Hypex NCx500 Class D Amplifier Review

[pma]

Some people claim that class D will suffer from intermodulation products caused by the 450 kHz (or so) switching-frequency. Basically I am open to this argument, but why don't I see these intermodulation products then in the 32-tone test? What is it I don't understand?

The usual concern about high-frequency intermodulation effects is because the tweeters (in the speaker systems) generate IMD when presented with such high signals, modulating "electromechanically" to create distortion that falls into the audible frequency range. An electrical test at the output of the amp will not show that.

About 10mA(rms) 400kHz flows from my NC252MP to my speaker in a "room listening" configuration. With all the cable and speaker inductance.

 

[DonH56]

To clarify: You can see the high-frequency switching signal at the amp's output, as @pma demonstrates above, but the IMD generated by the tweeter itself would have to be measured at the output of the tweeter to see if it is audible. I have no idea if that actually happens in the real world, or if the tweeter (or its crossover) effectively blocks such signals. The tweeter may just refuse to move and turn the HF energy into heat, or its inductance could block it but cause voltage ripple through the crossover, etc. I can think of several ways such HF signal could, or could not, cause problems, but not something I have ever measured. That said, @pma's measurement shows there is current flow into the speaker, so that energy is going someplace.

 

[Matias]

Probably heating the heat coil with a few miliwatts. Pma can get a very high frequency response mic, put it right in front of the tweeter, record high sample rate and see if something shows up with his NC252MP compared to other amps. I doubt it. If anything, the NCore probably has LESS energy being sent to the tweeter because of lower IMD.

 

[Endibol]

About 10mA(rms) 400kHz flows from my NC252MP to my speaker in a "room listening" configuration. With all the cable and speaker inductance.

View attachment 259393

Interesting! I recently measured the output of an Audiophonics HPA-S400ET (purifi-based) amp. It showed a constant 280 mV(rms) output voltage at approx. 500 kHz (without load).

 

[Mnyb]

The situation on the electronic components market, both active and passive, is really disappointing, worldwide. Lead time often 8 months, depending on 1 - 2x a year production batch. We need to find substitutes on a regular basis. Really disappointing.

exactly. the same here. we even get 80 weeks lead-times sometimes. just crazy.

Yes , we have that to often I get 52 weeks lead time from within our own company and have to postpone industrial projects or substitute temporary with stuff from our lab or design around the problem and fit the design of the plant to what hardware we can get. We have bought used stuff of our own brand from closed industries to use as temporary measures.
Can’t imagine what making consumer products on a small scale is like.

 

[cavedriver]

So Amir dedicated an entire graph to show the on and off "pop" noises. I have an NC502MP and honestly the noises seem quite loud (perhaps at the volume of normal conversation or just below). Perception-wise, how loud are these noises supposed to be? Without a frequency measurement I was worried these noises could damage my speakers and so I've taken to disconnecting the speaker cables (very carefully!!) before turning the amp on or off. Also, earlier I found the noise was even louder if I used XLR-to-RCA adapters so I found a DAC with XLR outs to drive the amp (an SMSL M300 mk II). Given this sensitivity to the input design it seems to me there ought to be more that could be done to reduce or eliminate these pops?

 

[Doodski]

I have an NC502MP and honestly the noises seem quite loud (perhaps at the volume of normal conversation or just below). Perception-wise, how loud are these noises supposed to be? Without a frequency measurement I was worried these noises could damage my speakers and so I've taken to disconnecting the speaker cables (very carefully!!) before turning the amp on or off.

Dependent on your speakers efficiency the power ON/OFF noise that you hear does not require much voltage to be audible. I recommend not disconnecting the speakers while powering ON/OFF because you are running the risk of causing damage to the amp or speakers. Be very careful.

 

[pma]

Given this sensitivity to the input design it seems to me there ought to be more that could be done to reduce or eliminate these pops?

To fix this, amp must have a circuit for delayed connection of speakers, with the long enough delay after turn on. This needs a relay or solid state relay (MOSFETs). Re input sensitivity to pops, hypex modules are DC coupled and they amplify input DC voltage!!! IMO this is a great design fault and they should have had DC servo instead. It happened to me that one of the tested preamps fell to oscillations with resulting DC output. This resulted in high audible click from the NC module followed by AM demodulation and whistles. There is a big space for improvements in these modules.

