Topping PA5 Review (Amplifier)

[Kevinfc]

The PA5 thread is sizzling hot, the Topping people must be delighted.

 

[Kevinfc]

Superb. Thanks for the reply Amir.
It seems to be a really competent machine.

PA5 feels like a game changer. I'm waiting, like others (I have read this thread), for Topping to come out with someone similarly good but with preferably twice as much effect.

However, this one with fairly sensitive speakers (depends on the type of music and how loud you listen), smaller rooms. Secondary system, summer cottage (easy to take with you). Hmm. Just the imagination that sets the limits .Very interesting!

I agree. This fills a niche in the midrange power category, but unfortunately my needs are filled for such a piece. The next step, something in the 80wpc or more at 8ohms interests me more.

 

[Kevinfc]

I totally agree with your opinion.

Nonetheless, and very objectively, the manufacturing quality of the sample shown by Genfreeciv is subpar compared to what Topping used to deliver. (exemple : EX5 or DX7 pro are both very well made, the PCB is super clean and the housing is partially made using CNC aluminum).

The engineering part is amazing, this is one thing and everyone will agree to say that John and Topping's team have done a very good job regardless of the chip used. For that, thumbs up and well done.

However, this unit seems to be very badly manufactured (solders, flux, attention to details, maybe component sourcing ?) and when you buy something, especially when the manufacturing process is already mastered by a number of Chinese factories, it is normal to evaluate the product also on the manufacturing characteristics and in comparison to what we already have on the market.

Here we have a 350€ class D amplifier, yes it is higher priced than what we can buy on AE for the same kind of layout and same TI chipset and that is due to the R&D behind the product, resulting in a higher performance product. BUT the problem is, for 350€ we have a unit that is less well produced than a sub-100€ class D chinese amp, and THAT is not acceptable, especially when you have an amazingly well engineered product.

I already bought a PA5 unit, I'm currently waiting for its delivery, but I'm quiet disappointed by the build quality.

So you‘re assuming yours will be bad based on a few earlier reports? Could you clarify? Is your issue with build quality asthetic or functional?

 

[Michal]

Can you try with orchestral recordings? please.

And "Hotel California"

Eagles - Hotel California (1976) Vinyl, Asylum Records, Original US

Analyzed: Eagles / Hotel California {US}
--------------------------------------------------------------------------------

DR Peak RMS Duration Track
--------------------------------------------------------------------------------
DR15 -1.23 dB -18.96 dB 6:30 01-A1 Hotel California
DR13 -1.53 dB -18.50 dB 5:05 02-A2 New Kid In Town
DR15 -0.01 dB -17.36 dB 4:46 03-A3 Life In The Fast Lane
DR13 -2.38 dB -20.72 dB 4:57 04-A4 Wasted Time
DR11 -6.44 dB -22.97 dB 1:24 05-B1 Wasted Time (Reprise)
DR14 -1.16 dB -17.66 dB 4:10 06-B2 Victim Of Love
DR16 -1.85 dB -21.91 dB 3:58 07-B3 Pretty Maids All In A Row
DR14 -0.47 dB -17.22 dB 5:11 08-B4 Try And Love Again
DR13 -1.12 dB -18.71 dB 7:22 09-B5 The Last Resort
--------------------------------------------------------------------------------

Number of tracks: 9
Official DR value: DR14

Despite the fact that every person has a different musical sensitivity, the 3255 chip does not sound good anyway.

 

[Papaya_X]

So you‘re assuming yours will be bad based on a few earlier reports? Could you clarify? Is your issue with build quality asthetic or functional?

I don't assume anything, noone knows and that's why I asked people who already got their units to verify.
I hope Genfreeciv has an early unit and that Topping are producing the PA5 with better overall build quality.

My issue is clearly about aesthetic and manufacturing execution, I have already done very bad looking solders and DIY electronics projects and they are working very well, still they are of a very bad build quality.

The PA5 is performing tremendously well, but what we have seen from its guts *in terms of build quality* is the same kind as my AIYIMA A03 which cost me ~40$. When I compare *once again : the build quality* of similarly priced amps/DACs, most of them have a better manufacturing execution, even in Topping's product line such as the EX5 for exemple. Anyway the build quality can only be compared for a specific price point, and that's why *I think* it is expected to have nice solders and a relatively nice technical overall execution at this price point.

Other than that, honestly I don't care, the performance is top notch, the power seems plenty for my application and that is the result of a good engineered design that I'm ready to pay for.

 

[timing3435]

Can you try with orchestral recordings? please.

And "Hotel California"

Eagles - Hotel California (1976) Vinyl, Asylum Records, Original US

Analyzed: Eagles / Hotel California {US}
--------------------------------------------------------------------------------

DR Peak RMS Duration Track
--------------------------------------------------------------------------------
DR15 -1.23 dB -18.96 dB 6:30 01-A1 Hotel California
DR13 -1.53 dB -18.50 dB 5:05 02-A2 New Kid In Town
DR15 -0.01 dB -17.36 dB 4:46 03-A3 Life In The Fast Lane
DR13 -2.38 dB -20.72 dB 4:57 04-A4 Wasted Time
DR11 -6.44 dB -22.97 dB 1:24 05-B1 Wasted Time (Reprise)
DR14 -1.16 dB -17.66 dB 4:10 06-B2 Victim Of Love
DR16 -1.85 dB -21.91 dB 3:58 07-B3 Pretty Maids All In A Row
DR14 -0.47 dB -17.22 dB 5:11 08-B4 Try And Love Again
DR13 -1.12 dB -18.71 dB 7:22 09-B5 The Last Resort
--------------------------------------------------------------------------------

Number of tracks: 9
Official DR value: DRWill

 

[tmtomh]

Agree that NC122MP and PA5 are roughly even on power output and I would consider both rather power limited. As far as I can tell the NC122MP is not very popular and I imagine power output (for the price) is a primary reason. Agree that PA5 will outperform NC122MP in terms of noise but I imagine NC122MP has less frequency response variability in to complex loads and has better higher frequency distortion performance than the PA5. I am sure that both sound great when used within their power limits.

