Any hints as to the next KMPKT product?

[Thehack]

In all fairness, and I greatly respect what TheHack is doing, I don't think open frame supplies are a good solution for the average user or broader market. In the hands of the novice I would actually argue that open frame power supplies are dangerous. I also have no interest in selling non-rated/certified supplies to the public due to the liabilities that can incur.

In order for me to do 12V seriously I would need to have something as plug and play as the HDPlex solutions with a significant reduction in AC-DC converter size. If you look at the Bel 180W AC-DC open frame supplies (which aren't GaN) this gives you a bit of an idea of where I think these units are headed. This is ultimately what I am working towards longer term.

Open frame power supply are no different to passive atx PSU, with their massive vents. There are issues with being plug and play, and I heavily rely on the certified meanwell units to be safe, and provide the correct connectors so users can easily hook it up. The Pure case has been warmly received because it is actually straight forward to build in.

I don't intend to chase the novice market at all. As soon as you chase the novice market, a big company will steal your ideas and mass produce it and kill you off.

It is no more dangerous than running the super open S4M, where a bad spill can short 400W worth of DC current, or the full wall of AC current in an ungrounded AC-DC unit (I heavily criticize this more than anything).

DC-DC units don't have the same liability issues as AC-DC units, since the ac-dc unit is the only thing that really require certifications, and they often come with many protection features. At least in the USA. Any off the market ac-dc is already fully certified so that keeps small run certification down.

I have two primary goals:

1. To bring small form factor computing even smaller. This means I have to fully integrate power solution and chassis design. Since the beginning I have supported the idea that 12V is the future of sff.

2. To bring total cost of uSFF. Which means I have to fully integrate the power and the chassis.

Keep in mind that our Chinese market competitors are bringing flex atx units and chassis. If they ever figure out good customer support and a good flex atx unit is released, we'd be in trouble.

At the time being, I think 19V serves a unique purpose in that the external units are powerful and compact compared to 12V. Should a GAN unit ever come out, it will certainly remove that advantage.

-----

For the dynamo 360, have considered 12V pass through instead? It'll bring the total cost of a kit down.

 

[Legion]

Some interesting views being aired here: @Kmpkt @Thehack Thank you for your insight and how you see your power solutions progressing.
I really appreciate the clarity

 

[Phuncz]

It's mature discussions like these that bring fuzzies to my SFF heart.

 

[Kmpkt]

For the dynamo 360, have considered 12V pass through instead? It'll bring the total cost of a kit down.

I actually hadn't, but that is an intriguing idea. Might look into in for a future version of the unit

It's mature discussions like these that bring fuzzies to my SFF heart.

Your mom @Phuncz

 

[Kmpkt]

As an example of how potent GaN will be, check this baby out:

https://www.us.tdk-lambda.com/lp/products/pfh-series.htm

4" x 2" x 0.5" (101.6mm x 50.8mm x 13.5mm) @ 504W of AC-DC output. 85-265V input and the 12V version is incoming shortly. Unless I'm mistaken on how these units work (the evlauation boards have additional components for function), a similar device with PC compatible inputs/outputs could give us a 300W+ AC-DC converter that is roughly the size of a 2.5" SSD, no?

 

[Thehack]

As an example of how potent GaN will be, check this baby out:

https://www.us.tdk-lambda.com/lp/products/pfh-series.htm

4" x 2" x 0.5" (101.6mm x 50.8mm x 13.5mm) @ 504W of AC-DC output. 85-265V input and the 12V version is incoming shortly. Unless I'm mistaken on how these units work (the evlauation boards have additional components for function), a similar device with PC compatible inputs/outputs could give us a 300W+ AC-DC converter that is roughly the size of a 2.5" SSD, no?

Not quite. It'll be half the size of current ac-dc yes, but the evaluation board contains the reference circuit, two fat capacitors and two inductors for AC filtering, a relay for some reason, and a regular fuse for protection. You can easily rearrange them more compactly though.

GAN is certainly the future though. It just needs to scale up to be cheaper.

 

[Kmpkt]

Yeah I know I was overreaching with the 2.5" SSD but the point being a 400W AC-DC solution will be vastly smaller than the already compact HDPlex 400W AC-DC with better efficiency, thermals, etc. I'm excited for that day.

 

[Analogue Blacksheep]

Seeing how the conversation has gone, I'll add some context as to why I asked about a future 400W mini.

At the time of the post I was exploring the option of a build in an Antec VSK2000 U3. However, it only could take TFX PSU's and the highest I could find was 300W (Seasonic used to do a 350W but they seem to be missing). I then looked at the possibly of using the two HDPlex 400W parts as a work around, but I wasn't sure about how they could fit together in the case. I wanted it to be brickless. I then noticed direct plug DC-ATX models and I felt that if they came in a higher wattage it would make the build easier.

