• 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!

Review and Measurements of Purifi 1ET400A Amplifier

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,303
Likes
233,691
Location
Seattle Area
This is a review and detailed measurements of the Purifi 1ET400A Class-D amplifier module. Company was kind enough to send me a complete, assembled kit consisting of two modules, power supply and case, ready to go. They have not disclosed the cost of the unit. Modules will be available to DIY channel in Q4 of 2019.

As cases go, the prototype unit came in an attractive one:

Purifi 1ET400A Class-d Amplifier Module Audio Review.jpg

A touch of color on the lid goes a long way to break the monotony of DIY aluminum cases. Someone should take this case, put two giant VU meters in the front and make me happy! :)

Here is a quick teardown:

Purifi 1ET400A Class-d Amplifier Module Teardown.jpg

A single Hypex SMPS1200 is used to power both channels. That feeds two 1ET400A amplifier modules.

A single "Amp Connector Board" feeds both channels using balanced input through an optional gain stage of 13 dB. Here is more detail from documentation on the features of this EVM board:

1562181895454.png


I was pleased to see high-quality Rubycon capacitors used in the amplifier modules:
Purifi 1ET400A Class-d Amplifier Module Capacitor.jpg


The power supply seems like the stock Hypex SMPS1200:
Hypex SMPS1200  Switching Power Supply Review.jpg


It is remarkable that a high-performance amplifier could be built in such a small enclosure with so few subcomponents.

In use, the evaluation unit would barely get warmer above room temperature. Heat dissipation is simply not an issue here. You could easily put these amps in an enclosed cabinet and not worry.

Note that the main offering here is the 1ET400A. The Amp Connector board is offered as convenience. In that regard, I tested the amplifier with and without the extra gain that the connector board provides. Alas, I wish there was something in between. You get full power with as little as 1.5 volt with the extra gain but need 11+ volts without it. I like to see manufacturers build a gain stage that produces full power at 4 volt which we routinely get out of source products (e.g. DACs).

Amplifier Audio Measurements
As provided, the extra gain stage was active producing this dashboard view at 4 ohm producing 5 watts:
Purifi 1ET400A Class-d Amplifier Module Audio Measurements.png


Focusing on FFT spectrum of a 1 kHz tone, we see stunningly low distortion product. 2nd Harmonic is barely visible which may actually be the contributions from my source signal. The third harmonic hovers around -130 dB which again, is almost at the limit of what we can measure. Given the ear's 116 dB or so dynamic range, you are assured zero audibility of these distortion products.

THD+N and hence SINAD (signal over noise and distortion) is substantially lower though at 104 dB. Since THD+N is sum of distortion products and noise (+N), the latter is the culprit here. We can see this reflected in SNR measurements:

Purifi 1ET400A Class-d Amplifier Module SNR Audio Measurements.png


As you see on the left, our noise contributions is essentially what we see for SINAD. Fortunately at full power (right), performance shoots way up to fantastic level of 123 dB.

We can do better though by disabling the pre-gain stage:
Purifi 1ET400A Class-d Amplifier Module No Pre-gain Audio Measurements.png


We gain 4 dB in SINAD resulting this summary ranking:
Best Audio Amplifiers Reviewed 2019.png


The Benchmark in its low gain mode maintains the championship status still but the 1ET400A gives it serious competition here.

SNR naturally improves with lower gain:

Purifi 1ET400A Class-d Amplifier Module SNR no -pre-gain Audio Measurements.png


Drilling into the distortion (with pre-gain active) we see the vanishing harmonic distortion more plainly:
Purifi 1ET400A Class-d Amplifier Module THD Audio Measurements.png


Power is everything in amplifiers so let's see how we do with 4 ohm load:
Purifi 1ET400A Class-d Amplifier Module Power into 4 ohm Audio Measurements.png


The tables are turned and the 1ET400A produces 257 watts of exceptionally clean power. More could be had if I had run this test with finer resolution and/or allowed the distortion to climb more.

We see the effect of lower gain with reduction of noise at lower power levels although even with the extra grain stage, we are talking very quiet here.

Switching to 8 ohm, the results are similar:

Purifi 1ET400A Class-d Amplifier Module Power into 8 ohm Audio Measurements.png


Notice the big difference between a commercial amplifier (QSC DAC2422). They have more brute-force power but nowhere as clean as the 1ET400A amplifier module.

