• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Review and Measurements of Accuphase E-270 Amplifier

svart-hvitt

Major Contributor
Joined
Aug 31, 2017
Messages
2,375
Likes
1,133
#61
The measurements Amir performed are extremely limited. (no doubt due to time constraints). I certainly feel a categorical recommendation or lack thereof is a very unwise position to take.

What we can see, is the performance exceeded rated specifications by a wide margin (factors of 10+) for the parameters Amir tested. This was as expected and as I predicted. Frequency response tests appear to exceed spec also, but Amir has not specified the power output level he tested FR at. The standard and the spec is 1W. We can only guess what he used.

The high level preamplifer inputs were driven with a signal to give the same level out as in (2V). There is absolutely no point using a preamplifier that is specified with an 18dB gain and then winding back the attenuator to give unity gain for testing purposes. All preamplifiers are tested and specified at the wide open position and have been forever. The rated output of the preamplifier is 1.07V from a 134mV input. He fed it 2V and wound back the attenuator. You test a preamplifier at its rated input and output, not some made up number that suits you. We may be merely seeing non-linearities and noise (hum) etc, from being buried in the multiple stages of the AAVA volume control (see below).

View attachment 20348
I had hoped Amir would investigate this unusual implementation and test it at various increments/levels to give some insight into its performance or lack thereof. I'm not convinced piling VI buffers in line is a good idea, but who knows. Accuphase have been touting their AAVA for several years now.

View attachment 20347

One cannot characterize the performance of an amplifier with only one resistive load test (4 ohms), which is half the standard, and make no comparisons between different loads and different numbers of channels driven. There is no reactive testing, transient (IHF/EIA) testing, power bandwidth, shorted input residual, comparisons of various inputs, crosstalk

There is zero discussion on the topology as the cover was likely not removed (understandable if it was on loan) and therefore we can only surmise as to the reasons for the burst of HF hash on the FFT. I have offered a possible reason in a previous post, but Accuphase is unlikely to have released a product into the wild with such an issue. It may be operator influenced perhaps.

The power amplifier stage is conventional and of a decent standard according to what I see on Accuphase's brochure.

View attachment 20349

What we can see is Accuphase appear to have decoupled supplies for the VA, the driver stage and the output stage and we know they will have high quality regulated supplies for the preamplifier stages. Amir makes the statement the preamplifier supplies are being affected by the output. This is perfectly normal in integrated amplifers with single transformers when operated at high powers into low impedances.

Comparisions to the Hypex are flawed. The Hypex units have not demonstrated they can hit their specifications. Nor have they demonstrated they can be trusted not to shut down at inopportune times. Their power supplies overheat and are failing in the marketplace. The Accuphase wasn't pre-condtioned, or tested for 5 minutes at full rated power either.

I take the whole review with a grain of salt.

That said, further investigation would clear up a whole lot of the above. Time contraints are a b#tch.
Hmmm...seems like lack of measurements standardization - which is for everyone to see and debate - is haunting @amirm ’s review?

Such standardization would consume more time. But it would probably be worth it, wouldn’t you agree?

Standardization is a theme that I come back to so often, I wanted to add that «furthermore, I consider that Carthage must be destroyed» ;)
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
20,594
Likes
25,379
Location
Seattle Area
#63
The high level preamplifer inputs were driven with a signal to give the same level out as in (2V). There is absolutely no point using a preamplifier that is specified with an 18dB gain and then winding back the attenuator to give unity gain for testing purposes. All preamplifiers are tested and specified at the wide open position and have been forever. The rated output of the preamplifier is 1.07V from a 134mV input. He fed it 2V and wound back the attenuator. You test a preamplifier at its rated input and output, not some made up number that suits you. We may be merely seeing non-linearities and noise (hum) etc, from being buried in the multiple stages of the AAVA volume control (see below).
As I have explained in the past, for Dashboard view I try to give the benefit of doubt to products by optimizing their parameters fair bit. This was no exception. Giving the unit much lower signal as you state, is liable to produce worse results, not better due to rise in noise levels. Here is that outcome precisely as you state:
Accuphase E-270 Integrated Amplifier Pre-amp RCA with Max Volume Measurements.png


This was the same test from the review at unity gain and 2 volt input:



With the lower input as you suggested, and asking the unit to provide nearly 18 dB of gain, the level of noise and background spikes rises way up, causing a nearly 20 dB reduction in SINAD.

