Class D amp long term reliability

[dlaloum]

No doubt. However there is a big change from historical class AB amps which is that pretty much anyone could design and build a class AB amp with a linear PS so there were a huge variety of different designs, some of which failed quickly and some which last for decades, and everything in between. With class D amps, because they are much harder to design and build, there are only a handful of manufacturers making modules that go into every brand of amplifier. If one of these manufacturers makes a mistake it is going to show up in dozens if not hundreds of different brands. With modern design and quality control these generic modules are going to be designed to last for the warranty period but not much longer. This means high end brands don't really have the same control over their quality as they did in the past.

Most of these modules use a baseplate for thermal dissipation, and most manufacturers merely bolt it onto the base of their case and call it a day.

Bolting the modules to true heatsinks, mounted vertically, would potentially reduce the operating temperature of the modules, and increase their operating life substantially....

 

[dlaloum]

I think the modules are probably designed to last ant least until the warranty period is over in the worst case conditions and usage. Which means they will last much longer for most people.

The modules seem adequately engineered - what could be improved is how they are mounted by manufacturers that use them, and whether heatsinking is also used...

 

[TheWalkman]

This five year old thread popped up in my notifications today, and I’m happy to report that virtually all of my Class D amps - Aiymas, Fosi, Purifi (and Topping, more or less we all know that story) are still going strong five years later and with multiple hours of play per day.

Audio life these days is good, especially with Class D amps.

 

[Roland68]

This five year old thread popped up in my notifications today, and I’m happy to report that virtually all of my Class D amps - Aiymas, Fosi, Purifi (and Topping, more or less we all know that story) are still going strong five years later and with multiple hours of play per day.

Audio life these days is good, especially with Class D amps.

In the case of the Topping PA5, however, this had nothing to do with Class D amplifiers, but rather with the improper potting of small-signal circuits to conceal them. This would have led to failure in DACs, preamplifiers, headphone amplifiers, etc. as well.

 

[Bruce Morgen]

This five year old thread popped up in my notifications today, and I’m happy to report that virtually all of my Class D amps - Aiymas, Fosi, Purifi (and Topping, more or less we all know that story) are still going strong five years later and with multiple hours of play per day.

Audio life these days is good, especially with Class D amps.

Yup. My first pair of Aiyima A07s arrived in June of 2020, and have been running pretty much 24/7/365 on 48 VDC ever since with zero issues. My Kali IN series monitors with their multiple channels of Class D amplification have also been completely trouble free for many years. The only Class D failure I've experienced was the "Sledge" plate amp in my SVS sub -- that was under warranty, and SVS sent a replacement amp and a return label for the dud pretty much immediately. Interestingly, that sub was the only Class D based product in my system that relied on external MOSFETs. I wonder if there are any stats indicating that "chip amp" Class D -- my monitors also rely on TI (Class D plus on-die DSP) silicon, so IMO they're in that category -- is more reliable than solutions involving discrete output devices.

 

[dlaloum]

I will note that my (class D) Crown XLS2500's have been running 24/7 for about 15 years now... with no issues at all.

It all comes down to the engineering of the individual manufacturer / design

The Crowns have fans for cooling if/when under stress - the fans do not seem to run at all when running normally (for my setup) and as 440W@8ohm amps, running at around 1W most of the time - they aren't stressed!!

 

[EJ3]

Choose wisely: This was my choice in the late 1980's and I am still using them.
NAD 2200 Power Amplifier PE (an AB amp)
Date of manufacture : Jan 87 - Feb 89
(I own 5 of these & one earlier non PE [Power Envelope or for those down under: Power Tracking]). I typically run one for my pair of subs in a 2 ohm configuration. And one on each channel in a bridged mono 4 ohm configuration. Yes, that takes up 3. Sometimes I also run one per sub in a bridged mono 4 ohm configuration. The spare amp rotates in & out of the system or does duty on other projects. All are subject to being replaced by the spare for various periods of time.
These have traveled the world with me (USA, OCEANA, Islands in the Indian Ocean, Saipan, Guam and many other places and operated under many different conditions). In 2016 I decided to have a look see, figuring that they were long in the tooth. Many things were done to bring them to the best specs possible (since I had commissioned to have them gone through) without outright changing any of the circuitry. The results can be seen in Amir's NAD 2200 Review. The only outright failure that I ever had was a speaker relay (from wear). There are the questions of will parts for whatever amplifier that you choose be available in 10 or 15 years? At what cost? And will there be technicians that can repair it?
IMHO: It is worth it to spend more money up front to have better quality/reliability of the parts than to be down & waiting to upgrade the lesser parts in an inexpensive amp. As to the box it is in, if you are not going to use it as a seat, it doesn't need to be billet.
I don't own a class D yet, as I am concerned about reliability.