Intermodulation distortion versus power gives us another shot at analyzing distortion, this time what is in audio band:
Purifi 1ET400A Class-d Amplifier Module IMD Audio Measurements.png


The Benchmark AHB2 maintains its less noisy baseline but as we reach full power, the 1ET400A catches up to it and keeps going with more power.

Distortion+noise versus power and frequency yields this (with pre-gain stage on):

Purifi 1ET400A Class-d Amplifier Module THD vs Frequency vs Power Audio Measurements.png


I was surprised to see the rise in distortion with frequency. I had hoped that the super high gain-bandwidth of the 1ET400A would do away with this. Not an audible concern though as the distortion products here are all in ultrasonic range.

I ran this test multiple times, tuning the input voltage. This eventually upset the amp and caused it to shut down when producing 274 watts into 4 ohm before our sweep completed (in green). Documentation clearly states that extra cooling is needed for continuous power.

The Benchmark AHB2 does a lot better since it has much cleaner ultrasonic spectrum:

index.php



Frequency response is exceptionally flat for a class-d amplifier:

Purifi 1ET400A Class-d Amplifier Module Frequency Response Audio Measurements.png


The output filters in class-d amps can cause variations as we get close to 20 kHz but we see none of that here. Response smoothly rolls off. At 50 kHz, we are down 1 dB.

Crosstalk using my non-optimal setup is still exceptional:
Purifi 1ET400A Class-d Amplifier Module Crosstalk Audio Measurements.png


I had to pull up the graph to show its crosstalk at 1 kHz. The 1ET400A is that good!

EDIT: here is the broadband spectrum of a 1 kHz tone:
Purifi 1ET400A Class-d Amplifier Module 1 khz FFT Spectrum Audio Measurements.png


We have our switching frequency around 500 Khz and its harmonic at 1 Megahertz. Otherwise pretty clean.

Conclusions
It was just a few years ago that people scuffed at class-D amps not being very clean or good for audiophile use. How the situation has changed. First with Hypex modules and now with the Purifi 1ET400A. Audiophile myths are shattered with use of large amount of feedback and high bandwidth to produce an amplifier which brings transparency to anything you throw at it.

Importantly, the 1ET400A does all of this while producing a ton of power and staying cool and efficient to boot.

There is a subjective aspect to measurements that doesn't come across in the graphs. When I run these tests with switching amplifiers, I often watch the analyzer struggle to get reliable reading, or there are jumps and glitches in measurements. None of that was here. The amplifier basically acted like a traditional class AB amplifier. Indeed, I measured it with and without my AES-17 40 kHz filter and the analyzer was happy both ways. This is when I know there is quality engineering that has gone into design of this amplifier.

Overall, it is my pleasure to strongly recommend the Purifi 1ET400A to DIY and OEM manufacturers.

-------------
As always, questions, comments, recommendations, etc. are welcome.

The pink panthers can be pretty lazy at times. I asked them to help clean up the lab and they rolled their eyes and went back to sleep! So I need to hire a housekeeper to keep the walk way clean to my test station. Please donate generously as the hourly rate can be high for such work:

Patreon: https://www.patreon.com/audiosciencereview), or
upgrading your membership here though Paypal (https://audiosciencereview.com/foru...eview-and-measurements.2164/page-3#post-59054).
 
Last edited:

Matias

Master Contributor
Forum Donor
Joined
Jan 1, 2019
Messages
5,011
Likes
10,757
Location
São Paulo, Brazil
Class D for the years 2020+. :)

@amirm could you show the wide spectrum no filter high frequency switching noise, as you did with other class D amps too? Thanks.
 
Last edited:

MZKM

Major Contributor
Forum Donor
Joined
Dec 1, 2018
Messages
4,237
Likes
11,434
Location
Land O’ Lakes, FL
At 509 kHz, we are down 1 dB
I think you mean 50.9kHz.

Also, their interview said 9Vrms for full output, did you mean 11Vrms or 11Vp (7.78Vrms, with a 5x buffer means 1.556Vrms for max output).

The NC400 amps are much more powerful, and is ~$1400 USD shipped for a monoblock pair, so hopefully it’s priced adequately.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,303
Likes
233,691
Location
Seattle Area
I think you mean 50.9kHz.
Thanks. Corrected.

