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.

UCC28950: Problem with Synchronous Rectifier

Part Number: UCC28950
Other Parts Discussed in Thread: UCC28951

Hello

We are developing DC-DC PSU with input voltage range 600-800VDC and output voltage 60VDC@10А using UCC28950. With asinchronous rectifier system works without problems, but power dissipation is too high.

With sinchronous rectifier we got a problem: when SR starting to work we gets a huge noise on all nets - on output, on control nets and even UCC28950 power input. As result, it kills controller. What can cause it?

Oscillograms with enabled SR:

  1. SR transistors gates with SR enabled before stabilization started

2, 3. Output with asynchronous rectifier without load

4, 5. Output with asynchronous rectifier loaded with 1A

6, 7. Output with SR enabled after stabilization started (load 1A)

With input voltage increasing noise level significantly increasing.
If necessary, we can send project files to the email.

Pic1

Pic2

Pic3

Pic4

Pic5

Pic6

Pic7

  • Hi Roman Zavyalov,
    Please could you send your schematic to my Email address j-leisten@ti.com.
    I am not an expert on this part, but I work with a man who is, and can ask for his help.
    Thanks
    Joe Leisten
  • Hello, Joe
    Done. Check your email.

  • Hello Roman

    I was out of the office for a week and this is the reason for the delay in getting back to you.

    I'm fairly sure that the problem is that you have swapped the SR drives so that the OUTE and OUTF signals are being delivered to the wrong SR. This will put a short circuit on the transformer and cause destructive high currents.You have been using OUTF to drive VT6 but you should be using OUTE to drive this SR. OUTE should be used to drive VT7. You should be able to check this easily by 'tombstoning' R62 and R63 and wiring the signals to the correct transistors. I marked up part of the Typical Application Diagram to show what I mean.

    Please let me know if this solves the problem.

    regards
    Colin

  • Hello Roman
    There is a typo above, I should have said 'OUTF should be used to drive VT7'

    Regards
    Colin
  • Thanks for your advise. We has swapped pins Out E and Out F, after that output voltage has dropped to zero. After that we additionally swapped primary winding's pins and PSU has started to work. Output voltage noise has significally reduced, but after SR starts to work, VD23-VD26 (1.5KE300CA) has started to heat. If we manually stops SR, they are stops to heat. Also, noise in UCC28950 power cicuit has kept and it tied with input voltage (increasing when input voltage increasing). Any ideas about it?

  • Hello Roman

    I double checked your schematic and my marked up schematic above and my markup is correct. Please change your circuit accordingly.

    Please send me your PCB layout files (Gerber or native files) and I'll review them - you can send them to me at colingillmor@ti.com

    In the meanwhile, I suggest that you go back to diode rectification and we can diagnose and correct whatever is causing this misbehaviour.

    Remove R62 and R63 - this will disable the MOSFET gates and they will behave as diodes.

    Send me a plot showing Pin7 of U3, Pin 5 of U3, Pin 7 of U1 and Pin5 of U1. Please show 5 or 6 cycles across the horizontal axis. This will allow me to be sure that we are delivering the correct signals to the switches.

    I'd also like to see a plot of the OUTA, OUTB, OUTC and OUTD signals at the UCC28950 controller. Please look for dropouts or other defects in these waveforms and send me a plot of whatever you see. The switching pattern from the UCC28950 can be upset by external noise sources, for example an OUTx line running under the IC body will cause problems.

    If you can, it would be useful to see the voltage across the primary of the transformer and at the drains of the SRs. A plot showing VT1 drain, VT3 Drain, VT6 drain and VT7 drain would do this nicely.

    Finally, please try to source some UCC28951 devices. These are a 100% compatible drop-in replacement for the UCC28950 but they work better at high duty cycles. This may or may not be the solution to this particular problem but I'd like you to be in a position to try the newer device if our early checks do not fix the problem.

    Regards

    Colin

  • Hello, Colin

    1. Email with PCB have sent.
    2. UCC28951 have ordered.
    3. Currently we are prepare picture of waveforms.
  • 1- OUTA, 2 - OUTB, 3 - OUTC, 4 - OUTD

    1- OUTA, 2 - OUTB, 3 - OUTC, 4 - OUTD

    1- OUTE, 2 - OUTF

    1- OUTE, 2 - OUTF

  • Hello Roman

    Thanks for the added information.

    Plot 1, OUTA,OUTB,OUTC,OUTD looks good, the controller is running at Dmax because Vout is below its setpoint but other than that it is fine.