Personally I am not seeing value in chasing manufacturer specs for measurements. How would we then compare one piece of equipment to another? I have picked unity gain before in my headphone tests and tried to do that here. A DAC will produce full 2 volt output and then you use the volume control to lower than level. Why is that worse than what you suggested above?
 

restorer-john

Major Contributor
Joined
Mar 1, 2018
Messages
2,359
Likes
3,750
Location
Gold Coast, Queensland, Australia
#64
Hmmm...seems like lack of measurements standardization - which is for everyone to see and debate - is haunting @amirm ’s review?
He has the gold standard test equipment (the APX-555). He clearly needs a decent load bank and a reactive load setup for proper amplifier testing. There have been offers in other threads to help him with this.

Clearly, there are not enough hours in the day and amplifier testing done properly, takes a long time. The reviews would be huge. Take a look at amplifier reviews from HiFi publications and realize they are only a summary of hundreds of connections, cables, loads and burnt fingers. I would offer my help too, but I'm on the other side of the world. If I could afford an AP and wasn't in Australia, I'd be happy to test amplifiers for ASR and leave all the digital stuff to Amir.

Such standardization would consume more time. But it would probably be worth it, wouldn’t you agree?
Absolutely.
 

sergeauckland

Addicted to Fun and Learning
Patreon Donor
Joined
Mar 16, 2016
Messages
924
Likes
1,665
Location
Suffolk UK
#65
I think this review and John's comments above indicate that measuring amplifiers is a whole lot more complicated and consequently time consuming than digital products like DACs. There are far more variables, like load impedance, mains input voltage, volume control setting (I agree testing should always be done with the volume control at full) and at different power levels. Unless one has some form of automation, I expect fully testing an amplifier to take several hours, possibly a whole working day if RIAA phono stages and tone controls are tested too.

At very least, amplifiers should be tested at their permitted loads, 2,4,8 ohms both channels operating, and with 2uF in parallel to simulate a reactive load. Test reports in the 70s and early 80s, before subjective reviewing became the fashion, used an 8ohm/2uF load as their standard, and measurements at 1w and just below clipping.

As amplifiers to drive conventional loudspeakers will still be required in the foreseeable future, ASR could do well to test these as in depth as it does with digital products.

S
 

restorer-john

Major Contributor
Joined
Mar 1, 2018
Messages
2,359
Likes
3,750
Location
Gold Coast, Queensland, Australia
#66
A DAC will produce full 2 volt output and then you use the volume control to lower than level. Why is that worse than what you suggested above?
Because the correct test you performed above verifies the performance of the unit against its rated specifications.

Using standard levels mean you can compare one tested product with another. Tweaking individual products for best performance and then measuring that is meaningless.

I understand you like 2V, but it is an output arbitrary voltage from CD players that has become common. It is not an input voltage that manufacturers specify for their preamplifiers. Power amplifiers have input ratings from 500mV to several volts. Line level inputs have always been 150mV for full rated output. Preamplifiers have had the defacto standard of 1.0V rated output. Many of course go way higher, but they are rated and specified at 1.0V.

Why is it so hard? Your reviews are wonderful, but if they are not adhering to standard practices, people will ultimately dismiss, discount or argue with them. I don't want that and I don't think you do either. Pull out any respected HiFi magazine amplifier review from the 70s, 80s or 90s and follow those practices all the while with an infinitely better piece of automated test gear at your disposal.

From my perspective, ASR reviews could become an ultimate authority in time, but only if the reviews are beyond criticism or at least approaching it.
 
Last edited:
Joined
Jan 11, 2019
Messages
79
Likes
76
#67
Hmmm...seems like lack of measurements standardization - which is for everyone to see and debate - is haunting @amirm ’s review?

Such standardization would consume more time. But it would probably be worth it, wouldn’t you agree?

