UCC28019: distorted current at the startup time for few seconds

Hi Vasudev,

Thanks for the reaching out.

We cannot access the video with the link to the google drive. It will be better if you can attach some waveforms (or another method to share video) describing the problem along with the necessary schematics/ input & output specifications.

Regards,

Harish

pfcdesignvals.xlsx

I have shared the design calculator spreadsheet . Input voltage is 230 volts,500watts and output voltage is 380vdc . Inductor(1.3mh) is wound on a gapped ferrite core and has been designed using area product method. the distortion in startup current can be better observed in a video file. I have tried to share mp4 video file below. please let me know if more details or required. also please let me if you cannot see the video even after downloading, I will try capturing waveforms and share it to you. 

Hi Vasudev,

I tried viewing the mp4 file online and also downloaded it, but was not able to view the video due to some encoding issues. Please share the waveforms.

Thank you

Harish

1) initial current waveform-1   2)current after 1sec or less  3) current after 2nd image     4) current shaping to form sinewave after few seconds  5) and next  finally better current after 2 to 3 sec approximately

Hope you can see the images attached. here you can see the progression of current from t=0  till t = 2 or 3 sec.  

Hi Vasudev, please share a pdf of your schematic so we can better assist you.

Thank you,

Ray

Hi Vasudev,

Do you have a bypass ddiode to handle the inrush current? The startup behaviour could be due to this. I think the FETs are turning on due to this and the inductor hitting the peak current (saturation) limit during this process. Please confirm if you already have this diode or let us know your observation after making this change.

Thank you

Harish

Hi harish 

It dose have an bypass diode. I have followed the standard recommended circuit in design calculator of this IC. I have also used a autotransformer to slowly increase the voltage which had same behavior at startup time once the voltage hit the set threshold point (1.6v @ 4th pin).

Thank you

vasudev        

I have followed the same schematic recommended in design calculator. have shared the excel file below where you can find the schematic and components values. let me know if you want any other details. 

thank you 

8880.pfcdesignvals.xlsx

Hi harish 

Do you think I have to remove the bypass diode and try to see the waveforms at start-up? 

Hi Vasudev,

I see that the input range is between 200/250 with 230 nominal. Is this phenomena at all voltages or at low line voltage? Bypass diode is recommended and I don't think we should remove it.

Regards,

Harish

Hi Harish

This phenomena is observed in nominal voltage range too. I have also tested the inductor saturation current which is 5.5 amps where it begins to saturate. Does loop compensation has something to do with it? 

Regards,

Vasudev 

Hi Vasudev,

Can we try increasing the value of ICOMP capacitor. Current calue is 800pf. Can we increase it to 1400pf or so to see if it slows down the current loop and provides better robustness to this startup behaviour?

Regards,

Harish

sure will try and let you know 

Hi Harish 

I did try adding another 800pf cap in parallel but the results are same as before. There is no noticeable improvement. also I have to let you know that I am using through hole current sense resistor of 0.1 ohm 2w which has less than 1uh inductance. please let me know if you need more details. 

Thank you  

Hi Vasudev,

Thanks for the feedback. Can you please send the schematic as I happended to see Rvins1 not using the recommended value from excel sheet calculator.

Regards,

Harish

Hi Harish 

I have attached the circuit PDF below. Please let me know if anything else is required. I got the same results at startup after replacing it with new IC. 

 Schematic_pfc circuit_2022-07-15.pdf

Thanks regards

Hi Vasudev,

Thanks for sending the schematic. I will go over it and get back to you early next week.

Regards,

Harish

Hi Ray, I have attached the circuit schematic the design calculator file below. For your better understanding of the issue I have attached a YouTube video link too. 

Thanks you 

7713.Schematic_pfc circuit_2022-07-15.pdf                                  8880.pfcdesignvals (1).xlsx                                                                    

Hi Harish, I have shared the YouTube link of the video at start up for you to better analyze the issue. Please try to watch the video in 0.5 at the time of startup this might give better understanding of the issue. Hope you can access the link below. 

Regards 

Hi Vasudev,

I have a couple of observations from your schematic.

1, The voltage loop compensation elements seem to have an issue. The closed loop bandwidth is designed for 10Hz crossover, but it crosses at 3Hz (as seen from your bode plot). You might need to tune your compensator properly. Try the following values to start with. Cvcomp-5uF, Rvcomp-55k, Cvcomp_p - 0.15uF

2. Then try running with Cicomp -800pf and 1400pf or higher to see if there is any difference.

Other optional changes

3. Consider adding a schottky diode at the gate as shown in the datasheet schematic.

4. Choose Rvins1 and Rvins2 close to the recommended values from the excel calculator to ensure AC turn on/off thresholds are accurate.

Thank you

Harish

Hi Harish

1) I tried it with the values which you suggested. There is an noticeable improvement in startup where current quickly (a bit faster than before) settles to sinewave. I have also observed better transient response than before. But the start up is not as expected. Also I tried the 2nd point in your suggestion by increasing the Cicomp to 1400pf and 2200pf I wasn't able to notice any significant changes. But overall I see a improvement.   

2) I wonder why the design calculator gave wrong results at first place ? 

Thank you 

Hi Vasudev,

Good to know it improves the startup situation.

I think the problem with the design calculator in your case is that, error amplifier needs to add +5db gain at 10hz but it was rather attenuating gain by that value. For a quick fix you can reverse the sign for Gvldb(fv) at line 263 equation and this should address the issue while we will take action to update the excel sheet online.

Thank you

Regards,

Harish

Hi Harish 

The spread sheet will ask for the password to unprotect. how to edit it ?

Thank you  

8880.pfcdesignvals_ul.xlsx

See attached.

