UCC28740: Device not starting, Gives one pulse per second

Hi

2 Questions

1) What is the voltage output on Drive Pin supposed to be? I only get 5V.

2) The FET I chose is in the list of alternatives in WeBench, Can it be a problem? Specs are very close to the original FET in WeBench

(1) DRV will follow VDD when VDD < 14V, and when VDD > 14V, DRV is clamped below 14V. So 5V does not seem correct.

(2) What is the FET Vgs threshold,  your test shows more than 10us DRV pulse, while CS does not show a typical triangular rising up voltage. It looks the FET is not on, or your converter has connection problem. When the FET Vgs > threshold, it Vds should be low, you need to probe both Vgs and Vds to see if the FET works then decide what else to check.  

Hi

Thank you for the reply.

Looks like the FET is not functioning properly. I will need to get another part number.

If I remove the FET from Circuit, The DRV pulse is much higher.

Hi

Thank you for the information.

I received my new components today and will start testing.

One question, I guess the Diode needs to face the UCC24636? (Cathode facing pin 6 of IC)?

Hi Hong

I completed some testing. It still does not start properly

In this image yellow = FET source

                     Blue = Drive pin

   Pink = FB pin on chip

  Green = Vdd on chip

In this image yellow = FET source

                     Blue = Drive pin

   Pink = CS pin on chip

  Green = Vdd on chip

R13 is currently at 600R

Is R11 at correct value?

I have a DC multi meter and a light load on the output of the circuit. The meter jumps between 0V and 0.6V DC

Can you provide your updated schematics for us to review? Did you add a diode in series connection with R3?

Please find the updated Schematic V2.1

Forgot to include the fileAC-DC Converter V2.1.pdf

The diode D4 connection is not correct in your schematics. Please correct it as below. R3 = 10-ohm seems a bit too small. You may need re-calculate its value. To be safe, you may change it to 100-ohm and test to see VDD voltage and make sure VDD not exceeding the abs max specified in the datasheet.

Do you have any document with examples on what the output pulses should look like

For the UCC28740 and the UCC24636?

You can refer this below document for UCC24636 DRV and MOSFET switching waveform. 

You can refer to the below for UCC28740 DRV and MOSFET switching waveform.

Thank you for the documents.

I had problems with the STD16N65M5 FET's I ordered.

The circuit still does not continue to work after startup. But one thing I see today is that as soon as I put my Scope probe on Pin5 (CS) on the UCC28740, I only get one clock pulse on the drive pin. As soon as I lift the probe, I get more pulses, about 26ms. Then the circuit stops and starts again.

If I measure on R11, on the FET side, the circuit continues to work

Is this behavior correct? Is there any way I can de-sensitise it?

It looks you have an issue of ac input grounding, or your oscilloscope grounding issue, or your oscilloscope probe has some issue. Can you try a different probe, oscilloscope?

If you probe MOSFET source pin and your converter continuous to work, it means, there may be high current spikes over there so scope probe filtering the noise. You can try to add a small capacitor (say, try 100pF first) between CS pin and GND pin without scope probe on R11 FET side to see if the circuit can work continuously.

As no furhter communication, we need to close this thread. You can open a new thread if further support is needed.

I ran into a problem with my oscilloscope and the FETS.

I still need to do some more tests and I may need to ask some more questions

Please keep this open for a little while longer

You can open a new thread when you have additional questions and referring to this thread so we can know the history and make help accordingly.

Per the E2E policy, we need to close this thread if no immediate support need.

Good day

Today I am using a different Scope. I managed to get the voltage on the VS pin down so it does not reset the chip.

I now have a problem with the VDD voltage that keeps on dropping until the chip resets and restarts.

In the above image I have the following signals

Green - FET Drain

Pink - DRV pin on chip

Yellow - Aux Winding signal  before R14

Blue - VDD at that stage

The next image is just showing the total time the DRV pulses are available before reset. It resets when VDD becomes to small, as per Datasheet.

