Hypex NC500MP capacitor failure

[TimoJ]

My DIY Hypex NC500MP dual module amplifier started to make loud "pop" noise coming inside the amp from the amplifier PCB. I started to investigate and noticed burn marks around one capacitor. After removing the capacitor I noticed that the PCB under the capacitor was damaged. Looks like the capacitor had leaked and started to corrode the PCB between the PCB layers. Both 10uf capacitors also had bulged tops.

The other NC500MP module also showed similar bulged capacitor problem. Original 10uf values were now 4.3uf esr 33 ohm and 0.8uf ja esr 496 ohm. I have since fixed the modules, replacing all (except the two caps after the AC rectifier) capacitors with Rubycons and also fixed the PCB issue. Working fine now.

But I wonder if this is a common issue? The modules were only 1 year old. All other capacitors appeared to be fine and still measured OK (but with little worse values than new Rubycons).

If you have an amplifier using these modules, please check those two capacitors and report what you see. I noticed that also other NCxxxMP modules are using almost similar circuit, but I'm not sure if this effects them, probably not.

 

[pma]

High dV/dt across the capacitor, resulting in high Ic = C.dV/dt, my guess. It is alarming. Have you measured the DC voltage across the capacitor? Is it within specs?

 

[sigbergaudio]

Did you ensure proper cooling?

 

[TimoJ]

High dV/dt across the capacitor, resulting in high Ic = C.dV/dt, my guess. It is alarming. Have you measured the DC voltage across the capacitor? Is it within specs?

I haven't measured the voltage, Hypex support says it's 75V and those capacitors are 100V. I replaced those two with 130C RX30 series capacitors, others with 105C ZLJ series.

 

[TimoJ]

Did you ensure proper cooling?

The modules are attached to an aluminum case from Modushop, bottom has is 2mm + extra 2mm aluminum sheets. And the amps were used with low wattage bookshelf speakers. I will modify the case and make the top panel fully vented.

 

[Davide]

The modules are attached to an aluminum case from Modushop, bottom has is 2mm + extra 2mm aluminum sheets. And the amps were used with low wattage bookshelf speakers. I will modify the case and make the top panel fully vented.

If you want advice, having dealt with the issue of thermal dissipation of my NCXXXMP, try to measure the temp near FETs. In theory you can use the built-in thermistor in NCXXXMP module but you have to build the measurement circuit.
I used a PRT of and old pc fan controller, placed attached to a FET of the PSU and amp part, and with NC502MP just screwed to the case I measured up to 98°C after 30 min of normal load (PSU FETs work hotter in my case, although switching lower frequency than the amps should have less loss...).
Clearly I could not estimate the measurement uncertainty... but +/-5°C is realistic.
For sure the components around don't reach the same temperatures but they are not so much colder probably (I didn't measure this because it was an unrepresentative measurement and I wanted to dissipate the heat at the root in the expected way).
In fact I replaced the thermal pads and put a 10 mm aluminum plate with thermal paste in between, plus an aluminum heatsink with fan.
Now I measure 55°C at maximum.

Note that these modules use the PCB to press the FETs onto the thermal pads (bad choice in my opinion), and it has already been documented that sometimes the PCB deforms by stressing the components, or by reducing pressure on the pads, therefore decreasing the thermal transmission.
So there are several things that can go wrong…

These modules require more mechanical than electrical engineering skills to be integrated.

 

[dfuller]

Hypex has an unpleasant tendency to use junk electrolytics. This is not unexpected, unfortunately...

 

[TimoJ]

My other NC500MP module's capacitors:

 

[Hipocrates]

My other NC500MP module's capacitors:

View attachment 306324
View attachment 306325

Those AISHI caps are good for prototype, they are extremely cheap caps, cost a fraction of a first tier good quality capacitor, I get them on tayda electronics.
I would not use them on a restoration or an amplifier for daily use.

For example
47UF 35V 105C ALUMINUM ELECTROLYTIC CAPACITOR 5X11MM AISHI LONG LIFE... $0.08 (tayda)
47UF 35V PANASONIC EEU-FC1V470H.... $0.41 (mouser)

 

[restorer-john]

Hypex has an unpleasant tendency to use junk electrolytics. This is not unexpected, unfortunately...

And predicted several years ago, by me, here on ASR.

The chickens are coming home to roost and there will be plenty of people with egg on their faces...

 

[iLoveCats]

And predicted several years ago, by me, here on ASR.

The chickens are coming home to roost and there will be plenty of people with egg on their faces...

You mean there's no free lunch?

 

[Doodski]

And predicted several years ago, by me, here on ASR.

The chickens are coming home to roost and there will be plenty of people with egg on their faces...