I've moved onto a different idea since then. I'm toying with the possibly of going with an S4 mini, but the iPad Pro might make more sense for what I do (Art). Plus I'm still a bit of a novice myself when it comes to PC builds.

---

A PSU that would in theory fit into a 2.5 or 3.5 drive bay/mount would be amazing.

 

[smitty2k1]

Can someone tell the uninitiated what "GAN" means? Is it a certain technology or a brand?

 

[Legion]

Can someone tell the uninitiated what "GAN" means? Is it a certain technology or a brand?

https://www.electronicdesign.com/power/gan-power-transistors-master-strokes-power-supply-canvas

GaN = Gallium-Nitride

Using components based on that instead of silicon mosfets (hopefully not too far away now) will enable a vast reduction in the size of power supplies. There are also other positives in conversion efficiency and thermals (without being too technical)

 

[smitty2k1]

https://www.electronicdesign.com/power/gan-power-transistors-master-strokes-power-supply-canvas

GaN = Gallium-Nitride

Using components based on that instead of silicon mosfets (hopefully not too far away now) will enable a vast reduction in the size of power supplies. There are also other positives in conversion efficiency and thermals (without being too technical)

Nice. Thanks!

Kinda like back in the mid '00s when some stereo amplifiers started coming out with Class-D digital amplification. Allowed equivalent power in a much smaller package.

 

[Legion]

Nice. Thanks!

Kinda like back in the mid '00s when some stereo amplifiers started coming out with Class-D digital amplification. Allowed equivalent power in a much smaller package.

https://www.theverge.com/2018/11/1/...e-anker-material-silicon-semiconductor-energy

It's not a brilliant article but it has information in there that may be easier to digest than technical schematics and diagrams and a better overview than my brief description ( I honestly CBA with walls of text anymore, sorry)

 

[Aluminyum]

As of right now the Merge may be getting baked into the Dynamo 360 which would make it 16-24V.

There are however improvements planned for the next version of the Mini that will hopefully increase overall rating of the unit significantly.

Any updates you could share?

 

[Kmpkt]

Yeah actually. 360 version 2.0 is complete and tested and I am in the process of placing the order. The new unit will differ from version 1.0 in that the SATA function is gone and the 4 pin header is now the DC-OUT (ie. one end has all outputs, 2 x 8 pin, 1 x 4 pin passthrough for the mini). The other end of the unit will have the same two 6-pin headers and a modified jumper array. The two 6 pins will both be DC-IN headers. The leftmost header will supply the whole unit from a single AC-DC unit. When wanting to use two AC-DC sources, then the rightmost header will supply only the passthrough while the leftmost will only supply the 8 pins. This layout is intended to easily facilitate a true merging of both DC inputs on future revisions (will require significant further development to make this work properly).

As far as a new Mini goes, I will be color matching the new Dynamo 360 to the HDPlex 200W DC-ATX so that they can be used as a pair. Seeing as the Dynamo Mini was simply a rebranded 160W DC-ATX, I've decided not to do another custom heatsinked run at this point. The main reason for this is that I needed to reallocate about half of my funds finance another project (unrelated to SFF) which, if successful, will fully finance the production of my heatsink as well as providing much needed funding for other products including the GaN 200W AC-DC bricks I would like to offer.

 

[goose1969x]

Any estimate to preorder date of the new Dynamo?

 

[iaomw]

Yeah actually. 360 version 2.0 is complete and tested and I am in the process of placing the order. The new unit will differ from version 1.0 in that the SATA function is gone and the 4 pin header is now the DC-OUT (ie. one end has all outputs, 2 x 8 pin, 1 x 4 pin passthrough for the mini). The other end of the unit will have the same two 6-pin headers and a modified jumper array. The two 6 pins will both be DC-IN headers. The leftmost header will supply the whole unit from a single AC-DC unit. When wanting to use two AC-DC sources, then the rightmost header will supply only the passthrough while the leftmost will only supply the 8 pins. This layout is intended to easily facilitate a true merging of both DC inputs on future revisions (will require significant further development to make this work properly).

As far as a new Mini goes, I will be color matching the new Dynamo 360 to the HDPlex 200W DC-ATX so that they can be used as a pair. Seeing as the Dynamo Mini was simply a rebranded 160W DC-ATX, I've decided not to do another custom heatsinked run at this point. The main reason for this is that I needed to reallocate about half of my funds finance another project (unrelated to SFF) which, if successful, will fully finance the production of my heatsink as well as providing much needed funding for other products including the GaN 200W AC-DC bricks I would like to offer.

This plan sounds great. Would you consider the pass through could support 12V when using 2 PSUs (one 19V and one 12V)?
I mean, it's just a pass through. Logically, it could be any voltage, could be both 12V and 19V. That will offer much more flexibility for the 24pin DC-ATX unit.