• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 31 subscribers
  • Views 1258 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TAS5613A: Amplifier based on TAS5613A external protection Schottky diodes failures

Les_K
Prodigy 110 points
Part Number: TAS5613A

Performing thermal testing of a 2-channel audio amplifier based on TAS5613A in PHD package I encountered a problem of failures of the external protection Schottky diodes on the chip outputs.

Amplifier Configuration

The amplifier schematic is similar to the TAS5613PHD2EVM evaluation board schematic with two channels in BTL mode powered from 36V 300W SMPS. The amplifier has been designed to work with 4Ω nominal loads.

Each of the four TAS5613A half-bridge outputs OUT_x has two Schottky diodes with reverse polarisation connected. One between the output and the board ground and the other between the output and the chip PVDD supply rail. The diodes used are MBRS360T3G (SMC package) from ON Semiconductor.

The two amplifier outputs drive two step-up transformers loaded with two 50Ω resistive loads. The transformer primary windings are connected to the amplifier outputs trough coupling capacitors (2 x 1000µF in series). The amplifier board, the SMPS and the two transformers with coupling capacitors are placed in a fully enclosed metal enclosure (19” rack type).

The Test

In order to estimate the enclosure internal temperatures corresponding to the chip two over-temperature alarm thresholds (100°C and 125°C) the amplifier was driven from continuous 1kHz sinusoidal source so it would produce 71W on each of the two resistive loads. Accounting for the transformers and coupling capacitors efficiencies this corresponds to around 82W (18.5VRMS, 4.44ARMS) per each of the chip two BTL outputs (164W total).

The Failure

The alarm output corresponding to the TAS5613A junction temperature of 100°C was activated at around 57°C air temperature inside the enclosure and the amplifier continued to work as expected. Just before or right after the second alarm output corresponding to the chip 125°C junction temperature was activated one or two of the protection Schottky diodes failed (they went short circuit). At that point, the temperature in the amplifier enclosure reached around 62°C.

The diodes are mounted on the board bottom side, hence they are located in the small area between the enclosure bottom and the board (around 5mm gap).

It seems the diodes failed due to their thermal inability to dissipate the required power at such temperature.

The Problem

How to estimate the power each of the eight Schottky diodes needs to dissipate assuming in a practical implementation a user could drive the amplifier continuously to its maximum power limited only by what the power supply can provide?

Situation may be more complicated if saturation of the output transformers at low frequencies is possible (i.e. no high-pass filters on the amplifier inputs).

  • 0
    Prodigy 110 points
    Update.

    After inserting a heat conductive material (Gap Pad VO Soft) between the board and the metal enclosure bottom the Schottky diodes were able to work even with higher continuous audio output power (2 x 100W) without failure.

    The second TAS5613A junction temperature threshold of 125°C was reached and the amplifier continued to work as normal. After that the temperature inside the enclosure stabilised and the chip thermal shutdown didn’t occur. Nothing abnormal was found regarding the amplifier operating at 125°C or above .

    This suggests that the currently used MRS360T3G Schottky diodes are not suitable for this application.

    Could TI suggest how to select Schottky diodes for a D class amplifier.

    Regards,

    Les
  • 0 in reply to Les_K
    TI__Guru 65390 points

    Les,

    Thanks for contacting us. Looking into the EVM, I don't see a diode model number listed, where did you get that diode spec recommendation?

    I also don't see the datasheet listing need for Diodes between OUT and PVDD, am I missing something?

    Lastly, which diodes were failing? The ones between OUT/PVDD or the ones between OUT/GND?

    Regards,

    -Adam

  • 0 in reply to Adam Sidelsky
    Prodigy 110 points
    Hi Adam,

    Just to clarify, I didn't designed the amplifier board.

    I have noticed that the EVM only has one Shottky diode per output (between the output and ground).

    My understanding is that both diodes were added after the chip failures due to over-voltages created when the output transformer saturates and the chip over-current protection is activated. Not, sure how this specific diode type was selected. I have seen some Schottky diodes suggested on this forum, but usually without any rationale behind it.

    On some boards only one diode failed (the one to ground) and on another both diodes failed shorting the PVDD supply.

    Regards,

    Les
  • 0 in reply to Les_K
    Prodigy 110 points
    Another case why the second Schottky diode is needed is when the amplifier audio transformer output is lightly loaded (as it may be the case in normal application). As the amplifier LC filter has been designed for a 4Ω load it will exhibit large gain peak at 40kHz or so (and ringing) with no load or small load on the transformer secondary.