12V / 6A Power supply. Help with Buck and Boost converters + OR'ing circuit

[ronsimpson]

Coupled Inductors
https://coilcraft.com/smpwrcoupled.cfm

 

[bountyhunter]

I didn't understand what you said.
The voltage range of two Li-Ion batteries in series would be 6V to 8.4V (8.4V when both are fully charged and 6V when both are almost discharged). They can't buck to obtain a 12V output

You should not be running the Li cells down to 3V/cell. Very little energy gained past about 3.4V/cell it just wears them out fast. Stack four Li cells and you will have a source voltage you can buck to 12V.

If you think adding two cells is "too expensive", wait until you try to build a boost circuit to try to get 72W load power out of two Li cells. And you would be lucky if you get ten minutes of battery run time.

As a rough ballpark estimate, I would try a four cell series stack "two wide" which means each cell is parallelled. Allowing for converter losses, you will need to deliver about 85 Watts.

 

[esm.]

I was rethinking about this design.
The output current will be reduced to 4A at the maximum load, not 6A as the initial idea
I think now I can have a single solution based on a buck-boost topology, right?
My only experience with a buck boost IC was the TPS63000 from Texas, but it doesnt fit the requirements of this design.
Some other recomendations of buck boost ICs now?

Best regards

 

[bountyhunter]

I was rethinking about this design.
The output current will be reduced to 4A at the maximum load, not 6A as the initial idea
I think now I can have a single solution based on a buck-boost topology, right?
My only experience with a buck boost IC was the TPS63000 from Texas, but it doesnt fit the requirements of this design.
Some other recomendations of buck boost ICs now?

Best regards

I can't figure out why you are not comprehending this part... from an earlier post you said:

Using 4 li-ion cells the cost would increase a lot i think. Until now, I just know the "18650" size for li-ion batteries (18650 is a standard format),

have you even looked at the spec for one of these batteries?

https://www.amazon.com/Tenergy-Cyli...9731634&sr=8-1&keywords=18650+lithium+battery

It's about 3.7V @ 2.6A-hr which gives you about 9W-hours per cell.

Yet, you are trying to supply a 12V output at 4A which is 48W. With converter losses, it will be close to 60W of load power.

And you are stuck on the idea you can do this with two of those batteries? You would be drawing about 10A (DC) from the batteries and peaks would be probably 13 - 14A.

Your design can't progress until you get hooked into reality. At MINIIMUM I would recommend you stack four of those cells in series and then buck down to 12V @ 4A. Forget about boosting or buck-boost, you can use a straight buck. You could even use a simple-switcher type which would save you a lot of trouble and development time.

 

[Ubergeek63]

I have a doubt about the SEPIC topology. This topology automatically works like a Buck-Boost (VIN can be lower and higher than the output voltage) ??? Or does it need to be "configured" to work just for buck operation or to be "configured' to work just for boost operation ???

Do you know if I can find this kind of inductors (dual inductors on the same core) on Digikey, Farnell, Mouser or others??? Or if I would need to produce them specifically for my design?

it by nature will put out more or less than the input. in truth ANY flyback will, but there is more restrictions on flyback magnetics

Dan

 

[esm.]

I can't figure out why you are not comprehending this part... from an earlier post you said:



have you even looked at the spec for one of these batteries?

https://www.amazon.com/Tenergy-Cyli...9731634&sr=8-1&keywords=18650+lithium+battery

It's about 3.7V @ 2.6A-hr which gives you about 9W-hours per cell.

Yet, you are trying to supply a 12V output at 4A which is 48W. With converter losses, it will be close to 60W of load power.

And you are stuck on the idea you can do this with two of those batteries? You would be drawing about 10A (DC) from the batteries and peaks would be probably 13 - 14A.

Your design can't progress until you get hooked into reality. At MINIIMUM I would recommend you stack four of those cells in series and then buck down to 12V @ 4A. Forget about boosting or buck-boost, you can use a straight buck. You could even use a simple-switcher type which would save you a lot of trouble and development time.

Hi.
I understand that.
The Li-Ion batteries that will be used are for 4Ah or 5Ah. But maybe stacking 4-cells and using just one Buck converter could make the design be not so expensive.

Following your idea of stacking 4 ce-lls and just doing a Buck, do you have some recommendations of Buck converters capable of 4,5A to 6A output? I have less experience with SMPS design, mainly when for power above 10~15W. Below these power range, there are several options to choose. Many ICs that can be found easily at digikey.

