LM51561-Q1: How to design for SEPIC & CUK ？

Hello Jung Wei Tsai,

Thanks for reachig out to us via e2e.
Your inquiry is reaching me on the weekend. I will have a look at your questions at the beginning of next week.

Best regards,
Harry

Hello Jung Wei Tsai,

I am sorry that I kept you waiting.

I had a look at different implementations how we can generate negative voltages with our parts.

Our dedicated CUC device, the LM2611 cannot handle the requirements.
When using a pure Boost device, you may use an additional amplifier like the TLV271 to convert the negative output voltage into a positive feedback voltage.
It is also possible to use a Buck controller for the generation of a negative voltage.

But as you are mentioning the SEPIC architecture, I wonder why you do not consider a (non-isolated) Flyback topology.
This could cover all the listed requirements and it would also be possible to combine multiple output rails.

I have a few questions:
Why do the listed output voltages vary across such a wide range?
Should they be adjustable (maybe even during operation)?
Or are those voltages meant for different designs?

What are the requirements in terms of accuracy (when multiple output rails get combined, the accuracy will be somewhat lower).

Has the mentioned part (LM51561-Q1) already been selected /determined ?
For which reason?

Thanks and regards,
Harry

Hi Harry

Why do the listed output voltages vary across such a wide range? By different firmware to set in our current design.
Should they be adjustable (maybe even during operation)? Currently, our design used the current feedback power ic, and use externally connect a DAC circuit to increase or decrease the current for this feedback to achieve the purpose of changing the output voltage. We know this architecture of feedback is not popular. So, we think to change this design to a voltage divider of feedback now, and we will use a different BOM(divider resistor) instead of an externally DAC.
Or are those voltages meant for different designs? Yes, you can also think so.

What are the requirements in terms of accuracy (when multiple output rails get combined, the accuracy will be somewhat lower). Normally, the accuracy of each group of circuits requires +-1V

Has the mentioned part (LM51561-Q1) already been selected /determined? Not yet, please provide the more advice
For which reason?

By the way, we list such a wide range because we want to know whether each set of voltage outputs, the minimum and maximum values, can meet our needs

Hello Jung Wei Tsai,

I am not certain if I understand correctly.
Once the device gets sold, shall the voltages remain stable at certain settings?
In that cas some fixed feedback dividers can be used.

Or shall it be possible to adjust the output voltages while the device is in use?
Then you would need the setup with some kind of DAC and program different voltages.

Can you please clarify.

Also, in case it is necessary to adjust the voltages, will all four need to be changed individually?

I am asking because in case of a flyback topology you could create multiple voltages with a single controller.
And the turns ratio of the windings of the transformer will define the output voltages.

But if you modify one voltage settings, the others would also change accordingly.

Please let me know.

Thanks,
Harry

Hi Harry

We think that our current design is not the point, because we just want to change the current design.
Our application is the timing control board of the electronic paper, it's just like LCD TCON board.
Because of the limited space, we hope that each group of power does not need too much space.
We need you to help to tell us, is TI has like ADI_LT3471 dual power output and can be used to boost or invertor capability solution (SEPIC/CUK), and meet our specifications. If you have other suggestions, please tell us which part to use and how to start designing.

Let me re-describe the specification requirements:
Q1: Vin: DC+12V, Vout: -4V~ -21V /IccMax: 2A
The description of the voltage is that the power ic must be able to output a maximum of -4V, a minimum of -21V, not changes in a certain period of time. Other specifications have the same meaning as this.

Thanks.

Hello Jung Wei Tsai,

A dual output controller would be a new requirement.
I will ask my colleague and come back to you on that one.
Is it a must?

I was more thinking of a flyback topology with multiple windings for multiple outputs.
But, as mentioned before, the turns ratio of the windings of the transformer will define a fixed multiplicator/ratio between two output voltages.

Example:
Q1: Vout= -20V,    Q2: Vout = -5V    =>  Ratio = 4:1
Therefore, if Q1: Vout = -4V,  Q2: Vout  will be -1V    (same ratio)
If a different ratio is required, a different flyback transformer will need to be populated on the board.

Would this limitation be acceptable?

