Should I market this thing, and how?

[tszaboo]

About marketing, and other stuff: Nobody likes power plug adapters. I would never buy a power brick, which has a flimsy cabling, which will be the first part to break, coming from a third party manufacturer. What you can target is USB type C powered laptops. Just have a USB type C output, bring your own cable, and support type C power delivery.

Both cords are already detachable (and therefore replaceable).

Sure. Show me a shop which sells whatever to Thinkpad X220 power input cable.

[AndyC_772]

Have you built a unit yet? Would be interesting to see how simulation compares to real life.

No.

I'm surprised nobody else has jumped on this yet.

You're a long way from being ready to consider production. *Much* further than I think you realise.

It's great to have an idea of how a product might work, and even better to build one and be able to demonstrate it - but that's just the beginning of the development process.

Beyond that, there's still a ton of work to do. A great deal of functional testing, design for manufacture, mechanical design and packaging, enclosure design, regulatory approvals, finding a manufacturer, securing a supply of components (with sensible lead time, pricing and MOQ), shipping and logistics. That's without any commercial considerations.

None of these things individually are beyond you, I'm sure, but I'd urge you to treat the project as a learning exercise, and not to get too excited by the prospect of making any money out of it.

I've done several of these types of project over the course of my career so far. I've designed things that interest me, up to the point of having a working prototype that could - with the involvement of outside expertise - be turned into commercial products and sold.

I've never actually sold them, though - but I've taken parts of the designs and used them elsewhere. They've taught me new and useful skills, and they've been really good adverts to show prospective customers the kinds of things I can do (which is difficult when your other work is generally kept confidential).

[FrankBuss]

Have you built a unit yet? Would be interesting to see how simulation compares to real life.

No. It's still in early stage.

So how do you know that it is 96% efficiency? You do know that the numbers in the datasheets of power supply ICs are only true with better than perfect parts and layout, and only on full moon?

[nctnico]

Have you built a unit yet? Would be interesting to see how simulation compares to real life.

No.

I'm surprised nobody else has jumped on this yet.

I think everyone is still doing this: 
@Blueskull: make a prototype first and don't try to make the first version ultra compact because that will make measurements difficult. Simulations and ideas only go so far.

[mikeselectricstuff]

Have you built a unit yet? Would be interesting to see how simulation compares to real life.

No. It's still in early stage.

So how do you know that it is 96% efficiency?

Or Thermally viable?
or EMC Certifiable?
or stable?
or will even work.?

[AndyC_772]

I'd expect a design that only wastes 4% of the input power as heat should be pretty good from a thermal point of view, since there's just not that much heat being generated. Probably awful from an EMC perspective, though, given that it'll need to switch very hard and very fast (edge rate, not kHz) to get the efficiency up.

I'm really interested to know how that efficiency could possibly be achieved, though. Aren't most supplies of this type more like 85% efficient? That's a really, really major improvement in efficiency over currently available commercial product, so the design is either a) radically better in some fundamental way, or b) hopelessly optimistic.

Place your bets. I'd really like to hope it's (a).

[Kjelt]

I'm really interested to know how that efficiency could possibly be achieved, though. Aren't most supplies of this type more like 85% efficient? That's a really, really major improvement in efficiency over currently available commercial product, so the design is either a) radically better in some fundamental way, or b) hopelessly optimistic.

No they are commercially available for EU PSU's check out the term "80 plus titanium"
The catch is that this efficiency is only available for a certain load condition, so you should tweak the supply to the load.
There is no way you can build a PSU that has 96% eff. over the entire range of possible loads.

[rollatorwieltje]

XP Power makes some supplies that are around 95% efficient. They do come at a price though.

[AndyC_772]

Cool, I've actually learned something useful from the internet!

(I wonder what got pushed out of the back of my brain to make room? No doubt I'll find out in time...)

[mikeselectricstuff]

I'd expect a design that only wastes 4% of the input power as heat should be pretty good from a thermal point of view, since there's just not that much heat being generated. Probably awful from an EMC perspective, though, given that it'll need to switch very hard and very fast (edge rate, not kHz) to get the efficiency up.

And the resistive losses in the EMC filter then eat into your efficiency.

[amspire]

If you do not have transformers designed and tested, you do not have much. In designing small high efficiency designs, it is not uncommon to have many attempts at getting the transformer right, and it is usual to get results much poorer then expected. There is often signifigant losses in the transformer, and there is often more leakage inductance then expected. If you have to add snubbers to cope with the leakage inductance, you can easily find yourself below the 90% or even 80% efficiency level in spite of the great LTSpice numbers. Unless you design can actually somehow reabsorb and use the energy in the leakage inductance, it is energy that has to go somewhere.

The hardest part of switching supply design is the transformer design. You have variables of copper shape and cross section, wire insulation, core material, core gap, shape, winding arrangement, insulation gap between layers, former design and so on. If this is your first attempt, it will be much harder then you expect. High performance transformers are very imperfect devices and their design is a real art that can take years to master.

[amspire]

If you do not have transformers designed and tested, you do not have much. In designing small high efficiency designs, it is not uncommon to have many attempts at getting the transformer right, and it is usual to get results much poorer then expected. There is often signifigant losses in the transformer, and there is often more leakage inductance then expected. If you have to add snubbers to cope with the leakage inductance, you can easily find yourself below the 90% or even 80% efficiency level in spite of the great LTSpice numbers. Unless you design can actually somehow reabsorb and use the energy in the leakage inductance, it is energy that has to go somewhere.

The hardest part of switching supply design is the transformer design. You have variables of copper shape and cross section, wire insulation, core material, core gap, shape, winding arrangement, insulation gap between layers, former design and so on. If this is your first attempt, it will be much harder then you expect. High performance transformers are very imperfect devices and their design is a real art that can take years to master.

My design is not sensitive to leakage inductance -- actually, the more the better.
I'm using PL160 from CoilCraft, it is designed for 48V-72V input, but its rated Vs balance is okay for my application case.

Looks like you are using two transformers to get the 240W. It is an impressive transformer. You cannot make transformers yourself like that using off-the-shelf parts. The negative is you will be paying something like $15 for transformers from a single supplier and you will be competing against companies paying $2 for their single transformer.

It is a transformer that normally has 48V between primary and secondary, and you have to allow for something like 400V. I am assuming your supply can run off 240V mains. If it doesn't, I cannot see much of a market as a laptop supply. You will actually want to pull one of these transformers apart to see what the actual insulation between the windings consists of. It might be only a layer of enamel coating on the coils. I would be nervous about that for use with the mains.

The fact you do not care about leakage inductance is impressive. I am not going to ask you to reveal your design, so I will leave that to you. I assume you have added the extra leakage inductances into your LTSpice simulations.

[mikeselectricstuff]

Efficiency numbers are based on simulation and calculation.

And  have mininmal credibility until you actually build & test it.