• 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!

Vera Audio Class-D Amp Build Quality

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,161
Location
Tarragona (Spain)
Vera Audio P400/1000 Power Amplifier

https://www.vera-audio.com/product.php?product_id=1

Hypex NC500, own buffer, stereo or bridged selection and... silent forced ventilation (two Noctua I think) -> more power.
Gain adjustment on both channels making it easy to match with different preamps/DACs or to incorporate to active speakers with amps to each driver.

Some key performance data:

- Selectable gain in very accurate 3dB steps from 12dB to 30dB as well as mute function
- CMRR of at least 110dB@100Hz and better than 95dB@1kHz
- Very low noise of less than 15uV
- 680W into 4 ohm with less than 0,01%THD+N (stereo)
- 1200W into 8 ohms with 0,01% THD+N (bridged)
- S/N of more than 130dB (stereo)
- S/N of more than 132dB (bridged)

Pictures resized, optimized and clarified to see details:

Vera-Audio-P400-1000-Power-Amplifier-inside.jpg


Vera-Audio-P400-1000-Power-Amplifier-inside-rear.jpg


Vera-Audio-P400-1000-Power-Amplifier-inside-rear-2.jpg


Vera-Audio-P400-1000-Power-Amplifier-rear.jpg


https://www.google.com/search?q=Vera+Audio+P400/1000+Power+Amplifier
 
Last edited:
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,161
Location
Tarragona (Spain)
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,161
Location
Tarragona (Spain)
https://www.diyaudio.com/forums/vendor-s-bazaar/281361-hypex-ncore-nc500-build-post5957553.html

by Armand
We have based our buffer on the newest opamps from TI on a four layer PCB with very short signal paths. This is mounted inside a separate compartment behind the XLR inputs for optimum shielding.

A special arrangement of low noise operational amplifiers and differential amplifiers with 0,004% resistors provides a very robust input signal stage for both balanced and single ended inputs...

BTW, th Big Hypex SMPS is positioned in such a way that the heatsink acts as a shield in front of the NC500 modules too.


Hypex SMPS3kA700 I think. 2 x 85 Vdc, 3600 watts.

https://www.hypex.nl/product/smps3ka700/13

Optimized

Hypex-SMPS3ka700.jpg
 
Last edited:

somebodyelse

Major Contributor
Joined
Dec 5, 2018
Messages
3,682
Likes
2,962
It looks like more thought has gone into this one than in many. It's unusual to see 3d printed parts in production, but probably a good option for low volume production. Using 'heatsink as shield' in this case seems unnecessary - it's in a milled pocket with metal between it and the amps anyway, and the power leads are similarly between walls. Filtering the fan inlet is a nice touch. Where does the exhaust air go? A perforated bottom plate perhaps? It looks like the fans are to cool the output filter coil, but there's no similar cooling for the PSU coils which were the hot bit on @amirm's thermal pictures. I look forward to seeing measurements to back up the claims of improved performance.

Any idea what the microcontroller board is, out of curiosity rather than any performance issue?

There's one real oddity though - what's going on with the uninsulated -ve speaker leadout in front of the ribbon to the inputs? It's not clear in the photo and is somewhat strange.
 
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,161
Location
Tarragona (Spain)
We can not see the top cover. If the decision is mine, the logical thing is to put a grid of small round holes to prevent them from acting as a multitude of RF emitting antennas, in the SMPS side, until the end of the heatsink or something more. The new Cambridge Audio CXA81 Integrated Amplifier has this cover:

Cambridge-Audio-CXA81-top-grill.jpg


Maybe the SMPS side of chimneys of the fans are not closed.

Vera-Audio-P400-1000-Power-Amplifier-inside-cut.jpg


It seems that has a kind of felt in the both sides.

Vera-Audio-P400-1000-Power-Amplifier-inside-rear-cut-2.jpg


The other questions and the one in the chimney that Vera Audio answers.
 
Last edited:
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,161
Location
Tarragona (Spain)
But... there are not any kind of holes on top cover :confused:

Vera-Audio-P400-1000-Power-Amplifier-top-rear.jpg


BTW, the outlet fan holes have a kind of filter / felt like chimneys.

From Google Images cache:

Vera-Audio-P400-1000-Power-Amplifier-top-google-cache.jpg
 
Last edited:

Armand

Member
Audio Company
Joined
Jun 25, 2018
Messages
93
Likes
321
We are excited about the release of this amp. It has been under construction for a long time and we have put a lot of effort into finding optimum solutions throughout the amplifier. We do not believe in snake oil theories or old audiophile myths and there is nothing mystical going on here. Just hardcore engineering using the best components available today.
Our goal is a lot of power, compact design, a clean look with no visible screws or seams, low noise, high CMRR and low THD+N. We are confident this is a well engineered product that will last for many years.
We also want to be transparent and privide ALL measurements and technical information about the product.

