Fresh From the Test Bench: Purifi 1ET9040BA Balanced Class-D Amplifier Module

[Robocop]

Some may find this an interesting read. At the cutting edge of measurement. Robert

Fresh From the Bench: Purifi 1ET9040BA Balanced Class-D Amplifier Module

The Purifi 1ET9040BA balanced Class-D amplifier module, a Bridge-tied-load (BTL) output stage, is the second product in the company's Eigentakt range. Following a conversation with members of Purifi about the amplifier's audio performance being hard to measure even with the best available test...

audioxpress.com

 

[amirm]

It is good set of measurements but hard to compare due to lack of standardization between us. Using non-zero references makes it hard to interpret as well:

Measurements like this need to be zero referenced (dbRA in AP parlance) so that one can interpret the noise/distortion floor easily. And of course there is the issue of what do with 100 watt measurements when another amp can't do 100 watts.

I tried unsuccessfully to get JA at stereophile to adopt at least one of my tests. So I adopted a couple of theirs (60 Hz into 600 ohm and 19+20 Khz). As is one needs fair bit of mental gymnastics to compare measurement even when they are made using the same analyzer.

 

[Sokel]

Review aside, Didden is one of the most helpful members at diyaudio ever, the frequents there are accustomed with its charts.

Great results by the way, as expected, although for the base module only.
Assemblers will be judged by their input boards I guess.

 

[pma]

Interesting. I wonder why there are those mains frequency multiples in Fig. 1. Though they are low, they should not be there. Probably sub-optimal measurement setup, per my guess.
I am also not amazed by HF linearity of the amplifier.

 

[Sokel]

Interesting. I wonder why there are those mains frequency multiples in Fig. 1. Though they are low, they should not be there. Probably sub-optimal measurement setup, per my guess.
I am also not amazed by HF linearity of the amplifier.

If he tested it as shown in the pic mains noise can be justified,cable routing is a mess, no lid on, etc.
He could also install the modules at the far side of the mains plug.

What I don't get is this chart:

It measures 1,10 and 100W at 4 and 8 Ohm THD (only) vs frequency and the results are odd.
Even if we rule out the HF distortion rise, the rest don't line up at all with the chart exactly above (THD+N vs frequency) for the measured range (1-100W)
I mean the result vary from 0.0002% to 0.2% THD ratio and that's across the whole 20Hz - 20kHz range.

 

[boXem]

What I don't get is this chart:

Neither do I.
Would have appreciated proper spelling of the brand of the tested product too...

 

[Sokel]

Neither do I.
Would have appreciated proper spelling of the brand of the tested product too...

It looks diy, I would expect that from Didden, the fact that it has no input board* suggests so too.
About the chart:

Separating the measurements (for example the worst case red ones, all of them) one is for two channels at 1 watt at 4 Ohm and the other is again two channels at 1 watt at 8 Ohm.
These two sets have about 35-40dB difference!

Edit: it has, didn't have a good look.

 

[boXem]

These two sets have about 35dB difference!

Dashed lines are probably the relative noise, not the THD.
Then the graph kind of makes sense if we also consider that the AP ADC noise changes with auto scaling.

 

[Sokel]

Dashed lines are probably the relative noise, not the THD.
Then the graph kind of makes sense if we also consider that the AP ADC noise changes with auto scaling.

Still confusing, I see 14 measurements there (7 two channel ones) and non is marked as noise.
Maybe using each other channel as noise reference? *

The chart's summary below it does not say so either.
We also don't know measurement test conditions, had to look to find a maybe permanent link to it it but none found.I have an idea from diyaudio but they should be stated.

*But still,what could cause noise in the ~80's? Strange stuff.

 

[Rednaxela]

I feel it’s actually the brand name that has the spelling mistake.

 

[Sokel]

I feel it’s actually the brand name that has the spelling mistake.

\I now see what you both mean, "Purify"