2nd generation EIGENTAKT PURIFI 1ET6525SA (successor to the 1ET400A)

[Oniiz86]

It appears @Audiophonics had revealed the 1ET6525SA's module specifications in a bit more detail a few months ago, I especially like the significant improvement in output noise at 6.6uV over the 1ET400A's 11.5uV, it truly rivals the 1ET9040BA & Hypex's latest Nilai & NCx500 efforts

 

[rsg]

If this is all true, then between the Hypex NCx500 and Purify 1ET6525SA, this may be the end of Class A/B amplifiers for audio. I'm anxious to try this all out considering I've spent several thousands on 60-100 pound amplifiers over the last 25 years.

 

[rsg]

It appears @Audiophonics had revealed the 1ET6525SA's module specifications in a bit more detail a few months ago, I especially like the significant improvement in output noise at 6.6uV over the 1ET400A's 11.5uV, it truly rivals the 1ET9040BA & Hypex's latest Nilai & NCx500 efforts

From 11.5uV to 6.6uV would correspond to a 2.4 dB reduction in output noise. This is likely in a low gain vs. a high gain setting. It would be interesting to see what happens in a high gain setting.

 

[pogo]

The heat from an Eigentakt based amplifier is mainly caused by the power supply, not by the modules.

That's correct.

 

[boXem]

This is likely in a low gain vs. a high gain setting. It would be interesting to see what happens in a high gain setting.

That's the module specification, not the amplifier. So 12,3 dB gain, not settable.

 

[atomek]

I'm thinking of replacing my current amp with this amplifier, would SMSL SU-X be a good match as a DAC for it? It has 5.2 Vrms - does it mean I should match it with amplifier with low gain?

 

[boXem]

I'm thinking of replacing my current amp with this amplifier, would SMSL SU-X be a good match as a DAC for it? It has 5.2 Vrms - does it mean I should match it with amplifier with low gain?

As long as the amplifier you choose has a sensitivity below 5.2 V RMS, i.e. a gain above 18,4 dB, you will be able to get full power out of it.
These modules are to be integrated with an input/gain stage that is specific to each manufacturer. There is no direct link between the upstream component (DAC or pre) and the module.

 

[Oniiz86]

From 11.5uV to 6.6uV would correspond to a 2.4 dB reduction in output noise. This is likely in a low gain vs. a high gain setting. It would be interesting to see what happens in a high gain setting.

@rsg I wasn't aware that the difference is only a reduction of 2.4dB in output noise, so only a modest improvement then, as @boXem had already pointed out both figures refer only to the module's residual noise unbuffered, the 1ET6525SA's 6.6uV at 12.3dB Gain & 1ET400A's 11.5uV at 12.8dB.

 

[boXem]

11.5 to 6.6 is a reduction of 4.8 dB.

 

[Oniiz86]

11.5 to 6.6 is a reduction of 4.8 dB.

I wasn't aware that you could convert uV to dB, what's the formula?

I thought 11.5 to 6.6 is a reduction of 4.9dB?

 

[rsg]

11.5 to 6.6 is a reduction of 4.8 dB.

That's correct. Sorry for using power instead of voltage.

 

[rsg]

I wasn't aware that you could convert uV to dB, what's the formula?

I thought 11.5 to 6.6 is a reduction of 4.9dB?

20log(11.5uV) - 20log(6.6uV) = 4.8dB

 

[boXem]

20log(11.5uV) - 20log(6.6uV) = 4.8dB

Although it is mathematically correct, your formula hurts my engineers eyes since it implies that it allows converting uV to dB.
dB is the ratio of two physical quantities, whatever their unit, as long as it is the same.
dB = 20 log (quantity1/quantity2)
Then the math will dérive the formula you use.

 

[rsg]

Although it is mathematically correct, your formula hurts my engineers eyes since it implies that it allows converting uV to dB.
dB is the ratio of two physical quantities, whatever their unit, as long as it is the same.
dB = 20 log (quantity1/quantity2)
Then the math will dérive the formula you use.

Again, you're correct. I was quickly providing the reference to reference to Oniiz86.

Where Vin = 1V or 1,000,000uV

Many engineers who are working in RF, telecommunications and audio will use Decibel for logarithmic expression in reference to voltage, power, field strength and others, but often leave off the suffix for what the reference is. I am one of the worst abusers.

 

[DonH56]

Again, you're correct. I was quickly providing the reference to reference to Oniiz86.
View attachment 404898
Where Vin = 1V or 1,000,000uV

Many engineers who are working in RF, telecommunications and audio will use Decibel for logarithmic expression in reference to voltage, power, field strength and others, but often leave off the suffix for what the reference is. I am one of the worst abusers.

Using ratios as @boXem said the units drop out. dB is defined as a ratio, not difference of absolutes (which as you noted must use the same reference to provide the equivalent ratio). One advantage of using ratios as is normally done is that you don't have to worry (as much) about taking the difference of two very large (or very small) numbers to calculate the result. Since it all drops out in the end, for your example I would use 1 uV as the reference ("Vin").

There are many engineers of all types on ASR, FWIW.

 

[Streamc]

I'm thinking of replacing my current amp with this amplifier, would SMSL SU-X be a good match as a DAC for it? It has 5.2 Vrms - does it mean I should match it with amplifier with low gain?

Direct connection of SMSL SU-X to ET400 module is my configuration. I use this bundle. High gain.

 

[rsg]

Since it all drops out in the end, for your example I would use 1 uV as the reference ("Vin").

Yes, you can do that, but there is a difference. For example, 20log(11.5uV/1uV) = 21.2 dB as opposed to 20log(11.5e-6V/1V) = -98.8 dB
If you use 1uV as the reference, then it would be best to use the suffix of u or uV. For this example it would be ideal to be written as 21.2 dBu or 21.2 dBuV.

Again, my apologies for being pedantic.

 

[DonH56]

Yes, you can do that, but there is a difference. For example, 20log(11.5uV/1uV) = 21.2 dB as opposed to 20log(11.5e-6V/1V) = -98.8 dB
If you use 1uV as the reference, then it would be best to use the suffix of u or uV. For this example it would be ideal to be written as 21.2 dBu or 21.2 dBuV.

Again, my apologies for being pedantic.

Obviously I was not clear. To subtract the logarithmic (dB) values as you initially did, you need to use the same reference, or add (or subtract) a constant to account for the different references. And yes you have to state your reference, hopefully any engineer (etc.) knows that, and how it influences the result. The choice of reference often depends on the application (I have used dBU, dBV, dBuV, dBW, dBm, dBf, etc. etc. etc.) and usage, e.g. amplifier gain is usually expressed as a positive number, whilst signal level is often absolute relative to 1 V or whatever. You can assume I am reasonably familiar with various units and their logarithms. But at this point it's just trying to see who's most pedantic.

 

[Matias]

Here is March Audio P482 implementation of this module.

New P482 Stereo Power Amplifier

Hi everyone Exciting times at March Towers. The new Purifi 1ET6525 modules have finally arrived. A special thanks to Fedex for losing them for 4 weeks!!! Anyway, we have been busy testing them in our new P482 amplifier. This replaces the P422 in a few weeks. The new amp has a revised power...

forum.marchaudio.com

 

[boXem]

I think k Boxem posted a quick measurement of their version in another thread, but it wasn't close to this.

It was indeed close. You are comparing apples to oranges.

115 dB is obtained by summing the THD measured on one side and the noise measured with inputs shorted on the other side.

The March equivalent to what I measured is a bit below 112 dB while I measured around 110 dB with a different generator.

I would also like to know at which gain these measurements are done, since it can make a huge difference.