Two questions:

1) Is the Aux signal supposed to be in phase with the Primary(Green)

2) If it is correct, do I need more winding's on the Aux as the max positive voltage seems to be only 8V

Please refer to next post

sorry..previous post was out of sequence.

Please refer to on above , the one with the images

Two questions:

1) Is the Aux signal supposed to be in phase with the Primary(Green)

Yes, they are in phase.

2) If it is correct, do I need more winding's on the Aux as the max positive voltage seems to be only 8V

You need increase UCC28740 VDD capacitor value and re-test.

I partially solved the problem by adding a new winding onto the transformer. I then used the inverse pulse(inverse from the Primary) to generate a voltage high enough for the UCC28740 to run.

The board started working and I had a constant 16V dc @ 0.7A on the output. I was hoping for 20VDc. At that tatage the input AC voltage was approx 190V AC.

I started to increase the input AC voltage and as I got to about 210V AC, every thing stopped.

I still have not been able to get the circuit restarted. I now only get one pulse, sometimes 3 pulses.

The questions I have is why did I need to add the extra winding? Why do other circuits work?

What could be the reason why the circuit stopped working at 210V AC input. Is there anything I need to check.

I am now waiting for new UCC28740's to replace the one currently on the board.

How did you determine AUX winding number of turns?

The current AUX winding is 5 turns. So I started by adding 5 Turns and tested the voltage output.

I then removed one winding as the Voltage on the inverted pulse was too high.

So currently I have 4 windings as AUX2 feeding the VDD pin. VDD pin was not higher than 26V at the time the circuit failed

How did you determine AUX winding number of turns? How did you make this design?

Can you answer the above questions?

If you are making a design just by trial-and-error, it will take very much time before you can get your design target. Please let us know how you have made your design including the number of turns for AUX winding.

The windings on the AUX on the transformer is from the data in the WEBENCH design.

The transformer was made for me according to those specs. I could measure the voltages on the oscilloscope. 

I then used the transformer winding formulas to check that the voltages I measure is according to the formulas. I used that to calculate the new windings I added to the transformer.

I have 2 sources of design info.

 1) WEBENCH design

2) Calculations in the excel sheet sluc487b.xlsxCoreCoilformer31.pdf

 The actual windings is as per below document

IMG_20190417_0001.pdf

As per previous posts, using the design as it was , I could never get the VDD pin high enough to keep the chip working. I used your suggestions and using a larger cap on VDD pin did not help.

That is the reason I made another winding. The AUX is still as it was feeding the VS pin. I just added another winding onto the existing transformer core.

NA/NS = 6:6 based on webench which should be good. Since your Vout = 20V, with cross -regulation, VDD should be at about 20V.

Can you increase your VDD capacitor value to see if your converter can start up into regulation with NA/NS = 5:5?

If you increase NA to 5+5, VDD can reach > 40V which will damage UCC28740.

I only received new UCC28740 yesterday afternoon and will continue testing.

Increasing the value of the VDD capacitor Did NOT solve the problem.

Will continue testing today

Good day

Webench suggests the STD16N65M5 as the first choice for the switching FET.

Is there any other FET that I can use?

Can you maybe guide me on which parameters to watch out for while choosing a replacement?

I get a very high value of ringing on the transformer when the FET switches off.

That ringing propagates through the whole circuit and pushes the voltages on VDD, VS, CS too high. 

Yellow is the Gate pulse - The puls does not look clean to me. I have swapped the UCC28740.

Blue is the FET  Drain. During this test the input was only 70V DC

Pink - Not measuring

Green - VS

Can the FET have this effect, or is there any other possibility?

I also do not understand why there is noise on the Gate pulse rising edge. As a test I moved the FET to right next to the UCC26740 to make the path to FET gat as short as possible. There was no change.

Can the transformer have an effect on the UCC28740? They are close together.

