Class D and/or SMPS reliability (MTBF)

[Ren Kitchener]

During a general class D chat with a high-end 'showroom', they stated that they have had problematic reliability issues with Hypex and the Purifi based amplifiers - so they are not really keen on stocking them for that reason.

When purchasing a $2,000+ amplifier, I can really understand that reservation - even if it does get repaired, it's still an annoyance for the customer and the stockist. When the guarantee runs out, it's more than just an annoyance - the amplifier instantly becomes an expensive 'paper-weight' - unless you are a Surgeon or an Attorney, whereon you can get it repaired or bin it and order a new one.
Getting to the point:
I remember asking Class D amplifier company 'X' for their MTBF figures - they didn't have any data. So, looking at the cards, and knowing general MTBF figures for switchers, I would be surprised if the MTBF extended beyond 20 Years:

This means that for every 20 sold, one of them will fail within the year, and that one is going to be mine - it always is.

So, those with Hypex or Purifi or other similar Class D based amplifiers (also amplifiers with switch-mode power supplies), are you seeing any failures within the year, or are they still hanging on after say 3 or 5 years?

Sometimes I think that having a nice simple 50 Watt class AB, with a toroidal transformer and a 0.02% THD spec, is going to be the best value amplifier you could ever purchase, and that 0.02% THD is not as bad as you are led to believe (compared to a short lived 0.000001% THD, 500 W blast, followed by a puff of smoke)...

 

[TankTop5]

During a general class D chat with a high-end 'showroom', they stated that they have had problematic reliability issues with Hypex and the Purifi based amplifiers - so they are not really keen on stocking them for that reason.

When purchasing a $2,000+ amplifier, I can really understand that reservation - even if it does get repaired, it's still an annoyance for the customer and the stockist. When the guarantee runs out, it's more than just an annoyance - the amplifier instantly becomes an expensive 'paper-weight' - unless you are a Surgeon or an Attorney, whereon you can get it repaired or bin it and order a new one.
Getting to the point:
I remember asking Class D amplifier company 'X' for their MTBF figures - they didn't have any data. So, looking at the cards, and knowing general MTBF figures for switchers, I would be surprised if the MTBF extended beyond 20 Years:

This means that for every 20 sold, one of them will fail within the year, and that one is going to be mine - it always is.

So, those with Hypex or Purifi or other similar Class D based amplifiers (also amplifiers with switch-mode power supplies), are you seeing any failures within the year, or are they still hanging on after say 3 or 5 years?

Sometimes I think that having a nice simple 50 Watt class AB, with a toroidal transformer and a 0.02% THD spec, is going to be the best value amplifier you could ever purchase, and that 0.02% THD is not as bad as you are led to believe (compared to a short lived 0.000001% THD, 500 W blast, followed by a puff of smoke)...

Following.... I do think class D is the future as it’s much lighter which will save in shipping and much more efficient. On to your concerns, I don’t know but I’mgoing to stick around for other’s opinions.

 

[Ren Kitchener]

I guess that there is a very small clue hidden between the lines - reading company X's returns policy, there is a 'prominent' request to remove the heat-sinking compound before returning. I can see that to clean them at the factory, would be an annoying process if there was a constant stream of returns with sticky white-grease left on them. However, there may be a lot of amplifiers sold, so you would likely see a lot of returns. On the other hand, I don't really see this request with 'analogue' amplifiers (OEM or otherwise).

Why worry? Well, I have a product X on my workbench, ready to put into an enclosure. If this 'breaks' within the year, the replacement is expensive - worst still, I may not be able to get a replacement - meaning a lot of work to shoehorn a class AB in it's place. I have to ask myself, do I ditch that and revert to a class AB (I have a few toroidal transformer paperweights laying around)?

 

[highfigh]

I doubt most failures occur when the amp is connected to a load that's within its limits. I don't remember hearing or seeing that ClassD do as well with extremely reactive loads and extremely low impedance.

 

[highfigh]

I guess that there is a very small clue hidden between the lines - reading company X's returns policy, there is a 'prominent' request to remove the heat-sinking compound before returning. I can see that to clean them at the factory, would be an annoying process if there was a constant stream of returns with sticky white-grease left on them. However, there may be a lot of amplifiers sold, so you would likely see a lot of returns. On the other hand, I don't really see this request with 'analogue' amplifiers (OEM or otherwise).

Why worry? Well, I have a product X on my workbench, ready to put into an enclosure. If this 'breaks' within the year, the replacement is expensive - worst still, I may not be able to get a replacement - meaning a lot of work to shoehorn a class AB in it's place. I have to ask myself, do I ditch that and revert to a class AB (I have a few toroidal transformer paperweights laying around)?

Why not use an outboard amplifier?

