Apollon Hypex NC2K Amplifier Teardown

[andreasmaaan]

Oh I just referenced NL because of Bruno/Hypex. Very interesting note about German custom--I'm now quite curious to ask my German colleagues about it!

Would be interested to hear what they say My info comes from a (non-audio industry) patent law firm I do some freelancing for.

 

[Deleted member 65]

And using better quality capacitors...

This picture on their website shows Panasonic capacitors. It may be a very early version. I'm pretty sure the unit I saw I had Rubycon on the amp board.

https://purifi-audio.com/eigentakt/

View attachment 88969

The SMPS had rubbish caps on it however.

Not started my Purifi assembly yet, weather is still good enough here for mountainbiking ... ;-)

Purifi module:

Neurochrome buffer:

Hypex SMPS:

 

[Francis Vaughan]

The usual issue with patents is that only the inventor can actually patent. The employer can't patent something an employee did without the employee. That is why you always see the inventor listed. But at the moment of a patenting, the patent ownership can be assigned. And as noted, that usually goes to the employer. It would be a very odd company that allowed an employee the own a patent on IP that the company paid that person to research.
Famously, Kary Mullis invented the polymerase chain reaction. One of the most significant advances in biochemistry. His employer gave him a $10,000 bonus. They went on to make billions. Kary got the Nobel prize. But he never owned the patent.

 

[andreasmaaan]

Ok, after some further research, I've clarified the situation in Germany somewhat. Legal title to the invention generally vests in the employer, but a statutory scheme (aims to) ensure that the inventor also receives a share of the profits flowing from the invention.

There are some interesting factors that are considered in determining the level of remuneration:

1. Was the inventor solving a specific task assigned by the employer or did they commit to solving the problem on their own initiative?
2. To what extent did the inventor fall back on the employer's operational resources and know-how?
3. What was the position of the inventor in the company? (A younger engineer in a position in which they would not be expected to come up with patentable inventions will be entitled to more.)

Based on this model, typical entitlements of an inventor range between 8 and 21%, according to this article (in German).

Not sure if it's the same in NL, although I'd expect it to be similar.

 

[Karbunkle]

Is it just me, or do the connections and connector from amp board to speaker terminals look a bit under-sized for the average audiophile?

 

[Willem]

ot sure if it's the same in NL, although I'd expect it to be similar.

From the little I know, the Dutch situation is similar. It is also in the interest of the employer: incentives are a real incentive.

 

[Willem]

Anyway, to return to this amplifier: I think it and the Hypex modules on which it is based are a spectacular achievement. The technology obviously owes a lot to the massive research investment by Philips in the preceeding years, and the final product here is a satisfying testimony to the success of people in former Yugoslavia to build a new life for themselves.

 

[andreasmaaan]

It is also in the interest of the employer: incentives are a real incentive.

+1

 

[Apollon Audio]

Is it just me, or do the connections and connector from amp board to speaker terminals look a bit under-sized for the average audiophile?

Why would these connectors be under-sized? Every pin of this connector is rated @ 20A/600V AC. We use 4 pins from the module connected with 4x12 AWG pure occ copper wire to the speaker terminals.

 

[Universal Cereal Bus]

Ok, after some further research, I've clarified the situation in Germany somewhat. Legal title to the invention generally vests in the employer, but a statutory scheme (aims to) ensure that the inventor also receives a share of the profits flowing from the invention.

Ah that makes a lot more sense.

The employer can't patent something an employee did without the employee.

Minor technicality: though the correct inventors need to be named on the patent application, they need not be involved in the patenting, so to speak. In the USPTO, an applicant can declare an inventor is uncooperative and consequently the applicant need not provide that inventor's signature on various forms. This can happen when an inventor-employee leaves the company and becomes unreachable.

 

[FrankF]

This guy, John Curl, did a video back in June on what makes a good amp. His main point is it's all about quality capacitors. I know nothing about them since it is not in my capacity to understand how they work::

 

[Koeitje]

This guy, John Curl, did a video back in June on what makes a good amp. His main point is it's all about quality capacitors. I know nothing about them since it is not in my capacity to understand how they work::

I doubt they have an effect on sound quality. It mostly matters for life span.

 

[andreasmaaan]

I know nothing about them since it is not in my capacity to understand how they work::

Pun intended?

 

[oldsysop]

I don't like class D amps, but if I had to buy one by reading this thread I know which one I would buy and which one I wouldn't.

 

[Lorenzo74]

I tested some caps in a piece of hifi equipment from the mid 70s just yesterday, the caps were all nippon chemicon ... and found that all the caps measured like new, seriously, . Didn't find a single bad cap, ..., no corrosion visible.

