Why a linear power supply?

[miguelito]

So that got me to thinking—instead of a high quality linear power supply for the Mac Mini, why not a HIGH QUALITY purpose built SMPS?

Indeed. My biggest surprise was when I saw the Linn Linto (a phono stage no less) using a SMPS!!! Amazing. Proof they can be incredibly silent. My CD player/DAC uses a SMPS as well. Yes, it's a matter of good design.

[JohnSwenson]

Very impressive case and seems to have selectable output voltages, but any idea why 19V is not an available option for even more universal flexibility ?

 

This supply was designed to be used over a range of 5V to 12V, the raw supply has to be high enough in voltage to handle 12V at 5A, but that means that at 5V a huge amount of heat is generated and has to be dissipated in the heatsinks. To produce a 19V output would require a much higher raw supply voltage which would more than double the heat dissipation at low voltage. Not only would that require a much larger heatsink it would require a different regulator, the current ones we are using simply cannot get that much heat out, they would fry even with a super large heatsink.

 

In order to do 19V with the existing case and regulators, it would have to be a dedicated high voltage supply with 12V as the lowest voltage, which would take a different transformer and different passive components. The existing design was carefully optimized for what it is, changing the voltage that much would significantly change the optimizations resulting in a different design. A fair number of parts in the existing design could be kept, but a fair number would also have to change.

 

John S.

[tboooe]

This supply was designed to be used over a range of 5V to 12V, the raw supply has to be high enough in voltage to handle 12V at 5A, but that means that at 5V a huge amount of heat is generated and has to be dissipated in the heatsinks. To produce a 19V output would require a much higher raw supply voltage which would more than double the heat dissipation at low voltage. Not only would that require a much larger heatsink it would require a different regulator, the current ones we are using simply cannot get that much heat out, they would fry even with a super large heatsink.

 

In order to do 19V with the existing case and regulators, it would have to be a dedicated high voltage supply with 12V as the lowest voltage, which would take a different transformer and different passive components. The existing design was carefully optimized for what it is, changing the voltage that much would significantly change the optimizations resulting in a different design. A fair number of parts in the existing design could be kept, but a fair number would also have to change.

 

John S.

 

Now that is a good answer!! Thank you sir!

[miguelito]

This supply was designed to be used over a range of 5V to 12V, the raw supply has to be high enough in voltage to handle 12V at 5A, but that means that at 5V a huge amount of heat is generated and has to be dissipated in the heatsinks.

Why not a fully separate 5V regulated linear supply instead? Isn't the 5V line drawing very little current?

[sandyk]

My CD player/DAC uses a SMPS as well. Yes, it's a matter of good design.

 

A CD player that doesn't use a cheap Asian SMPS, WITHOUT the obligatory high voltage capacitor between primary and secondary sides ?

 

Why not a fully separate 5V regulated linear supply instead? Isn't the 5V line drawing very little current

 

Unless John used a transformer with another lower voltage secondary winding, there would be a lot of needless heat dissipation in the 5V regulator. Say that the voltage supply needed 5V headroom above 12V for reliable operation, then you would need to drop 12V across the 5V regulator.(17V minus 5V) With say 2A being supplied by the 5V regulator, that's 24 Watts of heat.

Another option could be to supply the 5V regulator from the +12V regulator, but that may result in degradation of the 12V rail, as well as the need for additional heatsinking for the 12V rail.

 

Alex

[Superdad]

In order to do 19V with the existing case and regulators, it would have to be a dedicated high voltage supply with 12V as the lowest voltage, which would take a different transformer and different passive components.

 

Thanks for chiming in John and saving me the time to answer (sounds better coming from you anyway).

 

I've tried to explain in my 'newsletters' that offering a supply with selectable voltages places considerable demands on the heatsinking capabilities of the chassis and the thermal limits of the devices. That is why most other LPS units are supplied with one or two fixed outputs.

 

I do see that the Keces unit offers 19V, but it looks like they are using a transformer with multiple secondaries. That's also probably why they limit the 5V current to 3A (a trans of a given physical size/VA rating can only handle so much and that gets split up among the secondaries).

 

Being that the choke filter is a key design element of the JS-2, we would really need a second big choke and another PCB for a higher DC output range. That would require a bigger chassis, and as above, an even larger R-core trans since I'd not want to sacrifice output current capability in the other range when adding secondaries to cover a new higher range.

 

BTW, I was running in a JS-2 on the bench last night at 9V/7.3(!) amps with no complaints. Pretty good since we are only rating it for 5A continuous (not that any gear draws that much on a continuous basis) Got nice and toasty after 1/2 hour. Wish there was a way to come up with an instantaneous/peak current specification.

 

In the end I am not too concerned about offering 19V output as it seems that most newer CAPS designs are based on motherboards taking 12V (either via external jack or via PicoPS). And of course motherboards are not actually running on those high voltages anyway--they all have DC-DC converters taking it down to 5V, 3.3V, 1.8V or less. The only reason computer makers are specifying 14.5 to 19V (I'm tossing in a lot of laptops in that range) is to lessen the current requirements of the supplies and chargers.

