Isolated DC power supply ideas

[M6YRU]

my project will need 3 isolated power supply circuits, i know there is DC-DC isolated devices but these are lacking somewhat with the ones i seen at a measly 2watt

i need a minimum of 1amp at 5v but i am flexable in voltages 3.2-42v and i dont mind if i need to convert to AC and such,

just looking for better ideas a the moment than going to the AC and multiple transformers route...

my issue is most common/ground is shared, the system i would be using it from is multiple tap off points of batteries so a common on one will be a live on the next, thus shorting out the cells...

does anyone know how a DC-DC isolator device works and that it could be built?

[mariush]

What's your budget?

Lots of power supplies would fit your needs, here's for example isolated dc-dc converters from 5w output power and up : https://www.digikey.com/short/qqcmb8

[Jeroen3]

does anyone know how a DC-DC isolator device works and that it could be built?

You need to drive a transformer. Push-pull configuration is fairly common.
You can make this quite easily with some fets/transistors and PWM. Common small driver IC is the SN6501.

[T3sl4co1l]

Why do you need so much power in a battery tap application -- presumably, a battery monitor or equalizer application..?

There are better ways to do that; it's an already-solved problem.

Tim

[Someone]

i know there is DC-DC isolated devices but these are lacking somewhat with the ones i seen at a measly 2watt

Look harder, there are isolated DC-DC modules available in tens and hundreds of watts (and probably larger).

[M6YRU]

Why do you need so much power in a battery tap application -- presumably, a battery monitor or equalizer application..?

There are better ways to do that; it's an already-solved problem.

Tim

the project is a multiple tap off with a mix of buck/boost converters and they run different LED arrays that have different voltages, like a small 3w led then a 12v 10w COB then bigger n bigger, the 10w draws about 1amp at 12v....
 however to charge these cells i would like to use tp4056 per cell to charge each to the full value...
a BMS wouldnt work as its the  full 10 pack cells an there be too many buck converters

[mariush]

I don't understand.

There are actual led driver ICs which take some input voltage and limit the current going through the led to the value you set (either through a resistor, or through an analogue/pwm signal or through i2c/spi commands ) so you can connect a 3v led, a 12v led any led or series of leds with forward voltage lower than input voltage in the case of a step-down led driver.

If only ONE led needs to be turned on out of a bunch of leds, then you could even use a single led driver and you can just disconnect leds by shorting them out using a relay for example see the picture below for some example circuit

Here's an example of a chip, A6217 from Allegro : https://www.digikey.com/product-detail/en/allegro-microsystems-llc/A6217KLJTR-T/620-1816-6-ND/6189121
6v to 48v input voltage, up to 3A current, you can set the maximum current using the sense resistor, and then use pwm to dim the active led
Lots of other chips that would fit your needs.
Why do you need isolated power supply i don't get it...

Or MP3302 : https://www.digikey.com/product-detail/en/monolithic-power-systems-inc/MP3302DD-LF-P/1589-1159-1-ND/5298246
3v to 6v input, up to ~ 36v and 1.3A internal current limit

[David Hess]

An inverter or flyback converter will support as many outputs as you care to wind secondaries on the transformer.

The least expensive standard option appears to be multiple 1:1:1 pulse transformers from Murata.

[macboy]

So if I understand, you have a 10 cell lithium battery pack, and it drives various high power LEDs at various voltages, using multiple buck and/or boost conveters to get those various voltages? Correct?

I REALLY hope you are not tapping the battery bank at different points to get various voltages. This is a disastrously bad idea. Use the the battery bank as a single high voltage source to power all of the buck converters. You absolutely can find 10 cell balance/protection devices. For output voltages that are near to or higher than the battery bank (when it's nearly discharged) you will need to use buck/boost to generate the output, but that's not a problem. If the fully charged battery bank has a higher voltage than some of the buck converters can handle at their input, then find a buck converter which can handle the input voltage, and use it to create an intermediate voltage rail for the multiple smaller low voltage buck converters. Or if the 12 V output one is running directly from the battery bank voltage, you could use that 12 V as the input to the 3.3 V output buck converter.

All LEDs will have a common ground in this case. You have not convinced me that you need multiple different ground points for different LEDs.

[M6YRU]

So if I understand, you have a 10 cell lithium battery pack, and it drives various high power LEDs at various voltages, using multiple buck and/or boost conveters to get those various voltages? Correct?

I REALLY hope you are not tapping the battery bank at different points to get various voltages. This is a disastrously bad idea. Use the the battery bank as a single high voltage source to power all of the buck converters. You absolutely can find 10 cell balance/protection devices. For output voltages that are near to or higher than the battery bank (when it's nearly discharged) you will need to use buck/boost to generate the output, but that's not a problem. If the fully charged battery bank has a higher voltage than some of the buck converters can handle at their input, then find a buck converter which can handle the input voltage, and use it to create an intermediate voltage rail for the multiple smaller low voltage buck converters. Or if the 12 V output one is running directly from the battery bank voltage, you could use that 12 V as the input to the 3.3 V output buck converter.

All LEDs will have a common ground in this case. You have not convinced me that you need multiple different ground points for different LEDs.

thats exactly what im doing lol 10s tapped off in different places, no BMS, charging each cell seperately...

lets say i need to run 4 led's, cell 1, for one led, cell 2 for second led but the higher led needs cell 1,2 and 3 to get to 10+v but the 4th led i wont run from cell 1 or 2 due to current drain but would run from 4,5,6,7,8 and 9 to get to 20/24v
but one COB led likes to be 36v! so i have to run all  cells

i know this is confusing somewhat, just trying to save on the buck converter conversion efficency lol

[mariush]

Proper buck (step-down) led driver ICs are up to 97-98% efficient. You can't reach that with isolated power supplies since they require transformers and other crap to isolate.

You'd only want to adjust input voltage if you use basic linear regulator style led driver chips, which adjust voltage by dissipating the excess as heat. LED driver ICs of switching type don't have this issue.

In your case, it may be a good trade-off to use only 6 cells and a commercial battery charger off the shelves with cell balancing like this one for example : https://hobbyking.com/en_us/imax-b6ac-v2-professional-balance-charger-discharger.html  which works with either AC in or DC in and has standard JST connectors to plug the battery pack and charge it safely.

So you'd have 6 x 3.7..4.2v = up to 25v or so ... you could use a single step-down led driver for all leds below ~ 21v and maybe a SEPIC (step-up and step-down) for 21v and up leds ... this one would be slightly less efficient, but you won't care that much... you can always have two series of 6 batteries connected in parallel for a total of 12 batteries and you have more current.

You could use for example MP2483 , step-down (buck) or buck-boost regulator (you have example circuits for both in datasheet) : https://www.digikey.com/product-detail/en/monolithic-power-systems-inc/MP2483DQ-LF-P/1589-1123-1-ND/5298210
Up to 95% efficient with ~20v input, but works with up to 48v , over voltage protection, adjustable frequency for better efficiency, dimming, the works.

[james_s]

It's pretty easy to make a Royer oscillator using a couple transistors and a transformer. A few years ago I was discussing these with a friend and in the process I lashed one together out of parts I literally found on the floor near my bench. It worked on the first go and was easily capable of supplying >1A. I just wound some turns on a random ferrite toroid that was laying around and then used a pair of NPN power transistors and a couple resistors and tacked it together in free air.