Apollon Hypex NC2K Amplifier Teardown

[March Audio]

Oh on the contrary. He not only has broad, relevant experience with it, so do I. Remember my power supply story while at Sony? We evaluated a bunch of brands and my senior design engineers who qualified them, would always, always look at the capacitor quality and how close they were to heat source. And manufacturer would always agree with us and offer higher priced units with better designs.

It is not like these caps are some unique part that only Hypex uses them. They are common components in countless other products and failures there, read on potential reliability in this application.

Can we predict that these power supplies fail in 8.25 years? No. Our data is far more general than that. We just know that sleep less easy when the part quality is not what we like to see.

No thats quite obviously wrong and you have another electronics repair engineer directly contradicting it and you.

He has narrow specific experience. He knows absolutely nothing of the reliability of the Hypex products and has not had a failed one on his workbench.

Yes they are used in different products, with different designs and different operating conditions.

How many million of these caps are in use and how many has John seen?

Again a specific simple question which I would be grateful if you could directly answer without commentary.

Out of the millions of these specific caps out there how many has John seen and replaced? Is that quantity of any statistical significance?

Yes or no?


It would also be really helpful if you could also directly answer the Yes/no question I posed earlier.

 

[March Audio]

To clarify my position, I am an external consultant at Hypex for several years and do a lot of beta testing on their products.

The problem in this discussion is mainly the reliabity of Suscon cap's, well I can't find any proof on the net they are shit. Hypex is pooping out amp modules and smps modules in such quantaties that a problem would show directly. To be frankly, if you produce more than 100.000 modules a year and 1 person can keep up the warranty claims ( and failure research) (and repair)......

Thank you for this information.

 

[stunta]

Wow! They only stand behind the product for 5 years???

And "provided that the product is still in our product range". So if there are widespread issues with a product, they can, theoretically, pull it out of their "range" and walk away. Given they have a reputation to maintain, I doubt they will actually do that though.

 

[TimF]

Sorry, but I disagree.

Given the manufacturer's rated lifetime, and the actual operating conditions of the capacitors in question, one can predict the mean lifetime of the capacitors in these Hypex modules. That's not a matter of opinion.

What is a matter of opinion is what the mean lifetime should be. @restorer-john thinks that, for a TOTL product, like the one that is the subject of this thread, the mean lifetime should be 50 years or more. @March Audio thinks it perfectly reasonable for the mean lifetime to be much shorter.

Personally, I'm sympathetic to the latter point-of-view. Whereas Class AB is pretty mature, the technology of Class D is still advancing. 10 years from now, I expect to be able to replace the Hypex amplifier that I currently own with a much superior unit for a price smaller than Apollon would charge today for upgrading the capacitors on the Hypex amplifiers they sell.

But your opinion could differ on this ...

What planet are you living on? Fifty years? Billion dollar sports stadiums get torn down in ten to twenty years. You guys must be the last last few people in the United States who expect electronics to last more than ten to fifteen years, if even half that long. Science textbooks have freshness dates in months, not years. Maybe, just maybe refrigerators and dish washers last a decade but why in hell should a piece of audio equipment last longer than the average marriage and four times longer than the average position with a corporation. Does anyone really expect to be using a modern day class d amp more than a few years? This is just stupid. You are talking about mainstream mass made commercial product. Who the hell is this John to dictate the specs for such? Get real. The whole city of Detroit was depreciated and written off. That's the world we live in. I suppose John wants to dictate standards and specs for point and shoot cameras, too. My point is that ten years or more for these products, or maybe even eight years, is good quality these days relative to consumer electronics. Besides, for many of the members of this website (as I imagine) a ten to fifteen year perspective is more than adequate given our age. It is unlikely that your survivors will share your enthusiasms and hobbies.

 

[March Audio]

Alan, the problem as I see it is the following. You are vigorously defending a position where one of your suppliers (Hypex) is using third tier components from a low cost manufacturer in a range of products touted as "State of the Art", premium, exceptional etc.

