Class D amp long term reliability

[Marc v E]

How does that solve the problem of people naturally being more likely to report a problem than report not having a problem? Look at the topping PA5 thread. There are enough failure reports to indicate there is an issue, but we still have no idea what the failure rate is since anyone with a problem will goolgle for it, find that thread, and post their issue. Even then we are only able to detect a serious problem. If the failure rate was a more typical <2% we'd not have any data on it, other than one or two disgruntled members which could easily be a random distribution variation.

With 7000 people you'd have no meaningful data ever. Product sales life is generally only a few years at the volume end of the market. People are generally not changing their kit every few years, and we are a tiny part of the market. Even if they do change their kit regularly then you get no data on long term reliability.

You can't get any information on reliability without data only the manufacturer has - how many units have been sold - how many have been returned failed.

I understand your point.

What I'm suggesting is that instead of guessing and getting all emotional in a thread about 1 certain issue, we will get more dependable data if a langer number of ASR members have a central place where they can deposit their data all the time. It might be anecdotal evidence in the beginning, but better in the longer run.

It's a shame to dismiss this chance in the name of accuracy if we can get more accurate data than we have now.

 

[pma]

Also, the board is clearly under a lot of strain (it bends) due to the screwing down of the PCB to the heat-spreader near the heat producing SMPS switchers:

All, my 12 cents - my NC252MP after 1 year operation (since 12/2021). Sometimes >8 hours a day, sometimes nothing when I live with another amplifier. Some quite severe tests that the common user would not do, like this one:

Meaurements of Hypex NC252MP in bridge-tied load (BTL)

Following the discussion in the NCx500 review thread, I was curious how my PMA-NC252MP amplifier based on NC252MP module would perform in a bridged configuration. Original Hypex datasheet allows for this configuration, as we can see in the following part from the datasheet: It declares...

www.audiosciencereview.com

Despite of the bent PCB, the module has been working perfectly, so we shall see. My request for reliable operation would ask for at least 10 years of reliable service.

Edit: adding a 3D model

 

[arvidb]

I understand your point.

What I'm suggesting is that instead of guessing and getting all emotional in a thread about 1 certain issue, we will get more dependable data if a langer number of ASR members have a central place where they can deposit their data all the time. It might be anecdotal evidence in the beginning, but better in the longer run.

It's a shame to dismiss this chance in the name of accuracy if we can get more accurate data than we have now.

Respectfully, but no, you are missing the point. The point is that we wouldn't get more dependable data. We'd get a larger number of failures and non-failures, but no meaningful statistics. It would all be anecdotal. In fact it would be worse than nothing since many would think it was meaningful and draw completely invalid conclusions from it - whatever those conclusions would be.

Edit: I don't draw any conclusions from the one failure shown in this thread; some failures are to be expected from the best of electronics, so no problems that some show up. However the shoddy heatsinking solution with the bent PCB is worrying to me, because I am "pretty sure" that in combination with heat cycling that will have an adverse effect on the solder joints. Which could potentially lead to such "Chernobyl" moments as shown in the thread. I was willing to accept that perhaps Hypex sells enough boards (and thus has enough statistics) to optimize the choice of capacitors so that they are "good enough" to generally last at least 10-15 years of normal use - but seeing that bent PCB I'm afraid they simply designed the module to be as low cost as possible and not generate too many failures during the warranty period, with no thought whatsoever about longevity after that. I hope I'm wrong, for the sake of less landfill... but I'm not ready to bet my money on it.

 

[antcollinet]

.... getting all emotional...

Where is the "getting all emotional"? Why do you feel the need to throw an accusation of emotionality into the discussion? It is no different (especially on a science based forum) than bandying about the word "snowflake". It is essentially an ad hominem.

 

[sigbergaudio]

I must agree that trying to collect data on this will at best give misleading results since we would have no information about errors against total sold amps/modules. So as long as we can't get a percentage of failures, the data will be useless, and as others mention potentially worse than useless since they can give people the wrong impression of (lack of) reliability.

 

[Willem]

Quite simply, it would not be a representative sample because those who did not have problems did not report that. Conversely, if you had the total numbers that were manufactured you would underestimate the problem because not every problem would be reported. The only more or less reliable statistic would be the proportion of warranty claims known to the manufacturer or to a particular large dealer. But even that will only give you data for the warranty period, and not for the first ten years or whatever.

 

[pma]

@sigbergaudio , just a short question, do you use the modules mentioned here in your products? Thank you.

 

[Marc v E]

Where is the "getting all emotional"? Why do you feel the need to throw an accusation of emotionality into the discussion? It is no different (especially on a science based forum) than bandying about the word "snowflake". It is essentially an ad hominem.

It wasn't intended that way. I'm sorry if it came across as such.

What I meant was that a thread on one manufacturer or specific items tends to focus on the negatives only, without any perspective to the ratio of the devices that are functioning well.

