Stereophile has started calculating EPDR.

[peng]

Once upon a time, Matti Ottala presented an Audio Engineering Society paper showing that an "8 ohm" speaker could eat up to "1 ohm" of current.* Later someone published a pushback, saying that Ottala's paper was based entirely on simulation and that an empirical study had shown no evidence for current draws greater than the D.C. resistance of the voice coil would indicate.
I don't know if there was ever further study made. I wanted to do that with some kind of simultaneous current and voltage readings into a data logger but have never gotten around to it. Another way would be a long term trace on a storage scope and looks at the trace slopes. Someday...

In theory, and on transient basis (not steady state), I can see that happens, and I guess it is understandable that to make his point, Ottala had to use simulation lol.. It is naturally hard for people, including many EES to understand the negative impedance effect, or "behavior".

*note this is different than EPDR, since no dissipation is factored in, it's just equivalent resistance.

That's very much my point, unfortunately is is very evidence on ASR that Stereophile and other experts in the field might have unintentionally misled many hobbyists who do not have sufficient understanding of the topic to believe "current/amperes capability" was the issue, when it is in fact not always current, but dissipation, that is just as important to consider but then how often would hobbyist, even amp designers look at transistors SOA? It's just much easier to push the hearsay that it's not power, but those speakers with low EPDR needs high current amps (mostly true anyway I guess..).

 

[Head_Unit]

those speakers with low EPDR needs high current amps (most true anyway I guess..).

Yes...though someone (I think here at ASR) posts that is nonsense and "you just need more power" (which supplies more current of course). They're probably correct, the beefiest 100W amp probably will clip sooner than a 400W amp unless the latter is super cheap ha ha. I started saying you need to get the most 4 ohm or even 2 ohm power if rated, since speakers have gotten lower and lower in impedance per Stereophile's testing.

 

[MaxBuck]

"One ohm of current?" WTF?

 

[peng]

Yes...though someone (I think here at ASR) posts that is nonsense and "you just need more power" (which supplies more current of course). They're probably correct, the beefiest 100W amp probably will clip sooner than a 400W amp unless the latter is super cheap ha ha. I started saying you need to get the most 4 ohm or even 2 ohm power if rated, since speakers have gotten lower and lower in impedance per Stereophile's testing.

When people said not power, but current, they usually meant to say don't just look at power, but current as well because a 200 W 8 ohm amp may not have enough current to meet the demand of many speakers in some use cases. Lots of misconception on the internet/forums, but lots of time it is also just the wording, emphasis etc...

So, I am all for using voltage and current specs. If an amp is rated say, 12 A, at 50 V, continuously for not less than 1 minute, 20 A for not longer than 1 second, then it is more informative than one that is rated 600 W 8 ohms.

 

[antcollinet]

When people said not power, but current, they usually meant to say don't just look at power, but current as well because a 200 W 8 ohm amp may not have enough current to meet the demand of many speakers in some use cases. Lots of misconception on the internet/forums, but lots of time it is also just the wording, emphasis etc...

So, I am all for using voltage and current specs. If an amp is rated say, 12 A, at 50 V, continuously for not less than 1 minute, 20 A for not longer than 1 second, then it is more informative than one that is rated 600 W 8 ohms.

But there is another factor that probably causes most of the confusion.

For any given voltage/current/impedance - which results in a certain current and resulting dissipation in the amp, with most of the V*A power going into the speaker there is also the phase of the speaker response. As that phase goes from 0 degrees towards 90 degrees (90 never actually reached), more of that power - which would normally be dissipated in the speaker now needs to be dissipated in the amp.

So even if the current is not getting bigger the heat in the amp can increase dramatically.

(Note this only applies to linear amp designs - class D designs don't operate in the linear region so don't dissipate this power - instead dumping it in and out of the PSU.)

So, not only current capacity is needed, but also cooling capacity. This of course is never specified.

 

[peng]

But there is another factor that probably causes most of the confusion.

For any given voltage/current/impedance - which results in a certain current and resulting dissipation in the amp, with most of the V*A power going into the speaker there is also the phase of the speaker response. As that phase goes from 0 degrees towards 90 degrees (90 never actually reached), more of that power - which would normally be dissipated in the speaker now needs to be dissipated in the amp.

So even if the current is not getting bigger the heat in the amp can increase dramatically.

(Note this only applies to linear amp designs - class D designs don't operate in the linear region so don't dissipate this power - instead dumping it in and out of the PSU.)

So, not only current capacity is needed, but also cooling capacity. This of course is never specified.

Agreed 100%!

 

[Head_Unit]

"One ohm of current?" WTF?

It's the new math!
Answer B: it's part of the updated FTC Amplifier Rule!

 

[Head_Unit]

"One ohm of current?" WTF?

