Apollon Hypex NC2K Amplifier Teardown

[welsh]

I always wonder about long term "sustainibility" of your work pace, IMHO this is pretty crazy for one man. Despite every expactations, "pressure" and support from community I think you have always right to be "selfish" about your health and feel free to consider time out, be it several weeks, or months...
Now I'm ready to be stoned but I think every sane person understands. Take care.

I second that. Wishing you well.

 

[Willem]

Desoldering the caps however well done does not seem a promising strategy for greater reliability.

 

[hansik]

I haven’t said that the MP series are not recommended. 65% of our sales are amps based on ( nc500, nc1200 and nc2k). The fail rate of these modules is very very low. The highest fail rate is in MP series modules. Although I haven’t seen a failed module from the newer batches in a while now. But still it is not close to the fail rate of icepower modules. From around 200 installed 1200as2 modules 20 of them failed. The fail rate of icepower modules is higher although they use better caps.

Wow, didn't expect the icepower to fail like that, is it always the same issue? I was hoping the 1200as2 would be tested here as it's an interesting option.

 

[Apollon Audio]

Wow, didn't expect the icepower to fail like that, is it always the same issue? I was hoping the 1200as2 would be tested here as it's an interesting option.

Yes the first batches failed quite often. Although the last few batches are good for now and they also managed to reduce the hum of the transformers in latest revisions.

We can send a 1200as2 for a review no problem.

 

[LTig]

I last suffered a cap failure in 2009. Since then, I've gotten in the habit of powering my gear down and, if necessary to totally power it down, disconnecting it from mains electricity.

I trust that the Hypex caps will last a long time in domestic environments. It is difficult for me to envision an application, even within a pro environment, in which high-fidelity amplifiers would need to be powered 24/7.

Well, the real target of my question was actually an industrial application of class D power amps where they would run 24/7 over the lifetime of the instrument (15-20 years).

 

[amirm]

All the caps are glued together which makes the job even harder.

Yeh, this aspect makes replacement miserable. Usually the process is easy as you heat up the two leads as you pull up on the part. With them being glued, that won't work. So you have to work hard to get rid of the solder in the vias (holes) while taking care to not lift pads and such. And then wrestle with the glue. Having to do this for so many parts will be miserable work!

For 500 Euros, i would use the modules as is and put that money toward the eventual replacement if needed.

 

[Apollon Audio]

Yeh, this aspect makes replacement miserable. Usually the process is easy as you heat up the two leads as you pull up on the part. With them being glued, that won't work. So you have to work hard to get rid of the solder in the vias (holes) while taking care to not lift pads and such. And then wrestle with the glue. Having to do this for so many parts will be miserable work!

For 500 Euros, i would use the modules as is and put that money toward the eventual replacement if needed.

Yes indeed.

I have talked with Hypex few hours ago and unfortunately custom modules with better caps are not an option from them. So better caps are only possible with a later upgrade replacement from us.

 

[hansik]

Yes the first batches failed quite often. Although the last few batches are good for now and they also managed to reduce the hum of the transformers in latest revisions.

We can send a 1200as2 for a review no problem.

Would be great, I wonder how close they come to Ncore.

 

[Apollon Audio]

Would be great, I wonder how close they come to Ncore.

The 1200as2 module measures a bit worse than a nc500 module. Compared with nc1200, nc2k or Purifi it is not on the same level. We will send a stereo 1200as2 based amp for a review.

 

[Dialectic]

Well, the real target of my question was actually an industrial application of class D power amps where they would run 24/7 over the lifetime of the instrument (15-20 years).

Would anyone use these amps in an industrial application?

 

[John1959]

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

I think that's very much the way Hypex, as a pure technical oriented company, designs their products. No "audiophile" components, especially not just for the sake of it. I remember a discussion about their FusionAmp plate amps with, according to several people, too cheap ADC and DAC's (don't known exact which ones) but they are fully up to the task in that application. Hypex don't sell the lion share of their products to typical high end users too, not withstanding their excellent performance. That also explains a lot I think.

 

[John1959]

Would anyone use these amps in an industrial application?

That's not recommended by Hypex (read their spec sheets )

edit:

"DISCLAIMER: This subassembly is designed for use in music reproduction equipment only. No representations are made as to fitness for other uses. Except where noted otherwise any specifications given pertain to this subassembly only. Responsibility for verifying the performance, safety, reliability and compliance with legal standards of end products using this subassembly falls to the manufacturer of said end product."

 

[LTig]

Would anyone use these amps in an industrial application?

Not the completed audio amps but the class D amplifier modules and power supplies. There is a lot you can do with power amps in industrial applications, and when numbers are low an own development is too expensive.

