At this point, I'm leaning towards a Push-Pull-PWM Transformer Output.
Much more stable and predictable.

A transformer for 60A primaries/20A secondaries is unlikely to be off the shelf.

 

Here is a small 12 to 36 boost using a current mode UC3842.
Updated the sch. Now running at 100khz.
View attachment 245501

Make that a UC3843, a '42 won't start unless the supply exceeds 16V.

 

The higher the ripple current the greater the CORE loss in the output inductor. It varies as the peak-to-peak ripple current raised to the power 5/2.
The higher the inductance, the greater the COPPER loss in the output inductor. You can work out the copper loss from the length of wire and its thickness.
Iron powder inductors give more CORE loss than ferrite for the same peak-to-peak ripple current, but they will take more current than ferrite before saturation for a similarly sized core.

Vishay's 1.5uH and 3.3uH parts quote 6.5W and 2.8W as total (heat + core) losses respectively @ 70A, which suggests core losses > copper losses (resistances 0.4mOhm & 0.65mOhm respectively).

 

A transformer for 60A primaries/20A secondaries is unlikely to be off the shelf.

I'm completely aware of that, and I keep pushing on
having a Custom Solution designed by a reputable Transformer manufacturer.

This whole problem is like lifting a ~20lb. Concrete-Block with one finger,
you can do it,
but it ain't easy or simple.

It would also appear that this is a Commercial endeavor,
and in that case, a Custom-Part is viable when ordered in quantity.
But truthfully, I don't think the Thread Starter knows what he's asking for.
.
.
.

 

Thank you both for your answers. If i decide to wind a new inductor i wont raise the inductance. But how can i achieve a good current accuracy? Are wire gauge specifications enough?

For current carrying capacity the AWG 'ampacity' charts are sufficient and manufacturers data for the ferrite/iron ore will give guidance on # of turns. Final check for inductance is with LCR meter for crude measure of inductance, a resonance test against a known value of capacitance if you have a tracking spectrum analyser, or a switching test rig built on a small bit of perf-board or a purpose-made PCB (incuding potentially re-purposing the one you already have) and using an oscilloscope/pulse generator configuration similar to my simulation - the latter tests it in a near 'real-world' scenario, is probably the easiest to set up and is arguably the most useful in that you can do saturation current testing with an appropriate power supply and load. In any case you need some test equipment.... what do you have available?

There are many 'turns' calculators on the web that, given the data for the core, will calculate the # of turns needed, but read here for caveats/guidance on saturation inductance and other issues with using these approaches for power inductors...

 

Vishay

coilcraft
It is hard to find off the shelf inductors near the size you want. Another reason for using multiple supplies.

 

coilcraft
It is hard to find off the shelf inductors near the size you want. Another reason for using multiple supplies.

Wurth or Bourns, but not everyone stocks the big high-current stuff. Its slow moving inventory and quite expensive - 15 - 30GPB1-off, around 6GBP 100up.

 

not everyone stocks the big high-current stuff

That is why I have 100 of pounds of wire and winding machines and cores.....
NOT volunteering to make some for you. lol I see this weekend I got an order for ten 220uH 100A monsters. The price of CU is out of sight.

 

For current carrying capacity the AWG 'ampacity' charts are sufficient and manufacturers data for the ferrite/iron ore will give guidance on # of turns. Final check for inductance is with LCR meter for crude measure of inductance, a resonance test against a known value of capacitance if you have a tracking spectrum analyser, or a switching test rig built on a small bit of perf-board or a purpose-made PCB (incuding potentially re-purposing the one you already have) and using an oscilloscope/pulse generator configuration similar to my simulation - the latter tests it in a near 'real-world' scenario, is probably the easiest to set up and is arguably the most useful in that you can do saturation current testing with an appropriate power supply and load. In any case you need some test equipment.... what do you have available?

There are many 'turns' calculators on the web that, given the data for the core, will calculate the # of turns needed, but read here for caveats/guidance on saturation inductance and other issues with using these approaches for power inductors...

Thanks, i did my research on coil32.net, it is a very useful site. And i think i will be going with ferrite torroid and i can get ferrite ring supplies it is not a problem. But i still need to do research about them. The CU is also a problem.

I dont have right now but i will buy frequency generator and an oscilloscope. These should be enough for measurements. For real-world tests i will use battery and motor and prepare the worst scenerio motor will face.

 

If you're still going with the original boost converter circuit, then a ferrite toroid isn't going to work, because it will saturate unless it is gapped.
Gapped ferrite toroids are not easy to wind - unless you fill the gap with epoxy the wires drop through the gap as you try to wind it.
So, if you want to use ferrite, use a EI or an ETD wound on a bobbin xo you can set the gap, or, if you want to use a toroid, use iron powder as I suggested back in post #80

 

If you're still going with the original boost converter circuit, then a ferrite toroid isn't going to work, because it will saturate unless it is gapped.
Gapped ferrite toroids are not easy to wind - unless you fill the gap with epoxy the wires drop through the gap as you try to wind it.
So, if you want to use ferrite, use a EI or an ETD wound on a bobbin xo you can set the gap, or, if you want to use a toroid, use iron powder as I suggested back in post #80

Thanks, i will use iron powder but it still will be ring, and my supplier calls ferrite ring to all of them, sorry about that. But i have a question, only permeability changes when it comes to calculations in ferrite vs iron powder core right?

 

Thanks, i will use iron powder but it still will be ring, and my supplier calls ferrite ring to all of them, sorry about that. But i have a question, only permeability changes when it comes to calculations in ferrite vs iron powder core right?

Saturation flux density is much higher for iron powder (1.6T vs. 0.3T).
The saturation is much more gradual for iron powder than ferrite. An iron-powder core is often run partially in saturation. This has the advantage that the inductance varies with current, so when you run it at low current the inductance is usefully higher.
Micrometals' software works all this out for you. Their cores (or copies, some of which are good, and some aren't) seem to be available everywhere. If you see a two-colour core, it's probably Micrometals'. The most common one on a switched-mode power supply is a type #26 (yellow/white) or a type #52 (Green/blue)
Core loss is much higher, so make sure that there is not too much ripple current.

 

Saturation flux density is much higher for iron powder (1.6T vs. 0.3T).
The saturation is much more gradual for iron powder than ferrite. An iron-powder core is often run partially in saturation. This has the advantage that the inductance varies with current, so when you run it at low current the inductance is usefully higher.
Micrometals' software works all this out for you. Their cores (or copies, some of which are good, and some aren't) seem to be available everywhere. If you see a two-colour core, it's probably Micrometals'. The most common one on a switched-mode power supply is a type #26 (yellow/white) or a type #52 (Green/blue)
Core loss is much higher, so make sure that there is not too much ripple current.

Thanks, i will keep these in mind.

 

coil32.com

coil32.net

 

coil32.net

Edited, thanks.