Class D amp long term reliability

[dlaloum]

Without schematics, self repair is pretty challenging. Even then, access to some key connections on these boards is tight. Notably ones right around some of the power transistors.

When I cannot fix readily, I have a local shop that repairs.

Do any repair shops actually do component level repairs on these kinds of highly packed SMD boards?
My impression is that repairs are always complete board replacement and that component level repairs are either not even considered, or financially non-viable...

 

[PristineSound]

Count me as one of the unlucky ones for my Buckeye Amp 4-channel NC252MP. Purchased March 2021, plugged into SurgeX SX-AX20E. It’s used for ATMOS with ATI AT4000 as main amplifier. Recently it randomly powers off after being on for 15-30 minutes. I’m surprised as I don’t use often, 3-4 hours per month.

Sounds like stage one of failure as described by Tibor:

The random power cycling you’re experiencing is typically the first symptom of a failing NC252MP module. In our experience, the next stage is that the unit may eventually stop powering on altogether. We’ve seen this exact failure pattern before, starting with the same power-cycling behavior.

You all see, Tibor was only spitting knowledge from experience, and some here were hating on him and trying to shoot the messenger.

 

[Buckeye Amps]

Sounds like stage one of failure as described by Tibor:


You all see, Tibor was only spitting knowledge from experience, and some here were hating on him and trying to shoot the messenger.

Who was hating on him or refuting the behavior of how these fail?

The only disagreement I remember was on observed failure rates by different companies...

 

[dlaloum]

I wonder if, for ecological reasons, a regulation might solve the problems, requiring repair by the manufacturer (or exchange for another refurbished part) for max. 50% cost of a new one, for 10 years? And publication of repair manuals etc. after a defined period of time. Maybe we would have fewer designs, but better longevity (IMHO).

I could see EU going yes, and USA protesting vociferously!

 

[PristineSound]

Who was hating on him or refuting the behavior of how these fail?

The only disagreement I remember was on observed failure rates by different companies...

This is an old thread so I had to review what you said, which is disagreement on failure rate from your own experience compared to his experience, whereas others have been questioning Tibor's reporting, even grilling him.

In any case, for Tibor to be able to clearly articulate exactly how the failure patten looks like, clearly he's came across enough failures of these MP units to know the pattern so precisely. So yes, I completely believe his report on failure rates.

 

[Buckeye Amps]

This is an old thread so I had to review what you said, which is disagreement on failure rate from your own experience compared to his experience, whereas others have been questioning Tibor's reporting, even grilling him.

In any case, for Tibor to be able to clearly articulate exactly how the failure patten looks like, clearly he's came across enough failures of these MP units to know the pattern so precisely. So yes, I completely believe his report on failure rates.

I must remember wrong because I don't remember anyone saying he was wrong or such. I think people were suggesting maybe the higher rates had more to do with usage of early revisions of the module vs. later revisions, which would explain the discrepancy between what he and I have seen in terms of failure rates.

 

[PristineSound]

I must remember wrong because I don't remember anyone saying he was wrong or such. I think people were suggesting maybe the higher rates had more to do with usage of early revisions of the module vs. later revisions, which would explain the discrepancy between what he and I have seen in terms of failure rates.

This hobby is a hobby of pride, people have a pride of ownership. When I looked through this thread again before, I see some MP owners coming to the defense of their MP amps, some are questioning Tibor's data, others grilling him on design. That's when I was like why are people attacking Tibor, he's just a messenger.

I just did a scan again, Tibor said the new revisions still have a 3 to 4 time higher failure rate than the upper model Hypex and Purifi.

That said, even with these improvements, the failure rate is still approximately three to four times higher compared to Hypex’s high-end modules like NC500, NCx, NC1200, or Purifi platforms.

Bottomline, you as a manufacturer don't have this issue, so just state your stats and keep doing what you're doing, such as what you are doing now, going above and beyond to offer warranty even after warranty period is over.

I personally have own two Hypex amps (NC and NCx) and I am very happy with them. I was considering the MP for a multiple-channel setup but no longer, partly because it's not as cost effective as I originally thought as there are some incredible AB amps out there and yes, I'll be honest, these reliability reports are raising concerns for me.

 

[JSmith]

FWIR many were from high power subwoofer duties? Ah, here it is;

“We’ve noticed the higher failure rates in setups where the modules are pushed harder, such as subwoofer applications or other high-demand, high-output environments. Even though we mostly build custom fan-cooled amplifier solutions for pro applications, which are thermally optimized, the MP modules still showed noticeably more failures than NC500, NC1200, NC2K, NCx, or Purifi platforms.”

JSmith

 

[Sokel]

FWIR many were from high power subwoofer duties? Ah, here it is;



JSmith

To be honest here, naming does not help.
People see this 250 or 500 (in case of hypex) or1200 (in case of icepower) and presume a boatload of W infinitely available.
Nope.

 

[sigbergaudio]

Do any repair shops actually do component level repairs on these kinds of highly packed SMD boards?

Yes. We've had I think three cases of Hypex amps dying from power surges (thunder storm in one occasion), two was repaired locally here in Norway, one by Hypex since the customer was actually in the Netherlands.

In all three cases I've provided the customer with a new Amp so they don't have to wait, and used the repaired as a spare or in demo products.

