Hypex Ncore

Status
Not open for further replies.
I have a dual mono nc400 (+2x smps600) box, and I have a couple of questions. A while ago, one of the channels started to exhibit some strange behavior. The volume would drop, and simultaneously either white noise or ~100-200 Hz noise plus high pitched noise would occur. This happens also if the source is turned off. Also, this seems to happen more and more often. Has anyone else had this happen?

I haven't yet tried to swap the modules around to isolate the cause. Will probably do so soon.

Another question: seeing as I'll probably have to send the module (+ probably also smps) of the offending channel to Hypex for repair, will I be able to just put both speakers on the good channel while I wait for the repair? If so, how should I make the connection? I have 8 ohm speakers.

I already sent these questions to hypex, but I guess they've shut down for the weekend and I'm slightly impatient, hope someone can help!
 
wiring speakers in series is safer than wiring in parallel

I have a dual mono nc400 (+2x smps600) box, and I have a couple of questions. A while ago, one of the channels started to exhibit some strange behavior. The volume would drop, and simultaneously either white noise or ~100-200 Hz noise plus high pitched noise would occur. This happens also if the source is turned off. Also, this seems to happen more and more often. Has anyone else had this happen?

I haven't yet tried to swap the modules around to isolate the cause. Will probably do so soon.

Another question: seeing as I'll probably have to send the module (+ probably also smps) of the offending channel to Hypex for repair, will I be able to just put both speakers on the good channel while I wait for the repair? If so, how should I make the connection? I have 8 ohm speakers.

I already sent these questions to hypex, but I guess they've shut down for the weekend and I'm slightly impatient, hope someone can help!

Whilst your speakers are 8 ohm nominal, they may well drop down to a 4 ohm impedance at some frequencies. If you wired them in parallel then the impedance would be 2 ohms.

Better play safe and wire them in series so the nominal impedance will be 16 ohms.
 
Whilst your speakers are 8 ohm nominal, they may well drop down to a 4 ohm impedance at some frequencies. If you wired them in parallel then the impedance would be 2 ohms.

Better play safe and wire them in series so the nominal impedance will be 16 ohms.

The amp described by the poster is safe to 1 Ohm load (no power rating @ 1 Ohm, rated 600W @ 2 Ohms). Bruno replied re. Ncore driving the huge Apogee full range ribbon with 1 Ohm minimum impedance.

Compared to 4 Ohm 16 Ohm results in less heat (1/4 current requirement), lower distortion magnitude, less offensive distortion components, and 1/4 current capacity for the speaker wire. The only down side is 1/4 the power of 4 Ohm.
 
Last edited:
Compared to 4 Ohm 16 Ohm results in less heat (1/4 current requirement)

True, but only because of:

The only down side is 1/4 the power of 4 Ohm.

(and of course the heat generated when driving 16 ohm won't be 1/4 of the heat generated when driving 4 ohm, so in that sense you lose more in available power (in proportion) than you gain in less heat).

lower distortion magnitude, less offensive distortion components

Curious about what that is based on.
 
True, but only because of:



(and of course the heat generated when driving 16 ohm won't be 1/4 of the heat generated when driving 4 ohm, so in that sense you lose more in available power (in proportion) than you gain in less heat).



Curious about what that is based on.

IIRC independent test by Audio Circle member "mgalusha" revealed NC400 distortion increased and distortion components worsened (both very little) with falling load impedance. Sorry no link. Bruno's spec sheet shows only 2 Ohm load for THD+N...I presume readers are to infer 2 Ohm is worst case scenario and all higher impedance loads yield better results.

AFAIK this is universal SS rule. Please correct if wrong. Audibility obviously varies per many variables.

My system employs two high resolution stand mount monitors per channel, each monitor is 8 Ohm nominal, 5.3 Ohm minimum. I auditioned at length NC400 mono driving two loads identical except for 2.7 Ohm minimum (parallel) vs. 10.6 Ohm minimum (series...speaker wire was the only other variable...in every other respect identical loads...I adjusted the preamp for 6 dB gain difference in the amp).

Trying to avoid going OT, but this is noteworthy (I discussed it with Bruno): With too small AWG (larger numeral) the latter performed better. With appropriately large AWG the former worked better. Results were not marginal, but rather very audible. Even with monitors active HP crossed 2nd order @ 80 Hz.

NC400 has huge current capacity greater than anything else lb for lb. Some applications (low Z) require much larger AWG than expected to hear the amp perform properly.

For speaker wire on the 2.6 Ohm load I used short runs of high quality fine strand copper 4AWG (car audio power cable, super soft touch rubber jacket, highly pliable) with a parallel 30AWG wrapping wire to compensate for high self inductance.
 
Last edited:
IIRC independent test by Audio Circle member "mgalusha" revealed NC400 distortion increased and distortion components worsened (both very little) with falling load impedance. Sorry no link.

I see you edited your response. I have seen mgalusha's FFT plot for 8 ohm, but not ones showing lower impedances.

AFAIK this is universal SS rule.

I am always sceptical towards "universal" rules. What, in your view, would be the cause for higher distortion (and especially for a different distortion spectrum) as current increases (in a class D amp such as nc400)?

Trying to avoid going OT, but this is noteworthy (I discussed it with Bruno): With too small AWG (larger numeral) the latter performed better. With appropriately large AWG the former worked better.

I find AWG a very confusing measure - partly because it is *american* wire gauge and incompatible with any other measuring systems, partly because it is an arbitrary scale, but mostly because it is an inverse scale - smaller numbers mean thicker wires. Wire thickness (be it in inches or mm) or area (mm2) are much better descriptors.