There is a fundamental discord between what they say, and what they do:

View attachment 89406

The purposes of these teardown threads is to hold manufacturers, and by extension, their suppliers, to account by exposing any, and all sins, for the readership to view and comment upon.

You cannot convince anyone here with a knowledge and experience with SMPSs or amplifiers that inexpensive, third tier capacitors, are anything other than a bad choice for long term reliability. Be that mains side rectification or secondary side filtering.

Perhaps our friends at Su'scon have improved their processes and quality. They have been in business for around 40 years and I have written to their head office in Taiwan to ascertain exactly what they have done to address their reliability issues, particularly in HV mains rectification/SMPS applications where I would see huge numbers of failures in the recent past. Their aluminium electrolytic capacitor production is now coming out of China, Dong Guan City.

Hopefully they will reply with something other than a standard response and I'll happily post their response or the lack thereof, in this thread.

Im vigorously defending the intent and spirit of this forum to present factual information to its audience.

No there isnt any discord.

The point you have spectacularly missed is that you havent demonstrated any compromise in the components and product performance/reliabilty or life.

You just keep extrapolating your narrow experience.

 

[March Audio]

Here is some actual data, it is supplied as is with no further intent just to show whats going on inside this amp.

I have a P451 on my bench that has just finished its 24 hour soak test (5 watts 4 ohms). So as thermally stable as it will get. Hypex psu and Purifi amp. Its just a convenient subject to open and see what cap temps are

Ambient temp is 21 deg C
External case temp is 27 deg C

I pulled the lid off and imediatly measured the PSU cap temps. They were all around 46 deg C

So using the basic calculation of life v temp (there are some other factors involved) the life of these caps is:

3000*2^6= 192000 hrs

Thats 22 years with it on 24/7!

Is that right? That even surprises me

Also can I point out this extract from the Rubycon (I presume everyone accepts that as a premium cap) datasheet

http://www.rubycon.co.jp/en/products/alumi/pdf/life.pdf

The equation to estimate life expectancy is based on our accelerated test data; however, the Estimated life calculated using the equation includes a margin of error. Therefore, we recommend that designers allow an ample safety factor between the result of the equation and the designed service life.

In the event that the result of the estimation exceeds 15 years, we recommend that 15 years be taken as the maximum service life.


So clearly even premium cap manufacturers dont have confidence past a 15 year life.

 

[Bruce Morgen]

And "provided that the product is still in our product range". So if there are widespread issues with a product, they can, theoretically, pull it out of their "range" and walk away. Given they have a reputation to maintain, I doubt they will actually do that though.

This points to mid-price hi-fi amplification going in the direction long ago established by the desktop PC sector: module-level replacement rather than component-level troubleshooting and repair. Only the most expensive of amps will be repaired via traditional means requiring highly-skilled techs like John and the guys I managed back in the day -- and, at the low end, sub-$100 Class D "chip amp" class products will simply be discarded and replaced if they fail. I personally have repaired a PC SMPS, but only to replace an obviously seized-up cooling fan -- for the most part, you correct that sort of failure by replacing the SMPS with an inexpensive aftermarket equivalent. I once attempted to repair the video card in my son's old PC by replacing a couple of obviously vented little electrolytics -- only to find out the caps failed because of an intermittent problem in the VLSI chip at the heart of the card that took out my replacements after only a few hours use.

 

[amirm]

You dont know that Amir, thats wild speculation.

I love to watch you in front of your doctor accusing him of wild speculation when based on his years of experience and a bit of evidence, is able to predict what is wrong with you.

So no, an informed opinion is not wild speculation. A mechanic going with you to examine a car can far better predict if it is reliable than you can without said experience. His opinion there is not a guarantee of future outcome, but an informed one you can count on. You best refuse it at your peril.

 

[amirm]

Does anyone really expect to be using a modern day class d amp more than a few years?

A pair of these amps cost US $5000. If they last 5 years, then you are renting them for $1000/year. My class D amps are 12 years old I think and retail for $50,000. If they only lasted 5 years, I would have slashed my wrist.

 

[March Audio]

I love to watch you in front of your doctor accusing him of wild speculation when based on his years of experience and a bit of evidence, is able to predict what is wrong with you.

