UCC28701: WEBENCH POWER DESIGNER Result does not match with TINA-TI simulation.

Qingxuan Ma

Part Number: UCC28701
Other Parts Discussed in Thread: TINA-TI

Hello,

I am using UCC28701 to design a 20V/2A output flyback with a switch frequency around 100kHz.

I downloaded its TINA-TI model from website and used the WEBENCH Designer shown below to design the system.

The WEBENCH gives a detailed schematic design shown below and I copied all these parameters into the TINA-TI to do the simulation.

(Transformer info generated by WEBENCH is also attached.)

However, the simulation shows that the gate signal provided to device is not constant and lower than 100kHz.

I understand that the simulation may not 100% match with the calculation. But the variable frequency is what confused me.

I am not sure whether this is a TINA-TI simulation setting issue or not.

The DRV2 signal in the following figure is the gate signal sent to device.

The simulation file is also attached.

40W_Charger_New2.TSC

My question is that: Which result should I trust?

The TINA-TI simulation result or the design generated by WEBENCH POWER DESIGNER?

Thanks for your help.

over 1 year ago

0 Mike O' over 1 year ago

TI__Guru 66135 points

Hello,

I am reviewing your inquiry and will get back to you soon.

Regards,

0 Mike O' over 1 year ago in reply to Mike O'

TI__Guru 66135 points

Hello,

Does the Tina model and your Webench schematic have the same primary magnetizing inductance, turns ratio, input, output capacitances; as well as, current sense resistors?

Also not that the design/s should only be running at maximum frequency under, maximum load. As the load decreases the controller will use frequency modulation to control the duty cycle. As the load decreases the frequency decreases. So I would double check your load conditions as well.

Regards,

0 Qingxuan Ma over 1 year ago in reply to Mike O'

Prodigy 50 points

Hello Mike,

Thanks for the reply. Yes, I am using the same parameters and transformer turns ratio in my TINA-TI simulation as the one generated by Webench. Please see the screenshot below. You can also take a look at it in the.TSC file attached.

My load is a 10Ohm resistor which will not change during the simulation. Thanks.

0 Mike O' over 1 year ago in reply to Qingxuan Ma

TI__Guru 66135 points

Hello,

I have received your latest inquiry and will get back to you shortly.

Regards,

0 Mike O' over 1 year ago in reply to Mike O'

TI__Guru 66135 points

Hello,

Could you check your transformer models that you are using in Tina? They may not be ideal and may have magnetizing inductance that is causing your model to behave differently than inspected.

Regards,

0 Qingxuan Ma over 1 year ago in reply to Mike O'

Prodigy 50 points

Hello Mike,

I am using ideal transformers in my simulation and paralleled an ideal inductor with the primary side winding to act as a magnetizing inductance.

But I am not very sure that whether this is an appropriate way to simulate a 3-winding transformer. Thanks.

0 Mike O' over 1 year ago in reply to Qingxuan Ma

TI__Guru 66135 points

Hello,

I have received your inquiry and I am evaluating presently and will get back to you shortly.

Regards,

0 Mike O' over 1 year ago in reply to Mike O'

TI__Guru 66135 points

Hello,

Still working on this.

Regards,

0 Mike O' over 1 year ago in reply to Mike O'

TI__Guru 66135 points

Hello,

Could you send me your Tina model and the output from Webench?

I would like to review it.

Regards,

0 Qingxuan Ma over 1 year ago in reply to Mike O'

Prodigy 50 points

Hello Mike,

The files are attached here. Thanks.

8880.WBDesign6.pdf

1856.40W_Charger_New2.TSC

0 Mike O' over 1 year ago in reply to Qingxuan Ma

TI__Guru 66135 points

Hello,

I was unable to get this today and will look into it Monday for you.

Regards,

0 Mike O' over 1 year ago in reply to Mike O'

TI__Guru 66135 points

Hello,

I had time to look at your Webench file. According to the file the design was done for 13 W and not 40 W. I think this might be what your issue is.

I would recommend entering your design parameters in Webench and then adjust the design accordingly and rerun the simulation.

Regards,

0 Qingxuan Ma over 1 year ago in reply to Mike O'

Prodigy 50 points

Hello Mike,

Thanks for the feedback. However, I guess the 13W number comes from the Pd (power dissipation).

I think that may be the power loss instead of the power delivery.

As you can see on the top right corner, Vout=20V, Iout=2A. This means that the output power is 40W.

If you put the following parameter in WEBENCH and change the switch frequency for 100kHz. you would get the same number as it shows here.

We are proposing this UCC28701 based design to a customer and the quantity is quite good.

I understand the TINA-simulation may not 100% match with the real-world operation and the WEBENCH design is based on calculation.

Let's focus on the following questions.

1.Which result should we trust more? The TINA_TI simulation or the WEBENCH design?

2.If we use the parameter provided by WEBENCH, will we have a system operation around 100kHz with a 40W output?

If a simplified answer could not be provided here and more detailed discussion is needed, would you please provide me an e-mail address that we can reach out to? Thanks for your time and help.

0 Qingxuan Ma over 1 year ago in reply to Mike O'

Prodigy 50 points

Hello Mike,

I just noticed from the UCC28701 datasheet that the IC has frequency-jittering functions.

I guess the variable gate signal frequency in TINA-TI simulation is caused by this feature in the IC model.

If my guess is correct, this will solve all my questions.

Please feel free to correct me if I am wrong.

Thanks for your time.

0 Mike O' over 1 year ago in reply to Qingxuan Ma

TI__Guru 66135 points

Hello,

Thanks for the feedback. It does look like it gives 176 uH for Lpm. Let me review this some more to see what I can figure out.

There is an application note that you can also use to check the design. You can find it at the following link.

https://www.ti.com/lit/pdf/slua653

Regards,

0 Mike O' over 1 year ago in reply to Qingxuan Ma

TI__Guru 66135 points

Hello,

The device does have frequency jitter. It should vary roughly 8 to 10% to reduce EMI.

Regards,

0 Qingxuan Ma over 1 year ago in reply to Mike O'

Prodigy 50 points

Hello Mike,

Thanks for the clarification.

But the frequency vary range in my simulation can reach up to 30%.

I guess this is due to unproper simulation setting and will not happen in real operations.

Please feel free to correct me if I am wrong. Thanks.

0 Mike O' over 1 year ago in reply to Qingxuan Ma

TI__Guru 66135 points

Hello,

You may not have enough output capacitance. Double check your voltage ripple to make sure it is less than 2%. Also double check you turns ratios to see if they are correct.

Regards,

Power management

Power management forum

UCC28701: WEBENCH POWER DESIGNER Result does not match with TINA-TI simulation.