Standardization is a theme that I come back to so often, I wanted to add that «furthermore, I consider that Carthage must be destroyed» ;)
There's no standard, but there is a pretty good method for doing them to determine how an amplifier would meet the old FTC "20Hz to 20kHz, full power, not to exceed" standard. SoundStageAV has good test sets that are more useful than those found elsewhere. https://www.soundstage.com/measurements/amplifiers/bryston_2bsst/. Stereophile adds to this the very important 19+20kHZ IMD test. Do both of those test and you have a very good visual representation of how the thing is going to perform.

Now, if something like the Hypex excels on the IMD test but falls on its face on the full power 20kHZ THD test, that's when I would start investigating just what is going on. Traditional amps perform predictably. These things? Not so much. I haven't had the time to read much in depth to understand just why they behave like this. There is a lot going on in those things, including huge amounts of gain, gobs of feedback (not a bad thing) and an insanely complex 6 pole feedback loop to make good use of it. But it's pretty clear that once you measure one of them, measuring very many more of them is a waste of time since it's an amp on a chip.

One area I would disagree with some other posts is that it's really all that necessary to test 8 ohms, 4 ohms, 2 ohms, etc, unless you have a big need to look at power output levels. 4 ohms is a good compromise point. If the amplifier does well into four, it will blow it out of the water. 2 ohms is basically a power supply test.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
20,594
Likes
25,379
Location
Seattle Area
#68
Did you test the headphone out?
Just did. Here are the results:

300 Ohm Load:
Accuphase E-270 Integrated Amplifier Headphone Output at 300 Ohm Measurements.png


Output level is quite good for high impedance. What is not of course is much higher noise level as compared to Massdrop THX AAA 789.

And 33 ohm:

Accuphase E-270 Integrated Amplifier Headphone Output at 33 Ohm Measurements.png


We have quite a shortfall here. Seems there is plenty of voltage to do well at 300 ohm, but not enough current to do well with 33 ohm load. Power output before clipping is just 45 milliwatts.

Output impedance is very good though:
Accuphase E-270 Integrated Amplifier Headphone Output at Impedance Measurements.png


If you are spending this kind of money on an integrated amplifier, I would spend a few hundred dollars more and get the THX AAA 789 to add to it. But in a pinch, the one that is built-in would do well with high impedance headphones.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
20,594
Likes
25,379
Location
Seattle Area
#69
From my perspective, ASR reviews could become an ultimate authority in time, but only if the reviews are beyond criticism or at least approaching it.
I want to make sure my implicit goal here is clear.

My intention is to sufficiently test a device until the right conclusion pops out of it. Specifically should it be on your buy list with no engineering faults or not. I do a lot of work behind the scenes to get this answer and will not post a review until it nets out this way.

It is not my goal to produce comprehensive library of tests. To do that would mean reviewing one piece of equipment a week if we are lucky. Let's remember that it is not just the test time but documentation that is resource intensive. Graphs need to be prepared, annotated, explained, etc.

I have a room full of equipment waiting to be tested and more is constantly on the way. I rather get 100X broader coverage than test a few pieces of gear extensively.

Back to your statement then, what I like to be known for is have sufficient hard data to allow consumers to perform buying decisions. By definition then there will be criticism of why he didn't test this or that. To which, will let them read the above :).

This said, I am flexible and will refine the testing for this new category of product (pre-amps/integrated). For that, I need suggestions for tests that have good justification behind them. I am not beholding to old "standards" such as testing 1 watt distortion+noise. I am not a lemming just following old methods. No one listens at just 1 watt and power is a lot cheaper than it used to be when those standards were established.
 

restorer-john

Major Contributor
Joined
Mar 1, 2018
Messages
2,359
Likes
3,750
Location
Gold Coast, Queensland, Australia
#70
But in a pinch, the one that is built-in would do well with high impedance headphones.
Amir, in your experience, with sensitive med/high impedance cans, what is the approx threshold of audibility in uW? Have you used your AP direct into cans to generate ever lower signals until you cannot perceive a signal?
 

svart-hvitt

Major Contributor
Joined
Aug 31, 2017
Messages
2,375
Likes
1,133
#71
I want to make sure my implicit goal here is clear.

My intention is to sufficiently test a device until the right conclusion pops out of it. Specifically should it be on your buy list with no engineering faults or not. I do a lot of work behind the scenes to get this answer and will not post a review until it nets out this way.