 

[marcom22]

To fix this, amp must have a circuit for delayed connection of speakers, with the long enough delay after turn on. This needs a relay or solid state relay (MOSFETs). Re input sensitivity to pops, hypex modules are DC coupled and they amplify input DC voltage!!! IMO this is a great design fault and they should have had DC servo instead. It happened to me that one of the tested preamps fell to oscillations with resulting DC output. This resulted in high audible click from the NC module followed by AM demodulation and whistles. There is a big space for improvements in these modules.

Correct, mine 502mp has a relay switch button power and no pops happens (but my Elac dbr62 are not so sensitive)

 

[pma]

To clarify: You can see the high-frequency switching signal at the amp's output, as @pma demonstrates above, but the IMD generated by the tweeter itself would have to be measured at the output of the tweeter to see if it is audible. I have no idea if that actually happens in the real world, or if the tweeter (or its crossover) effectively blocks such signals. The tweeter may just refuse to move and turn the HF energy into heat, or its inductance could block it but cause voltage ripple through the crossover, etc. I can think of several ways such HF signal could, or could not, cause problems, but not something I have ever measured. That said, @pma's measurement shows there is current flow into the speaker, so that energy is going someplace.

Yes. As you probably know, I made some acoustical near-field measurements at tweeter output (my regular 3-way speaker), with NC252MP. The worst situation was, maybe (maybe not) paradoxically when the amp was driven from analog FM radio. The 19kHz pilot frequency was the highest spectrum component during low volume listening. However, as you can see, there is an ultrasound content above 20kHz, coming from NC252MP. Possible audible intermodulation effect still waits for thorough investigation.

 

[Burning Sounds]

To fix this, amp must have a circuit for delayed connection of speakers, with the long enough delay after turn on. This needs a relay or solid state relay (MOSFETs). Re input sensitivity to pops, hypex modules are DC coupled and they amplify input DC voltage!!! IMO this is a great design fault and they should have had DC servo instead. It happened to me that one of the tested preamps fell to oscillations with resulting DC output. This resulted in high audible click from the NC module followed by AM demodulation and whistles. There is a big space for improvements in these modules.

IME the turn on/off pop issue is not completely straightforward.

I have 3 different 8 channel DACS and the 4 NC502MP modules in my amp behave differently at turn on and off depending on which DAC preceedes them. The loudest pop at turn on is with the M Audio M-trak 8. The next loudest is the Mytek ADDA8x192 where there is a small turn on pop and a louder turn off pop (almost a loud click actually). With the Okto Research DAC8 PRO there is no turn on noise at all. There is usually a small turn off click, but it can vary in loudness. I've always presumed that these differences are down to the different DC offset levels in each DAC channel.

However, there is a still a further complication. As my speakers are 4 way actives each amp channel is connected directly to its own driver. In the Okto Research example the turn off click can vary somewhat in loudness and it does not happen in all the drivers. In one instance at turn off it may happen in only one driver or a couple of drivers and only in the left speaker. The next time at turn off it might be in the right speaker or both speakers, but again the driver(s) might vary.

Any thoughts?

 

[pma]

Any thoughts?

The power amplifier must be turned on as the last component of the audio chain (or its output protection relay must have delay time long enough). In case of Hypex modules, none of the preamps or DACs used before the power module may have DC voltage at the output. DC must be close to 0mV. Please take into account that even "small" DAC output DC of 10mV will make 200mV at Hypex output and it will be a huge audible pop/click.

Worse, Hypex modules have not very well defined output DC, itself. With my module, L channel has 33mV output DC (permanent) and R channel has -8mV DC. Hypex tech support says it is OK when less than 40mV. But, the turn-on click of permanent 33mV is audible.

 

[Burning Sounds]

The power amplifier must be turned on as the last component of the audio chain (or its output protection relay must have delay time long enough). In case of Hypex modules, none of the preamps or DACs used before the power module may have DC voltage at the output. DC must be close to 0mV. Please take into account that even "small" DAC output DC of 10mV will make 200mV at Hypex output and it will be a huge audible pop/click.

Worse, Hypex modules have not very well defined output DC, itself. With my module, L channel has 33mV output DC (permanent) and R channel has -8mV DC. Hypex tech support says it is OK when less than 40mV. But, the turn-on click of permanent 33mV is audible.

Thanks, very aware that the power amp is on last and off first.

But what is puzzling me is the inconsistency at turn off. Why would the turn off click appear to be random and at varying loudness levels?

