Review and Measurements of Purifi 1ET400A Amplifier

[rajapruk]

Why show without the pre-gain? Will anyone ever be running this with 12.8dB gain?

Exactly! We want to see them compared with the same gain. I think nc400 wins there (lower THD+N).

 

[boXem]

Why show without the pre-gain? Will anyone ever be running this with 12.8dB gain?

Why not? In example, for an integrated system, you can set the gain of the DAC output high enough to use the full power range.
Measuring with pre gain would also measure the front end performance, which in my mind is not really the subject of these tests.

 

[rajapruk]

Why not? In example, for an integrated system, you can set the gain of the DAC output high enough to use the full power range.
Measuring with pre gain would also measure the front end performance, which in my mind is not really the subject of these tests.

The reason here was to compare with nc400, then the same kind of gain should be used.
But it might be awkward showing nc400 winning?

 

[maty]

Optimized picture

There is a good distance between the SMPS and the modules. The power cable is covered with a mesh. The breathing grilles are far from the audio circuitry. You can tell they know what they do when they mount the amplifier so well!

I can not see a RF/EMI inlet filter like Hypex NC1200 -> Probably the PSRR must be higher.

CMRR? If you could measure both I would be bery grateful.

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

https://en.wikipedia.org/wiki/Common-mode_rejection_ratio


Question:

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

 

[amirm]

Why show without the pre-gain? Will anyone ever be running this with 12.8dB gain?

Because it is a pain to screw the case back together. I had taken out the gain and am too lazy to open the case again and put it back the way it was. Once you take the face panel off, everything gets out of square making it hard to fit 8 screws to the front. I will fix tomorrow and re-measure together with fixed speaker cable....

 

[amirm]

The reason here was to compare with nc400, then the same kind of gain should be used.
But it might be awkward showing nc400 winning?

What??? Why would I care which one "wins?"

Per my edit, one of my speaker cable connections had become intermittent causing the distortion to rise in one channel. I have to take it apart and re-build it. Will do that tomorrow and post with pre-gain enabled.

 

[audioBliss]

Why not? In example, for an integrated system, you can set the gain of the DAC output high enough to use the full power range.
Measuring with pre gain would also measure the front end performance, which in my mind is not really the subject of these tests.

Because it doesn't represent the real world performance. I'm sure some people will be able to do a special setup that takes days to configure just right but when I buy an amp I just want to plug it in and run it both with my AVR(with it's built in 8-channel DAC for movies) and with separate DAC (for two channel) with no issues. Correct me if I'm wrong but wouldn't a low gain like that mean that you'd need a really high performing preamp and DAC?

 

[audioBliss]

What??? Why would I care which one "wins?"

Per my edit, one of my speaker cable connections had become intermittent causing the distortion to rise in one channel. I have to take it apart and re-build it. Will do that tomorrow and post with pre-gain enabled.

I think this is pretty interesting. Cables really do matter Even though most cables work about the same when everything is just right I think one should not underestimate the importance of really good connectors that don't become loose/degrade in performance in any way with time. This is also the reason I use electrical wire in my speaker cable which doesn't have strands but is just one solid bit of copper - no chance of a strand coming loose or causing issues because it's just one single fat strand.

Good speaker connectors with a strong spring and knifes inside that cut into the copper. These do not come loose. https://www.conrad.se/p/banankontakt-kontakt-hane-rak-stift-4-mm-roed-731356

Cable: https://shopdelta.eu/elkabel-flat-ydyp-2x2-5_l22_p12800.html

 

[boXem]

Because it doesn't represent the real world performance. I'm sure some people will be able to do a special setup that takes days to configure just right but when I buy an amp I just want to plug it in and run it both with my AVR and with separate DAC with no issues. Correct me if I'm wrong but wouldn't a low gain like that mean that you'd need a really high performing preamp and DAC?

I understand your logic, but
- with the standard voltages +/-15V used in pres, this, in theory, should not be a problem
- even if it's not how the industry thinks today, working with voltages as high as possible as early as possible in the chain is a good way to minimise noise
- and the most important, here you just have a demo product for a module designed to be integrated in a variety of products from the AVR to the monoblock. Each implementation will have a common point, the module, and different results depending on how the module is used. Thus the interesting point is the module performance by itself, since it will not be possible to do better.

