• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Hypex NC1200: Quality of the implementations

OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
https://www.audiosciencereview.com/...-of-purifi-1et400a-amplifier.7984/post-194469

Purifi 1ET400A Amplifier DIY kit


index.php


Question:

How thick are the aluminum walls of the DIY box you received?
 
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
Hypex NC1200

PSRR and CMRR.

Hypex-NC1200-PSRR-CMRR.png



PSRR Min: 75 dB Typ: 80 dB Either rail, f<1kHz

CMRR Typ: 55 dB All frequencies


Like PSRR < 90 dB -> better to try with a RF/EMI filter. One or two stages? I do not know.
 
Last edited:
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
Francesco (Rouge Audio) has sent me a picture. It already came with that reddish color, which I suppose must be a silver box. I like it! It is not so difficult to implement the audio electronics as it should be, you just have to have the will.

Rouge Alauda MB3 mono, optimized

Picture updated

Rouge-Audio-Alauda-MB3-Hypex-NC1200-mono-inside-w.jpg


The pictures of the FB are custom jobs -> Great News: you can have the amplifier that you want with your modifications.

https://www.audiosciencereview.com/...ality-of-the-implementations.7918/post-194444


PS: Colin (Nord) still does not answer my request for measurements.
 
Last edited:

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,368
Likes
234,381
Location
Seattle Area
Rouge Alauda MB3 mono, optimized
Those heavy cables can break or disconnect their connectors which moved around. They all need to be securely fastened to the chassis. The one on the left has screws on the bottom already. Shame they didn't put a clamp half-way in the middle. Hate to see that cause a break with mains wiring during shipping.
 
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
Obviously it is improvable, as I would have it. I was comparing with the photos of other implementations that can be seen in this thread.

Obviously, the first thing would be to try with a RF/EMI Schaffner filter. New inlet or DIY cable with the inlet filter and the IEC C13 connector (better from Schurter).

The beautiful case is Japanese. The picture deceived me. The thickness of aluminum is only 2 mm. Or new box or add thickness in the upper and lower panels. Galvanized steel sheet of 1mm, stuck with cheap silicone, avoiding contact between the two metals. Or, if it will be inside a piece of furniture, put the steel sheets -top and bottom- in it instead of in the case, so pretty it.

I do not think that these thick cables move much, but it would not be bad to add fastening, logically.
 
Last edited:

wynpalmer

Active Member
Technical Expert
Joined
Dec 14, 2018
Messages
175
Likes
214
https://www.audiosciencereview.com/...ng-on-an-eigentakt-amplifier.7671/post-191851

[ https://www.miveraaudio.com/forum/main/comment/5d09e8d47668730061a7b5b2

The Purifi input board buffer uses TI OPA1612 opamps. If you have a look at the datasheet, based on the information they provide, the objective performance is outstanding:

http://www.ti.com/lit/ds/symlink/opa1612.pdf

TI made a newer version of this opamp, with higher current output called the OPA1622:

https://www.audioxpress.com/news/Te...eneration-OPA1612-Audio-Operational-Amplifier

And after that they released an enhanced version with built in thin film resistors, and EMI suppression provisions, that is actually capable of delivering better real world performance in a final implementation. The data sheet specs are not as good, but due to the internal laser trimmed resistors, and EMI provisions, real world performance should be better:

http://www.ti.com/lit/ds/symlink/ina1620.pdf ]
Really? I'm an ex- ADI etc. opamp designer.
My reading is that the specs of the 1612 are superior in many regards to that of the 1622. The 1622 is superior in some ways, but is optimized for low quiescent power and the ability to be shut down and in a number of ways is a very different beast to the 1612.
The INA1620 includes trimmed 1k TF resistors. Perhaps if you're building an instrumentation amp input or a unity gain differential to SE output driver it might be handy, but not generally and I really don't see what audiophile benefit there will be from this. As for EMI suppression- is that really much of a problem, particularly since, as far as I can tell, the protective action doesn't present itself until GHz frequencies...
 

wynpalmer

Active Member
Technical Expert
Joined
Dec 14, 2018
Messages
175
Likes
214
Obviously it is improvable, as I would have it. I was comparing with the photos of other implementations that can be seen in this thread.

Obviously, the first thing would be to try with a RF/EMI Schaffner filter. New inlet or DIY cable with the inlet filter and the IEC C13 connector (better from Schurter).

The beautiful case is Japanese. The picture deceived me. The thickness of aluminum is only 2 mm. Or new box or add thickness in the upper and lower panels. Galvanized steel sheet of 1mm, stuck with cheap silicone, avoiding contact between the two metals. Or, if it will be inside a piece of furniture, put the steel sheets -top and bottom- in it instead of in the case, so pretty it.

I do not think that these thick cables move much, but it would not be bad to add fastening, logically.

I have to ask, why does the Al have to be 4mm or thicker? A 4mm skin depth occurs at c. 500Hz for 4mm Al. Are you particularly concerned with field penetrations at sub 500Hz frequencies?
 
