• State
  • Locked Locked
  • Replies 16 replies
  • Subscribers 63 subscribers
  • Views 4575 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.

UCC2893 Active Clamp Forward Efficiency Problem (<80%)

guilherme.muller
Intellectual 270 points
Other Parts Discussed in Thread: UCC2893, UCC2897A, TPS2410

Hi,

 

My name is Guilherme Muller and I’m developing a step-down SMPS using UCC2893 from Texas. My benchmark is ~90% of efficiency, so I’m using an active clamp forward converter with synchronous rectification, similar to UCC2897A EVM.

This is my first SMPS, so I’m having a lot of troubles and doubts within this project.

Our SMPS is a 60W 5V output, with nominal 125Vdc input. Our temperature rating is 85ºC, so the dissipated thermal energy should be minimum (I’m considering 10W max).

During the validation of the first revision, I got caught in a series of problems within the control loop and other configurations. Adjusting them later, my maximum efficiency today is 82%@100V input, and 78% @ 125V.

I’m working on this product for at least a month trying to raise this efficiency, with no success. I’ve followed to the letter the slua535 (Understanding and Designing an Active Clamp Current Mode Controlled Converter Using the UCC2897A) and found no clue about what is happening.

I´ve already tried a few of solutions, but none helped, cited below.

 

I’m looking for something I have missed in this project, to find out why our efficiency is so low. I’m trusting on TI Engineers to help me find a solution.

 


 

Schematic

 

The schematic is attached on 01400641.pdf.

1258.01400641.pdf

The BOM is attached too, although, some components have changed.

 

Transformer

 

I’ve analyzed the power transformer in detail, looking for a project error. This transformer is custom made, since our isolation requirements are 3750VAC. The real transformer model is below:

 

 

 

V1 and R3 are test parameters for simulation.

My first assumption was that the secondary series resistance was causing this low efficiency, since the I^2*R power dissipation would be 2,88W at maximum output power (12A). A new transformer was made, with lower series resistance (5mOhms) and no change was noticed.

 Reading the slup081, I started to consider the leakage inductance (9.14 uH) as the main problem. So, I’ve changed the transformer just in case and started using the PA0810NL from the EVM. The result is worse (76% eff max and a lot of instability) for now, since I have not changed all the parameters.

 

 

Switching Mosfets

 

Checking the waveforms from the switching FETs, the shape of the gate of the four main FETs seems a little rounded, which could be from a high input current. Although I’ve tried to reduce this value, this was not possible.

 The primary FETs are SiHF840LCS-GE3 and the Synchronous Rectifiers are IRFH5302. Their output capacitance, which could result in switching losses, are minimum (170pF for the primary FETs and 860pF for the SR FETs)

 

 

Synchronous Rectifiers and Test Points used

 

 

 

Yellow - PT1, Blue - PT6, 5% Load

 

 

 

Yellow - PT1, Blue - PT6, 100% Load

 

 

Yellow - PT1, Blue - PT6, 5% Load

 

 

Yellow - PT1, Blue - PT6, 100% Load

 

   

Yellow - Qf (TZ8) VDS, Blue -  PT1, 5% Load

 

 

Yellow - Qf (TZ8) VDS, Blue -  PT1, 100% Load

 

 

Yellow - QR (TZ9) VDS, Blue -  PT6, 5% Load

 

 

Yellow - QR (TZ9) VDS, Blue -  PT6, 100% Load

 

Yellow – PT1

Blue – PT6

Purple – Main FET Gate

Green – Aux FET Gate (clamp)

5% Load

 

Yellow – PT1

Blue – PT6

Purple – Main FET Gate

Green – Aux FET Gate (clamp)

100% Load

 

 


I've already changed and recalculated the whole components lots of times. Also, I experimentally changed them to see the effects. I’m just hoping a tip or a north to my           problems.

Thank you very much in advance,

 Kindly Regards

  • 0
    Intellectual 270 points

    Also, why is the Current Slope resistor changing the effciency at light load so much? If I use a lower than the calculated value, the efficiency goes from 7% to 18%, although the heavy load efficiency remains the same...

  • 0 in reply to guilherme.muller
    TI__Mastermind 25130 points

    Hi,

    I would expect your switching waveforms to be much sharper. Look for capacitive loading on the transformer secondary (TZ8 and TZ9 ???), secondary leakage inductance??

    The waveforms do not look right.

    John

  • 0 in reply to John Bottrill
    Intellectual 270 points

    Hi John,

    Thank you very much for your reply. I would expect that too, but I could not find a way to get a sharper wave.

    The secondary output capacitance is two low ESR 1800uF capacitors, since this SMPS need to handle a high inrush current. 

    TZ8 and TZ9 are TVS needed to suppress high voltage surges (our requirement is 4kV differential mode surge). I removed them in my tests to ensure that they are not acting and reducing the efficiency.

    The waves that refer to TZ8 and TZ9 are measurements made on the secondary, since they are parallel.

    The first 5 waves are the SR gates voltage with a 22R charging the gate capacitor. I've reduced the resistor up to 2R2 and still got no positive result.

    The forward rectifier seems weird to me. Maybe the SR mosfet capacitance (~1nF) is affecting the waveshape.

    One more time, thank you very much for your reply!

