Ampapa D1 Stereo Amplifier Review

[MCH]

Just dial-in "GaN Audio Power supply 52V 9A" at aliexpress.com
View attachment 513242
First hit was ^this^....
I haven't been levied any tariffs from them, to the WestCoast-USA, since Potus#47.

It is not only the amp chip, you need to make sure there aren't other components (typically regulators) that could be not rated for more than 48V. Maybe they wrote it for a reason besides the socket, or maybe not.

 

[GGroch]

.......Your reviewed 3E-A7 stereo-amp had the 52V@9A(465W) GaN power brick:
Would substituting that in place of the 48V@5A(@40W) improve the headroom (etc.), w/o causing excessive heat or triggering protection circuitry?

I went the other direction with my D1. I replaced the 48v 5A stock PS with the 38v 5A one that came with my 3e Audio A5.

The reason: the A5 with its 38v PS provides far more wattage than I ever use, it has watts to spare. Larger power supplies in no way improve performance at lower output levels. Plus every test I have seen comparing lower and higher voltage supplies with TPA3255 amps show they run a lot hotter, both under load and quiescent.

A 52v PS will not increase headroom, because Amir's test of the A7 with the 52V supply has no headroom.

Running hot will lessen lifespan and of course, leaves the world with less energy to feed the AI Data Centers we so desperately need

 

[Bleib]

Note that the 3e Audio A7 mono is meant to use 52v but D1 is max 48v

A7 mono:

 

[NTK]

Somewhat disappointing that the HPF filter is just another "check the box" feature. The filter order is first order which means it is barely better than not having one.

The HPF filter characteristics match the textbook response of a 1st order filter with a 132 Hz cutoff perfectly. The top plot used the same axes scales as Amir's plot. The bottom plot expand the y-axis bottom limit to show this filter only attenuates by 10 dB at 40 Hz.

I'd much rather see a high order fixed frequency HPF at 80 Hz than this one.

Spoiler: Pyhton code to generate the 1st order filter response plots below

import numpy as np
from scipy import signal
import matplotlib.pyplot as plt

def dB(x):
    return 20.0*np.log10(np.abs(x) + 5*np.finfo(float).eps)

fc = 132.0
order = 1
butterHPF = signal.butter(order, fc, 'high', analog=True)

w = np.logspace(np.log10(10), np.log10(2e5), 200)
_, Hb = signal.freqs(*butterHPF, w)

fig, (ax0, ax1) = plt.subplots(2, 1, figsize=(8, 8), sharex=True)
ax0.semilogx(w, dB(Hb))
ax0.grid(True, which='both', axis='both')
ax0.set_xlim(w[0], w[-1])
ax0.set_ylabel('Response Magnitude, dB')
ax0.set_ylim(-5, 5)
ax0.set_yticks(np.r_[-5.0:5.5:0.5], [ '' if x%1 > 1e-8 else '{:3.0f}'.format(x) for x in np.r_[-5.0:5.5:0.5] ])
ax1.semilogx(w, dB(Hb))
ax1.grid(True, which='both', axis='both')
ax1.set_xlabel('Frequency, Hz')
ax1.set_ylabel('Response Magnitude, dB')
ax1.set_ylim(-25, 5)
fig.suptitle('HPF Response Magnitude, Filter Order: {:d}, Cutoff: {:.1f} Hz'.format(order, fc))
fig.tight_layout()
plt.show()

 

[john61ct]

Just dial-in "GaN Audio Power supply 52V 9A" at aliexpress.com

Thanks, but...

I am asking for listings specifically recommended not at random from Ali.

And ideally from sellers/platforms where there is no uncertainty about potential tariff issues

 

[pablolie]

Geez what a spoiled bunch ASR voters are!

"Fine" for a $199 amp producing 100W with a solid 90dB in SINAD and great behavior for anything other than exotic loads? :-D I say most definitely "great"!