Shameful using junk caps. Class D will get a bad rep from junk caps and we'll be back to class AB as the go-to stuff.

 

[dfuller]

And predicted several years ago, by me, here on ASR.

The chickens are coming home to roost and there will be plenty of people with egg on their faces...

I did kind of call this in that Apollon teardown thread didn't I.

 

[McFly]

This is the second hypex fail I've seen. I think the other was a nc122mp, on here too from memory.

If anyone know of any others, please post up.

I'm not disagreeing with anyone here, I agree some components are cheap asf.

But how many do you think they have sold?

 

[dfuller]

This is the second hypex fail I've seen. I think the other was a nc122mp, on here too from memory.

If anyone know of any others, please post up.

I'm not disagreeing with anyone here, I agree some components are cheap asf.

But how many do you think they have sold?

I've heard of a lot of failures. They sound great but their build is... not great. There's a reason very few active speaker brands use their plate amps, and it's not because they don't sound good.

 

[restorer-john]

I've heard of a lot of failures. They sound great but their build is... not great. There's a reason very few active speaker brands use their plate amps,. and it's not because they don't sound good.

3 dead Hypex amplifiers, all sent to me from ASR members. I may be able to save one of them, I hope. The bottom two are about 8 years old or so. Just a bullsh#t short life for a power amplifier. One was killed by an insect, but IDing the vaporized parts required another unit because Hypex don't give a damn and won't provide schematics or parts lists. I bet they don't even have any parts available at all, to anyone- never did. Just throw it out and start again.

A member here was able to point me in the right direction with parts ID with a previous repair of another Hypex.

But we will win and this type of behaviour will very soon become criminal in the EU. Every single part was available from the big Japanese HiFi makers. Every single component right down to the mounting screws had a part number you could order.

This stuff should not be going into landfill!

 

[Doodski]

But we will win and this type of behaviour will very soon become criminal in the EU. Every single part was available from the big Japanese HiFi makers. Every single component right down to the mounting screws had a part number you could order.

How many technicians could there have been on Earth servicing the gear and ordering parts from Japan? The numbers must be hefty!

 

[GXAlan]

@restorer-john
Have the failures all been in units running in hardware mode or have some been devices that run in software mode too? I wonder if the thermistors aren’t great because local temps may exceed what the thermistor is designed to trigger as a fault in hardware mode and you need a full software controlled setup for HypeX to be reliable.

I ask because Marantz has always implied that the default NC500OEM did not meet their internal quality standards when thinking about the next 10-20 years (yeah I know, that’s a big range) and that it was hard to get it to pass their own requirements and that this is somehow the reason why the Marantz HypeX products were really expensive.

Seems like it’s more than just statistical based control plots and they did delay the PM-10 by half a year compared to the SA-10 and they explicitly blame reliability concerns, though on record, about going relay-free for speaker protection.

I don’t think they had custom modules as they have done with ICEpower and their AMP 10 for home theater.

 

[DonH56]

3 dead Hypex amplifiers, all sent to me from ASR members. I may be able to save one of them, I hope. The bottom two are about 8 years old or so. Just a bullsh#t short life for a power amplifier. One was killed by an insect, but IDing the vaporized parts required another unit because Hypex don't give a damn and won't provide schematics or parts lists. I bet they don't even have any parts available at all, to anyone- never did. Just throw it out and start again.

A member here was able to point me in the right direction with parts ID with a previous repair of another Hypex.

But we will win and this type of behaviour will very soon become criminal in the EU. Every single part was available from the big Japanese HiFi makers. Every single component right down to the mounting screws had a part number you could order.

This stuff should not be going into landfill!
View attachment 307543

Very curious to know what you find wrong with these.

 

[Deleted member 48726]

@restorer-john
Have the failures all been in units running in hardware mode or have some been devices that run in software mode too? I wonder if the thermistors aren’t great because local temps may exceed what the thermistor is designed to trigger as a fault in hardware mode and you need a full software controlled setup for HypeX to be reliable.

View attachment 307561


I ask because Marantz has always implied that the default NC500OEM did not meet their internal quality standards when thinking about the next 10-20 years (yeah I know, that’s a big range) and that it was hard to get it to pass their own requirements and that this is somehow the reason why the Marantz HypeX products were really expensive.

Seems like it’s more than just statistical based control plots and they did delay the PM-10 by half a year compared to the SA-10 and they explicitly blame reliability concerns, though on record, about going relay-free for speaker protection.

I don’t think they had custom modules as they have done with ICEpower and their AMP 10 for home theater.

I don't think 10-20 years is a long life expectancy for an amp. Where did you see this info from Marantz?