So your logic is perfectly valid for any commercial amplifier that will use the 1ET400A, a little bit less here.

I will stop there, my secret fanboism for Mr. Putzeys may become public

 

[amirm]

OK, since I only needed one working cable, I put the gain back to high and measured the one channel:

We can see that while noise dominates, the two amplifiers perform the same. Once distortion becomes dominant, Purifi 1ET400A pulls ahead. Power output is the same for mono.

I will fix the other speaker cable tomorrow and run the stereo test. It is an elaborate Kelvin type connection so not a quick fix at 1:30 in the morning.

 

[amirm]

- even if it's not how the industry thinks today, working with voltages as high as possible as early as possible in the chain is a good way to minimise noise

This is the point John Siau from Benchmark makes and it is the right way to go. Otherwise, we just can't get to these extremely low noise and distortion levels. As a minimum, we need to let go of 1.6 volt generating max power over XLR connection. 4 Volt sensitivity will work with almost every DAC I have tested.

 

[amirm]

I think this is pretty interesting. Cables really do matter Even though most cables work about the same when everything is just right I think one should not underestimate the importance of really good connectors that don't become loose/degrade in performance in any way with time.

It is the connectors that can be the weak link. I have good speaker connectors and they were fine at the amp terminal point. Problem is that a tiny XLR cable wire needs to marry with the thick speaker cable before going to the speaker terming ("Kelvin connection"). I had married these in a butt connector but one end of it was not crimped properly and came loose. But it didn't look that way since the heat shrink portion was still holding. I switched cables and realized that was the problem and then saw the issue visually.

Here is what a kelvin connection looks like:

Using this scheme, the large currents that the dummy load creates on the output of the amplifier don't disturb the instrumentation scheme.

 

[maty]

PSRR and CMRR are in the data.

https://www.purifi-audio.com/eigentakt/

PSRR at 1 kHz is > 90 dB, OK

CMRR at 1 kHz is only > 60 dB. I would like above, > 90 dB.

 

[maty]

Neurochrome Modulus-686

https://www.neurochrome.com/modulus-686/

Common-Mode Rejection Ratio 91 dB 1 kHz

Common-Mode Rejection Ratio 67 dB 20 kHz

 

[maty]

Benchmark AHB2

https://benchmarkmedia.com/products/benchmark-ahb2-power-amplifier

Input CMRR

• 80 dB at 20 Hz, typical
• 80 dB at 1 kHz, typical
• 65 dB at 20 kHz, typical

https://www.audiosciencereview.com/...-and-measurements-of-benchmark-ahb2-amp.7628/

 

[Matias]

Bruno had encouraged me to run intermodulation test with 19 and 20 kHz tones. Alas, I forgot in my haste to post this review. So here it is:

View attachment 28896

I matched levels with Benchmark AHB2. I was surprised to see so much lower intermodulation distortion with 1ET400A. The first pair of sidebands is more than 10 dB cleaner and the rest almost don't exist! Well done. I let Bruno explain why this is so.

On his Q&A he did say that IMD is very low and it is something that we are sensitive to. Great results.

 

[Soniclife]

Bruno had encouraged me to run intermodulation test with 19 and 20 kHz tones. Alas, I forgot in my haste to post this review. So here it is:

View attachment 28896

I matched levels with Benchmark AHB2. I was surprised to see so much lower intermodulation distortion with 1ET400A. The first pair of sidebands is more than 10 dB cleaner and the rest almost don't exist! Well done. I let Bruno explain why this is so.

Impressive. Do you have the same test for hypex?

I second the idle power consumption request.

Output impedance as well please.

 

[Matias]

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 Hyex modules and now with the Purifi 1ET400A.

Typo

 

[Matias]

Added to the review. Here it is:

We see the switching frequency at 500 kHz and its second harmonic at 1 Mhz and not much else. You can also see the increased ultrasonic energy up to 100 Khz which messes up the baseline noise for THD+N when bandwidth is 90 kHz.

Amir, is the switching noise also this high when playing a lower signal, say 1kHz -90dB? I always wondered how it looks.

 

[rajapruk]

What??? Why would I care which one "wins?"

I was just trying to provoce you to show me what I wanted to see