As a real world example: modified NAD 2200 power amps were used as current amplifiers for gradient coils in analytical instruments in the 80ies. I grabbed two of those NADs out of the rubbish bin after a long work life and restored them to their original state. They worked fine for a few years, then the output relays want down. Now they are now waiting for a relay replacement (parts already here, waiting for a time slot).

 

[valerianf]

Between 1999 to 2007 there was the capacitor plague on computer mother boards because of faulty no Japanese EC capacitors. That said we are in 2020 and what is important is the whole board reliability.
We have to do with what OEM propose without any customizable option.

In my life I had the opportunity to design an industrial audio class D amp and I dream to buy a compact stereo one for my home. It is the technology evolution .
Why not an Hypex based one?
My only concern about class D amp is the high frequencies rendering.
In the past it was very bad.

 

[Billy Budapest]

We are an extremely small part of the market (and when I say we, I mean like minded individuals to ASR and companies like Apollon, March, Audiophonics, and Nord).

Add VTV to that list.

 

[John1959]

I haven’t said that the MP series are not recommended. 65% of our sales are amps based on ( nc500, nc1200 and nc2k). The fail rate of these modules is very very low. The highest fail rate is in MP series modules. Although I haven’t seen a failed module from the newer batches in a while now. But still it is not close to the fail rate of icepower modules. From around 200 installed 1200as2 modules 20 of them failed. The fail rate of icepower modules is higher although they use better caps.

That the MP series have a higher fail rate is more or less expected as they have the power supply integrated. It's a much more complex unit than the amp only units.

 

[restorer-john]

We are charging around 500 Euros to recap all modules with better caps for a stereo Hypex based amp (NC400, NC500, NC1200, NC2K). This procedure takes a lot of time and we use the best professional desoldering tools from Weller to do this job. It is very time consuming so a standard practice to all production would be impossible but I will talk with Hypex about the possibility of making custom modules for us with better caps.

That's the difference that will make your products stand out, command a higher price, last longer and perform better for longer.

 

[John1959]

That's the difference that will make your products stand out, command a higher price, last longer and perform better for longer.

I have a lot of respect for that philosophy. However, in this case I follow an other approach. I have a 8 channel amp with 4 NCxxxMP units. I only use three of them for a 3 way active system. I suspect those units don't have the longevity of a typical class AB amp with a lineair power supply. But when a unit fails, I have a "hot spare" already in place with no downtime. All those units are easy to replace by myself at low cost. Like a failed harddisk in your desktop computer.

 

[Milesian]

In paragraph 4 you describe the Hypex as ’tinny’, is that a subjective evaluation of the sound?

 

[Helicopter]

John only made the comment because he knows I used to work for Rolls Royce Aero engines. The comparison is false because the requirements are completely different. Its absolutely essential that an engine doesn't fail for obvious reasons. They still do fail from time to time BTW regardless of all the quality control and "premium" manufacturing/components.

A domestic amplifier is a different proposition. To design an amp that's absolutely reliable for 50 years would significantly increase cost. For the vast majority of consumers its of no relevance and a cost they don't want.

That's cool you worked for RR. I have personally certified a commercial airliner engine for flight, and done the same for parts of the installation of a military aircraft engine. I understand the point of aerospace specs too.

An aircraft radio should use $3 screws, can be reasonably priced over $30k, and is fine to have a THD spec of '<1%.' I would be impressed if one could meet all the specs an aerospace radio meets and all the specs a Hypex module meets under $100k. Unless it is for a commercial aircraft for a sultan or something, it would also be silly to attempt. It would be senseless for any home use. I wasn't advocating that.

The requirements have important differences, but they are not completely different. Just because you won't use a home amp to communicate where the good guys are, and where the bad guys are, like you might with an amp on a military aircraft, doesn't mean you don't want it to perform well for a long time. I am not suggesting anyone use National Aerospace Standard components in $2k music amps; that would be impossible; I am just suggesting they avoid short-lived LED lightbulb capacitors. I don't think a Hypex module needs to last 50 years either; but an amp at this price should last 10-20.

There is actually some crossover, as I mentioned regarding vintage gear components. It's not as if the components are made from different elements; it is just unusual that you would want something so expensively evaluated for life-and-death applications in something where the stakes are pretty low. I have boxes of military vacuum tubes that I use for music amplification. These categories of components have diverged a lot over the last 70 years, but they are not absolutely distinct.

Apollon checked to see if Hypex would offer a version with good caps and Hypex refused. The Hypex boards, with their glued-on LED-lightbulb caps are a component you swap out. Given the price and performance, they're fine. I wouldn't re-cap one of these modules until after it fails, and even then, I'd rather swap it out than upgrade the caps unless there is a big price difference, or unexpectedly short life to the LED-bulb caps.

What does it cost to swap out an NC2k module anyway?