Linear Power Supply for Class D amps - Issues? Idle loss? Efficiency? Benefits?

[Cbdb2]

Ive never seen an AB power amp with a regulated supply (they exist but there rate). Unregulated supplies are usually very efficient. And even more so by exchanging the diode bridge with a MosFET bridge rectifier. Its the output stage that burns up the power.
LPS vs SMPS cons: Heavy, expensive, if regulated, inefficient
pros: only a few parts so: easy to design, easy to fix, and the biggest pro, reliable.

 

[boXem]

Agree with @pma , weight is a major issue.
Another is that lower efficiency is not only an electrical consumption issue.
Modern modules control the voltage of their supply rails and shutdown when it's becoming out of range.
For an unregulated supply, lower efficiency means lower rails voltage under heavy load. Which means the need of higher voltage unloaded.
If the PS voltage working range of the module is too narrow, it just won't work.
Not just an hypothesis, I have in mind one of the latest modules on the market when writing this.

 

[mcochris]

You know guys - I prefer facts and I prefer facts over opinions. That said, below please see measurement of efficiency of my UcD180 implementation with linear power supply. Power consumed at idle is

You know guys - I prefer facts and I prefer facts over opinions. That said, below please see measurement of efficiency of my UcD180 implementation with linear power supply. Power consumed at idle is 8W.

View attachment 448724
Note: “total output power” on X-axis is in fact power delivered into 4ohm load at amplifier output.

Is this a unregulated LPS?

You know guys - I prefer facts and I prefer facts over opinions. That said, below please see measurement of efficiency of my UcD180 implementation with linear power supply. Power consumed at idle is 8W.

View attachment 448724
Note: “total output power” on X-axis is in fact power delivered into 4ohm load at amplifier output.

Is this an unregulated LPS?

 

[gberchin]

Benefit: I have resurrected a couple of vintage Class AB amplifiers with blown electronics and unobtainium replacement parts by stripping-out everything but the linear power supplies and installing TPA3255 Class D amplifier modules.

 

[pma]

Is this an unregulated LPS?

Yes. The photo is here. The transformer has quite high winding resistance, so it is "soft".

 

[pma]

Some occasional hum





View attachment 448691

That plot - is there a source, a link to the source? Or is it a marketing plot without any consequences?

This is what I measure:

 

[Cogito]

Early on Hypex supplied linear power supplies and toroidal transformers for use with its UCD amplifier modules, before moving to SMPS.

View attachment 448730

SMPS are overall cheaper to make (No large transformer and large capacitors). They are also cheaper to ship. Personally, I love moving my 74 lb power amp. It provides much needed exercise.

 

[pjug]

That plot - is there a source, a link to the source?

Yes @BeeKay where the plot in Post #4 come from?

Some info on this subject can be found here, including efficiency of transformers and linear power supplies:

Linear Power Supply Design - Part 2

ESP - The Audio Pages. Linear power supply design information - Learn how to design your own high performance amplifier power supply.

sound-au.com

 

[BeeKay]

Yes @BeeKay where the plot in Post #4 come from?

Some info on this subject can be found here, including efficiency of transformers and linear power supplies:

Linear Power Supply Design - Part 2

ESP - The Audio Pages. Linear power supply design information - Learn how to design your own high performance amplifier power supply.

sound-au.com

Was a leftover from a search on the same topic I did quite some time ago. Not sure wether can find the exact source again. Overall quite reasonable IMHO.

 

[AnalogSteph]

It certainly does not look realistic. 40% efficiency is something I'd be expecting from a cheap regulated 5 V wall-wart supply or something.

If you have a look at Talema's product catalog of toroidals, you will find this as a graph of purely transformer efficiency:

With DC voltages of 50ish volts, even the voltage drop of conventional silicon rectifiers is going to be comparatively small, although you can still employ Schottkys or active rectification if you so desire. Smoothing cap leakage current will contribute to idle losses a bit. And that's about it in terms of things that could reduce efficiency on an unregulated supply. So I would think that you could definitely get into the mid-high 80s if you so desired, decent SMPS terrain.

Idle losses in these toroidals actually aren't terrible either, improving upon conventional EI jobs by about an order of magnitude:

And all that with the robustness of a linear power supply. Seems like a no-brainer. Yes, you will have to watch out for inrush current (soft start) and keep DC away from the bigger sizes in particular.

 

[voodooless]

40% efficiency is something I'd be expecting from a cheap regulated 5 V wall-wart supply or something.

I think the keyword here is "regulated".

A transformer/capacitor PSU will obviously have a higher efficiency. As soon as you add linear regulation, the efficiency goes down the drain.

 

[pjug]

I think the keyword here is "regulated".

A transformer/capacitor PSU will obviously have a higher efficiency. As soon as you add linear regulation, the efficiency goes down the drain.

Or could be a wall wart like Analog Steph said. Who knows? There isn't any helpful info given with the curve. Here are efficiency measurements from the Elliott Sound products paper I linked a couple posts above.
---------------------------------------------------------
13 - Efficiency
Linear supplies are thought to be inefficient, but that is simply untrue. I tested two transformers (300VA toroidal and 212VA E-I types). The two test circuits used a pair of 6,800µF caps in a full-wave dual supply (nominally ±35V [toroidal] and ±40V [E-I], and I tested with 180Ω, 60Ω and 16Ω loads. It will come as no surprise that the toroidal transformer had higher efficiency, but both were far better than 'common wisdom' would have you believe. Try as I might, I was unable to find any published figures on-line for the efficiency of a simple transformer, bridge and capacitor power supply, so this is likely to be the first time you've seen these measurements.

Any time you try to find this info, you'll get countless pages talking about SMPS, but little or nothing for 'conventional' power transformer based supplies (sometimes referred to as 'heavy iron'). It's worth noting that it's actually quite difficult to get the results by simulation. Because iron losses aren't taken into account, the result will generally be optimistic. It's also rather a chore to get every parameter exactly right, and that's why I took measurements. Note that this high efficiency only applies for an unregulated supply. If the output is regulated with a linear regulator

Table 13.1 - Transformer Efficiency Measurements

300VA Toroidal
Load	DC Volts	Power Out	Power In	VA	PF	Efficiency
No Load	69.5V	0	2.1 W	3.22	0.65	0%
180 Ω	67.8 V	25.5 W	29.6 W	34.04	0.87	86.1%
60 Ω	66.0 V	72.6 W	79.6 W	93.84	0.85	92.5%
16 Ω	59.2 V	219.0 W	235.8 W	326.6 *	0.67	92.9%
212VA E-I
Load	DC Volts	Power Out	Power In	VA	PF	Efficiency
No Load	82.2 V	0	11.5 W	16.5	0.70	0%
180 Ω	78.4 V	34.1 W	42.4 W	50.6	0.84	80.4%
60 Ω	74.9 V	93.5 W	113.8 W	126.7	0.90	82.2%
16 Ω	62.8 V	246.5 W	330.0 W	365.0 *	0.90	74.6%

 

[voodooless]

Here are efficiency measurements from the Elliott Sound products paper I linked a couple posts above.

That's also all unregulated...

 

[pjug]

That's also all unregulated...

Yes it says right in the description and makes the same point you are making. I am certainly not arguing against that!