• TI Thinks Resolved

PMP22557: Change the design to 5V output

Prodigy 20 points

Replies: 3

Views: 32

Part Number: PMP22557

My design input voltage is range from 70V~120V, and the output expected to be 5V, load is about 50mA, applied for automotive

1.   Is it  ok to bear from 70~120V input for PMP22557

2. Is it ok to change the PMP22557 to 5V output ? how to change

3. I see UCC28730 is for a flyback controller , but it looks like the PMP22557 is not designed as a flyback, is that ok?

  • Hi Rick,

    I can give you suggested component changes to modify PMP22557 for your application.

    My design input voltage is range from 70V~120V, and the output expected to be 5V, load is about 50mA, applied for automotive

    1.   Is it  ok to bear from 70~120V input for PMP22557


    Yes, this design "floats" the UCC28730 controller on the switch node and can easily tolerate 120Vdc in this application.

    2. Is it ok to change the PMP22557 to 5V output ? how to change


    The controller must operate in DCM only. To change the design for your specs, I recommend the following changes to PMP22557:

    1) L1 > 120uH, so I recommend using 150uH. The inductor value must increase because your output power is much less. The expected Ipeak = 0.3A and Irms = 0.1A, so chose an appropriate rated inductor.

    2) R6 < 3.2 ohms to prevent hitting OC, so I recommend using 2.5 ohms max. Irms ~ 26mA, so an 0603 or 0805 is adequate.

    3) Q1 is ok and will work but is overkill for this low output power. Conduction and switching losses should be quite low. If you change FETs, a FET with 1 ohm Rdson should not be a problem. Select a FET with very low gate charge (10nC or less) and rated for >Vin max.

    4) C4 should be >50uF and have an ESR<0.1 ohm. A 100uF alum (low ESR) should be fine.

    5) C1 = 10uF min is recommended. I recommend keeping the 25V rating to account for cap reduction due to DC bias.

    6) R1 = 200k (this also sets the start voltage)

    7) R2 = 845k

    8) R8 + R9 ~ 30K. You can use one resistor here, not two Their total value also sets the converter's turn on/off levels.

    9) R7 = 7.50k. This is a preload to prevent Vout from increasing when under no load.

    10) R5 ~ 4.2k.

    3. I see UCC28730 is for a flyback controller , but it looks like the PMP22557 is not designed as a flyback, is that ok?


    Since the controller's GND pin is on the switch node, all control signals are referenced to either Vin or GND (actually -0.7V). The HV pin can easily tolerate your input voltage for startup. But once running, the internal HV FET turns off and bias power is taken from the output voltage through D1 + D2 to charge C1 (when Q1 is off, D3 conducts). Regulation is maintained (sensed) through D2/R1/R2. D2 forward drop will mostly cancel with the D3 forward drop. VS senses the output and controls the PWM. The VS pin is very sensitive and should NOT be probed because it will effect regulation. It is recommended that a PCB layout similar to the one used for PMP22557 be used. 

    Best regards,

    John Betten

  • In reply to John Betten:

    Hello John,

    Many thanks for the detail guidelines and explanation.

    I would like to confirm one more question , since the case is battery supply so the standby power loss would be very sensitive,so what is the standby power loss? Also is it possible to provide SPICE model to do the simulation for the new change ?

    Thank you very much

    Rick

  • In reply to rick Yu:

    Hi Rick,

    Based on the PMP22557 design, no-load input power and input current was measured as the following (Vout = 12.2V):

    Pin =23mW, Iin = 381uA, (Vin = 60Vdc)

    Pin =29mW, Iin = 319uA, (Vin = 90Vdc)

    Pin =37mW, Iin = 304uA, (Vin = 120Vdc)

    Keep in mind that the converter is operating and the output is regulated to 12.2V. There is a 10K 'preload" resistor on the output which dissipates ~15mW alone, so the actual converter power requirement is about 1/2 of the above. However, the preload limits the output voltage rise with no-load and the actual value should be determined during testing.

    I do not have a PSPICE model for this circuit, but you can download "PSPICE for TI" from the link below for free. There is a transient model of the UCC28730 controller in the library.

    regards,

    John Betten