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
Really? I'm an ex- ADI etc. opamp designer.
My reading is that the specs of the 1612 are superior in many regards to that of the 1622. The 1622 is superior in some ways, but is optimized for low quiescent power and the ability to be shut down and in a number of ways is a very different beast to the 1612.
The INA1620 includes trimmed 1k TF resistors. Perhaps if you're building an instrumentation amp input or a unity gain differential to SE output driver it might be handy, but not generally and I really don't see what audiophile benefit there will be from this. As for EMI suppression- is that really much of a problem, particularly since, as far as I can tell, the protective action doesn't present itself until GHz frequencies...

There are not my words.

I only extracted it because it was striking.

As I have said before, it will depend on the best or worst implementation of the op-amps. If the buffer is expressly developed for the OPA1622 it is expected that the measurements are very good.
 

wynpalmer

Active Member
Technical Expert
Joined
Dec 14, 2018
Messages
175
Likes
214
There are not my words.

I only extracted it because it was striking.

As I have said before, it will depend on the best or worst implementation of the op-amps. If the buffer is expressly developed for the OPA1622 it is expected that the measurements are very good.
Then I have to ask, with all due respect, what differentiates the implementations?
When parameters such as distortion, bandwidth etc are measured the bench tests are performed using circuits that are more or less exactly as described in the data sheets and using quality but not exceptional bench power supplies. This implies that meeting data sheet specifications such as noise, distortion, PSRR etc. for most opamps in a given application (excluding exotics like the AD797 which is a bit more demanding) is fairly straight forward and barely requires more than some technical knowledge and good reading comprehension.
In many cases where parameter distributions are presented the data has been generated using automated test equipment using a board which is loaded into a large automatic tester unit, and no additional care is taken, or indeed can be taken, in comparison to a bench set up.
These are hardly your normal audiophile environments.
 
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
Sorry for the delay and the short response from before. It was time for lunch (in Spain we eat much later than the rest of Europeans). I have already taken the coffee, I will try to expand a little more.

As for the op-amp. In another thread, in the NC500 thread already commented. Initially the buffer designed was for the Sonic Imagery. Later it was intended to work for other op-amps, too, but some measured quite badly at first, in the case of Burson and Muse. That is why the logical thing, with the NC500, was to use the Sonic Imagery, my choice.

https://www.audiosciencereview.com/...ements-of-nord-one-nc500-amp.7704/post-193449

If a buffer is designed specifically for an op-amp that already measures well it is assumed that it will also measure well, provided that good measurement tools have been used during the process.

Yesterday Bruno Putzeys participated for the first time in these forums, in the thread of the Purifi 1ET400. We could take to ask what was the change of the op-amp with respect to the NC400 and NC1200, the LM4562, if I am not mistaken.

https://www.audiosciencereview.com/...-of-purifi-1et400a-amplifier.7984/post-194984

And you could ask him about 1612 vs 1620.
 
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
Skin effect

We do not care here. They are using modules that do not touch the metal walls, keeping the safety distance. Another thing would be those devices that have transistors on the aluminum heatsinks, which many do not implement properly, and I do not mean just heat dissipation.

https://en.wikipedia.org/wiki/Skin_effect

Depht of penetration

It is the distance required for the wave is attenuated 1/e times. About 38 % of the initial value.

https://en.wikipedia.org/wiki/Penetration_depth


Absorption losses for the depth of penetration by the skin effect are about 9 dB. It is the main shield for low frequency magnetic fields (you know, transformers). Double the distance -> 18 dB.
 
Last edited:

wynpalmer

Active Member
Technical Expert
Joined
Dec 14, 2018
Messages
175
Likes
214
Sorry for the delay and the short response from before. It was time for lunch (in Spain we eat much later than the rest of Europeans). I have already taken the coffee, I will try to expand a little more.

As for the op-amp. In another thread, in the NC500 thread already commented. Initially the buffer designed was for the Sonic Imagery. Later it was intended to work for other op-amps, too, but some measured quite badly at first, in the case of Burson and Muse. That is why the logical thing, with the NC500, was to use the Sonic Imagery, my choice.

https://www.audiosciencereview.com/...ements-of-nord-one-nc500-amp.7704/post-193449

If a buffer is designed specifically for an op-amp that already measures well it is assumed that it will also measure well, provided that good measurement tools have been used during the process.

Yesterday Bruno Putzeys participated for the first time in these forums, in the thread of the Purifi 1ET400. We could take to ask what was the change of the op-amp with respect to the NC400 and NC1200, the LM4562, if I am not mistaken.

https://www.audiosciencereview.com/...-of-purifi-1et400a-amplifier.7984/post-194984

And you could ask him about 1612 vs 1620.
Well, my current opamp of choice is the OPA1612- I've swapped out a number of older duals with them as a number of specs are clearly superior to the LME4562 and similar to the LME49720 my old opamp of choice, and the inferior specs are, in my estimation, less important, so I'm not sure why I need to ask Bruno, even though his work on the analytic optimization of class D is spectacular... I have not swapped out the opamps on the Ncore400 amps that I have, and I believe that my other SS amp, an AHB2, already uses 1612s- although I may be wrong in that.
As for good measurement tools- what do you mean- are you referring to the tools themselves, or the actual kinds of measurements?
I will check out the NC500 thread and perhaps then I can understand your points.
 
