Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Class D Switching Power Amplifiers and Power D/A conversion

Hypex Ncore
Hypex Ncore
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Closed Thread
Thread Tools Search this Thread
Old 11th June 2011, 06:57 PM   #11
Bruno Putzeys is offline Bruno Putzeys  Europe
diyAudio Member
Join Date: Apr 2004
Location: BE/NL/RW/DK
I think the THD vs power graphs tell most of the story. 420W at 4 ohms is the minimum distortion point but by no means the clip level: at 600W distortion is still 0.01% which isn't clipping by any stretch of the imagination. Clipping power (1% THD) is about 700W which is perfectly in line with the loaded supply voltage, 100Hz ripple and a few volts of headroom. A 2.4dB trajectory between minimum distortion and clipping might sound like a lot until you consider the alternative, which would be to use a fixed frequency amplifier. Such an amplifier, if it used a ripple-compensated modulator, could be designed with sharper clipping (i.e. minimum distortion is closer to the clip point) but that would mostly be because the clipping point becomes *lower* on account of minimum pulse width. With the same power stage you might for instance get low distortion all the way up to 550W at which point pulse skipping sets in and distortion jumps aggressively. Given the choice I'd rather have gentle clipping with a higher maximum.

Last edited by Bruno Putzeys; 11th June 2011 at 07:07 PM.
Old 11th June 2011, 10:38 PM   #12
Javin5 is offline Javin5  Switzerland
diyAudio Member
Join Date: Mar 2006
Location: Switzerland
I fully agree. Some additional questions:

- Can you say anything on distortion and IM at very low levels, say in the range from a few mW to 1 Watt as a function of frequency?

- Is the HF at the speaker terminals under load similar to that of the UcD, about 200 mV RMS at 400 KHz?
Old 11th June 2011, 10:46 PM   #13
AP2 is offline AP2  Italy
Join Date: Dec 2009
Location: Italy
Thank, and Yes, I agree that thd vs. power you can see how the "engine". In fact, I asked the voltage of the PSU. Only one thing is not clear. it seems that measures 8-4-2R, have the same point on the ripple psu. this was my curiosity, since I know the psu unregulated.(or test with regulated PSU?) However, the measures have achieved great improvements over the traditional ucd700. Certainly, it would be interesting to test one.


Roberto P.

Last edited by AP2; 11th June 2011 at 11:10 PM.
Old 12th June 2011, 04:21 AM   #14
Trevor White is offline Trevor White  Australia
diyAudio Member
Join Date: Jun 2009
This N-core looks like a real game changer for class-D amplification as well as giving all other amplifier classes a run for their money. For the first time you have efficient operation and ultra low distortion amplification even into low impedance loads !!

Well done to Bruno and the crew at Hypex. It will be interesting to see how these are accepted in the market place, because traditionally class-D has been something that was usually relegated to driving subwoofers or in professional PA applications.

Old 12th June 2011, 10:16 AM   #15
Bruno Putzeys is offline Bruno Putzeys  Europe
diyAudio Member
Join Date: Apr 2004
Location: BE/NL/RW/DK
If the analyser had two low-distortion sine oscillators and two notch filters I could do low-level IMD plots as well but now I have to use the AD/DA path. As you can see from the grey plots even at high levels the converters pretty much dominate the result so all I could do was to crank up the level until something showed up that was caused by the amp.

You can do an estimate though: because the loop gain is more or less constant up to 20kHz, IMD and THD behave very similarly and furthermore readings will be largely independent of frequency. Under these circumstances, mix products of a 1:1 IMD test will have about the same amplitude as the HD products of a single tone test. If you mix 18.5k and 19.5k, the 2nd order mix product is at 1kHz, the 3rd at 17.5 (and 20.5), the 4th at 2kHz, the 5th at 16.5kHz etc.
This is only conditionally true (only if you don't get thermal modulation effects or slew related distortion for instance) but it holds pretty well for UcD and Ncore.
From the THD FFT plots at 1W you can then estimate that IMD levels will be similar, that is around -110...-120dB depending on the load.
Anyhow, low level IMD does not do anything crazy. There are a variety of fields in audio where people postulate fancy low-level IMD effects that fail to show up with single tones, but I've yet to see that happen in reality. IMD is mostly interesting to characterize performance at high power levels.