 

[Blumlein 88]

Quite simply, it would not be a representative sample because those who did not have problems did not report that. Conversely, if you had the total numbers that were manufactured you would underestimate the problem because not every problem would be reported. The only more or less reliable statistic would be the proportion of warranty claims known to the manufacturer or to a particular large dealer. But even that will only give you data for the warranty period, and not for the first ten years or whatever.

You know the OP wanted to know about reliability for a purchase. And the bottom line is not enough info on class D to make any general recommendations. OTOH, if class D amps blow up at high rates we would have heard of it by now. Plenty of 10 year old ones still run fine. 10 or more years is something you cannot really complain about. So basically, we don't have any reason to believe class D reliability is any big issue. Sort of a waste of time pursuing detailed info about it.

 

[restorer-john]

What I meant was that a thread on one manufacturer or specific items tends to focus on the negatives only, without any perspective to the ratio of the devices that are functioning well.

A good point for sure.

But, by the same token, in this day and age where technicians/repair persons don't actually repair on a component level and cannot communicate their experiences, you have a massive hole in the information being fed back.

 

[restorer-john]

You know the OP wanted to know about reliability for a purchase. And the bottom line is not enough info on class D to make any general recommendations. OTOH, if class D amps blow up at high rates we would have heard of it by now. Plenty of 10 year old ones still run fine. 10 or more years is something you cannot really complain about. So basically, we don't have any reason to believe class D reliability is any big issue. Sort of a waste of time pursuing detailed info about it.

Nicely summarized.

 

[EJ3]

In terms of LED home bulbs, I've been using them quite a while now. With one exception the three failures I have came from power blinks, in one case the power cycled on and off 4 times in less than a minute. I suspect it is over-voltage from spikes that got them. All were under warranty and replaced without issue.

I've done tear downs of a handful belonging to other people. Found the same as you primary filter caps and inadequate heat dissipation. I think nearly all of them were people using LED bulbs in a fully enclosed lighting fixture. I took apart two that had not failed, but had started flashing. Same thing evidence of over-heating.

Nonetheless they seem to last more hours than old bulbs on average. You gain in reduced electrical cost and the LEDs are not expensive anymore for the basic bulbs.

Reduced electrical cost: having gone from 100% CFL in my home to 100% LED, I believe that the reduced heat output has had a big effect on my Air Conditioning operating less. Since my A/C/Heater circulating fan is on always (due to my wife's COPD & the extra fine filtration [captures pollen, dander, mold spores, mildew, eliminates odors, etc]) and I have 90% usage of the A/C side annually, that is a definite effect on lowering the bill.
But I still use 2 100 watt incandescent bulbs (separately controlled so that I can operate 1 or both) for my water pump house for keeping the pump warm during the occasional freeze. 3 or 4 times a year, it goes to or below freezing for 1-4 days and sometimes a bit below freezing. When the weather warrants, I go turn 1 (or if a bit colder 2) of these on to keep my pump from freezing. I don't know of any other good use for these type of bulbs other than incubation for hatching chicks. But I also have not found any low wattage way to keep the pump warm enough to not freeze.

 

[fpitas]

Anybody know what type of capacitors, resistors, diodes, cabling and wiring etc was used on Voyager 1?

Probably very expensive aeronautical and space-certified stuff. Unless you want to pay considerably more for your amp than your house, may not be a good path.

 

[sigbergaudio]

@sigbergaudio , just a short question, do you use the modules mentioned here in your products? Thank you.

I use the Hypex plate amplifiers which I believe are made up of modules from the NCxxxMP series which I believe has been discussed in this thread, yes.

 

[fpitas]

I don't think anyone has addressed the elephant in this particular room; namely, we don't have any reliability data to compare garden variety class AB amps. They don't pay top dollar for electrolytics, either.

 

[fpitas]

Also, a dirty little secret I've been told by those who should know, is even the top manufacturers like Rubycon contract capacitor manufacturing to Chinese companies. I'm sure the quality control is in place and all that, but that same Chinese outfit can turn around and sell a completely similar cap under their own name.

 

[restorer-john]

we don't have any reliability data to compare garden variety class AB amps. They don't pay top dollar for electrolytics, either.

Well, actually they did and they do pay absolute top dollar for quality components. I can tell you what compromises were made, and name names. And also, what companies did not.

Happy to have a discussion on that in another thread if you want.

 

[fpitas]

Well, actually they did and they do pay absolute top dollar for quality components. I can tell you what compromises were made, and name names. And also, what companies did not.

Happy to have a discussion on that in another thread if you want.

No, thanks. I'll take your word. I really should get work done, as they do pay me for it.

 

[restorer-john]

Also, a dirty little secret I've been told by those who should know, is even the top manufacturers like Rubycon contract capacitor manufacturing to Chinese companies.

Rubycon never was tier one. They wished they were.

 

[restorer-john]

No, thanks. I'll take your word. I really should get work done, as they do pay me for it.

Absolutely. Make money while you can. I'm going to bed, it's 10:15pm here. Cheers.