Stacking Switchers for more volts

This was an interesting read.

I would be curious to see what folks think about it...

https://wtfamps.com/2019/01/02/stacked-smps-psu-for-tubes/

why not just scale up voltage? 

I recall in the early days of class D audio amps the switching devices were not up to the task for high power. If you can source a high enough voltage switch, just scale it.

JR

A buck boost set up?

pucho812 said:

A buck boost set up?

Click to expand...

A step-up or boost regulator.

JR

I think this is also a relevant idea to this thread:
https://www.ednasia.com/news/article/Switcher-regulates-HV-in-novel-way
It makes the observation that if you don't expect very wide variations in your high voltage supply, you can regulate them with a low voltage regulator, with the high-voltage supply stacked on top.

pucho812 said:

This was an interesting read.

I would be curious to see what folks think about it...

https://wtfamps.com/2019/01/02/stacked-smps-psu-for-tubes/

Click to expand...

So I actually just drew up and ordered some little PCBs made up to try this. The PCB is designed around the XP Power VCE5 footprint and holds three in series.  Connections are arranged so that I can daisy chain as needed.

Cost is comparable to an EI and basic filter but lighter weight and smaller footprint. Real motivation was to have something repeatable and scalable to various voltages. My projects are DIY-focused, so giving others the ability to source the same parts and expand on the ideas is a priority.

On purpose-built designs, Hagerman has a LM555 boost implementation in some of his new products. Millet has a 10W flyback example, too.

Yeah the units he used have a switching frequency of 40k. I think the VCE are 65k. Meanwell makes some nice encapsulated units (IRM series) but they don’t have 48V versions in the lower wattage sizes... gotta jump to 30W I think, which is pretty beefy for most applications. And nice CLC pi filter and you’re <1 mV ripple.

MeanWell higher voltage / lower wattage SMPS:

APC-16-350: 48V
APC-25-350: 70V
APC-25-500: 50V
APC-35-350: 100V
APC-35-500: 70V
APC-35-700: 50V
HVGC-65-350: 186V
HVGC-65-500: 130V
HVGC-65-700: 93V
HVGC-100-350: 285V
HVGC-100-700: 142V
HVGC-150-350: 428V
HVGC-150-500: 300V
HVGC-150-700: 215V
LPC-100-350: 286V
LPC-100-500: 200V
LPC-100-700: 143V
LPC-150-350: 430V
LPC-150-500: 300V
LPC-150-700: 215V

Note that part code tells you watts and mA. For example, APC-16-350 is 16 watts, 350 mA.

Also note that HVGC appears to be 220VAC input only.

Those Meanwell supplies are all constant current.  I'm no expert in LED supplies, but it seems to me these wouldn't  work as expected. My assumption is that they vary the output voltage to reach the rated current (with some range for adjustment). Maybe it could be used in Class A or in combination with some kind of shunt regulator? I've been corrected!

I also looked at the Meanwell EPS and PS series, which can be found in lower power ratings and 48V. I decided to try the XP Power part because I liked the fact that it was encapsulated and I didn't have to worry about safety earth on multiple PCB modules floating at different voltages.

sodacose said:

Those Meanwell supplies are all constant current.  I'm no expert in LED supplies, but it seems to me these wouldn't  work as expected. My assumption is that they vary the output voltage to reach the rated current

Click to expand...

No, that's not how they work. They are actually constant voltage up to the rated current and THEN they become constant current and the voltage drops.

The MeanWell LED supplies are known to be very solid and quiet. Making your own with some PCB is not recommended. SMPS are very sensitive to layout and parts. You're much better off buying one that was designed by professionals. There's no way to know if one is really good until you just try it. But if the datasheet has THD charts like some of the MW, that's a promising sign.

Ah, thanks for explaining. I didn't know that this is how they operate, but it makes sense.

Have you tested these as supplies for audio applications or are there any projects/experiments using them that you can point me to?

SMPS are very sensitive to layout and parts. You're much better off buying one that was designed by professionals.

Click to expand...

Exactly why I'm looking at stacking modules rather than designing something from the ground up

Edit: Nevermind, found the other thread https://groupdiy.com/index.php?topic=68437.0.  Looks like I'll have to add some of the LPC models to my shopping list...

sodacose said:

Have you tested these as supplies for audio applications or are there any projects/experiments using them that you can point me to?

Click to expand...

I don't do tubes if that's what you're interested in but if I did there's no reason not to just use the filter network for whatever the tube circuit is which is almost always multiple RC cutuffs below 1Hz. But I would still put common mode chokes close to the SMPS to block electromagnetic radiation from broadcasting inside the enclosure (or just make supply external).

One thing to experiment with is minimizing SMPS noise injected through the heaters. Traditionally heaters have been powered with AC and then heater wires were carefully balanced and physically twisted together to simply cancel mains hum. But many tube heaters can run on 12VDC. In this case, I would just try a 12V SMPS straight into the heaters (no reason not to keep wires balanced) and study that as a baseline. My guess is it would be fine.

Although I don't think I would use an SMPS with a tube power amplifier though. SMPS aren't designed for that.