Buckeye Purifi Eigentakt 1ET9040BA monoblock power amplifier Stereophile Measurements

[amirm]

That’s an interesting stance. I would have thought that vendor specs and price were two key contextual elements to any product review?

Product review or my personal recommendation? As I have explained repeatedly, I can't judge the value of a product for all members here. For some, $500 is nothing. For others, it is a ton of money. So unless a product is strictly budget priced, or hugely expensive for performance, I don't consider cost in my recommendations. There are exceptions to be sure but that is my stance.

On published specs, I usually note them when they are out of line in the review. With respect to watts specifically, my testing rarely matches whatever the manufacturer has used so I tend to ignore what they say unless it is hugely out of line. I especially ignore the watts claims in little desktop amps and in many headphone amp reviews.

You all then can use the data in the review and include cost, mismatch of spec, etc. and arrive at what you personally think. I created a poll for that reason so your voice can be heard.

 

[MaxwellsEq]

Sorry. Not. The issue had nothing to do with the case material. As I understand the explanation it was a result of a component manufacturing fault.

The best you might get in respect of the output inductor being wrong is that it was manufactured with the wrong material, gap or number of turns. That sort of mistake happens because production of such low volume items has humans in the loop. If it was the inductor then they probably build them and then test them at zero bias at a fixed frequency on a Simpson VOM. If so I guess that pre-production testing will have been tightened.

A self oscillating amplifier relies on operating on the point of instability. Part of that in terms of operating bandwidth will be the value of the output inductor.

Eigentakt drives that to it limits by including additional gain stages within the loop with those stages being dropped in and out according to operating conditions including output level. Even if you don't get it exactly right it can still be very forgiving to the extent that, in an early stage release, your product will 'look good' rather than immediately going futz before it gets out of the door.

Again rather than babbling about the case material being the issue you would be better to guess, and likely still be wrong, that level may have been the indicator given that the amplifier adjusts its loop gain to maintain oscillation at a particular, working/workable frequency without totally losing the plot.

Loss of loop gain, something got switched out prematurely, might go some way to explaining the higher distortion for the test that was being done. Cause and effect.... The amplifier exceeded its claimed distortion figures before it hit its claimed output power. Presumably if that was ignored and it was just driven harder then. assuming it did not hit thermal limiting, it would have met the power spec.

OK. An example of an early stage product as released for review did not hit the spec but they have been called for it, identified the issue and taken action to catch future mistakes before they get out of the door. Well done everyone. Thanks for the input. Good Catch. No we will not be using a different shade of blue on the cases unless you require that specific option.

In respect FTC requirements, which appears to be an aside to this discussion, assuming a manufacturer states the conditions under which they will claim specific performance for their product then they should be measured against such claims, as as happened here, rather than any poorly considered constraints placed upon them by a regulator trying to level the playing field. To me that just introduces a different opportunity for snake oil to the equation when it would be better to, rationally, challenge the snake oil that is already being used.

Interesting and informative. Thank you!

 

[amirm]

1) frequency response at mid power level,
2) THD, THD+N at 1kHz at -1dB below rated power,
3) CCIF 19+20kHz IMD at -1dB level below rated peak output voltage amplitude.

These seem quite random to me. You have done a effectiveness study to see how useful/coverage of these ad-hoc tests are?

Proper tests would be frequency response at low power levels. SNR. Power vs Distortion+noise., polarity, protection, etc. And then when you have some experience on specific failure modes, tests developed to catch those.

Stuff you talk about is design verification which would need to be backed by actual usage scenario of the amplifier. #3 certainly would not fall in that category.

 

[Axo1989]

Sorry. Not. The issue had nothing to do with the case material. As I understand the explanation it was a result of a component manufacturing fault.

The best you might get in respect of the output inductor being wrong is that it was manufactured with the wrong material, gap or number of turns. That sort of mistake happens because production of such low volume items has humans in the loop. If it was the inductor then they probably build them and then test them at zero bias at a fixed frequency on a Simpson VOM. If so I guess that pre-production testing will have been tightened.

A self oscillating amplifier relies on operating on the point of instability. Part of that in terms of operating bandwidth will be the value of the output inductor.

Eigentakt drives that to it limits by including additional gain stages within the loop with those stages being dropped in and out according to operating conditions including output level. Even if you don't get it exactly right it can still be very forgiving to the extent that, in an early stage release, your product will 'look good' rather than immediately going futz before it gets out of the door.

Again rather than babbling about the case material being the issue you would be better to guess, and likely still be wrong, that level may have been the indicator given that the amplifier adjusts its loop gain to maintain oscillation at a particular, working/workable frequency without totally losing the plot.

Loss of loop gain, something got switched out prematurely, might go some way to explaining the higher distortion for the test that was being done. Cause and effect.... The amplifier exceeded its claimed distortion figures before it hit its claimed output power. Presumably if that was ignored and it was just driven harder then. assuming it did not hit thermal limiting, it would have met the power spec.