A lot of power combined with compact design without compromising on life expectancy due to raised temperatures was a challenge and it became early clear that this is only acheivable by introducing fans.
Fans are often frown upon in in Hi-Fi communities. Who wants an amp with noisy fans that wear out and needs servicing every second year?
After testing many fan solutions we landed on two Noctua NF-A6x25 4-pins with PWM speed control. These fans can operate down to 300RPM and are truly 100% silent up to 1000RPM. At 1600RPM the noise is 8,2dB(A) according to Noctua. https://noctua.at/en/nf-a6x25-flx/specification

The fan tunnels are designed with unequal length to avoid amplification of coinciding frequencies if run at higher RPMs. It is impossible for the fans to create any vibration in the cabinet since it is milled out of a massive aluminium block and the fan tunnels are in fact several centimeters thick. To further safeguard from any possible vibrations they are decoupled from the chassis using rubber mounting screws.
1573855389006.png


The fans are controlled using intelligent fan control software that monitors how the user uses the amplifier and how high the environmental temperature is. The input to the intelligent fan control is taken from three temperature sensors. One inside the cabinet and one mounted on each output coil.
1573866476263.png

As long as the temperature is under a certain limit the fans will not start at all.
I have run torture tests of the amp at 1200W RMS continous into 8 ohms for two hours straight (a sine wave) and the fans kept the temperature under control when running at 2000RPM. While not truly silent anymore the sound emitted is well damped and really not noticable at a couple of meters distance. 1200W RMS sine wave has a crest factor of 3dB and is an impossible music signal. A music signal has almost allways a crest factor of more than 12dB and even when playing music with 1300W RMS in peaks (140V peak voltage) the fans will operate below a truly silent 1000RPM. And to be frank, at those levels it does not really matter anymore...
This a picture picture of the amplifier after two hours of torture. The hottest external temperature was 58 degrees and the hottest internal part was 79 degrees. I measured on 13 different locations inside the cabinet during the test. In the background we can see my dummy load dissipating 1200W of power.
1573860906297.png


No electronic device should not be left on when not in use. This also applies to Class D amplifiers and contradictory to old Hi-Fi myths this amplifier actually measures slightly better when cold. There is no gain in sound quality by leaving it on, and to have the longest life expectancy from the product we strongly advice to turn the amplifier off using the front panel button when not in use. The power supply and the NC500 modules are then shut down and only the microcontroller and front LED is active. These are powered from an extremely power effective small SMPS power supply using only about 0.25W.
The intelligent fan control allow for a certain raise in the temperature the first 48 hours of use without starting the fans at all if the environmental temperature is less than 30 degrees celsius. If the user decides to leave the amp on continously the fans will start a more active state and will activate at about 500RPM to make sure the temperature of the NC500 modules are held at low levels. All this ensures a long life of the product regardless of the use, environmental temperature or power delivery without compromising at all on the noise. Worst possible life expectancy is calculated to 15 years if used 6 hours a day every day.
The fans have a MTBF of 150.000 hours (17 years) at full speed so this should not really be a problem either.

The fan filters are attached using magnets and are easily removed and re-attached if cleaning is neccessary. They are prduced by Demciflex and are a quality product with high air flow and are easly rinsed. https://www.demcifilter.com/how-it-works-2

The fans air flow is concentrated around the hottest parts of the amplifier with 3D printed fan ducts. This enables even very moderate air flow to be effective. Nothing more to them than that.

The air exhaust is in front bottom of the cabinet through generously sized holes. When the fans are not active there is a moderate natural convection cooling effect from the air going the opposite direction and exiting through the fans.

Both fans RPM are individually monitored by the microcontroller. If a fan fails to maintain it's correct RPM this will be indicated on the front LED with RED flashes.

This is the MCU card.
1573867393926.png

With the DIP switches OPT1 and OPT2 is is possible to choose from four different fan "profiles"
1. "SILENT". In this mode slightly elevated temperatures are allowed before the fans increase speed. Of course at the expense of slightly increased temperatures and life expectancy. Otherwise equal to default.
2. "DEFAULT". The recommended setting ensuring silent operation under all normal conditions.
3. "COOL". In this mode the fans will increase RPM at a slightly lower temperature than default. Otherwise equal to default.
4. "POWER" This mode is intended for use in installations with elevated temperatures or in installations with high power requirement and where fan noise is not an issue. The fans are still intelligent and will not run at higher speed than neccessary, but main focus here is adequate cooling from the second it is turned on. The fans are allowed to speed up to their maximum 3000RPM and will keep the amplifier cool under ANY circumstance.