 

[Ren Kitchener]

Why not use an outboard amplifier?

I wanted an integrated amp - this could be an option @highfigh, but then I'll have a very large 15" high preamp - if you see what I mean from my 'CAD' design below:

 

[Verdinut]

Following.... I do think class D is the future as it’s much lighter which will save in shipping and much more efficient. On to your concerns, I don’t know but I’mgoing to stick around for other’s opinions.

QSC DCA series of cinema amplifiers operate in either Class AB or in 2 step-H topology. They all use SMPS power supplies. They all weigh only 21 lbs with rated power outputs up to 1100 watt/ch at 4 ohms. They are the standard product in cinemas all across North America:

DCA Series - Non-Network - Cinema - Products & Solutions - QSC

www.qsc.com

In my HT system, I am using four of them and I wouldn't use anything else. Moreover, you can't blow them.

 

[ryanosaur]

During a general class D chat with a high-end 'showroom',

There's your problem right there.
I haven't seen any reference to early failures of this gear.

and that one is going to be mine - it always is.

Self-fulfilling reality?

If you are having a custom doohickey built like that (and I did see your other post about this before you deleted it), then sure, I can appreciate your concern.
My interest is in clean, simple amplification. If I had to buy a new SMPS or NC400 module, I can easily switch them out on my own. The complexity you are looking at in your build would give me pause. I'd love to have some cool voltmeters, but whats the point?
And so I'm going for the simple external amplifier. Simple form factor, will fit in my future rack, have a good looking face plate, and not be mixed in with any other BS.
I'll let companies with a long history of quality electronics and Processors handle that aspect. Maybe it will be the ATI-built Monolith, maybe I stick with Marantz or switch to Anthem... who knows?

I do want to ask, though... do you have any links or documentation you can provide that sheds light on this questionable durability of Class D/SMPS technology?

 

[Ren Kitchener]

Self-fulfilling reality?

I do want to ask, though... do you have any links or documentation you can provide that sheds light on this questionable durability of Class D/SMPS technology?

Ah, it's an amplifier design that I want to have - that's down to my draw to the modular synthesiser (Eurorack) and studio outboard renascence - a simple aluminium box, with a single knob and a remote, is just not that appealing (to me).

I deleted the original post because I may stick with it (or not) - although the enclosure is still an expensive 'pain' - looks like a 'piano finish' takes a considerable amount spraying and drying time.

As for the links - alas, this is through a telephone conversation with a London based 'high-end' audio company - they were familiar with the Hypex modules in their brands that they sold. Normally, I would be wary of their 'reasoning' - but having lived with switch-mode power supplies, including inverters, I do know that they have a low MTBF - sometimes you are lucky, and sometimes not. My linear power supply has been 'hammered' in it's very long life, and it still works. During this time, I have gone through 2 decent (branded) switch mode bench power supplies and 3 PCs (power supply failures).

 

[ryanosaur]

Since my values are on my ears rather than my eyes, and its my intention to remove my gear from the front wall, you can understand where I'm coming from (perhaps ). Besides, I'm also known to have a relatively deep minimalist bent.
I'm pretty certain that in one of my posts to you I mentioned something similar in saying that it really comes down to your values what you decide on. If the risk of the SMPS is bothersome, then by all means, you should stick with your gut. Since none of this is really new technology, and I haven't seen much in the way of failure reports to make me nervous about it...
I guess that's why I'm slightly challenging you about it in a friendly manner. That said, and I suspect you have caught the gist of this already: I wouldn't trust what a Hi-Fi shop tells me about gear most audiophiles still demean in some manner or another. Obviously I don't know the shop of which you speak, and perhaps they are more objective than I am willing to give them credit for, but most Hi-Fi cat's here stateside have trouble getting past their embraced bias.

Regardless, I'm still in a holding pattern on my project while I wrap up my first speaker DIY modification. But I'll be forging ahead with my already in-hand Hypex gear very soon!

I do very much want to know what you choose to do, as we've engaged on this topic for a bit, now. I hope you do press forward with it!

Cheers,
R

 

[lovinthehd]

I don't think you're going to get a 5 year report on Purifi modules let alone any particular power supply used with them.....they haven't been around that long. The company was founded in what 2019? Hypex maybe?

 

[Ren Kitchener]

Since none of this is really new technology, and I haven't seen much in the way of failure reports to make me nervous about it...

I really do want the 'brand-X' module to 'do the job' - they are quite inexpensive, considering the number of components. But reliability is something I also place importance on, if not the looks. I know the sound is going to be good, so I'm not worried about that.
What makes me shudder a little is that the MOSFETs are compressed onto the heat-sink-plate using the resistance of the PCB only - which makes it bow considerably, between the two stand-offs. This may be common practice for the audio market - but placing a SMD PCB under stress after it's been flow-soldered is something the Military and Aviation industries try to avoid, unless they use specially designed flexible PCBs (only where they have to).