I agree that the nc400's benefit from low-resistance (thick) wires, and I do have a bit of a concern with some of the thinner "audiophile" wires.

NC400 has huge current capacity greater than anything else lb for lb. Some applications (low Z) require much larger AWG than expected to hear the amp perform properly.

I agree. I am using 4 mm (equivalent to AWG 6) for mids and tweeters, and 6 mm (AWG 2.5) for the woofers in my system.
 
Yes, I agree with your view of "universal." "Every rule has exceptions", as my friend used to say, "including this one!"

I could be wrong about Mike's test results; if so my sincere apology.

With SS amps, the distortion magnitude and presumed subjective audibility of distortion spectra both increase as load impedance falls, so says IIRC every Bascom King test in Audio Magazine and also John Atkinson tests in Stereophile. I've heard qualitative differences (unrelated to peak SPL) on at least two SS amps, at moderate SPL, comparing identical loads except for impedance, 2.7 vs. 5.3 vs. 10.6 Ohm.

With the two amps described above, into the higher impedance the sound was more relaxed, more dynamic, more musical, more transparent, and more enticing vs. the lower impedance. It's possible, with some SS amps, to yield little difference in peak SPL into 2.7 Ohms (runs out of current) vs. 10.6 Ohm (runs out of voltage).

Again, NC400 mono apparently passes so much current that it improved into the lower load impedance, provided the speaker cable resistance was low enough (results inverted with too high speaker cable resistance).
 
Last edited:
With SS amps, the distortion magnitude and presumed subjective audibility of distortion spectra both increase as load impedance falls

And why would that be?

Yes, there are reasons for why that might (often) be the case with traditional linear amplifiers, but class D the distortion characteristics are pretty much defined by the feedback circuitry, and there is nothing that I see in the ncore design that would be sensitive to current / load resistance (and that would cause any sort of change in distortion spectra).

so says IIRC every Bascom King test in Audio Magazine and also John Atkinson tests in Stereophile.

Measurements or subjective impressions? On class D amps?
 
Speaker Cable Gauge

Julf: you are generally a technically oriented guy, I am curious as to your use of such enormously over sized speaker cabling. Typical speaker cable gauge is anywhere from 18 AWG to 8 AWG running to a full range speaker. It appears that you believe that there could be advantages to much bigger cabling, considering that 5 AWG would be twice as big as 8 AWG.
My understanding is that the current required to drive most speakers does not dictate such high gauges, would you care to elaborate further on your choice of unusually high gauge cabling with your set up? I do not think you are using long runs either, right?
The idea that the Nc400s can deliver super high currents (compared to other well designed A and A/B amps) does seem to be supported by their specification, current ability seems about standard compared to other amps of similar power:
Is it your belief that most speaker cabling (18 AWG-8 AWG) sold is inadequate to the task?
 
Julf: you are generally a technically oriented guy, I am curious as to your use of such enormously over sized speaker cabling. Typical speaker cable gauge is anywhere from 18 AWG to 8 AWG running to a full range speaker. It appears that you believe that there could be advantages to much bigger cabling, considering that 5 AWG would be twice as big as 8 AWG.
My understanding is that the current required to drive most speakers does not dictate such high gauges, would you care to elaborate further on your choice of unusually high gauge cabling with your set up?

This is hi-fi, where "let's go beyond any technically reasonable specs" seems to be the norm, after all, based on the good old motto "if it is worth doing, it is worth overdoing" :)

Is it your belief that most speaker cabling (18 AWG-8 AWG) sold is inadequate to the task?

No. Most standard speaker cable is well up to the job (as long as it isn't some silly audiophile litz wire or something). It's just that with these specific speakers, I am obsessed with damping factor, and with the nicely low output impedance of the nCores, I wanted to minimize the additional cable resistance.

It also tends to impress visitors :)
 
Actually, i wanted to try the effects of low inductance with nCore and bought a used pair of Goertz Boa MI-3. AWG-7 solid copper, twisted inside a plastic. At least for now, it feels like the bass is now punchier and the sound stage is more open. I used to have ViaBlue SC-2 (awg 12, copper/silver hybrid), which has much higher inductance (10 fold, if not 100 fold).

Some amps have big problems with the Goertz MI cable capacitance (overheating quickly until the "smoke that makes things work" escapes), but nCore is not one of them, since it can take 200nF of cable, as john mentioned somewhere around page 29. Edit: page 28 - http://www.diyaudio.com/forums/vendors-bazaar/190434-hypex-ncore-28.html#post2659718

There is other ways to achieve low inductance, but wanted also try a solid copper cabling, just for trying something different. Actually I didn't even expect a change and i was positively surprised when i heard it.
 
Last edited:
I've finally got some ncores and wanted to double check the wiring of a single ended input.

For now I only have use of single ended cables (in the wall, running to each speaker). So I'm going to remove the RCA plug from the end and replace it with an XLR.

I've looked over the datasheets many times but there isn't a diagram for this. Summarising figure 3 and figure 4, I'm thinking...

Externally

Centre core of coax to pin 2 (hot) of the XLR.

Shield of coax to pin 1 & 3 (cold & ground) of the XLR.

Internally

Exactly as figure 3 - the normal way of using a balanced input. Pin 1 to chassis, shield of internal balanced cable to chassis at socket end and pin 4 (passive, ground) of the ncore.

And double insulated - no mains earth connected.

Could anyone comment if I've got this right? Thanks.
 
Status
Not open for further replies.