So no, an informed opinion is not wild speculation. A mechanic going with you to examine a car can far better predict if it is reliable than you can without said experience. His opinion there is not a guarantee of future outcome, but an informed one you can count on. You best refuse it at your peril.

Thats a strawman Amir and you know it. You can do better Thats just it, Johns is absolutely not an informed opinion based on generations of world wide medical knowledge. With the greatest of respect to John its just based on the tiny number (in the scheme of things) of components that has passed across his workbench. If it were anything more he would have presented it by now.

The mechanic will expect certain faults with SPECIFIC models due to previous experience ( or data from elsewhwere ) that shows it to be likely with that SPECIFIC model. He wont go, "oh this BMW has a water pump fault, therefore all other cars will suffer this water pump fault". Im afraid that is exactly what John is doing.

I note that you wont directly answer the yes / no questions I posed. With the greatest of respect you arent going to be able to spin this one

 

[March Audio]

A pair of these amps cost US $5000. If they last 5 years, then you are renting them for $1000/year. My class D amps are 12 years old I think and retail for $50,000. If they only lasted 5 years, I would have slashed my wrist.

and this is the point yet again.............If.............

Currently there is no indication that they (hypex) wont, quite the contrary in fact.

The indications are that the premium caps in your $50k amps might only last 15 years. See above. 3 years to go mate

 

[amirm]

No thats quite obviously wrong and you have another electronics repair engineer directly contradicting it and you.

What contradiction? Don't see him posting saying what I said is wrong. You are saying that.

He has narrow specific experience. He knows absolutely nothing of the reliability of the Hypex products and has not had a failed one on his workbench.

Narrow? He has repaired hundreds of pieces of electronics as I have. That is not narrow. What have you repaired?

All else being equal, we are seeing far more failures of these caps than Japanese top brands. This is a fact known among design engineers and repair technicians.

Really, you are challenging the ABCs of what is considered quality design and what is not. And what could fail. Here is the top electronics engineer and blogger from your neck of the woods, Dave Jones, troubleshooting a KRK active speaker. Watch it for just 1 minute from where I have marked here and you see him commenting exactly as we have:

 

[March Audio]

What contradiction? Don't see him posting saying what I said is wrong. You are saying that.


Narrow? He has repaired hundreds of pieces of electronics as I have. That is not narrow. What have you repaired?

All else being equal, we are seeing far more failures of these caps than Japanese top brands. This is a fact known among design engineers and repair technicians.

Really, you are challenging the ABCs of what is considered quality design and what is not. And what could fail. Here is the top electronics engineer and blogger from your neck of the woods, Dave Jones, troubleshooting a KRK active speaker. Watch it for just 1 minute from where I have marked here and you see him commenting exactly as we have:

You are just extrapolating Amir. you cant do that, its meaningless.

I also have repaired 100s of bits of kit Again citing a particular technician here and there is not indicative of anything, particularly about the Hypex products. Evidence has to be specific and directly related to Hypex.

"All things being equal" ? Thats the point they are not equal.

Another yes no question, please answer directly.

Do all pieces of equipment have the same design and operating conditions?

Yes or No?

No I am challenging John and your erroneous assumptions and extrapolations.


BTW Did you read the extract above from Rubycon where they state life should only expected to be 15 years?

 

[amirm]

How many million of these caps are in use and how many has John seen?

How many millions of patients has your doctor seen prior to diagnosing what is wrong with you?

Again a specific simple question which I would be grateful if you could directly answer without commentary.

Out of the millions of these specific caps out there how many has John seen and replaced? Is that quantity of any statistical significance?

Yes or no?

Yes, it is a statistically significant thing if one repairs caps in 50 products and 48 of them are non-Japanese/top brands. And when that experience is not only shared by countless other designers, but is the basics of knowledge about reliability of electronics.

What you are doing here is saying experience doesn't matter. Wisdom doesn't matter. All that matters is what is known right this minute. Well, if that is the case, you don't know anything to share with us. You have no knowledge of failure mode of these products in the future or even current. You have not examined the design to know the stress factor on these parts. Really, you are just debating with us using words against our real world experience and knowledge.