It's the "new ohms"

EPDR is always "for Class AB"

So how does it apply to Class D? Or what is the equivalent metric? How DOES phase angle affect Class D amps? Again understanding that it's a somewhat academic exercise and sometimes you just need a bigger amp.

 

[mhardy6647]

They, of course, meant one ohm of currant.
Best measured using Electric Prunes as the currant source...

 

[antcollinet]

It's the "new ohms"

EPDR is always "for Class AB"

So how does it apply to Class D? Or what is the equivalent metric? How DOES phase angle affect Class D amps? Again understanding that it's a somewhat academic exercise and sometimes you just need a bigger amp.

It doesn't effectively. In a class AB amp, the reactive "power" is burned in the amp output transistors resulting in significant extra heat. In class D - it is effectively just pumped in and out of the power supply. As long as the PSU bulk storage is sufficient it isn't a problem.

 

[DanielT]

This is what I'm curious about. A practical example with this amplifier. Under what conditions regarding impedance/EPDR will there be too tough a load, which will lead to the amplifier being destroyed or the protection mode being activated (if that function is on the amplifier)?
So: X minutes with Y power and Z impedance/EPDR= broken
If you can say something general about it? Maybe it depends on other things too?
Assuming an amp in fully working condition, no worn electronics.

Douk A5 Budget Stereo Amplifier Review

This is a review and detailed measurements of the Douk Audio A5 stereo class D amplifier. It was sent to me by the company and costs US $99.99 as of this writing on Amazon including free shipping. I really like the look of the unit. The control on the right turns the amp on and adjusts the...

www.audiosciencereview.com

A little tempted, if there's a good Black Friday deal/price I'm then I'm interested in that Douk Audio A5 I have to say.
($80 now with 32V/5A power supply)

 

[Roland68]

This is what I'm curious about. A practical example with this amplifier. Under what conditions regarding impedance/EPDR will there be too tough a load, which will lead to the amplifier being destroyed or the protection mode being activated (if that function is on the amplifier)?
So: X minutes with Y power and Z impedance/EPDR= broken
If you can say something general about it? Maybe it depends on other things too?
Assuming an amp in fully working condition, no worn electronics.
View attachment 492816View attachment 492817

Douk A5 Budget Stereo Amplifier Review

This is a review and detailed measurements of the Douk Audio A5 stereo class D amplifier. It was sent to me by the company and costs US $99.99 as of this writing on Amazon including free shipping. I really like the look of the unit. The control on the right turns the amp on and adjusts the...

www.audiosciencereview.com

A little tempted, if there's a good Black Friday deal/price I'm then I'm interested in that Douk Audio A5 I have to say.
($80 now with 32V/5A power supply)

Have you ever heard of a problem like this with Class D amplifiers?
Or any failures/shutdowns that might indicate something like this?

 

[Sokel]

Have you ever heard of a problem like this with Class D amplifiers?
Or any failures/shutdowns that might indicate something like this?

What @antcollinet describes is usually addressed with these extra capacitor banks that are around.
Other than that only at occasional reviews like Amir's , with all type of amps and their classes.

 

[DanielT]

Have you ever heard of a problem like this with Class D amplifiers?
Or any failures/shutdowns that might indicate something like this?

Well, I don't think so. Or I don't know. An amplifier can break down, but under the conditions I outlined? I'm mostly curious. Absolutely nothing I'll worry about if I were to buy a Douk A5.
Maybe it's the PSU that breaks down in the (extreme) case of class D amps? If I understand correctly what @antcollinet says in #110

Let's take a completely different type of amp and apply the same conditions. A tube amp, measured at 31 watts into 4 Ohms. When do you think it would break down?

Review and Measurements of Dynaco ST 70

This is a review and detailed measurements of the vintage Dynakit Dynaco Stereo 70. It is on kind loan from a member who bought it used online and repaired it. Previous owner hand upgraded it but could not get it working. Given that it is 60 years old, one can forgive the rust and dust...

www.audiosciencereview.com

 

[antcollinet]

Well, I don't think so. Or I don't know. An amplifier can break down, but under the conditions I outlined? I'm mostly curious. Absolutely nothing I'll worry about if I were to buy a Douk A5.
Maybe it's the PSU that breaks down in the (extreme) case of class D amps? If I understand correctly what @antcollinet says in #110

Let's take a completely different type of amp and apply the same conditions. A tube amp, measured at 31 watts into 4 Ohms. When do you think it would break down?

Review and Measurements of Dynaco ST 70

This is a review and detailed measurements of the vintage Dynakit Dynaco Stereo 70. It is on kind loan from a member who bought it used online and repaired it. Previous owner hand upgraded it but could not get it working. Given that it is 60 years old, one can forgive the rust and dust...

www.audiosciencereview.com

I don't think anyone can say without having done detailed design calculations on a specific design. And then it is only applicable to that design.