    Plot2, is labelled the same as Plot 1. I suspect these may be the waveforms at the output of the drivers U3 and U1. If this is the case, then they look ok too – with the same comment about running at Dmax.

    Plot 3 and 4 are labelled OUTE and OUTF. I’m very puzzled by this for two reasons –

    First, OUTE and OUTF should be different and second they should be in antiphase

    I don’t know at which points you took the OUTE and OUTF plots but they are definitely incorrect. I’d suggest you start at the OUTE and OUTF terminals of the UCC28950 and follow the signals along to the SRs.

     

    Here is a plot from a simulation I ran just now – it’s at 100kHz and close to Dmax – so it should give you a good general ideal of what you should be seeing.

    Here is a slide showing more or less the waveforms above but at a lower duty cycle.

    /cfs-file/__key/communityserver-discussions-components-files/196/3823.Waveforms.pptx

    Please check the OUTE/OUTF signals and let me know how you get on.

    Regards

    Colin

  • Sorry, with SR oscillograms I've took second signal from wrong point. Here's correct images for SR.

  • Hello Roman

    Ok - the waveforms now make sense. OUTE and OUTF overlap during the 'freewheeling' interval which is very short when the controller is running at close to Dmax. They are in antiphase and with some minimal overlap. I'd prefer to see a full cycle on these waveforms but what you have sent is fine for now.

    I'd like to see a plot which shows the relationship between the voltage across the primary of the transformer and the SR drains. VT1 drain, VT3 drain, VT6 drain and VT7 drain - ideally 4 or 5 cycles of the weaveforms.
    I'd also like to see a plot showing 4 or 5 cycles of the waveforms at U3 Pin 2 (OUTE), U3 Pin 4 (OUTF), VT6 drain and VT7 Drain

    Do you know why the system is running at Dmax and is this expected ?
    Normally this happens if Vin is too low to generate Vout. This can be a useful diagnostic tool sometimes.

    Regards
    Colin
  • Hello Colin

    That oscillograms was made with low input voltage. We can make previous oscillograms with high input voltage (around 500V, when SR starts to work; above 500V UCC28950 is dying) if needed. Can that plots be usefull?

    Today, when we tried to check one theory, we accidently killed UCC28950, so we will make oscillograms tomorrow. Also, unfortunately, rigth now we can't make oscillograms from primary side because we don't have high voltage probes.

  • Hello Roman

    I think that the first thing I would like to confirm is whether you have the OUTx signals connected correctly. Ideally, I'd like to see a plot showing the voltages I mentioned in the last post.

    However a plot showing OUTA DA1, OUTB DA1,  PIn 7 U3 and Pin 5 U3 would do that. They are all low voltage signals and have the same ground reference. You can take this plot at any input voltage where the SRs are active

    Regards
    Colin

  • Top to down lines:

    • OutA
    • OutB
    • U3-5
    • U3-7

    Vin 0V 4 periods

    Vin 40V 4 periods

    Vin 150V 4 periods

    Vin 250V 4 periods

    Vin 350V 4 periods

    Vin 450V 4 periods

    Vin 530V 4 periods

    Vin 550V 1 period

    Vin 550V 4 periods

    Vin 550V 6 periods

  • Hello Roman
    Thanks - these waveforms are quite informative.

    1/ The OUTA, OUTB, U3-5 and U3-7 waveforms are ok. I assume that you are have corrected the polarity issue I mentioned earlier.
    2/ The output signals from U3-5 and U3-7 get increasingly noisy as Vin is increased. It looks like there may be some false outputs coming from them, especially in the final image.

    For now, I'd suggest you continue to use diode rectification until we can convince ourselves that the SR drives are correct.

    I looked over the layout again and I see that you have a wide separation between the controller and the SR driver IC. This shouldn't be a problem in itself but long lines are prone to noise pickup.

    I'd suggest that you try to get a more stable trigger - perhaps OUTA at the controller - and look at the inputs and outputs from U3 to see if the part is triggering on noise or not. Maybe the outputs are not obeying the inputs correctly in which case we should look at the decoupling around the driver. Maybe the inputs to the driver are very noisy and this is causing the outputs to misbehave. Maybe the controller OUTE and OUTF signals are noisy. The aim would be to follow the signal chain and try to find out the point at which the system is getting disturbed.

    The OUTE and OUTF lines are on the top of the PCB, it may be necessary to put them on an inner layer with ground above and below them to reduce noise pickup - but let's try to understand where the problem is coming from first.

    Regards
    Colin
  • I've marked this thread as resolved. We can continue to work via email or you can open a new thread if you wish.

    Regards
    Colin