@Roberto, the test was done with the unregulated SMPS1200. You can see this from the fact that:
1) 2ohm clipping power isn't four times 8 ohm clipping power
2) clipping power is lower at 100Hz than at 1kHz.
If the power supply were regulated the tail end of the THD plots would overlap exactly and power levels would scale exactly with load impedance. There's something like 10Vpp ripple on the supply at 1200W.

It would be nice to see this thing in a PA amp. Imagine that, audiophile live sound. I find that live sound engineers are sensitive to sound quality (albeit only after a number of practical requirements) so it might well happen.

Last edited by Bruno Putzeys; 12th June 2011 at 10:29 AM.
Old 12th June 2011, 04:43 PM   #16
Bruno Putzeys is offline Bruno Putzeys  Europe
diyAudio Member
Join Date: Apr 2004
Location: BE/NL/RW/DK
I forgot to answer Jav's second question. HF output is identical to that of the UcD700.
Old 12th June 2011, 05:21 PM   #17
Javin5 is offline Javin5  Switzerland
diyAudio Member
Join Date: Mar 2006
Location: Switzerland
Bruno, thanks for answering my questions. Seems to me that you have a winner here. With the (probably?) coming EU restrictions on power dissipation, class A could potentially become a thing of the past and your circuit may well be a key element for future high-end power amplifiers. I hope you can get international patent protection; could be worth a fortune. Good luck!
Old 12th June 2011, 08:21 PM   #18
stinius is offline stinius
diyAudio Member
stinius's Avatar
Join Date: Oct 2008
Hi Bruno
Its interesting and nice to see that you have designed a new amp that is better than the UcD.
I have to agree in everything you are saying in the WP.

BTW: when you say that your new comparator is using 10 transistors, I guess that you dont include the CCSs, but only the comparator with the floating collector outputs.

BTW2: are you using one or two CCSs in the comparator?

Old 13th June 2011, 06:18 AM   #19
beun is offline beun  United States
diyAudio Member
Join Date: Feb 2010
Bruno, I have a question comparing UCD versus N-core in a very specific loading configuration.

Generally using UCD to drive electrostatic speakers work fine except when driving the stats transformer so hard that the core starts to saturate which is actually not that hard to do. It looks like that the resulting very low loading impedance collapses the feedback taken from behind the filter resulting in the sound cutting out for a second or two in a rather nasty way. A regular class AB amplifier only starts to distort heavily but doesn't mute.

My question is will the N-core technology, which seems capable do drive much lower impedances, behave in a similar way or would it be capable to handle this better.
Old 13th June 2011, 09:45 AM   #20
Bruno Putzeys is offline Bruno Putzeys  Europe
diyAudio Member
Join Date: Apr 2004
Location: BE/NL/RW/DK
Ooh you're trying to get hold of the fine details there. The comparator is a more traditional design compared to the old UcD one. Counting tail sources (two) it's 11 transistors.

The difference lies not in the feedback but in the way the overcurrent protection works. In many class AB amps the overcurrent / SOA protection just limits current. In most UcD modules and in the NC1200 it limits current (i.e. keeps playing with distortion) for about 50 to 100ms, depending on the signal. If the overload condition lifts before that, no muting occurs. If the overload condition persists, it'll mute for a second as a safety measure. In fact, you could disable the mute stage and force the amplifier to remain in the ordinary current limiting mode. I'm not comfortable giving instructions on how to do this, however, because some people might then try to drive our amps beyond their abilities for sustained lengths of time.
Actually the low impedance capability of the NC1200 speaks against doing this. The current limiter kicks in around 38A. Do you want an amplifier insisting on driving that amount of current into the primary of the step-up?
In short, the protection behaviour is not the result of feedback collapsing or any such thing but is done by protection circuitry that was explicitly added. The only thing I might consider is tweaking the overload mute timing.

Last edited by Bruno Putzeys; 13th June 2011 at 09:53 AM.

Closed Thread

Hypex NcoreHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Hypex UcD700 vs NCore Earbound Class D 27 20th October 2019 03:50 PM
Hypex Ncore vs. NAKSA AlliumPorrum Class D 1 26th May 2015 07:36 PM
Problem with Hypex Ncore Koifarm Class D 29 13th October 2014 07:32 AM
Any news on hypex ncore 2.0? JZatopa Class D 1 13th August 2013 11:30 AM
Building a Hypex Ncore... Jabbejokker Class D 0 7th March 2013 10:08 PM

New To Site? Need Help?

All times are GMT. The time now is 11:13 PM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.
Copyright ©1999-2020 diyAudio