We are using the SMPS3kA700 PSU. We started with two SMPS1200 PSU's but we were not happy with the heat generated in the transformers. When driven hard the temperature reached 120 degrees! The SMPS3kA is surprisingly much better and the hottest spot on the transformer reaches only 88 degrees under same conditions. In fact the SMPS3K is about 50% more powerful than two SMPS1200s.

3D printed parts are used many places in the amplifier. We use high temperature resistant PETG plastic. For instance is the input buffer card and the XLR connectors in the back fastened using only two specially modelled plastic pieces. No screws :)

The microcontroller is a AVR MEGA 2560. Not the most powerful MCU, but it has a lot of I/O for everthing we measure throughout the amplifer. More about this later.

About the unisolated speaker leadout: Nicely spotted :) This is a part of the optimization of the feedback system in bridged mode.

Nobody mentioned the aluminum cooling bar?
1573864141544.png

This cooling bar lowers the temperature on the NC500's coils by 15 degrees celcius.
Heat is transported away from the output coils and into the massive cabinet through the cable channel. Some heat is also transported down to the bottom lid through the alumunum spring that keeps this cooling bar securly in place. This arrangement gives us probably the coolest running NC500's in the market. At least NC500's capable of the same or near the same power output.

Since we are in the technical corner I want to talk a little about the things we measure and control using the microcontroller (MCU).
It continuously measures the status of the NC500 modules and the SMPS power supply.
* Speaker output current (up to 25A).
* NC500 overcurrent conditions individually on both channels (500.000 times per second).
* NC500 signal clipping individually on both channels (up to 20kHz).
* NC500 error status signal from both modules.
* SMPS rail voltages.
* SMPS error status.
These signals are used to give the user information on the front panel RGB LED. Different colours and combination of colours will inform the user if the output is overloaded. Both overcurrent conditions and clipping conditions are reported individually for both channels through different colour combination. The user can then easily indentify if the clipping is due to too high current or too high voltage and if this occurs on left, right or both channels.
It can also mute the two channels. This is important in bridged mode. If one channel fails, the other have to be muted very fast to avoid damage to connected speakers. Much faster than just disabling the SMPS and just wating for the SMPS filter caps to discharge.

Other jobs for the MCU is of course the operation of the front RGB LED, the fans, the on/off button and the on/off of the SMPS.

If you are interested I can talk a little about the design of the input buffer also. But that is for another day.
 

Attachments

  • 1573855347121.png
    1573855347121.png
    52.4 KB · Views: 206
  • 1573866450703.png
    1573866450703.png
    457.5 KB · Views: 216
Last edited:

Kvalsvoll

Addicted to Fun and Learning
Audio Company
Joined
Apr 25, 2019
Messages
878
Likes
1,643
Location
Norway
Excellent product, shows we can still make something useful in this -still- cold country.

Too expensive? Does not by any means seem overpriced to me. Yes, it costs more than buying a bare hypex module, but then you can't really play music on that, can you.
 
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,161
Location
Tarragona (Spain)
https://noctua.at/en/nf-a6x25-flx/specification

Airflow 29,2 m³/h
Airflow with L.N.A. 23,5 m³/h
Airflow with U.L.N.A. 16,7 m³/h
Acoustical noise 19,3 dB(A)
Acoustical noise with L.N.A. 14,5 dB(A)
Acoustical noise with U.L.N.A. 8,2 dB(A)

Two fans: + 3 dB

Acoustical noise with L.N.A. 14,5 dB(A) + 3 dB = 17,5 dB(A)
Acoustical noise with U.L.N.A. 8,2 dB(A) + 3dB = 11,2 dB(A)

Human ear perceives from 19-21 dBSPL >> 11,2 dB(A) [at 1 meter]

I saw the aluminum cooling bar, a very good idea.

I do not like Hypex SMP1200. Good idea to change to SMPS3kA700 PSU.


Only a problem: Norway is not in EU -> customs taxes, how many? Better to ship from Sweden like Abrahamsen Audio, factory near the border I think.
 