 

[amirm]

"All things being equal" ? Thats the point they are not equal.

You say that based on what design analysis? As I just mentioned, you don't even know how these products are designed. Sticking a temp probe on a cap is not analysis. You are just wasting time with word arguments. As I said, Hypex can put this argument to rest by providing stress testing data. Or full characterization of the circuit and component in use. Without it, you are just sharing lay arguments with us devoid of technical merit.

 

[amirm]

Let me say that likely these are good designs and products. We have that much faith in the company. The issue is that we see cost cutting in an area that has potential for issues and wishing that this was not the case. So I hope people do buy these products and support companies like Apollon that are doing a great job of handling the parts that are under their control.

 

[March Audio]

How many millions of patients has your doctor seen prior to diagnosing what is wrong with you?

Yes, it is a statistically significant thing if one repairs caps in 50 products and 48 of them are non-Japanese/top brands. And when that experience is not only shared by countless other designers, but is the basics of knowledge about reliability of electronics.

What you are doing here is saying experience doesn't matter. Wisdom doesn't matter. All that matters is what is known right this minute. Well, if that is the case, you don't know anything to share with us. You have no knowledge of failure mode of these products in the future or even current. You have not examined the design to know the stress factor on these parts. Really, you are just debating with us using words against our real world experience and knowledge.

The doctor comparison is a straw man and irrelevant/inapplicable as explained above.

So do you have that data for Suscon caps in Hypex products?

If not you are generalising again.

No I am saying experience of other products doesn't tell you about the reliably of this product.

If all that matters is "what is known at the minute" then what do we know?

We know that there is no evidence to support the claim that these Hypex amps are suffering from any kind of premature cap failure problem. Ncore have been on the market 10 years and UCD 15 or more years.

Do you have evidence to show otherwise?

Yes or No?

What else? Well we know the data indicates a 3000 hour life for the caps at 105 deg c. In our P451 that indicates the life may be as long as 22 years of 24/7 operation. See the calculations above.

From a premium cap manufacturer, Rubycon, their own data suggest that cap life expectancy should be limited to 15 years regardless of how benign the operating conditions are.

This directly contradicts the expectations voiced here by people with "real world experience and knowledge".

 

[Jinjuku]

That aside, I didn't ask you to provide extended warranty. I asked that products be built with better material

Are you asking this to the exclusion of increased price?

 

[March Audio]

Let me say that likely these are good designs and products. We have that much faith in the company. The issue is that we see cost cutting in an area that has potential for issues and wishing that this was not the case. So I hope people do buy these products and support companies like Apollon that are doing a great job of handling the parts that are under their control.

Is it cost cutting with implied reliability issues, or is the cap perfectly fit for purpose? Ie damn good engineering practice to not over specify and increase cost where it is not necessary?

Would you say a life of 22 years (15 according to Rubycon) in any way unreasonable?

Yes or No?

 

[GDK]

So do you have that data for Suscon caps in Hypex.

You are generalising again.

No I am saying experience of other products doeant tell you about the reliabitly of this product.

If all that matters is "what is known at the minute" then what do we know?

We know that there is no evidence to support the claim that these Hypex amps are suffering from any kind of premature cap failure problem.

Do you have evdience to show otherwise?

Yes or No?

I have no doubt that you are sincere in your views here, and I also don’t doubt that in the vast majority of use cases, there will never be an issue with the chosen components.

At the same time, rightly or wrongly, you will be perceived as having a vested interest. I don’t have an issue with that - people can make up their own minds on whether that is relevant.

However, you are also taking up all the oxygen on your side of the discussion and preventing others, who will not have any perceived conflict, from expressing their opinions on the point. In that way, you are likely doing more harm than good. I would advise that you take the same position as Apollon themselves and let people decide for themselves whether this is an issue. I am not convinced that most people are expecting 20+ years of service from their electronics anymore (again, not to say that these amps cannot deliver that).

JMO