Yokoso Japan

 

[Lorenzo74]

Could be a simple engineering decision. You don't put over specified and higher cost components in if the cheaper ones are perfectly adequate.

“Perfectly Adequate” from amp manufacturer!

I tried assessing the same, few pages ago but many listen with eyes...

Lucky who cannot withstand 5000h lifetime amplifier!
I just works 10 h/day, spend weekends with my family (young lovely daughter), and listen to music with wireless headphone before got asleep.
even if I can show 25-20kHz +-1db across all sofa seats, >105db @ less 1%THD
Time aligned with Dirac
Subs + floorstands
Minidsp
Blablabla
Take care-stay healthy
Best
Lorenzo

 

[restorer-john]

Not started my Purifi assembly yet, weather is still good enough here for mountainbiking ... ;-)

Purifi module:
View attachment 89012

Neurochrome buffer:

View attachment 89013


Hypex SMPS:

View attachment 89014

Thanks for those closeups, they identify the grade of Su'scon Hypex are using and the series of Rubycons Purifi are using.

The Su'scons are their HZ series. One step up from the bottom grade 105 degree caps. They are only a rated 3000hr cap. These are primary side (50/60Hz) filter caps for the bridge rectifier in the SMPS.

As discussed back in the thread, it is the internal I2R pwer losses (ripple current) that heat the capacitor internally and play a very significant role in premature failure. Su'son advise that a 5 degree internal temperature rise will halve their life. Coupled with localised heat sources from the heatsink and transformer in a compact form factor package, their life will be compromised.

The Rubycons in the Purifi are onboard local +/-V rail filtering/decoupling for the main rails. They are Rubycon ZL series. As the rated load life is a function of the ability to get rid of internal heat, the bigger the capacitor, the longer they last.

The ZL series have a load life of between 1000~5000hrs depending on package diameter. The 100V 100uF used on the Purifi board appear to be a 10x23mm package.

So, their rated load life is 4000hrs.

 

[DDF]

So just to be clear was the hypex failure the caps or something else?

Their remote diagnosis was cap, based on the symptoms. They seemed quite confident, given the symptom, and it made sense to me (my background is EE including analog and mixed signal electronics design). As I mentioned, they also zeroed in on the caps and had changed them in production, so there must have been an issue there

 

[March Audio]

Thanks for those closeups, they identify the grade of Su'scon Hypex are using and the series of Rubycons Purifi are using.

The Su'scons are their HZ series. One step up from the bottom grade 105 degree caps. They are only a rated 3000hr cap. These are primary side (50/60Hz) filter caps for the bridge rectifier in the SMPS.

View attachment 89090

As discussed back in the thread, it is the internal I2R pwer losses (ripple current) that heat the capacitor internally and play a very significant role in premature failure. Su'son advise that a 5 degree internal temperature rise will halve their life. Coupled with localised heat sources from the heatsink and transformer in a compact form factor package, their life will be compromised.
View attachment 89093

The Rubycons in the Purifi are onboard local +/-V rail filtering/decoupling for the main rails. They are Rubycon ZL series. As the rated load life is a function of the ability to get rid of internal heat, the bigger the capacitor, the longer they last.

The ZL series have a load life of between 1000~5000hrs depending on package diameter. The 100V 100uF used on the Purifi board appear to be a 10x23mm package.

View attachment 89096

So, their rated load life is 4000hrs.

View attachment 89097

So you have just demonstrated there is no issue then John

3000 hours v 4000 hours at 105 dec C.......and before you say it you have no idea whatsoever about the internal temp.

How does the other cap derate with internal temp?

Shall I measure the actual cap temps so we can get an idea of life? After a long soak the internal external temps of the can will equalise.

 

[March Audio]

Their remote diagnosis was cap, based on the symptoms. They seemed quite confident, given the symptom, and it made sense to me (my background is EE including analog and mixed signal electronics design). As I mentioned, they also zeroed in on the caps and had changed them in production, so there must have been an issue there

Caps do prematurely fail of course. This true of any brand. I have replaced my fair share over the years including "premium" brands. Amir showed a premium cap failure in his Mark Levinson dac earlier in the thread..

The issue here is that a certain individual is trying to erroneously extrapolate this into meaning that Hypex products have a fundamental reliability/longevity problem. This is not the case.

Hypex ncore amps have been on the market for 10 years and UCD for 15 years and there is no indication that there is an issue. Apollon have seen zero cap failures an we have seen zero cap failures.