It is not my goal to produce comprehensive library of tests. To do that would mean reviewing one piece of equipment a week if we are lucky. Let's remember that it is not just the test time but documentation that is resource intensive. Graphs need to be prepared, annotated, explained, etc.

I have a room full of equipment waiting to be tested and more is constantly on the way. I rather get 100X broader coverage than test a few pieces of gear extensively.

Back to your statement then, what I like to be known for is have sufficient hard data to allow consumers to perform buying decisions. By definition then there will be criticism of why he didn't test this or that. To which, will let them read the above :).

This said, I am flexible and will refine the testing for this new category of product (pre-amps/integrated). For that, I need suggestions for tests that have good justification behind them. I am not beholding to old "standards" such as testing 1 watt distortion+noise. I am not a lemming just following old methods. No one listens at just 1 watt and power is a lot cheaper than it used to be when those standards were established.
Peer critique is always a good thing. And it feels like a nuisance every time (at least to me...) to be at the receiving end.

:)
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
20,594
Likes
25,379
Location
Seattle Area
#72
Amir, in your experience, with sensitive med/high impedance cans, what is the approx threshold of audibility in uW? Have you used your AP to generate ever lower signals until you cannot perceive a signal?
People have such sensitive headphones/IEMs that can be impacted if someone passes gas in the house next door. :D A lot of what we ignore in home audio becomes an issue with sensitive headphones due to their high isolation capability and so little power they need to produce something audible.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
20,594
Likes
25,379
Location
Seattle Area
#73
Peer critique is always a good thing. And it feels like a nuisance every time (at least to me...) to be at the receiving end.

:)
My best manager had a mother that was an English teacher. So every time I wrote something, he would go after the English before even paying attention to what I had written. I would go multiple rounds just to get through that, only to be told at the end what he thought of the message. He gave me an excellent advice on this: "it is easier to critique than to create." So I learned to not worry about such things. Folks can provide the critique.
 

sergeauckland

Addicted to Fun and Learning
Patreon Donor
Joined
Mar 16, 2016
Messages
924
Likes
1,665
Location
Suffolk UK
#74
Your reviews are wonderful, but if they are not adhering to standard practices, people will ultimately dismiss, discount or ague with them.

From my perspective, ASR reviews could become an ultimate authority in time, but only if the reviews are beyond criticism or at least approaching it.

Very much this!

S
 

MZKM

Senior Member
Joined
Dec 1, 2018
Messages
330
Likes
240
Location
Land O’ Lakes, Florida
#76


You can see the comparison to recently reviewed Onkyo M-282. The E-270 outperforms it in every category from lower noise floor and distortion to more power prior to clipping. The Onkyo of course is a budget product at $300.
Right, but if we look at 0.1% THD, the Onkyo offers near identical wattage. My calculation might be wrong, but I calculated that if you have a total of 6 components in your chain (most won’t have more than 4: DAC + preamp + MiniDSP/similar + power amp), all having 0.1% THD would lead to a sum total of <1% THD (what has been pretty much accepted as audibility thresholds for treble for musical content).

Now, I get that aiming for absolute threshold of audibility rather than audibility when masked with music will separate the good from the great, but it’s nice to put things into perspective.
 
Last edited:
Joined
Nov 22, 2018
Messages
238
Likes
180
#77
There is absolutely no point using a preamplifier that is specified with an 18dB gain and then winding back the attenuator to give unity gain for testing purposes. All preamplifiers are tested and specified at the wide open position and have been forever.
True, we won't get very far if none of the tests are standardized, but I disagree with your first sentence.

Most of us use digital sources that produce between 1 & 4Vrms at 0dBFS. Putting that 1V signal straight into this power amp section produces 156W, with output almost exactly at Amir's marker in the power/THD graph. Most users will be on the verge of clipping with the preamp at unity gain.

How can there be no point to this test, when it's already at the upper limit of normal use? Sure, testing at this level isn't comprehensive, but it is meaningful. At least for me.
 

March Audio

Major Contributor
Manufacturer
Patreon Donor
Joined
Mar 1, 2016
Messages
3,910
Likes
3,729
Location
Perth Western Australia
#78
The measurements Amir performed are extremely limited. (no doubt due to time constraints). I certainly feel a categorical recommendation or lack thereof is a very unwise position to take.