OK. An example of an early stage product as released for review did not hit the spec but they have been called for it, identified the issue and taken action to catch future mistakes before they get out of the door. Well done everyone. Thanks for the input. Good Catch. No we will not be using a different shade of blue on the cases unless you require that specific option.

In respect FTC requirements, which appears to be an aside to this discussion, assuming a manufacturer states the conditions under which they will claim specific performance for their product then they should be measured against such claims, as as happened here, rather than any poorly considered constraints placed upon them by a regulator trying to level the playing field. To me that just introduces a different opportunity for snake oil to the equation when it would be better to, rationally, challenge the snake oil that is already being used.

No, the assumption referred to is per @pma, quoted by @SIY "huge difference in CCIF IMD compared to the datasheet parameters" which followed confirmation that case material (confirmed to be aluminium) was not the issue. That difference certainly existed, and is now attributed to the faulty Purifi board in that sample. Keep up.

 

[amirm]

Thank you John. While doing this it would be great to hear what you, Jim, or Kal think about the following:

I quoted in the other thread JA's stance on FTC tests. He said he was supportive which was very surprising to me especially since he said he had spoken to Raph (Atmosphere) about it. Surely class D amps Ralph sells would fail FTC tests at 20 kHz at full power for 5 minutes. And so would many other products that Stereophile reviews and gives high ratings to. The current review as we have noted, is not compliant with FTC requirements either in bandwidth or duration. Would be nice to see if JA finds amps that can pass these tests, assuming such a test can even be created.

 

[amirm]

The assumption referred to is "huge difference in CCIF IMD compared to the datasheet parameters". This was correct.

You think? Anyone could tell that. @pma though, went to forcefully suggest the case is the problem, that he knows better and post the same in the comment section at Stereophile. The assumptions of what could be wrong ranged from company wanting to screw customers to the aforementioned case. All were wrong.

 

[Axo1989]

You think? Anyone could tell that. @pma though, went to forcefully suggest the case is the problem, that he knows better and post the same in the comment section at Stereophile. The assumptions of what could be wrong ranged from company wanting to screw customers to the aforementioned case. All were wrong.

You've made the same error as @SMPSGuy. I refer to the specific exchange between those posters that occurred immediately after the case material issue was resolved. The way I read it, @SIY was contesting the assumption (and he used the singular, not the plural) that the device was faulty, and inferred that (for example) the measurements may be. That was reasonable, but wasn't the case here. I'm referring to that assumption, nothing else. If I misread @SIY's meaning, he can correct me of course.

 

[Sokel]

This assumes that the fault is audible. There's so much discussion recently on audibility of performance metrics. There are some things which should be SOTA, but which when flawed, we humans simply can't hear the flaw with music.

Yes.And that is fueled by the gazzilion of comments by people getting used to this astronomical depths (often beyond the ability of even the best measuring systems to measure them) and claiming improved listening experiences (like going from -100 to -105dB THD+N) .

And it's funny cause at the exact same time at the listening tests of one of the best measuring companies (Klippel) thousands of people took (and I expect these people is not the ramdom joe,but people who are involved with audio) the test and the results are maybe closer to the good,old pro standard (-60dB THD+N with distortion dominating and noise at check)

So...

 

[amper42]

You think? Anyone could tell that. @pma though, went to forcefully suggest the case is the problem, that he knows better and post the same in the comment section at Stereophile. The assumptions of what could be wrong ranged from company wanting to screw customers to the aforementioned case. All were wrong.

While @pma case guess on the HF performance issue with Buckeye 9040 may have missed the mark, he was the first one to suggest the poor performance with the initial Buckeye Purifi amp design was a result of steel parts in the binding post assembly. Amazingly, once those parts were replaced with brass the issue was substantially improved. I give him credit for that suggestion while the rest of us were scratching our heads.

 

[pma]

We either speak about audibility or about technical correctness of the design and implementation. If we stick with audibility only, then distortion measurements (and one of the reasons that ASR exists) are pointless, as any distortion below -80dBr is inaudible. But we obviously want more, we want proper implementation without tons of excuses.

 

[Sokel]

We either speak about audibility or about technical correctness of the design and implementation. If we stick with audibility only, then distortion measurements (and one of the reasons that ASR exists) are pointless, as any distortion below -80dBr is inaudible. But we obviously want more, we want proper implementation without tons of excuses.

We can very well say then (that's what I do) that we enjoy watching these lines down at -140db of the analyzers and call it a day.
But that's not exactly audio,that's pure engineering admiration (which I'm guilty too) .

It's the mixing that results at this mess.