Last edited:
OP
maty

maty

Major Contributor
Joined
Dec 12, 2017
Messages
4,596
Likes
3,161
Location
Tarragona (Spain)
There are customs! To Spain: about 4% + 21% (VAT).

From Sweden, not customs but... 25% VAT.

For some other reason AA ships from Sweden, I say.
 

Bjorn

Major Contributor
Audio Company
Forum Donor
Joined
Feb 22, 2017
Messages
1,286
Likes
2,562
Location
Norway
sound & acoustics (lyd & akustikk) who will be selling the Vera Audio amplifiers have a company in Sweden too. However, shipping them to Sweden from Norway (where they are built) in order to ship them to other countries again will simply raise the cost in a way that the 4% savings will be eaten up.
 

somebodyelse

Major Contributor
Joined
Dec 5, 2018
Messages
3,682
Likes
2,962
@Armand thanks for the detailed response. Is the status available on USB serial as well as coloured LED? My experience of Chord's LED colour to indicate bit rate is that it's visually neat but a poor way to convey information. It's even more frustrating for the colourblind.
 

Armand

Member
Audio Company
Joined
Jun 25, 2018
Messages
93
Likes
321
There exitst much better ways to convey information about statuses of course.
It is possible to send these data through USB, I2C and TX/RX but this must be a custom order.
There is a display connection on the PCB also. It is possible to connect a display there and display whatever.
1573904286502.png
 

Armand

Member
Audio Company
Joined
Jun 25, 2018
Messages
93
Likes
321
Here is a measurement of our amp intp 8 Ohms. The jump in measurements just above 1W is due to the measure instrument changing range introducing more noise. Following the linear path the minimum is probably around 0,0004% around 2W.
This is with both channels driven into 8 Ohms loads.
1574450775988.png
 

Armand

Member
Audio Company
Joined
Jun 25, 2018
Messages
93
Likes
321
Just finished two samples being sent for a subjective review in a Norwegian magazine tomorrow.

THD+N bridged into 8 Ohm load.
1574715945538.png

1210W into 8 Ohm with only 0.004% THD+N. 10W is at 0.0004% THD+N

Measured gain with gainsetting @ 21dB
1574716113928.png

Gainmatching is well within limits. Both absolute (0.03dB off) and between channels (0.0016dB off).
 

Armand

Member
Audio Company
Joined
Jun 25, 2018
Messages
93
Likes
321
More measurements:

THD+N 4 Ohm:
1574728475346.png

55.923V into 4 Ohm is 780W with 1% THD+N. About 650W with 0.006% THD+N

A closer look at what happens above 500W.
1574728612136.png

At 680W distortion starts to rise quickly.

S/N
1574728709088.png

Signal to Noise is 129,5dB with 670W 4 Ohm.

SINAD 5W 4 Ohm
1574728811424.png


THD+N Spectrum 5W 4 Ohm
1574728844676.png


IMD 5W 4 Ohm
1574728877420.png


THD+N vs frequency. Bandwidth 61kHz.
1W 4 Ohm
1574729001841.png



10W 4 Ohm
1574729105361.png


100W 4 Ohm
1574729136688.png


300W 4 Ohm
1574729158689.png


550W 4 Ohm
1574729181697.png


Same as above, but now with bandwidth 20kHz
1574729340375.png

We can see that the harmonic components fall outside the upper bandwitdh of 20kHz. At 7kHz, third harmonic is 21kHz and falls outside the measurement. Ar 10kHz second harmonic is 20kHz and is also removed.

CMRR.
We have put effort into having high CMRR.
Here I measure the CMRR with a 1V CM signal.
1574731335391.png

^This is without source resistance on both phases (hot/cold). We can see a relatively large discrepancy between the channels. This is due to resistor differences in the buffer. This is what you get with 0.004% tolerance. The right channel is extremely good at low frequencies so the resisor values must be much tighter the 0.004% in this channel.

Adding 10 Ohms to the HOT phase gives ut this:
1574731568094.png

^Left channel now rises above 1kHz too.

Adding 10 Ohms to the COLD phase gives ut this:
1574731680107.png

^Right channel did nok like this, but it is still at 100dB CMRR at 1kHz.

Worst result at 1kHz of the three CMRR tests above is used. This is left channel with 98dB CMRR.
Fine-tuning of the resistors on left channel could have led to better results on this channel, but 98dB is what you get with 0.004% resistors.
 

Attachments

  • 1574728970924.png
    1574728970924.png
    174 KB · Views: 145
  • 1574728980915.png
    1574728980915.png
    174 KB · Views: 154

March Audio

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