 

Frankly, if we were to think about a fancier LPS, I'd want to consider a try ATX-type supply with at least 3 rails and go down to 3.3V. Some newer motherboards are less picky about turn-on sequence timing, so that's not as big a deal as in the past. I'm waiting to see what shakes out with newer CAPS designs--like what Chris might pick as a mobo for v4. Maybe discrete DC inputs (and fewer DC-DC regs) will become popular.

 

I need to be careful not to sound like a product promo, but I really hope people will think of the LPS John and I have done for more than just a computer and peripherals. The R-core and the choke design benefit those SQ-wise, but the big excitement comes when a DAC, headphone amp, or USB>S/PDIF converter is powered by this mini-beast. Waiting for someone with an exaSound e20/22 to pair up with a JS-2…

 

Ciao,

 

--Alex C.

[Superdad]

Why not a fully separate 5V regulated linear supply instead? Isn't the 5V line drawing very little current?

 

It's hard to please everyone. Someone else wanted to power a CAPS at 12V and an array of 8 hard drives--all at 5V--and those would add up some current. Maybe it is my fault for choosing a small chassis, but between the R-core, the choke, the AC input module and the PCBs, there really is no space left for a small choke and another PCB (attached to a heatsink somewhere) to do an entirely discrete 5V supply.

 

Now if if you are asking for separate, smaller box with a nice 5V supply, that may happen. But I'm still ramping up with shipments of this model...

[dtb300]

Where can I find more info on the JS-2? Specs and cost and availability

[tboooe]

Where can I find more info on the JS-2? Specs and cost and availability

PM Superdad. I am definitely going to pick one up as soon as I build my new CAPS. I have high hopes it will outperform my Keces LPS. It better for almost 3x the price!

[kisjano]

New high end audio pc coming soon.

[miguelito]

It's hard to please everyone. Someone else wanted to power a CAPS at 12V and an array of 8 hard drives--all at 5V--and those would add up some current. Maybe it is my fault for choosing a small chassis, but between the R-core, the choke, the AC input module and the PCBs, there really is no space left for a small choke and another PCB (attached to a heatsink somewhere) to do an entirely discrete 5V supply.

 

Now if if you are asking for separate, smaller box with a nice 5V supply, that may happen. But I'm still ramping up with shipments of this model...

Fair enough.

[kisjano]

Your comment:000000000

[JJJ]

New high end audio pc coming soon.[ATTACH=CONFIG]13907[/ATTACH]

 

At a minimum, twist all those wires together…

[kisjano]

wires togetherr????hahahahahaha you are funny gay

[Superdad]

Your comment:000000000

 

Kisjano: I wish I could say that I understood any of your posts--but I do not. You pop up with random, out of context comments and photos. Please tell us if you are doing DIY or are are part of a small company (name?) producing a commercial product.

Thank you,

--Alex C.

[kumakuma]

wires togetherr????hahahahahaha you are funny gay

 

How do you know JJJ's sexual orientation?

[Superdad]

I don't want any hong kong monkey!!!!!!!!!!!!!!!!!!!!!!!!!

 

????

[miguelito]

wires togetherr????hahahahahaha you are funny gay

Errr. WTF?

[JJJ]

wires togetherr????hahahahahaha you are funny gay

 

Well I am not gay (as if that was relevant) and I wasn't trying to be funny. EMI 101: twist wire pairs together.

 

What's the PSRR of your regulator boards? Post some simulation results or, better yet, actual tests and circuit diagrammes. The pics are a bit boring...

[sandyk]

Personally, I doubt that it's a good idea to have those R Core transformers so close together. Sudden changes in the secondary side of one transformer seem likely to induce into the primary side of the next one. R-cores may be good, but are they that forgiving of poor practice ?

Perhaps a question for Jon Swenson ?

Alex

[miguelito]

wires togetherr????hahahahahaha you are funny gay

For the record, i am a homosexual and i find the sheer out-of-placeness of your comment mildly amusing in a sort of "how stupid is this guy?" sort of way.

[Jud]

For the record, i am a homosexual and i find the sheer out-of-placeness of your comment mildly amusing in a sort of "how stupid is this guy?" sort of way.

 

I thought it was just a typo (that or a wild spell check), exactly 'cause it *was* so out of place.

[sandyk]

Hi Jud

After seeing post 8 here http://www.computeraudiophile.com/f10-music-servers/hdplex-linear-power-supply-21175/

I am not so sure it was a typo or error.

Kind Regards

Alex

[JJJ]

Let's give the guy the benefit of the doubt for now. But Kisjano, share something interesting and informative with us.

 

PS: paying up for R-cores is a waste of money in my opinion especially when one doesn't twist together the primaries and secondaries

[4est]

I do not see an issue here.

Hi Jud

After seeing post 8 here http://www.computeraudiophile.com/f10-music-servers/hdplex-linear-power-supply-21175/

I am not so sure it was a typo or error.

Kind Regards

Alex