• TI Thinks Resolved

TAS5634: Unable to achieve output power at lower impedance load

Part Number: TAS5634

My customer encountered problem below while performing output power measurement:

Setting : PVDD=58.5V, Rosc=30kohm

PBTL mode:

3 ohm load : Vout about 27Vrms, /OTW & /Fault are activated

4 ohm load : Vout about 28Vrms, /OTW & /Fault are activated

5 ohm load : Able to achieve Vout about 50Vrms, /OTW is activated at low Vout too but /Fault remain high.

Questions :

1) Why /OTW is activated at low output?

2) Why /Fault is activated for 3 & 4 ohm load?

BTL mode: (5 ohm load)

Abnormal output signal when output power exceeds 120W 

Question : What is the possible cause?

Additional info : Customer is using their current design of TAS5624A circuitry and replace the IC by TAS5634 with surrounding components' value as stated in datasheet.

Thanks.

Best regards,

Chin

  • Chin,

    Could you send me the datasheet of the customer output inductor?

    Please send me a picture of the board they are using, I want to see the thermal design including heatsink.

    OTW signals temperature or clipping, at low input levels, this can be caused by improper thermal design, output LC filter issues, or clipping due to lack of voltage or current headroom.

    Higher load of 5 ohm requires PVDD to be lowered. With 30k ROC, PVDD in 5 ohm case should be max 53V, at 27k ROC, up to 56.5V PVDD can be used with 5 ohm.

    Regards,

    -Adam
  • In reply to Adam Sidelsky:

    Hi Adam,

    Please refer to attachment for the construction and inductor's specs.

    Construction & Specs.xlsx

    "Higher load of 5 ohm requires PVDD to be lowered. With 30k ROC, PVDD in 5 ohm case should be max 53V, at 27k ROC, up to 56.5V PVDD can be used with 5 ohm."

    ==> Customer actually wants to achieve output power with PBTL 3ohm load and BTL 5ohm load as stated in datasheet at PVDD 58.5V. 

    ==> Using higher load (4 & 5 ohm) at PBTL just to verify the shutdown symptom.

    ==> Originally Roc=27k is used and it seems not much different result compared to 30k.

    Do inform if you need more info.

    Thanks.

    Best regards,

    Chin

  • In reply to Wee Keat Chin:

    Chin,

    In all cases above, please try the test using a fan or cold air, we need to check if this is caused by improper thermal design.

    Please also use a scope to look at the PWM signal before the output filter and the VOUT signal at the load. If you can show me images of the output we can debug this.

    Regards,

    -Adam
  • In reply to Adam Sidelsky:

    Hi Adam,

    We used the EVM to verify the output power and below are the results:

    1) PBTL (3 ohm ): Result is OK. Can achieve output power (>600W) at PVDD=58V, Vo=43.5Vrms

    2) BTL (5 ohm):  At 58V, there is noise observed as below (signal is not clipped) [This is similar waveform, the actual waveform I took in my hp seems missing. If you need, I need to ask customer to measure again]

    So the noise disappeared when the PVDD is reduced to 54V(signal clipped as normal, Vo=37V), however, we can't achieve the desired output power due to lower PVDD. 

    And we also encountered the DC power supply (equipment) is shutdown(overload) when we tried to increase the Vo.

    [eg. at PBTL mode, further  increase the gain to get THD 30%. Same phenomena occurred for BTL mode]

    We check the current at the DC power supply, there is only drain about 14A. The equipment current capability is 60A max.

    Please advise for this phenomena and the noise.

    After that, customer may study using their own PCB.

    Thanks.

    Best regards,

    Chin

  • In reply to Adam Sidelsky:

    Hi Adam,

    We used the EVM to verify the output power and below are the results:

    1) PBTL (3 ohm ): Result is OK. Can achieve output power (>600W) at PVDD=58V, Vo=43.5Vrms

    2) BTL (5 ohm):  At 58V, there is noise observed as below (signal is not clipped) [This is similar waveform, the actual waveform I took in my hp seems missing. If you need, I need to ask customer to measure again]

    So the noise disappeared when the PVDD is reduced to 54V(signal clipped as normal, Vo=37V), however, we can't achieve the desired output power due to lower PVDD. 