What we can see, is the performance exceeded rated specifications by a wide margin (factors of 10+) for the parameters Amir tested. This was as expected and as I predicted. Frequency response tests appear to exceed spec also, but Amir has not specified the power output level he tested FR at. The standard and the spec is 1W. We can only guess what he used.

The high level preamplifer inputs were driven with a signal to give the same level out as in (2V). There is absolutely no point using a preamplifier that is specified with an 18dB gain and then winding back the attenuator to give unity gain for testing purposes. All preamplifiers are tested and specified at the wide open position and have been forever. The rated output of the preamplifier is 1.07V from a 134mV input. He fed it 2V and wound back the attenuator. You test a preamplifier at its rated input and output, not some made up number that suits you. We may be merely seeing non-linearities and noise (hum) etc, from being buried in the multiple stages of the AAVA volume control (see below).

View attachment 20348
I had hoped Amir would investigate this unusual implementation and test it at various increments/levels to give some insight into its performance or lack thereof. I'm not convinced piling VI buffers in line is a good idea, but who knows. Accuphase have been touting their AAVA for several years now.

View attachment 20347

One cannot characterize the performance of an amplifier with only one resistive load test (4 ohms), which is half the standard, and make no comparisons between different loads and different numbers of channels driven. There is no reactive testing, transient (IHF/EIA) testing, power bandwidth, shorted input residual, comparisons of various inputs, crosstalk measurements, etc.

There is zero discussion on the topology as the cover was likely not removed (understandable if it was on loan) and therefore we can only surmise as to the reasons for the burst of HF hash on the FFT. I have offered a possible reason in a previous post, but Accuphase is unlikely to have released a product into the wild with such an issue. It may be operator influenced perhaps.

The power amplifier stage is conventional and of a decent standard according to what I see on Accuphase's brochure.

View attachment 20349

What we can see is Accuphase appear to have decoupled supplies for the VA, the driver stage and the output stage and we know they will have high quality regulated supplies for the preamplifier stages. Amir makes the statement the preamplifier supplies are being affected by the output. This is perfectly normal in integrated amplifers with single transformers when operated at high powers into low impedances.

Comparisions to the Hypex are flawed. The Hypex units have not demonstrated they can even hit their specifications. Nor have they demonstrated they can be trusted not to shut down at inopportune times. Their power supplies overheat and are failing in the marketplace. The Accuphase wasn't pre-conditioned, or tested for 5 minutes at full rated power either.

I take the whole review with a grain of salt.

That said, further investigation would clear up a whole lot of the above. Time contraints are a b#tch.
I think you are being a bit harsh on the review. There is clearly an issue with mains pickup and other spuria on this unit which wrecks a lot of the numbers. Perhaps that isn't typical of Accuphase products but there is no doubt the NC400 is significantly cleaner than this particular example. More / different tests won't change that, unless it is considered faulty.

BTW the replacement power resistor arrived so I will be doing the 5 min test later on at least one of the hypex amps. Also I think one person replied to the forum about failed psus
 
Last edited:

AnalogSteph

Active Member
Joined
Nov 6, 2018
Messages
112
Likes
87
Location
.de
#79
What is really, really strange is that if I measure RCA but leave the XLR cables connected, performance drops way down just the same! You can see it easily in signal to noise ratio measurements:
For the record: XLR cable plugged or unplugged at source end for this test?

Do you happen to have some "ground-lifted" XLR cables (with the shield disconnected @ one end)?

Just a theory, of course, but I am not aware of any XLR jacks with a switching contact, so I very much suspect that it is the shield connection that is giving trouble for some reason... possibly a bad case of Pin 1 Problem, something to do with a ground loop in any case. Which clearly defeats the point of using a balanced input in the first place, of course. (Apparently it's unity gain, too. Completely silly when your input sensitivity is a mere 134 mV rms. A 40 kOhm input impedance also suggests a bunch of 20 kOhm resistors in the input stage... not exactly what you would call low noise. Overall, a surprisingly poor implementation. I would want to aim for -6 to -12 dB and take a look at where that pin 1 connection goes.)
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
20,594
Likes
25,379
Location
Seattle Area
#80

Similar threads

Top Bottom