Also, their interview said 9Vrms for full output, did you mean 11Vrms or 11Vp (7.78Vrms, with a 5x buffer means 1.556Vrms for max output).
That was a guesstimate on my part.
 

digitalfrost

Major Contributor
Joined
Jul 22, 2018
Messages
1,519
Likes
3,068
Location
Palatinate, Germany
Wow. Anybody know pricing yet? Not regretting my AHB2 purchase, but I still need an amp for my subwoofers :cool:
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,303
Likes
233,691
Location
Seattle Area
Class D for the years 2020+. :)

@amirm could you show the wide spectrum no filter high frequency switching noise, as you did with other class D amps too? Thanks.
Yes, I had run it but forgot to save it. Will do so later tonight.
 

Feyire

Active Member
Joined
Mar 9, 2019
Messages
272
Likes
314
Location
Netherlands
I see some improvements in SNR, frequency response flatness and power vs distortion when compared to the NC400, but nothing mind blowing. I'm curious how they will price this against the current NC400 DIY kit.

@amirm , did I miss anything obvious?
 

CRKebschull

Member
Forum Donor
Joined
Feb 7, 2019
Messages
87
Likes
94
Location
Germany
If only I had half a brain hemisphere capable of assembling my own amp.

>_>

Building the NC400 monoblock kits was very straight forward. I think anybody could do it, perhaps with a little guidance or help, but mostly it is just screwing the components into the chassis and connecting the wires.
 

Tarzan

Member
Joined
Dec 3, 2018
Messages
25
Likes
25
Location
Brisbane, Australia
+1 for the VU meters, I almost bought a Oppo HA1 for the meter. Then I saw one in person and wished I had.
 

orangejello

Active Member
Joined
Apr 18, 2019
Messages
232
Likes
354
I see some improvements in SNR, frequency response flatness and power vs distortion when compared to the NC400, but nothing mind blowing. I'm curious how they will price this against the current NC400 DIY kit.

@amirm , did I miss anything obvious?
Yeah. I have the AHB2 and the Nc400-based NAD M22 v2 at the moment and am trying to decide between the two. And here is a third option. So in what way does this amp distinguish itself from the other two?
 

digitalfrost

Major Contributor
Joined
Jul 22, 2018
Messages
1,519
Likes
3,068
Location
Palatinate, Germany
Yeah. I have the AHB2 and the Nc400-based NAD M22 v2 at the moment and am trying to decide between the two. And here is a third option. So in what way does this amp distinguish itself from the other two?
If you could write down your impressions between the two, that would be great. I was on the fence and decided to go with the AHB2.
 

reza

Active Member
Joined
Mar 27, 2019
Messages
110
Likes
131
Is Bruno still profiting from the Hypex sales? Because if not, then he might price these DIY kits to compete with the NC400. Otherwise, I imagine they will be priced well above the NC400.
As for the OEM modules, if the NC1200 is any indication, these are going to be costly.
 

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,530
Likes
38,127
Location
Gold Coast, Queensland, Australia
More could be had if I had run this test with finer resolution and/or allowed the distortion to climb more.

Great review. And wow, there's a massive improvement in just about all parameters. This could be the Class D that I might want.

For the THD vs power plot, I do think you should seriously consider either consider increasing the resolution (more measurement points-there's not enough) or place the crosshairs on discrete measurement points with a finite value (not somewhere random along the line between two points).

Here is the Purifi plot of same. How many points they are using is not determinable, but it is certainly a lot more than 5 or so above the 200W mark.

1562191985692.png


Your power determinations are arbitrary and actually, in this case are under reporting the 'clipping' point by a long shot. There's well over 300W@4R still under 0.01% and around 175W@8R under 0.01% based on your plots.

1562191467359.png


With amplifiers such as this, with extremely low measured THD, merely looking at the 'knee' as you see it, is not fair, and not comparable to other amplifiers. Even well up past its 'knee', this amplifier is still several orders of magnitude less THD than other products.

It's time to pick a number as a line in the sand as measure 'clipping' as being that figure. Everyone else has done it. Hypex do it. Stereophile does it. Audioholics do it. Magazines from the past did it. The power cube does it. Amir should do it.

Otherwise superbly measuring products are at a significant disadvantage. If the amplifier won't clip before muting (Benchmark maybe) that can be mentioned too.
 
Last edited:
Top Bottom