Hi Harish,

Here is the start up current capture. It can be observed after few surge current cycles there is clean sinewave for around 200ms. After which the current gets distorted heavily for some time. The next segment of capture could be see in 2nd image where the distortion starts to reduce gradually. Surely your previous suggestion has improved the start up and transient response.   

1) One thing which bothers me is why is the current getting distorted after the clean sine wave, How to get rid of this issue.   

Thank  

In the below figures we can see the gradual settling to sine wave.  

Hi Vasudev,

Can you please probe the inductor current waveform before the distortion and after and share some waveforms?

Thank you

Regards,

Harish

Hi Harish 

Is it enough if I probe the voltage across current sense resistor ? Since I don't have current probe with better bandwidth. Or if you want to see current measurement itself I will take some time and arrange for it.  

Thank you 

Regards  

Hi Vasudev,

You can share them but probing on the inductor will be necessary to  ensure that it is not getting close to saturation during operation.

Thank you

Regards,

Harish

HI Harish,

I think I am quite sure that the inductor isn't saturating in steady state even at full load ie 500w. Also I have tested the inductor in single pulse test setup where I gave a 15us pulse at 330vdc and observed the inductor, it wasn't spiking and the inductor will saturate at 5 amps but in my case the peak is 4 amps max. But it is a good idea to measure as you suggested. I will share the current asap. 

Thank you 

Regards,

vasudev 

Hi Harish,

I have observed one more thing , if I am using 0.1ohm (which is recommended by design calculator ) sense resistor the current tends to flatten and I can hear audible noise at full load ie 500w. with 0.066 ohm the current is fine without any distortion or noise in steady state. below I have attached few captures for your  reference. I don't understand this phenomenon. Can you please let me know why is it happening ? 

Thanks                                                           

0.08 ohm Rsense ,little audible noise could be observed. 

/resized-image/__size/600x240/__key/communityserver-discussions-components-files/196/20220719_5F00_230404.jpg   

 /resized-image/__size/400x400/__key/communityserver-discussions-components-files/196/20220719_5F00_230413.jpg

Rsense 0.1 ohm heavy noise

 /resized-image/__size/411x189/__key/communityserver-discussions-components-files/196/20220719_5F00_232401.jpg

https://e2e.ti.com/resized-image/__size/600x600/__key/communityserver-discussions-components-files/196/20220719_5F00_232410.

with 0.066 ohm  we see better sinwave and no audible noise at full load.     

/resized-image/__size/600x600/__key/communityserver-discussions-components-files/196/20220719_5F00_234558.jpg

/resized-image/__size/600x600/__key/communityserver-discussions-components-files/196/20220719_5F00_234606.jpg

Hi Vasudev,

The design calculator was specifically written to calculate the appropriate inductance and sense resistor for use with the UCC28019A.  The calculator takes the User inputs and calculates the Rsense based upon the internal parameters (Vsoc) of the device and the desired inductor peak to peak ripple current with a sufficient derating of 1.25. So, I do not doubt the sense resistor value of 0.1ohms recommended by the tool. If feel there are overshoots in your current waveform (which is above the calculated peak value by excel spreadsheet) and probably that is why 0.1ohm resistor seems more for your design and a lower value suits your requirement. As I suggested earlier please probe the inductor current. 

There could be a second possibility which could be mostly true in your case. ICOMP pin is most likely getting saturated at the peak of the input line voltage. The internal amplifier has a linear range only upto 5.6V (I think) and after which it is non linear upto 7V. So duty cycle might be irregular as compared to linear range which will result in distortion. The only way to prevent this saturation is by reducing the current sense resistor which lower the value of ICOMP.

Hope this helps.

Thank you

Harish

Hi Harish,

I have capture the current waveform of inductor. Also I observed one peculiar thing, If I probe voltage across Vsense pin and observe the startup current the settling time will reduce drastically. My compensation values are as suggested by calculator yet I am not getting to know where exactly is the issue.    

/resized-image/__size/600x600/__key/communityserver-discussions-components-files/196/20220722_5F00_151139.jpg

/resized-image/__size/600x600/__key/communityserver-discussions-components-files/196/20220722_5F00_151208.jpg

/resized-image/__size/600x600/__key/communityserver-discussions-components-files/196/20220722_5F00_151128.jpg

Thank you. 

Hi Vasudev,

In our previous conversations you had better input current waveform and the problem was at the center of AC line voltage for which you had described to reduce the sense resistor. Did you check if your ICOMP is saturating here as I had suggested earlier?. If that is the case you must reduce sense resistor. I also remember you telling about inductor saturation current of 5.5A (Rms or peak??). But your current seems to cross this and you might need to take care of this also.I feel compensation values are fine from your observation earlier on improved sttartup.

Thank you

Regards,

Harish

Hi Harish,

I reduced the current sense resistor and now at steady state the current peak wont get flattened .As you told  icomp was getting saturated. reducing the current sense resistor solved the issue. About inductor saturation , I tested with bigger inductor which will saturate at 11amps peak yet the start up current after the soft start Is getting distorted and it is not as (expected which should be sinusoidal). This is the only issue as of now. 

Thank you 

Hi Vasudev,

You will have to optimize the compensation network Cicomp, Cvcomp, Rvcomp etc. as per the new shunt resistor that you have chosen.

Thank you

Regards,

Harish

Hi Vasudev,

Thanks for the observation from the evm board. As you told, you can use UCC28180 as well. It comes with the calculator which takes care of the saturation issue of ICOMP. 

Regards,

Harish

Hi Harish,

Sure ,And once again thank you for your great support really appreciate it.

Thank you 

Thank you