 

[DonH56]

But what is puzzling me is the inconsistency at turn off. Why would the turn off click appear to be random and at varying loudness levels?

The relative offset ("DC equivalent") signal and overall average signal level depends upon the volume thus influences the turn-off "click". For example, a square wave has different average level than a sine wave, and music/movie sounds may vary widely in their effective average (offset) level. The click may depend upon the signal applied at turn-off and internal state of the amplifier (temperature, where it is in the PCM output waveform, etc.) Note whilst a coupling capacitor to block DC may change this behavior, it will not in general eliminate it, just add another time constant to the system. A DC servo loop as suggested by @pma is the usual fix, which may be incorporated into the overall feedback loop.

 

[Burning Sounds]

The relative offset ("DC equivalent") signal and overall average signal level depends upon the volume thus influences the turn-off "click". For example, a square wave has different average level than a sine wave, and music/movie sounds may vary widely in their effective average (offset) level. The click may depend upon the signal applied at turn-off and internal state of the amplifier (temperature, where it is in the PCM output waveform, etc.) Note whilst a coupling capacitor to block DC may change this behavior, it will not in general eliminate it, just add another time constant to the system. A DC servo loop as suggested by @pma is the usual fix, which may be incorporated into the overall feedback loop.

Ah, yes - "where it (the amplifier) is in the output waveform" when it is turned off makes sense to me. As there are four amplifier modules this would explain why the click appears at different drivers and with different intensities. Sometimes if I'm lucky there is no click from any of the drivers. Thanks @DonH56 and @pma.

 

[cavedriver]

When turning off the chain (which is the louder pop and Amir's measurement matches my experience in that regard), I wonder if it isn't likely the noise is louder by turning the amp off first because there might be this "offset voltage" from the preamp/DAC if it's still on? Of course there's no sound when the preamp/DAC and amp are still both on so I'm at a loss to understand what this offset voltage represents. I'll have to try it again but I thought during my earlier testing I found that turning the DAC off first resulted in the quieter amp pop...

Also, because the lack of RCA adapters made a difference I thought it might be that a different grounding strategy could make the difference. This made me very interested in the Q&A sequence about why Rick grounded the XLR connector the way he did.

 

[ivo.f.doma]

And is this behavior inherent in NcoreX and Purifi modules as well?

 

[Rick Sykora]

When turning off the chain (which is the louder pop and Amir's measurement matches my experience in that regard), I wonder if it isn't likely the noise is louder by turning the amp off first because there might be this "offset voltage" from the preamp/DAC if it's still on? Of course, there's no sound when the preamp/DAC and amp are still both on so I'm at a loss to understand what this offset voltage represents. I'll have to try it again, but I thought during my earlier testing I found that turning the DAC off first resulted in the quieter amp pop...

Since a lot has been covered already on the potential pop source, will add that the NC series has gotten better. My first amp was a VTV NC502MP and it had a pop. VTV could not resolve and so I returned it. EDIT (in bold): Later I talked to Hypex support and found they had improved. I subsequently acquired a NC502MP but unfortunately did not get the version with the improvement. It did seem worse with unbalanced to balanced connections. I later added a Neurochrome Guardian to the NC502MP and it is much improved (still gets an occasional pop). Hope this helps!

 

[multisport4me]

I tried the NC502MP and ended up having to return it to have amp rebuilt with NC500OEM. The popping on the NC502MP was terrible in my set up. The NC500OEM is silent.

I also built up some NC400s and they intermittedly have a very soft pop but mostly are silent unless I put my ear against the speaker. I had a Marantz PM10 in the past and it was silent. I have an AMP10 that I'll be installing tonight or tomorrow and will report on whether or not it pops.

I guess I'm surprised to hear that the NCx500 pops.

 

[Rick Sykora]

I tried the NC502MP and ended up having to return it to have amp rebuilt with NC500OEM. The popping on the NC502MP was terrible in my set up. The NC500OEM is silent.

I also built up some NC400s and they intermittedly have a very soft pop but mostly are silent unless I put my ear against the speaker. I had a Marantz PM10 in the past and it was silent. I have an AMP10 that I'll be installing tonight or tomorrow and will report on whether or not it pops.

I guess I'm surprised to hear that the NCx500 pops.

I auditioned with my Purifi SPKs before sending the unit to Amir and did not hear any power down pops.

Am very careful about power up sequence, not so much on power down.