  • 0 in reply to guilherme.muller
    TI__Mastermind 25130 points

    Hi Guilherme,

    I would try replacing the output sync rectifiers with either Shottky diodes or with High Speed silicon rectifiers (Shottky  diodes have high capacitance) and see how that works in correcting the waveforms.

    You may want to refer to the paper that the below link will lead you to. Self driven sync rectifiers have an issue when the unit is turned off. particularly with high capacitance on the output.

    http://www.bodospower.com/pe/restricted/downloads/bp_2007_05.pdf

    Regards,

    John

     

  • 0 in reply to John Bottrill
    Intellectual 270 points

    Dear John,

    Thank you again for your prompt reply!

    You mean the article The Effects of Turning off a Converter with Self-Driven Synchronous Rectifiers at page 34? We have a FET driven load share with a load share controller (TPS2410PW) right next to the output capacitors. Do you think that the SR is leading to a steady-state resonant effect?

    Anyway, I'll change the FETs with High Speed silicon rectifiers, since I've already tried with shottky diodes once and got no positive result.

    Kindly Regards

    Guilherme

  • 0 in reply to guilherme.muller
    TI__Mastermind 25130 points

    Guilherme,

    My initial concern with the output sync rectifiers was that on depowering the primary side the sync rectifiers will go into self generated oscillations leading to failures and damage to the FETs.

    Are you testing this with two converters in parallel? In my experience Load Share doesn't normally work well with synchronous rectifier outputs. You can end up getting back feed to the primary side in that configuration.

    John

  • 0 in reply to John Bottrill
    Intellectual 270 points

    John,

    Again, thank you for your reply...

    Yes, we are using a Load Share configuration with the TPS2410 to control. Since this controller measures the VDS of the controlling FET, I was not worried about getting back feeded. I've tested this configuration and until now I got no problems... Do you think I should get worried?

    Anyway, I tested the ultra fast diodes and found an interesting thing: The minimum load efficiency increased, but the high load decreased. Although, the minimum load waveforms went nuts.

    In the figures below,

    Yellow: Reverse Rectifier

    Blue: Forward Rectifier

    Green: Main FET gate

    Purple: Clamp

    The first one is at full load, and the second one is at light load.

    Beside the fact that it could bring me troubles on Load Share configuration, my guess is that the problem is not the sync rectifiers.

    Any other suggestion? Unfortunately, I'm out of them :(

    Regards,

    Guilherme

  • 0 in reply to guilherme.muller
    TI__Mastermind 25130 points

    Hi Guilherme,

    The upper waveform looks more like resonant reset than active clamp. Is the clamp FET working correctly?

    John

     

  • 0 in reply to John Bottrill
    Intellectual 270 points

    John,

    Thanks for your reply!

    My mistake... I was testing with a snubber before, and forgot to change. Here are the updated images, same points of measure:

    I agree with you, the clamp is a little bit weird. But it does not act this way when I use the SR (on the two last images from the first post, green line). My gate drive circuit is this:

    Regards,

    Guilherme

  • 0 in reply to guilherme.muller
    TI__Mastermind 25130 points

    Guilherme,

    It looks to me that the transformer is resonating with the total capacitance hanging on the windings at a frequency near the switching frequency. You need those transitions to be much faster.

    John

  • 0 in reply to John Bottrill
    Intellectual 270 points

    John,

    Thanks again!

    I've already changed the transformer three times (including using the Pulse PA0810) and saw no difference... Unfortunately, no changes (using the Pulse transformer, the efficiency went down).

    My guess is that the clamp FET is not saturating, or the active clamp is not working correctly. I get the same efficiency results using a dissipative snubber. I'm checking the gate drive circuit right now.

    Any other suggestion?

  • 0 in reply to guilherme.muller
    TI__Mastermind 25130 points

    Guilherme,

    No other ideas at the moment.

    John

  • 0 in reply to John Bottrill
    Intellectual 270 points

    Hi John,

    Sorry for the late reply and thank you very much for all the help you are providing me.

    Here are the new waves I've got. You see, they are much sharper now than later. But my efficiency went down to 74%. I have no other idea of what could help me get a higher efficiency.

    The first image is with 5% load, and the second at full load. The waves are:

    Yellow - Forward Rectifier Gate

    Blue - Reverse Rectifier Gate

    Purple - Active Clamp FET Gate

    Green - Output FET Gate

    Kindly Regards

  • 0 in reply to guilherme.muller
    TI__Mastermind 25130 points

    Hi Guilherme,

    When the main FET turns off under light load it takes a long time for the voltage on the forward rectifier to turn off. This should be fast. This possibly means that the primary side FET is not turning off properly.

    What is the voltage across the clamp capacitor druing the cycle? How much does it vary?

     

  • 0 in reply to John Bottrill
    Intellectual 270 points

    Hi John,

    Sorry for the late reply! 

    The capacitor is a 10nF 250V.

    Below, 

    1) Voltage

    2) Frequency

    3) Detail in AC Coupling

    4) Full Load

  • 0 in reply to guilherme.muller
    TI__Mastermind 25130 points

    Guilherme,

    Please attach a full schematic preferrably in PDF form. Mark where each of the scope shots above were taken on that PDF either by identifying a junction or a part number and terminal.