I also hope most amps will implement protection so they don't fail (thermal protection for class A or AB, over voltage for Class D) but I don't know if they all do. Perhaps for class D, it is not even needed - I don't know.

Clearly from the measurements shown above (see also the Fosi V3 Mono) the class D amps so measured operate quite happily with low impedance reactive loads. If you look at a typical TPAxxx25 amp, you see a couple (or more) of large capacitors across the chip supply - these are going to be what store, and release the energy from the circulating reactive currents. They seem to be able to do this without over-volting to the point that a problem is caused.

 

[Head_Unit]

Under what conditions regarding impedance/EPDR will there be too tough a load, which will lead to the amplifier being destroyed

Well jeez I would hope nothing gets DESTROYED, yikes! My take on EPDR is it was made as a metric to try and quantify how tough a load different speakers would be to an amp, seemingly really more about heat perhaps. I think that has been conflated into "OMG you need a 1Ω amp to drive these things!!!!" when in fact it rarely matters except when you crank it up, and if you do crank it up then as someone posted if you are running out of power you just need a truly more powerful amp.

A "100W/8Ω" amp that can drive down to 800W/1Ω is only going to clip maybe what 1-2 dB (??) higher than a weak "100W/8Ω" amp with regular speakers. Unless it is massively underrated at the 8Ω in order to give the appearance of doubling down, since I've only seen I think 3 amplifiers that truly double at their clipping points...and that is tested into resistors, not reactive loads either.

To me the significance of EPDR is more that is shows how speaker companies are lying more and more about impedance...labeling "4Ω" or even "2Ω" speakers as "8Ω" presumably so people won't be scared off buying them since their AVR says only "8Ω"

 

[Flaesh]

when you crank it up

It doesn't happen often, but it does happen.

In the near future I plan to connect a pair of these in parallel to the V3 mono.

 

[DanielT]

I don't think anyone can say without having done detailed design calculations on a specific design. And then it is only applicable to that design.

I also hope most amps will implement protection so they don't fail (thermal protection for class A or AB, over voltage for Class D) but I don't know if they all do. Perhaps for class D, it is not even needed - I don't know.

Clearly from the measurements shown above (see also the Fosi V3 Mono) the class D amps so measured operate quite happily with low impedance reactive loads. If you look at a typical TPAxxx25 amp, you see a couple (or more) of large capacitors across the chip supply - these are going to be what store, and release the energy from the circulating reactive currents. They seem to be able to do this without over-volting to the point that a problem is caused.

Well jeez I would hope nothing gets DESTROYED, yikes! My take on EPDR is it was made as a metric to try and quantify how tough a load different speakers would be to an amp, seemingly really more about heat perhaps. I think that has been conflated into "OMG you need a 1Ω amp to drive these things!!!!" when in fact it rarely matters except when you crank it up, and if you do crank it up then as someone posted if you are running out of power you just need a truly more powerful amp.

A "100W/8Ω" amp that can drive down to 800W/1Ω is only going to clip maybe what 1-2 dB (??) higher than a weak "100W/8Ω" amp with regular speakers. Unless it is massively underrated at the 8Ω in order to give the appearance of doubling down, since I've only seen I think 3 amplifiers that truly double at their clipping points...and that is tested into resistors, not reactive loads either.

To me the significance of EPDR is more that is shows how speaker companies are lying more and more about impedance...labeling "4Ω" or even "2Ω" speakers as "8Ω" presumably so people won't be scared off buying them since their AVR says only "8Ω"

I know, an extreme scenario. I'm just curious about what would happen if you took it to the extreme.

Hopefully, manufacturers perform endurance tests where they see under what conditions things break. Now, I'm not a developer, but intuitively, manufacturers should do that if they're going to install shutdown functions to protect the amplifier. See where the breaking point is, so to speak.
Fuse sizes...to figure out what needs, then check breaking point ...and so on.
Or do manufacturers just perform theoretical calculations when they install shutdown functions and/or fuses?

 

[antcollinet]

I know, an extreme scenario. I'm just curious about what would happen if you took it to the extreme.

Hopefully, manufacturers perform endurance tests where they see under what conditions things break. Now, I'm not a developer, but intuitively, manufacturers should do that if they're going to install shutdown functions to protect the amplifier. See where the breaking point is, so to speak.
Fuse sizes...to figure out what needs, then check breaking point ...and so on.
Or do manufacturers just perform theoretical calculations when they install shutdown functions and/or fuses?

When I was part of a power electronic manufacturing organisation (not audio related) - we did all sorts of tests to destruction. An incredible amount of time was taken to find ways we could break the products - and try to eliminate those mechanisms.

Circuits were initially designed theoretically (obviously) - but there was not a single circuit that wasn't tested to work in the lab also.


Any manufacturer who doesn't do this soon finds out the cost of releasing products that fail into the field. If they manage to survive the experience, they will generally learn to do better next time (example - Topping PA5)