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.

TPS650864: TPS6508640 efficiency improvement

Part Number: TPS650864
Other Parts Discussed in Thread: TPS650860EVM-116, PMP12004-HE, TIDA-01393

Hello,

My customer changed the U1 of the TPS650860EVM-116 to TPS6508640 and measured the efficiency of the TPS6508640 DC/DCs.

1. For the BUCK contoller 1/2/6, the efficiency is lower than PMP12004-HE test result for all available Iouts.

Vin=12V / Vout=3.3V / Inductor: blue=470nH, read=2.2uH

Vin=12V / Vout=0.85V / Inductor: blue=470nH, read=2.2uH

Vin=12V / Vout=1.2V / Inductor: blue=470nH, read=2.2uH

Please tell us what is the casue and how to improve it.

2. For the BUCK converter 3/4/5, the efficiency is lower than PMP12004-HE test result for low Iouts.

Vin=3.3V / Vout=1.4V / Inductor=470nH

Vin=3.3V / Vout=1.1V / Inductor=470nH

Vin=3.3V / Vout=1.8V / Inductor=470nH

Please tell us what is the casue and how to improve it.

Thank you.

JH

  • Hi JH,

    There are three conditions that could be impacting the difference between PMP12004-HE and the TPS650860EVM-116.

    1. PMP12004-HE board was designed and specialized for high efficiency and the component selection is quite different from TPS650860EVM-116 for that reason. It has larger inductors and larger FETs to help achieve this. If they wanted to match the efficiency, they would need to increase FET and inductor sizes as a starting point.
    2. PMP12004-HE was tested before TPS6508640 OTP spin was released and the BUCKx_MODE bits default changed from 0 to 1 during that time period. I would expect light load efficiency to be lower as a result of the bucks being in Force PWM instead of PFM mode at light load. BUCKx_MODE = 1 (Force PWM) was preferable to most customers as it has a more stable frequency than PFM at light loads as well as smaller ripple voltage - making it easier to minimize noise impact.
    3. Method of measuring efficiency - calculating efficiency for the individual rails can be a bit complicated. For example, to avoid counting the PMIC digital core quiescent current is counted multiple times, it is important to subtract that out for each individual measurement. That includes some current on VSYS and VANA, but shouldn't include any current on PVIN I don't think. Depending on how they measured (for example if they just monitored the VSYS current), there may be some mismatch here as well.

  • Hi Kevin,

    Thank you for your kind response.

    The customer has futher questions.

    1. What are the factors to consider for large inductors and larger FETs? Should the user shoose an inductor with a large inductance and a FET with a large current? Is there anything else?

    1-1 There is 'Sized-Down BOM Selection' sector in PMP12004-HE user guide.

            Could you please provide us with efficiency data shen applying samll sized parts under 5A load?

    2. Is there a reference value for the light load that the efficiency of Force PWM lower than PFM mode?

    Thank you.

    JH

  • Hi JH,

    Efficiency with a 3.2x2.5 inductor is available in the datasheet and TIDA-01393 has some efficiency data as well for smaller solution size example. TIDA-01393 also includes data with Force PWM by default.