Measuring 2-channel Buckeye Purifi Amplifier Performance

[Ajax]

Hey Rick,

Many thanks for the time you put into this review and for sharing. A lot of us are waiting in anticipation for Amir to finish his reno so we can enjoy his reviews again, however, yours has filled the gap nicely.

With regard to price I think we all need to look at the bigger picture, i.e. what are we paying today for excellent performance with plenty of power compared to the audio jewellery flaunted by the magazines at insane prices ($10k+) for far less performance just a few years ago. This Purifi amp will drive any speaker for only $1,100.

I have a Buckeye Hypex NC252MP on order for just over $500, which will produce 150W into 8 ohms and easily drive my speakers to the sound levels I require (90db is heaps) without any distortion or colouration .... add Tidal + Chrome-cast Audio streamer + Topping D50s DAC and you have everything you need for your front end for under $1K

....... life has never been better in the audio world.

​

 

[pseudoid]

Thank you @Rick Sykora,

I notice that the loop-back "noise floor" seems to average out to about -132dBV on your 1st plot.
Yet, this noise floor jumps to around -110dBV during [email protected] input (without contributions of M0/M1/M2 distortion products).

You had stated that QuantAsylum internal attenuation (10db?) was needed to match @amirm 5V measurements w/AP (but w/o overloading the QA401).
Does this average noise floor of the 2nd figure remain the same when the 1kHz signal is not present (removed) while still interconnected?
I ask because the front-end of the QA401 maybe still getting overloaded, hence the noise floor increase by 20+dB.

BTW: That QuantAsylum (and peripherals) seems to be a very nifty analyzer at an equally nifty pricing.

 

[KEFCarver]

Nevermind, just saw that the recent version is the 403, not the 401!

Due to parts shortages the QA404 is going to be the next revision, which has slightly better performance than the QA403, but is not isolated. I love my QA402 and highly recommend them unless your need to measure SNR's better than maybe 110dB, which I do not.

 

[Rick Sykora]

Thank you @Rick Sykora,

I notice that the loop-back "noise floor" seems to average out to about -132dBV on your 1st plot.
Yet, this noise floor jumps to around -110dBV during [email protected] input (without contributions of M0/M1/M2 distortion products).

You had stated that QuantAsylum internal attenuation (10db?) was needed to match @amirm 5V measurements w/AP (but w/o overloading the QA401).

My mistake, I have struck the higher attenuation statement as it was left over from a previous round of measurements. The higher attenuation (26 dB vs 6) is only needed when I run the higher power testing. However, when the QA451 is added, there is additional attenuation of 12 dB. I will check if I may be overloading the inputs, but the analyzer does usually warn when this happens.

Does this average noise floor of the 2nd figure remain the same when the 1kHz signal is not present (removed) while still interconnected?

Yes.

 

[Philipp Paul Klose]

Due to parts shortages the QA404 is going to be the next revision, which has slightly better performance than the QA403, but is not isolated. I love my QA402 and highly recommend them unless your need to measure SNR's better than maybe 110dB, which I do not.

Looking forward to the release of the QA404. Have been in contact with Matt from QA. Part shortages are still crippling them a lot.

 

[thin bLue]

Nice Job!!

Happy to see another Purifi-based amplifiers!

But, it seems load's nonlinearity appears in power sweep. Can you confirm wha kind of load did you use?

If it is determined that the load is causing the problems(the bulge above 30 W), it can be cleaned with higher capacity loads.

that bulge's tendencies look similar to distortion observed in my previous reviews!

 

[MacCali]

Have a buddy putting a system together around some new Monitor Audio speakers, suggested a Purifi-based amp. He was concerned that many of these "companies" are just guys assembling parts and that if he has an issue with the equipment down the road, will they still be in business to do repairs. He was favoring an NAD Purifi amp because of this...any guidance I can can share with him?

I just ordered the NAD M23 as my first “legit” class D amp even with my skepticism on the class. M

I don’t think I have heard the HypeX nc, which everyone says is great

It was measured by Audioholics and my subjective experience with the M33 was very pleasant during a decent demo and speaker swap. Besides bass was indistinguishable between an A/B design

If he waits a bit deals will come with the holidays. It really depends what level he is at and what he expects performance wise.

 

[MacCali]

Also good to keep in mind: for whatever reason, the Purifi implementations by the “big guys” do not appear to perform as well on the test bench as the ones from the smaller companies.

Billy I find that a bit odd, and that depends on the company. Some people just like the rest of the audio industry will by mass quantities of pre built modules for cheaper and cram it into a box with the very same way we can just DIY the damn thing. Plus or minus a couple tweaks that even we can do.

