 100 Ω load, following the Application Report, I find that a values that could work are L = 1000 µH and C = 0.047 µF, making Q = 0.68. You're welcome to do your own calculations, but I like these numbers because the cutoff frequency is around 20 kHz and the inductor is not as big as it could be. We had worked with one load of 1 kΩ before, and the filter we suggested used L = 3.75 mH and C = 0.015 µF.

    Good luck!
    Matt