    And we also encountered the DC power supply (equipment) is shutdown(overload) when we tried to increase the Vo.

    [eg. at PBTL mode, further  increase the gain to get THD 30%. Same phenomena occurred for BTL mode]

    We check the current at the DC power supply, there is only drain about 14A. The equipment current capability is 60A max.

    Please advise for this phenomena and the noise. After that, customer may study using their own PCB.

    Below is the waveform for the PWM signal before the output filter and Vout (CH3&4 is PWM before coil) when the Damp shutdown.


    Thanks.

    Best regards,

  • In reply to Wee Keat Chin:

    Chin,

    The "noise" you see on the blue VOUT waveform is the device trying to limit output current due to Cycle-by-cycle-current-control. This is typical that we need to reduce the PVDD slightly in this configuration when using 5 ohms.

    Please tell me the required output power in BTL mode with 5 ohm load. Also the max allowed THD at this output power level.

    Regards,

    -Adam
  • In reply to Adam Sidelsky:

    Hi Adam,

    Customer still yet to decide the output power target, however they may study to drive the IC at typical PVdd =>about 58.5V, THD=30%, Load=3 ohm(PBTL) & 5ohm(BTL).

    Customer is checking the output power performance based on datasheet specification which is guaranteed by TI.

    Refer to the 5ohm data provided by TI, it mentions that  'Typical 58V PVDD is not supported at 5Ohm BTL'.

    If customer wants to use 58V@5ohm load [due to sharing of power supply to PBTL & BTL IC], could you advise what is the maximum output power can be achieved at 30% THD?

    Can Roc be reduced to 25k in order to achieve higher current?

    Also do you have any idea about the shutdown of the power supply during increment of Vo?

    Thanks.

    Best regards,

    Chin

  • In reply to Wee Keat Chin:

    Chin,

    If the customer wants to use 3ohm in PBTL or 5ohm in BTL, they CANNOT use typical PVDD, it must be reduced as I have mentioned. The issue is that the class-D amplifier must use the full PVDD and load of 3 or 5 ohms with 58.5V causes too much current, this trips the over-current protection. The only solution is to increase the speaker load or decrease the PVDD.

    For the OC_ADJ resistor, please see table 3 in the datasheet.

    At any power level where clipping is present (for example 30% THD) the output will show the behavior you mentioned above where it is trying to reduce the output current.

    The amp does not have Over-voltage protection, the max recommended PVDD should not be exceeded.

    Regards,

    -Adam
  • In reply to Adam Sidelsky:

    Hi Adam,

    I am little bit confused with your explanation compared to the datasheet information.

    For PBTL: Datasheet (pg 9) stated that  RL = 3 Ω, 10% THD+N, PVDD=58.5V can achieve 600W.

    Question: Is this setting feasible for actual design or PVDD=58.5V is not recommended (must be lower)?

    For BTL : Understood that lower PVDD is needed if 5 ohm load used. So based on the 5 ohm Data provided, with PVDD=56V & Roc=27k , the device can achieve 315W 10% & peak power 440W 30%.

    Question: i) Is this a feasible design?  Or PVDD= 56V is not recommended (must be lower) ? 

                    ii) Customer target is 350W at 10%THD. Can Roc be reduced (eg. 25k ohm used) in order to achieve higher output power? 

    Hope to get clarification from you. 

    Thanks.

    Best regards,

    Chin

  • In reply to Wee Keat Chin:

    Chin,

    My apologies, I made an error.

    For PBTL: Yes the datasheet is correct for 3 ohm, the only change is that 30k ROC might need to be reduced to 27k or 22k depending on the board and heatsink design. PVDD of 58.5V is correct.

    For BTL: Yes up to 56.5V with 27k ROC is acceptable. This should give you up to 315W at 10% or 440W at 30% as you mentioned. Lowering to 26k or 25k might give more output power. At 30% THD there will be a lot of cycle-by-cycle behavior and the device will heat up very quickly, please ensure a robust heatsink.

    Regards,

    -Adam