On my schematic previously uploaded D1 is supposed to be an Unidirectional Voltage suppressor diode.

I currently have a Bi-Directional Voltage suppressor fitted. Can this cause the excessive ringing?

Please provide your updated schematics so we can be synchronized to what you indicate and understand which is which, and able to provide our suggestions.

The schematics has not changed

But here it is againAC-DC Converter.pdf

What is D1 part number? Can you provide D1 datasheet so we can have a look if ok to use?

D1 part number: SMBJ110CA

D2 part number STTH108A

I am currently awaiting new stock of the suggested component and | am using STB16N65M5 as a repacement. The input capacitance is different and RDSon. The suggested part is STB11N65M5(Webench)STMicroelectronics-STD16N65M5-datasheet.pdf

STMicroelectronics-STTH108A-datasheet.pdf

Littelfuse-SMBJ110A-datasheet.pdf

This is the pulse on the Gate without it being connected to the FET

This is the Gate pulse after being connected to FET

I have 2 questions. On the gate pulse rising edge there is a small flat area. This area sometimes has a oscillation on as well. Is this the Miller region. Does this mean the UCC28740 cannot drive the FET hard enough?

The next question is Why is there such a huge amount of ringing at the point where gate pulse is off? Does this have something to do with the Input capacitance?

The diode you used for D1 and D2 should ok.

The FET you are using should be good enough and should not show the long flat area.

The turn off spikes are usually due to a big loop. So you need to check your layout to see if the loop is too big. another possibility is the transformer leakage inductance is high.

I suggest you get an EVM of UCC28740, UCC28740EVM-525, so you can compare your test waveform with the EVM. You can also put your MOSFET on to the EVM and compare.

Have you contact your TI local sales office so you can bring your board for them to have a look. It is difficult to use E2E to debug a board. So if possible, please contact your TI local office to help.

Thank you for the response.

I have contacted the local TI distributor and am awaiting feedback. I am also looking at changing the transformer.

As the Dev kit seems to have a transformer from Wurth, I send a request to them for that part number

Will keep you updated

Great support from TI - Thank you

Hi The local distributor ( Arrow Altech or Arrow Altron) says they do not have development kits available for testing.

Can you maybe send me one either directly or via Arrow?

Who else is a distributor in South Africa?

The distributor, Altron Arrow just let me know I need to puchase a kit and it will take 2-4 weeks for delivery. Is there any way to expedite delivery.

Hopefully for free. I am using TI Battery charger as well.

I will see if there is an alternative way to support you. But before I find it, it looks we may just continue.

On the EVM, how did you get TI battery charger board, can you ask the same person to help you to get a UCC28740EVM?

To continue support you through E2E, can you please provide your PCB layout for review to see if something can improve over there?

I do not have a dev board on the Battery charger, I just bought the chips (BQ24650), designed a PCB and populated it. It works. Wish that chip had I2C.

On the AC-DC board, attached are the gerber files. I have requested a sample of the Wurth transformer used on the Dev board and I asked the local transformer manufacturer to have a relook at the transformer. I send him more data from Colimaker on WEBENCH.AW-9270-1200-020-AC-DC_Conversion.zip

Can you send the Altium design files to me for review - it is difficult to check your design on gerbel files.

AC-DC_Conversion.zip

The layout shows very long and thin connections between UCC28740 DRV (pin 6) and Q1. It looks this is why the typical MOSFET cannot be driven to turn on and off like that in the other applications.

Also, the traces are through underneath the transformer and can pick up noise on the 8mil long trace loop.

Can you move Q1 to the right of the transformer to improve this design? There is enough space on the right so Q1 can be moved to there.

C1 long and wide traces to connect to Q1 should be ok.

Also the transformer pins can be re-arranged to help your layout. But here Q1 driving traces need to be short and wide in order to improve the Q1 on/off issue.

You can send your changed layout to me for my review before you order the board.