All said, unless I spend a lot more, I'm going to have to live with the purchase, and hope that I have the module that was made on a Thursday. It may be worth that risk - after all, you do occasionally see some older Citroen cars on the road.

 

[lovinthehd]

What do you specifically have? Pics?

 

[Ren Kitchener]

What do you specifically have? Pics?

Yes - see below @lovinthehd - it looks more bowed when set against a ruler. Granted, the ruler may be bowed.

 

[highfigh]

I wanted an integrated amp - this could be an option @highfigh, but then I'll have a very large 15" high preamp - if you see what I mean from my 'CAD' design below:

I asked because your comments from the post I responded to (If this 'breaks' within the year, the replacement is expensive - worst still, I may not be able to get a replacement - meaning a lot of work to shoehorn a class AB in it's place.) would kill this project during the design phase if it was done by an actual manufacturer. Using a one-off design is great, if the economy of scale and demand for that product allow for it, but in a true one-off, that fact doesn't even look good on paper. Why would you want to consider something that can't be easily replaced?

Also, you would need to prevent vibrations from affecting the meters- they're very sensitive and could be damaged if not isolated.

 

[highfigh]

I deleted the original post because I may stick with it (or not) - although the enclosure is still an expensive 'pain' - looks like a 'piano finish' takes a considerable amount spraying and drying time.

Not necessarily- look into pre-catalyzed lacquer, polyurethane or other finishes. Nitrocellulose takes a long time to fully dry and it's very prone to showing scratches from simply wiping dust off of it. If you want piano black, many materials would work great and require less drying/curing time. Spraying doesn't take much time, it's the prep and surface correction that really requires time and patience.

 

[Ren Kitchener]

Also, you would need to prevent vibrations from affecting the meters- they're very sensitive and could be damaged if not isolated.

Thanks for the opportunity to brag a little here @highfigh ....

The meters movements are designed to take vibration - they have critical electro-magnetic damping - plus, each one is individually processor controlled so that overshoot is avoided and 'ideal' rectification can be achieved. The decay is also linear (not log like most) - which really adds to the 'show'.

When they move, the movement is hypnotising - all that 'erratic flicking' is taken out and it stops at true peak 'on a sixpence'. Of course, I'm sure that some amplifiers will have a similar 'quality' PID controlled movement, with a smooth decay.

If anyone wants the code (in C) for the Atmel MCUs - let me know - but it is a little 'narrow' for this specific pairing.

 

[ParleyW]

During a general class D chat with a high-end 'showroom', they stated that they have had problematic reliability issues with Hypex and the Purifi based amplifiers - so they are not really keen on stocking them for that reason.

When purchasing a $2,000+ amplifier, I can really understand that reservation - even if it does get repaired, it's still an annoyance for the customer and the stockist. When the guarantee runs out, it's more than just an annoyance - the amplifier instantly becomes an expensive 'paper-weight' - unless you are a Surgeon or an Attorney, whereon you can get it repaired or bin it and order a new one.
Getting to the point:
I remember asking Class D amplifier company 'X' for their MTBF figures - they didn't have any data. So, looking at the cards, and knowing general MTBF figures for switchers, I would be surprised if the MTBF extended beyond 20 Years:

This means that for every 20 sold, one of them will fail within the year, and that one is going to be mine - it always is.

So, those with Hypex or Purifi or other similar Class D based amplifiers (also amplifiers with switch-mode power supplies), are you seeing any failures within the year, or are they still hanging on after say 3 or 5 years?

Sometimes I think that having a nice simple 50 Watt class AB, with a toroidal transformer and a 0.02% THD spec, is going to be the best value amplifier you could ever purchase, and that 0.02% THD is not as bad as you are led to believe (compared to a short lived 0.000001% THD, 500 W blast, followed by a puff of smoke)...

Still use my Hafler XL-280 daily. Bought new in’88. Kids will use it when I’m dead.

 

[Ren Kitchener]

Still use my Hafler XL-280 daily. Bought new in’88. Kids will use it when I’m dead.

I can see many pictures inside - it may not be the 'neatest' wiring, but this is the point - the wiring is shorter, more spaced apart, with no tight radiuses and it's not stressed.
I'm led to believe that those power transistors were originally designed for a MkIII Panzer Tank gearbox

Apart from a vastly simple reservoir changeout at some point in it's old age, this will go on for another 75 years. THIS IS HOW TO DESIGN AND MAKE AN AMPLIFIER!

So this may be a rebuttal @TankTop5 - Class D may on paper be 'lighter and less expensive to ship', but after a few repairs, this benefit fades a little.