 

[amirm]

While @pma case guess on the HF performance issue with Buckeye 9040 may have missed the mark, he was the first one to suggest the poor performance with the initial Buckeye Purifi amp design was a result of steel parts in the binding post assembly. Amazingly, once those parts were replaced with brass the issue was substantially improved.

We have no evidence that "steel parts" were the problem. I routinely find increased distortion due to many factors related to how binding posts and cable plugs are interfacing with each other, regardless of material. They all make a difference at these extreme situations. On Benchmark amp for example, the banana jacks deliver worse performance compared to SpeakOn connectors and Benchmark uses brass binding posts.

 

[amirm]

We either speak about audibility or about technical correctness of the design and implementation. If we stick with audibility only, then distortion measurements (and one of the reasons that ASR exists) are pointless, as any distortion below -80dBr is inaudible. But we obviously want more, we want proper implementation without tons of excuses.

No one has made excuses. What we said is that you jumping the gun and claiming incompetence and case being the problem are out of line:

OK, let us wait for the answer from @Buckeye Amps related to the amplifier sample under test. Otherwise, there is no other explanation [case not being aluminum] for the very high skirts in the 19+20kHz IMD, which do not correspond with the datasheet.

The lesson you need to learn is to not let your emotions and competitive attitude get in the way of proper discussion of facts.

 

[amper42]

We have no evidence that "steel parts" were the problem. I routinely find increased distortion due to many factors related to how binding posts and cable plugs are interfacing with each other, regardless of material. They all make a difference at these extreme situations. On Benchmark amp for example, the banana jacks deliver worse performance compared to SpeakOn connectors and Benchmark uses brass binding posts.

Buckeye offered brass connectors as the fix. They were the only change between the initial Buckeye Purifi and the follow up corrected unit measured on ASR. The results were also verified by Rick Sykora's testing. Obviously, SpeakOn offers even better connections for sure and the original Buckeye Purifi with SpeakOn parts never had the issue.

 

[amirm]

Buckeye offered brass connectors as the fix. They were the only change between the initial Buckeye Purifi and the follow up corrected unit measured on ASR. The results were also verified by Rick Sykora's testing.

Once more, there is no evidence that being or not being steel was the problem. Just mere swapping the part for another is responsible for this. No analysis or data exist to prove that steel is the problem or you would be seeing me test for that. I have to routinely re-torque the screws on my locking binding posts. That makes a huge difference yet the material has not changed.

 

[SMPSGuy]

Second, another related thread here involves reported class D power specifications. If I understand the discussion correctly, many class D amplifiers can produce large amounts of power on demand but are only able to deliver a fraction of that power on a continuous basis. The bone of contention seems to revolve around the relevancy of this in real world listening environments. Again, hearing what you think in the pages of Stereophile would be welcome.

Also "Many not Class D amplifiers can produce large[r] amounts of power on demand but are only able to deliver a fraction of that power on a continuous basis." That Chicken had already hatched before the Class D egg became viable.

Ooops. Fast moving thread. Someone has already suggested that Class D might be no less blameless than other classes in respect of such gamesmanship.

Just saying with an add that Class D will probably allow you to better determine the limits and catch things before the magic smoke escapes. That's largely a result of the associated sums being easier to come to terms with and the ability to implement more precise protections.

 

[Multicore]

There's so much discussion recently on audibility of performance metrics.

Well, not just recently. But yes, and I'm pleased that scientific methods are being applied to examine exactly that question.

 

[IPunchCholla]

Yes.And that is fueled by the gazzilion of comments by people getting used to this astronomical depths (often beyond the ability of even the best measuring systems to measure them) and claiming improved listening experiences (like going from -100 to -105dB THD+N) .

And it's funny cause at the exact same time at the listening tests of one of the best measuring companies (Klippel) thousands of people took (and I expect these people is not the ramdom joe,but people who are involved with audio) the test and the results are maybe closer to the good,old pro standard (-60dB THD+N with distortion dominating and noise at check)

So...

I was in the 77th percentile for sine waves over speakers when I took the test. -42dB. On AirPod Maxs and music I was at -15dB.

 

[kyuu]

In fact the wide set of measurements in Stereophile, like that 50W/8ohm CCIF IMD, has discovered something that would probably remained hidden with a "soft" 5W CCIF test as performed occasionally at ASR.

To clarify, you think if ASR has been testing this particular simple, the issue wouldn't have been caught with this test:

The module issue should had been found by both the manufacturer and the assembler. Then, they would prevent that "publicity".

There's no way that every single unit can undergo a complete testing suite without seriously increasing costs. Things will inevitably slip through, and as long as it is rare and is handled appropriately, it's not something to get up in arms about. Certainly not an excuse to start throwing wild accusations and questioning people's integrity.

 

[Doodski]

There's no way that every single unit can undergo a complete testing suite without seriously increasing costs.

To have a tech or a laborer doing fast repetitive tests would not increase the unit cost very much.