I think that’s where the price saving is coming from. The bigger guys are going in and doing far more tweaks and can be measured as well. So something must be going on

Clearly as you point out somewhere between those lines is someone doing enough and still getting better performance but which one it is I can’t say and seems like a witch hunt

 

[olbobcat]

I am guessing that you are not going to hear much difference in any of these amps. I am sorry but I just can't buy into NAD, I have owned several of their products and have never thought they were the best thing I owned at the time.

 

[Rottmannash]

Billy I find that a bit odd, and that depends on the company. Some people just like the rest of the audio industry will by mass quantities of pre built modules for cheaper and cram it into a box with the very same way we can just DIY the damn thing. Plus or minus a couple tweaks that even we can do.

I think that’s where the price saving is coming from. The bigger guys are going in and doing far more tweaks and can be measured as well. So something must be going on

Clearly as you point out somewhere between those lines is someone doing enough and still getting better performance but which one it is I can’t say and seems like a witch hunt

I believe he was speaking of the Purifi implementations @amirm has measured on ASR-all of the devices which had any custom buffer boards had worse measurements than the original Purifi Eval-1 he measured way back in the beginning of the Purifi revolution. But I agree that unless the measurements are far worse than what we have seen so far it will most likely not be audibly different (unless someone gets the bright idea to put a tube buffer before the module).

 

[57gold]

(unless someone gets the bright idea to put a tube buffer before the module).

Believe that VTV offers that alternative...used to be Vacuum Tube Valley.

 

[Rottmannash]

Ah I forgot about that. I have the OG VTV but w/ standard buffer and it is as advertised.

 

[MacCali]

I believe he was speaking of the Purifi implementations @amirm has measured on ASR-all of the devices which had any custom buffer boards had worse measurements than the original Purifi Eval-1 he measured way back in the beginning of the Purifi revolution. But I agree that unless the measurements are far worse than what we have seen so far it will most likely not be audibly different (unless someone gets the bright idea to put a tube buffer before the module).

Interested in trying that as well even though that probably makes no sense, but a tube pre on these nice class D amps to inject some color.

We will see, as stated with NAD seems like there upper end purifi module does seem to possess a warm undertone but very slight and just enough to take the edge off. It may not be warmth but it does remind me of an A/B at minimum minus that meaty bass. Not that the bass was bad

So far besides the bel canto black and just yesterday I heard the much cheaper bel canto silver tiny mono amps and dac at an audio shop with I believe hypex modules it was doing alright but probably seem to function more towards M23/33 vs C298

 

[TabCam]

Hi Rick, nice review. What did you do to filter out ultrasonic normally present in class-d amplifiers?

 

[Rick Sykora]

Hi Rick, nice review. What did you do to filter out ultrasonic normally present in class-d amplifiers?

Mentioned in the OP, a 6th order filter is built into the QA451 I use.

 

[TabCam]

Mentioned in the OP, a 6th order filter is built into the QA451 I use.

Sorry, I checked but missed that one, thanks!

 

[DualTriode]

The filter may be the best feature.

 

[Ap007nyc]

Are we expecting Amir's review of the 2 or 3-channel models of Buckeye Purifi amps?

 

[DualTriode]

Hello @Rick Sykora and All,

I like your review with the QA analyzer.

For measuring power amplifiers and speakers the QA analyzer is plenty good enough. The QA analyzer has the sensitivity to measure the the distortion levels of most amplifiers and speakers.

A couple or 4 suggestions:

1) in settings of the QA software turn on the paper white background. The black background is hard to look at, plus if you print your plots you will empty you the black laser cartridge.

2) When you use the markers M0, M1, M2 ... place M0 on the fundamental or test frequency, M1 will be HD2, M2 will be HD3 and the dBr numbers shown in the upper right on the plot, that way M1 HD2 dBr ,M2 HD3 dBr, and M3 HD4 dBr will be shown relative to the fundamental test frequency. Plus the left vertical axis can remain labeled dBV. I like the dBr data of the harmonic peaks and dBV data of the sound floor being shown on the same plot at the same time. This is an improvement over the Audio Precession method.

3) You might consider using averaging to tame the noise floor just a bit.

4) @thin bLue is correct. Your load resistors may be generating more distortion than the amplifier under test.

Thanks DT

 

[Rick Sykora]

Nice Job!!

Happy to see another Purifi-based amplifiers!

But, it seems load's nonlinearity appears in power sweep. Can you confirm wha kind of load did you use?

As mentioned in the OP, the load is a QA451. Now that Amir has confirmed the same higher power distortion, it is clearly the amp. While have tried some troubleshooting, the input board designer is better suited to remediate. @pma has a hypothesis that ferrous metal content may cause, but the only significant iron content is on the switch housings.