I'm replying to myself here, as I am hoping that this reaches all who have looked at my post.
I finally had one blow up (one of my 3 sacrificed itself to apparently save everything else in the house). There was a big BOOM and seemingly simultaneously: a big FLASH of light in my living room. I was 3 steps down & 60 ft away in the den when (for an instant) there were reflections of light & shadows throughout the downstairs area INSIDE my home.
Then came the smell of burned electrical stuff. Upon investigation, the burned electrical smell was coming from my stereo area (which was not on at the time). And the boom came from the transformer on the telephone pole about 40 ft. from the house.
Once the repair crews came & we had power again, one of my stereo system (My system is Tri Amped) amps did not respond to being turned on.
Here is why (failure due to current surge, I imagine [it's a good thing that I have other NAD 2200's, one of which immediately took this ones place]).
I'll be sending out the damaged unit for an autopsy & possible repair (in a month or so). This means that in 37 years 1 failed dramatically but my other 5 (and the 2 lesser powerful NAD 2100's) lasted at least 37 years. (Yes, I had Quirk Audio go through them in the last 10 years but that was preventative maint. (not due to failure).

 

[JSmith]

lasted at least 37 years

Not long enough, warranty claim for sure.


JSmith

 

[EJ3]

Not long enough, warranty claim for sure.


JSmith

Damn, now I have to dig up that paperwork!

 

[Mike123]

Switching power supplies have been widely used in household appliances for over 50 years. The operating conditions of the power transistors, chokes, and capacitors in SMPS are much more severe than those of most Class D amplifiers (we ignore exceptions such as those operating 24 hours a day at 1200 watts ). TV with SMPS typically last 10-20 years. After routine maintenance to replace worn-out electrolytic capacitors, they continue to operate. This is a significant and objective statistic. I see no reason why properly designed switching amplifiers should necessarily have a short lifespan. In most cases, a lifetime of no more than 5-7 years will occur if the manufacturer intentionally designed the product to have a short lifetime.

 

[Bwaslo]

PS - my first class D amp is long out of warranty (purchased 2018) and working fine.

I've been running a pair of ICEpower amp modules I got in 2008 since then, powered up most of that time. Some Hypex UCD amps about the same time period as well. I had AB amps lose their supply capacity in less time. No longevity issues with class D if designed right.

 

[TheWalkman]

In the case of the Topping PA5, however, this had nothing to do with Class D amplifiers, but rather with the improper potting of small-signal circuits to conceal them. This would have led to failure in DACs, preamplifiers, headphone amplifiers, etc. as well.

To the contrary, considering the potted op-amp is in a PA 5 Class D amp and failure of said op amp causes the PA5 to fail, I would argue that it has everything to do with the topic.

By your logic, if the fuel injection system fails on a BMW, it has nothing to do with the automobile. If the car won’t run, the failure is a reflection on the car and how it’s perceived reliability isn’t it?

An amp is only as good as its weakest component and in my now rather extensive collection of Class D amps, the PA 5 is the only amp I’ve had any issues with. The performance of the other amps has been flawless.

I rest my case.

 

[Sokel]

TV with SMPS typically last 10-20 years.

I once gutted an Acer display which couldn't keep up at start-up and of course the know suspects were the guilty ones, some hideous Taikon and Capxon capacitors swollen like they had to many big Macs

To my surprise though it also contained some REALLY nice Okaya's in there (now obsolete) for mains noise suppression.

 

[Roland68]

To the contrary, considering the potted op-amp is in a PA 5 Class D amp and failure of said op amp causes the PA5 to fail, I would argue that it has everything to do with the topic.

By your logic, if the fuel injection system fails on a BMW, it has nothing to do with the automobile. If the car won’t run, the failure is a reflection on the car and how it’s perceived reliability isn’t it?

An amp is only as good as its weakest component and in my now rather extensive collection of Class D amps, the PA 5 is the only amp I’ve had any issues with. The performance of the other amps has been flawless.

I rest my case.

The potting compound had no technical purpose; it served solely to conceal Topping's circuitry. Some of these amplifiers failed (by no means all) due to leakage currents and short circuits under the potting compound.
This definitely wouldn't have happened without the potting compound.

After removing the potting compound and cleaning the circuit board, these PA5s function perfectly again, without any component replacements. I've done this with over 15 PA5s.

This thread is about the general durability of Class D amplifiers in relation to Class D technology.
This failure caused by the potting compound has absolutely nothing to do with Class D technology and would have led to the same failures in DACs, preamplifiers, headphone amplifiers, etc. Furthermore, the potting compound was located exclusively in the analog small-signal range, which definitely doesn't belong to Class A technology.

Of course, I agree with you that if we're talking about simple statistics on Class D amplifier failures, the PA5 should be included in the analysis.
However, when assessing the long-term reliability of Class D amplifiers and Class D technologies, the failures of the PA5 have no place.

 

[TheWalkman]

The potting compound had no technical purpose; it served solely to conceal Topping's circuitry. Some of these amplifiers failed (by no means all) due to leakage currents and short circuits under the potting compound.
This definitely wouldn't have happened without the potting compound.