 

[charlielaub]

My 2 cents, and FWIW, I own two 4-channel NC502MP based amps that I purchased from Buckeye about 3-4 years ago IIRC. I have not had a problem with them, however, when I first started using them I noticed that the bottom chassis panel seemed to get a bit hot for my liking, when the amp was placed on a shelf but otherwise open to the air (like on top of a cabinet, not inside). This was even the case when there was no audio playing and the amp was just in the "on" state. I believe it is this bottom panel that the amp modules are bonded to for heat sinking. My solution was to buy a wire baking rack and put the amps on it instead of a shelf. This was simply a cheap solution but it totally worked and the amp was definitely cooler since air could circulate around the bottom of the amp chassis and carry the heat away. If you need to put the amp on a shelf, a small fan moving air across the BOTTOM of the amp is likely a similarly good solution, as long as the fan is pushing cool air and the hot exhaust air can flow out into the room. I would not put these amps inside a cabinet, even if the front is open. There is just not all that much air circulation happening...

 

[Mort]

My answer.

 

[Sokel]

My answer.
View attachment 502797

Even the most silent fan is orders more noisy than what we measure here, they can even interfere electrically.

You could easily get away from them with a couple of heatshinks at the right size.
They are even cheaper than nice fans, stick them or better bolt them to the chassis and the cooling will be way better and without sudden changes in temp which are often worst then heat itself.

I know that the argue is that nice fans are inaudible, and they probably are at some distance but nothing beats total silence.

 

[Mort]

Even the most silent fan is orders more noisy than what we measure here, they can even interfere electrically.

Ha ha ha. You are nothing if not consistent. No Amp is ever good enough!

 

[Sokel]

Ha ha ha. You are nothing if not consistent. No Amp is ever good enough!

They are good enough!
My point is not making them worst!

The whole ASR narrative is about this damn noise, distortion is hard to hear any way.

 

[Mort]

They are good enough!
My point is not making them worst!

The whole ASR narrative is about this damn noise, distortion is hard to hear any way.

I can't hear anything below the 90db music they are amplifying.

 

[Sokel]

I can't hear anything below the 90db music they are amplifying.

I 100% agree but I'll give you my own strange example:

I have a fairly silent PC, right? Silent fans and the works.

If I get into my room and stand 10 meters away from it I can never bet if its On or Off.
But I can immediately tell you if it's On and I shut it down, even at that distance.

Hard to describe but you get what I say (I hope) .

 

[JRowland]

Even the most silent fan is orders more noisy than what we measure here, they can even interfere electrically.

You could easily get away from them with a couple of heatshinks at the right size.
They are even cheaper than nice fans, stick them or better bolt them to the chassis and the cooling will be way better and without sudden changes in temp which are often worst then heat itself.

I know that the argue is that nice fans are inaudible, and they probably are at some distance but nothing beats total silence.

Adding a .5mm thermal pad between the heatsink and the case helps too.

 

[PristineSound]

Even the most silent fan is orders more noisy than what we measure here, they can even interfere electrically.

You could easily get away from them with a couple of heatshinks at the right size.
They are even cheaper than nice fans, stick them or better bolt them to the chassis and the cooling will be way better and without sudden changes in temp which are often worst then heat itself.

I know that the argue is that nice fans are inaudible, and they probably are at some distance but nothing beats total silence.

And they suck in a lot of dust.

Not a fan of fans, pun intended.

 

[Scrivs]

Bought myself a pair of FA503 (Hypex Fusion plate amps) for Christmas to replace two failed 2019 vintage NC252MPs.
I went with the FA503s because of a 25% off sale at Hypex, and because the more powerful version of the MP Fusion series incorporates a proper cooling solution (a fan configured for proper evacuation of heat from the platen of the module). Hypex is pretty clear about cooling in the range’s datasheets, and I’m confident that this factory engineered cooling solution will out outperform the Hypex-Ghent-in-a-box solutions
FWIW, one of NC252MPs that failed did so due to the thin PCBs warping over time, resulting in a driver transistor departing the board, which somehow caused a small cascade of failures around said component. The other one, just gave up the ghost- one relay energizes, but not the second (PS is ok). No visible damage (that I can see with a magnifier). These are horrible to troubleshoot and repair unless you’re properly set up for it. The second board also had a warped PCB. The pair of them lasted about 2.5 years or so.
That being said, I do see quite a few UCD boards with separate Hypex PS in PA equipment, and they seem to hold up pretty well (and are cheap to replace). They seem to do ok in NAD stuff too, but it’s almost always an NC, or UCD module with separate PS. Haven’t come across any failed NC500 This maybe because of the reduced mechanical stress on those much smaller boards. It sucks that Hypex OEM boards are unavailable for sale to the general diy crowd, but I 100% understand why.

I also own several ICE boards, everything from 50ASX2, 200ASC, 200AS2 (mains powered TPA3255), to a pair of 300A2s, and a 1200AS2. The oldest board is 12 years old. Failure rate for my ICEpower boards is 0 out of 12. They’re used in my Bass guitar amps, hi-fi duty, as gifts for family, etc. There’s a reason why a lot of high-volume,high-power Bass guitar amps use the ICE stuff. Never seen an Hypex MP product in any of the musical instrument amps I’ve repaired over the years, Fender knows better

While Hypex’s standalone PS are excellent, I feel ICEpower does a better job with the integrated power supply/amp modules. Thicker PCBs (my plate mounted models are all “perfectly” flat still, and their BOM throughout the range is vastly superior than the NCxxxMP series, although IME, the Hypex components are perfectly fit for purpose). Only replaced one or two low value Su’scons over the years.

Hopefully the new NCXxxxMP series have thicker, more robust PCBs.