 

[xanalog]

The OLED display looks nicer in person than the above picture.

Would there be a long-term issue w/an always-on OLED?
The OLED (2.5x5cm) display on my SMSL, turns-off after 62[?] seconds.

 

[amirm]

Would there be a long-term issue w/an always-on OLED?
The OLED (2.5x5cm) display on my SMSL, turns-off after 62[?] seconds.

As noted, the background of the VU meter is rather gray which would indicate it is probably an LCD, not OLED. When I saw that, i assumed they did that on purpose. But maybe not. But I agree that if it is OLED, it should have a time out.

 

[amirm]

On this being a "golden sample," as I mentioned in the G7 thread, this package was lost last year and resent out of warehouse in California. That rules out any hand picking of samples.

 

[john61ct]

But you're accepting their word? If a company was being intentionally deceptive...

Not saying I believe it happens, but a secondary sanity check with a unit sent from trusted members would help stop the speculation.

 

[Mort]

But you're accepting their word? If a company was being intentionally deceptive...

Not saying I believe it happens, but a secondary sanity check with a unit sent from trusted members would help stop the speculation.

Sure, but do you think that's easy to manage and execute? I don't. I would not want fewer products tested to eliminate the very slim possibility that all of the testing is bogus.

 

[john61ct]

Sure, but do you think that's easy to manage and execute? I don't. I would not want fewer products tested to eliminate the very slim possibility that all of the testing is bogus.

I think encouraging members to send in units for that purpose is enough, just once per model tested.

Amir of course is the judge as to whether this would cause any problems.

 

[amirm]

But you're accepting their word? If a company was being intentionally deceptive...

I get to apply my judgement and experience to such matters. You can choose to accept that or not. There have been people who try to game the system and I can tell from a mile away. Those people are on my "do not ever review" list.

On the second part, as has been explained, there is nothing in there to game here. Hand picking one vs the other could result in a couple of dBs of SINAD and take a ton of work to do. I don't see that being worthwhile effort.

You can't live your life by being paranoid about everything. But maybe you tell me that you send your milk carton to a lab every time you purchase it to make sure it is the same thing....

 

[amirm]

I think encouraging members to send in units for that purpose is enough, just once per model tested.

Indeed. We have done this a number of times and all it has done is cost us money in the form of shipping, and my time. For now, FUD is not something I like to entertain just for the sake of it.

 

[john61ct]

Again

> Not saying I believe it happens

So maybe, IF a member is suspicious of a given company, or thinks their QC is not consistent

The onus is on THAT member to "sponsor" a secondary spotcheck?

Of course they should DM you and see if you would be open to doing that before spending their money

 

[Mort]

Again

> Not saying I believe it happens

So maybe, IF a member is suspicious of a given company, or thinks their QC is not consistent

The onus is on THAT member to "sponsor" a secondary spotcheck?

Of course they should DM you and see if you would be open to doing that before spending their money

You can't test inconsistent QC with one unit, even a random one.

 

[john61ct]

The point is the SECOND, followup check, shows inconsistency in comparison to the original.

And if the original measures MUCH better than the random followup, suspicions are strengthened it may have been "golden"

 

[Mort]

The point is the SECOND, followup check, shows inconsistency in comparison to the original.

And if the original measures MUCH better than the random followup, suspicions are strengthened it may have been "golden"

Is it inconsistent QC or a fraudulent sample? You are equivocating.

 

[amirm]

The point is the SECOND, followup check, shows inconsistency in comparison to the original.

That is outside the scope of the job of a reviewer. We are not the police on production consistency of a company. As is, I don't have remotely enough bandwidth to test products once.

If there is reason to assume that the review process has been gamed, such as someone else getting different test results, then I am open to investigating. Otherwise, no.

You are welcome to assume that some percentage of my review results may not be 100% correct. I accept that as a trade off to get a ton more coverage of products out there.

 

[john61ct]

OK fair enough, that's that then