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
To finish, in audio it is considered that the artifacts are not audible from 90 dB.

Before I commented that for the civil electronic circuitry, in general, a protection of 60 dB is required. Maybe in audio we should raise that amount. How much? I do not know.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Hypex NC1200

PSRR and CMRR.

View attachment 28977


PSRR Min: 75 dB Typ: 80 dB Either rail, f<1kHz

CMRR Typ: 55 dB All frequencies


Like PSRR < 90 dB -> better to try with a RF/EMI filter. One or two stages? I do not know.

Did you read Bruno advice above? To NOT use filters?

When you are looking at these numbers you need to have understanding so you can apply context. Blindly comparing numbers, as you are doing, tells you little and sends you down rabbit holes.
 
Last edited:

wynpalmer

Active Member
Technical Expert
Joined
Dec 14, 2018
Messages
175
Likes
214
Skin effect

We do not care here. They are using modules that do not touch the metal walls, keeping the safety distance. Another thing would be those devices that have transistors on the aluminum heatsinks, which many do not implement properly, and I do not mean just heat dissipation.

https://en.wikipedia.org/wiki/Skin_effect

Depht of penetration

It is the distance required for the wave is attenuated 1/e times. About 38 % of the initial value.

https://en.wikipedia.org/wiki/Penetration_depth


Absorption losses for the depth of penetration by the skin effect are about 9 dB. It is the main shield for low frequency magnetic fields (you know, transformers). Double the distance -> 18 dB.
Yes, I understand what skin effect is. What I really don't understand is why EXACTLY 4mm? What is so important about 4mm rather than 2mm?
What frequencies are you concerned about? 4mm of Al will not have much effect at all on 60/50Hz fields from a transformer where skin depth is what, 11mmm, (you need 2mm of steel for that) and 2mm Al will have the same skin effect depth at 2kHz as 4mm has at 500Hz, so what are we really concerned about?
 
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
All that headache is avoided if we use galvanized steel cases, of 1 mm thick.

As aluminum is more expensive and cool and many only buy with their eyes ... Luckily in the laboratory material you do not have those preventions and you bet for sure, which is also much cheaper.
 
Last edited:

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Well, my current opamp of choice is the OPA1612- I've swapped out a number of older duals with them as a number of specs are clearly superior to the LME4562 and similar to the LME49720 my old opamp of choice, and the inferior specs are, in my estimation, less important, so I'm not sure why I need to ask Bruno, even though his work on the analytic optimization of class D is spectacular... I have not swapped out the opamps on the Ncore400 amps that I have, and I believe that my other SS amp, an AHB2, already uses 1612s- although I may be wrong in that.
As for good measurement tools- what do you mean- are you referring to the tools themselves, or the actual kinds of measurements?
I will check out the NC500 thread and perhaps then I can understand your points.
On what basis do you says its clearly superior? What do your measurements show?
 
Last edited:
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,160
Location
Tarragona (Spain)
index.php


4 mm or more
.

With this graph it is evident that 2 mm is insufficient if we are in an environment contaminated by radiofrequency. If it is not the case, with 1 mm will be enough, I say. But we have to foresee adverse situations.

Aluminum is expensive and it seeks to lower costs in production and transport.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
All that headache is avoided if we use galvanized steel cases, of 1 mm thick.

As aluminum is more expensive and cool and many only buy with their eyes ... Luckily in the laboratory material you do not have those preventions and you bet for sure, which is also much cheaper.

What headache? The Rfi problem you are imagining?
 

wynpalmer

Active Member
Technical Expert
Joined
Dec 14, 2018
Messages
175
Likes
214
Did you read Bruno advice above?

When you are looking at these numbers you need to have understanding so you can apply context. Blindly comparing numbers, as you are doing, tells you little.
I understand his specs. I also don't disagree, in principle, with the idea that -90dB of PSRR is a desirable thing, and indeed I add LCR filtering to my SMPS units to drive a phono stage which are simulated/measured to achieve that performance and higher as the gain in the stage is very high- c. 4000 and the output SINAD has to be better than 75dB weighted with power supply artifacts included.
To illustrate this, at a closed loop gain of 100 an AD797 or OPA1612 only has, if my memory serves me well, about 40dB of PSRR at 20kHz and a signal level of c. 0.25mv for a MC cartridge, so for a PSU with say c. 2mv of noise you would like 30dB or more of extra supply rejection to be safe.
I use damped LC filters to achieve this.
However, the opamps cited all have similar PSRR rejection figures, so this is hardly a blind comparison.
 
Last edited:
Top Bottom