 

[ChrisP58]

Make sure that you read and understand all of the application details of the cells you plan to use. A 5 AH rated cell probably won't deliver 5 Amps for one hour. The AH rating of most batteries is usually specified at far less than the 1C discharge rate. For instance, lead-acid batteries are usually spec'd at a 20 hour discharge rate.

As far as choosing a buck converter, go to https://www.ti.com/ and enter your operating details their Webench designer. It will recommend parts for you, and give you a full design to start from.

 

[esm.]

Hi. Thanks for the help of everyone.

As I said on previous posts, I can also do the following:

- When a voltage between 17 to 30VDC is present at the input, the output should be capable to deliver 12V @ 4,5V ( = 54W)

- When there is no voltage present at the input, the circuit could be supplied from 2-cell Li-Ion (5Ah), and the output should be capable to deliver 12V @ 3A or maybe only 2,8A ( = 33,6W)

There will be some devices connected to the 12V output, but some of them don't need do be ON when the power supply is running from the batteries. I can do this just by using some resistors and MOSFETs to switch-off some of these devices from the output when the circuit is running from the battery (2-cell Li-Ion), so that the MAX currrent at the output will be 2,8A (or maybe 2,5A, I will check). I will also check the possibility to reduce the output voltage to 10V.

- Considering I am using a circuit that switch-off the Li-Ion cells from the circuit when the series voltage of these 2 cell-s are 7V (3,5V at each cell), and considering a 100% efficient converter, the mean current consumption on the baterries would be around 4,8A (33,6W / 7V).

Detail 1: The batteries rarely will be used (power the circuit). They will be used for example, just one time during a one month period. Its a very low "duty cycle".

Detail 2: There is no need the batteries have long useful life.

Detail 3: Considering both cells are fully charged, and in a next moment, they start to supply the circuit input, they just need to maintain 12V at the output for a period of time of 5 to 10 minutes (if more, better)

So this way, I think I could use only 2-cells and use just a Buck-Boost / SEPIC topology, also.

 

[bountyhunter]

do you have some recommendations of Buck converters capable of 4,5A to 6A output?

5A

https://www.electro-tech-online.com/custompdfs/2012/06/lm2678.pdf

 

[esm.]

Hi.
Have you ever seen a single Li-ion charger IC, capable of charging 4 Li-Ion cells connected in series? In a single chip?

 

[ronsimpson]

I have not followed along for a while.
Have you thought about ORing the input voltage and the battery voltage together then having one PWM in SEPIC mode.
The PWM will get power from the input or the battery, which ever is larger. You will need to choose a diode that has low forward voltage.

 

[esm.]

I have not followed along for a while.
Have you thought about ORing the input voltage and the battery voltage together then having one PWM in SEPIC mode.
The PWM will get power from the input or the battery, which ever is larger. You will need to choose a diode that has low forward voltage.

Yes, someone here already had the idea to make the ORing circuit at the input, the diode model is not a problem, I can find a suitable schottky easily.

The problem is when the battery is powering the circuit, because the current would be very high and could damage them. To solve this problem, somebody here gave a suggestion to use 4 Li-ion cells connected in series to reduce this current, but I don't know a single charger IC capable of charging 4-cell connected in series.

The original idea of this SMPS was a 12V output capable of 6A, but now I reduced to 4,5A and maybe I will reduce to 3A, by switching-off some devices that doens't need to be connected to the output when the circuit is running from the batteries. I am studying that. And also studying the possibility to reduce the output voltage to 10V, I need to check this yet.

Following the previous posts, the main ideas are:

1) Connect 4 Li-ion cells in series, use a ORing circuit at the input, and use just a Buck converter capable to output 12V @ 4,5 to 6A. I also think this is the most correct solution.

2) Connect just 2 Li-ion cells connected in series, and use a Buck-Boost / SEPIC topology converter capable to output 12V @ 4,5 to 6A. Using this method, a Buck-Boost converter suitable I found is the LM25118 from Texas. There are Dev. Boards for this IC also.

3) I am also open to new ideas.

LM25118 is a Buck-Boost topology IC and has the required (and also extra) features, like:

- Input voltage range of 3V to 42V
- ENABLE input
- Programmable switching frequency from 50 to 500 KHz
- Undervoltage lockout
- Soft-start function
- Adjustable output with 1.5% feedback reference accuracy
- Integrated high and low-side gate drivers
- Only available in SMD packge TSSOP-20 (OK)

 

[bountyhunter]

Not from memory. You would need to check on the net.