Thanks,
Harry

Hi Harry

I'm sorry for misunderstanding you.
Let me re-explain my requirements, I need separate sets of power outputs as follows:
Requirement 1: DC IN +12V / Vout(max)= -0.3V (Through the feedback divider resistor, I can use different resistor values ​​or Variable resistor, adjust Vout= -0.3v ~ -21V)
Requirement 2: DC IN +12V / Vout(min)= +5V (Through the feedback divider resistor, I can use different resistor values ​​or variable resistors to adjust Vout = +5v ~ +28V)
The IccMax mentioned earlier is instantaneous, not continuous, so please have at least 3 amps of capacity, and I will use capacitors to compensate for other insufficient parts.

Since the timing control board of electronic paper, a positive voltage, and a negative voltage are regarded as a group, so we want to find a product like LT3471 from TI, which happens to be that SEPIC outputs a positive voltage, and CUK outputs a negative voltage.

Thanks.

Hello Jung Wei Tsai,

Somehow this discussion is going in circles and I want to apologize for that.
My take is that you would want to be able to configure each output voltage individually and that fixed dependecies between two outputs would not be acceptable.

I can not understand how the LT3471 could handle the current requirements that you have listed and how this would be a solution for your application.
The CUK converters that TI offers are low power and do not fulfil the power requirements that you are asking for.
You could use buck converters and connect them in this way to generate negative output voltages:

Or, as I had mentioned, many of our Boost converters could be used in a flyback topology.
If no fixed dependencies between the output voltages are allowed, you may end up with four individual converters.

Best regards,
Harry

Hi Harry

I can't find the calculation related to the output voltage from the datasheet.
Are there different equations for positive and negative voltages?
Please provide us with the equations.
and TPS54202 seems like just only Vout(max) +26V, our requirement Vout(max) is 28V.

if you have other suggestions, please tell us the model and how to design it, please consider requirement 1 and requirement 2 as independent cases.

I feel that you still misunderstood my requirements. The power IC outputs a group of positive voltages and a group of negative voltages. There should be no relationship between the two in circuit design. They should be regarded as independent individuals.

Hello Jung Wei Tsai,

I understood that all voltage outputs shall be treated individually, so a flyback with a fixed dependency between the output voltages is not an option.

Therefore, as our CUK converter cannot handle your requirements, I had proposed using a Buck to generate a negative output voltage.

I want to apologize for the confusion with the image.
The image was just to show the concept how a Buck converter can be used to create a negative voltage.
I was thinking about removing the part number but forgot about that in the end.

For sure you would need to chose a different Buck converter IC that can fulfill the expected ratings.

I am not responsible for BUCK converters, nor do I have enough knowledge about the ratings of the different parts we offer.
If this concept can be an option for you, please let me know and we will need to find another expert to help you out.

Best regards,
Harry

Hi Harry

I am not responsible for BUCK converters, nor do I have enough knowledge about the ratings of the different parts we offer.
If this concept can be an option for you, please let me know and we will need to find another expert to help you out. Understood. Please help us to tell the expert what we need.

Thanks

Hi Richard

Our requirements as below, Webench can not find our requirements like -0.3V out

Requirement 1: DC IN +12V / Vout(max)= -0.3V (Through the feedback divider resistor, I can use different resistor values ​​or Variable resistor, adjust Vout= -0.3v ~ -21V)
Requirement 2: DC IN +12V / Vout(min)= +5V (Through the feedback divider resistor, I can use different resistor values ​​or variable resistors to adjust Vout = +5v ~ +28V)

Thanks

Hi Jung Wei,

I've received notice that your power requirements have changed. Your converter will need to be able to support the sum of the absolute value of the input voltage and the absolute value of your output voltage. 

For example, if you have an input voltage of 12V, and an output voltage of -21, you would want a converter that can support at least 33-V operation. I've also attached an application note: https://www.ti.com/lit/an/snva856a/snva856a.pdf?ts=1660345461695&ref_url=https%253A%252F%252Fwww.google.com%252F#:~:text=A%20unique%20DC%2FDC%20converter,in%20line%20and%20load%20connections.

For one, I may suggest the LM5013.

Regards,

Richard

Hi Richard

Is LM5013 can meet Vout = - 0.3V ？

I saw LM5013-Q1 product parameters, the Vout (min) =1.2v,

If we have misunderstood, please let us know, thank you!