UCC28950: Problem with starting

Andrej Dobrovic

Prodigy 140 points

Part Number: UCC28950

Tool/software:

Hello. I created prototype by this:

8372.webench_ucc28950_design_forum.pdf

It works, but with load above 1A there has been heating of diodes D17 and D18.

I set som delays, no effect. Last I try to decrease frequency from 150kHz to 100kHz with Rt resistor changing. From this time nothink works. No outputs, only Vref is 5V and EA+ is 2.5V. Put Rt resistor to before value, no effect. Checked all connections, everything. Disconnect trafo, uses only Vbias, nothink. I change device UCC28950 for new one, the same. Some days ago, when I start it first time, with Vbias above 8V, there was pulses from OUTA ... OUTD. Now nothink. Checked DCM, RSUM, CS, still no outpus. What can cause this? What minimum must be connected for OUTA..OUTD get pulses? I'm frustrated, whole day without real effect.

23 days ago

0 Mike O' 22 days ago

TI__Guru* 86206 points

Hello,

D17 and D18 are output rectifier clamp diodes. If they are hearing up it is most likely because R23 is too small. I would increase the size of this resistor to reduce the heating of the diodes. You should not have to change the switching frequency or other controller timing components.

There is another way to protect the SR FETs using a more traditional RCD clamp. The following link will bring you to an e2e thread that shows how to set this RCD clamp. D4 and D6 in this thread can be replaced with D17 and D18. Also, R23 would be disconnected and then you can follow the instructions in the thread on how to setup the RCD clamp for this design. This technique will dissipate less power compared to what is recommended in the Webex design.

https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/1132532/ucc28950-ucc28950-senkron-rectf-snubber-question/4202839?tisearch=e2e-sitesearch&keymatch=UCC28950%2520Clamp#4202839

Regards,

0 Andrej Dobrovic 22 days ago in reply to Mike O'

Prodigy 140 points

And second question, which components must be minimally connected to UCC28950 for check that controller is not broken? I have prototype, where is no output to OUTA .. OUTF, and after controller replace, no output too. I had only two pieces of controller, now I wait for next pieces. I want to prevent break new one after replacing.

0 Mike O' 22 days ago in reply to Andrej Dobrovic

TI__Guru* 86206 points

Hello,

If your DCM comparator threshold is lower than the peak current seen at the CS pin Out E and Out F will not deliver any signals. Check the data sheet on how to setup.

As a basic check. You can power up VDD. You should be able to check to see if VREF and SS come up. You also can check the SYNC pin to make sure it is putting out pulses. Also, Out A through D should come up as well. If those things do not happen more than likely the IC is damaged.

Regards,

0 Andrej Dobrovic 9 days ago in reply to Mike O'

Prodigy 140 points

Hello. I change wiring by this recomended link https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/1132532/ucc28950-ucc28950-senkron-rectf-snubber-question/4202839?tisearch=e2e-sitesearch&keymatch=UCC28950%2520Clamp#4202839 . i change diode s5m for ultrafast mur460. Diodes no heats now, But still heating R8 (in link scheme). I increase to 10k from 1k, but heats over 100 °C in minute after stars. Output is 80V, load 0.5A

Here are voltages in anodes of diodes (yellow and blue) and caathodes (red). it is 1:10

0 Mike O' 8 days ago in reply to Andrej Dobrovic

TI__Guru* 86206 points

Hello,

Your cathode looks like it has to 10 to 20 V across it. If you used a 10 kohm resistor and there was always 20 V across it the resistor would dissipate 40 mW. I would not think this resistor should be running hot.

You might want to check differentially across the 10 k ohm resistor to see what the voltage is across it to double check to make sure there is not to much voltage across it or current running through it. The power dissipation of the resistor should be (Vrms)^2/R. To keep heating to a minimum I would size it for 2 to 3 times the rated power dissipation to reduce heating.

Regards,

Power management

Power management forum

UCC28950: Problem with starting