After removing the potting compound and cleaning the circuit board, these PA5s function perfectly again, without any component replacements. I've done this with over 15 PA5s.

This thread is about the general durability of Class D amplifiers in relation to Class D technology.
This failure caused by the potting compound has absolutely nothing to do with Class D technology and would have led to the same failures in DACs, preamplifiers, headphone amplifiers, etc. Furthermore, the potting compound was located exclusively in the analog small-signal range, which definitely doesn't belong to Class A technology.

Of course, I agree with you that if we're talking about simple statistics on Class D amplifier failures, the PA5 should be included in the analysis.
However, when assessing the long-term reliability of Class D amplifiers and Class D technologies, the failures of the PA5 have no place.

No argument that potting was do no purpose than to obscure what was underneath but it impacted the reliability of a PA5 which is a Class D amp which is the topic of this thread, right?

Reread my entire post with an open mind and I think you’ll see I am absolutely correct in my position and comment.

 

[Roland68]

No argument that potting was do no purpose than to obscure what was underneath but it impacted the reliability of a PA5 which is a Class D amp which is the topic of this thread, right?

Reread my entire post with an open mind and I think you’ll see I am absolutely correct in my position and comment.

I think you're the one who should read up on it again.
I already agreed with you that the reliability of the PA5 was limited by this, and that it should also appear in normal, general statistics.

But the fact is that these failures have absolutely nothing to do with the reliability of Class D technology. And that's what this thread is actually about.

You're welcome to see it differently, but that doesn't change my statement.
And everyone can form their own opinion on that.

 

[EJ3]

I think you're the one who should read up on it again.
I already agreed with you that the reliability of the PA5 was limited by this, and that it should also appear in normal, general statistics.

But the fact is that these failures have absolutely nothing to do with the reliability of Class D technology. And that's what this thread is actually about.

You're welcome to see it differently, but that doesn't change my statement.
And everyone can form their own opinion on that.

For the general public that makes the purchase, it appears to be the reliability of class D is in question. (so, in effect, it is).
Few would take it apart and understand that that the PA5 was killed by the use of potting material and would be fine without it.
What they do understand is that what they bought failed.
And that what they bought was CLASS D.
Now, that is what the manufacturers have to be on point about. Not creating failures by bad implementations.
Because those failures are part of CLASS D's reputation, like it or not.

 

[Decisionmaker]

Having researched for weeks, I was ready to jump on a class D with the new Hypex NCx252.
However, this excellent forum has made me have doubts.
My understanding is that the og NC252 and others in the range offer exceptional quality for the price.
However, in addition to the multiple user reported problems - which some may claim are anecdotal, but offer a very unified narrative, I've read the very honest accounts from @Apollon Audio about a high rate of failure and a precise analysis from @restorer-john on the shortcomings of Hypex's design.

I've also read counter statements from smaller retailers, whose businesses perhaps relies a bit more heavily on these products...

Am I correct in understanding that the brand new NCx252 does not address any of the design concerns that have been highlighted from the previous model?
Would it be fair to say that longevity beyond 5 years of use is highly questionable?
Aside from from upgrading to more expensive modules, any other recommendations for more reliable class Damps in the £600 range?
The Chinese brands see to offer similar quality sound at an even lower price, but also questions regarding long term operation.

 

[boXem]

Am I correct in understanding that the brand new NCx252 does not address any of the design concerns that have been highlighted from the previous model?

No
- PCB flexing seemed resolved on the sample I had in hand
- out of this there was no "design concern" identified
These modules are mainly designed for plate amps applications. Slapping them at the bottom of small boxes without consideration for thermals is a receipt for problems.
Where are the reports of bad reliability for the NAD M10 that sold by thousands and use the NC252MP?

 

[wwenze]

It's 2026, low sample size, population differences and interpretation subjectivity continues to make claims anecdotal at best.

Note that a product with 1% fail rate vs a 0.1% fail rate is a big world of difference in quality (and price), but to one consumer that's 99% vs 99.9% chance of not encountering a fail. And some people think 1% fail rate is acceptable and others not, that's the subjectivity I'm talking about.

The concerns mentioned are generic to any electronics. Non-Japanese capacitors and inadequate cooling reduce lifespan. Maybe it's 10 years reduced to 5 years. Maybe it's 50 years reduced to 25 years. Nobody knows. Nobody can define what is acceptable, again due to subjectivity. We just advice doing things known to be good, while designers try to see what they can get away with.

And some $0.02:
- Computers and other electronics in general are good sources of example of cooling solutions required at different power levels
- Some things / some design decisions are done by big companies for some reason, some things are also not done by big companies for some reason. Transistors slapped against the amp chassis seem like free and effective cooling, enough to cool an NC252MP which dissipates well over 100W at max power. Yet big company like NAD, Yamaha, Marantz still use a relatively big internal heatsink to cool a 40Wpc amp, granted those are class AB but the heat output is similar so they could have used the chassis method, yet they don't. Food for thought.