Personally I would opt for a NC252MP over either as it is not that much more expensive, I like to have more power on tap, I run DIY active speakers so having an adjustable volume knob is a liability, prefer an in-chassis power supply and I know it will give me excellent system noise performance as I long as I use a reasonably low noise DAC (>108 dB DR at 2 V).

Michael

I'm sure you are right about the popularity of the 122MP - I have wondered about that myself too, given that the price per watt seems better on the 252 and other modules. I too would get a 252MP over the 122MP or the PA5. However, while I also share your preference for an in-chassis PSU, I can understand why the PA5's small size would be a major benefit for folks using it for near-field listening on desktops and such where space is at a premium - and I can also see where $349 vs $500 or $600 would be a no-brainer if one did not see any added value in the extra money spent given one's particular use-case.

As for load invariance, I would guess you are right given that "chip amps" generally are more load-dependent than the Hypex and Purifi modules. But I don't know - Amir's tests seem to show that the load variance displayed by the PA5 is outside the audible range. As for higher-frequency distortion, yes, the PA5 does seem to have notably higher distortion at 10 and 15kHz as you get above about 3 watts of power - but like you I am confident it would likely be audibly indistinguishable from the Hypex units, especially given that as far as I understand some of that rising distortion at higher frequencies are harmonics above the audible range (since the bandwidth of the test goes up to 45kHz).

At any rate, this does seem like a very cool little amp and for the moment pretty unique in the market, given its price location above the chip amps and below the NCores, and its overall performance roughly on par with the lower NCores. I hope more amps like this come onto the market!

 

[Rottmannash]

I hardly think D90 is an amplifier.

Well, it is a Topping product and inferring they only sell "cheap" products is a bit unfair. I feel in 6 months that statement will not be valid anymore.

 

[timing3435]

Will do but do not know orchestral recording wells. Do know some Paganini so I can compare it to a tube amp and a Jeff Rowland 535 Want to give it a day or two to run. I can say it is very clean( full power nothing out of the speakers), the sound stage is wide with X3's. The X3's have a plate amp on them so power is not an issue and they are rated at 97db.
I am not real nitpicky about things other than I like the details on the high-end, tight bass, and the timbre of a standup bass. For my ears, the PA5 does what I want it to do. For the money, it is well worth it. Buy the thing and compare - world-class specifications and see if the PA5 matches your sound preferences. Will be interested in seeing how it matches with a tube preamp.

PS- I can hear Diana Krall's lips on Sway



Paganini

 

[Kevinfc]

I don't assume anything, noone knows and that's why I asked people who already got their units to verify.
I hope Genfreeciv has an early unit and that Topping are producing the PA5 with better overall build quality.

My issue is clearly about aesthetic and manufacturing execution, I have already done very bad looking solders and DIY electronics projects and they are working very well, still they are of a very bad build quality.

The PA5 is performing tremendously well, but what we have seen from its guts *in terms of build quality* is the same kind as my AIYIMA A03 which cost me ~40$. When I compare *once again : the build quality* of similarly priced amps/DACs, most of them have a better manufacturing execution, even in Topping's product line such as the EX5 for exemple. Anyway the build quality can only be compared for a specific price point, and that's why *I think* it is expected to have nice solders and a relatively nice technical overall execution at this price point.

Other than that, honestly I don't care, the performance is top notch, the power seems plenty for my application and that is the result of a good engineered design that I'm ready to pay for.

Sounds reasonable, after all, this is a forum where even the smallest details of a product are discussed. I’ll admit that I’d prefer very exact board work myself. Being a watch enthusiast, I certainly appreciate that the discussions can extend beyond what is entirely practical and extend into parts unseen.

 

[Bruce Morgen]

Diehard anti-"toy" Class AB fans might want to buy one of these for Amir to measure.

Emotiva BasX A2M

 

[BoredErica]

An important thing to consider is that once you get to this level of amplifier performance it is quite possible your DAC noise will have a much greater impact than amplifier noise. Fortunately the low gain of the PA5 actually helps quite a bit here because it will amplify the DAC noise less than a higher gain amplifier. See below for some calculations demonstrating the combined impact of DAC and amplifier noise.

Let's take a DAC like the MOTU Ultralite Mk5 which has a dynamic range of around 116 dB at 4.2 V. This implies 4.2 V x 10^(-116/20) = 6.7 uV of noise at the DAC output.

If you multiply this noise by 19 dB amplifier gain for the PA5 you end up with a noise contribution from the DAC at the amplifier output of 6.7 uV x 10^(19/20) = 59 uV.

Now let's look at the PA5 itself which has dynamic range of around 107 dB at 5 W in to 4 ohms. This implies sqrt(4 x 5) x 10^(-107/20) = 20 uV at the amplifier output. Clearly the DAC has a much larger noise contribution.

You can then sum these values to determine total residual noise at the amplifier output. Assuming the noise is uncorrelated they will combine as root of sum squared. So total noise at the amplifier output is sqrt (59^2 + 20^2) = 63 uV. This combined noise value is the most helpful in understanding whether you will have hiss issues. For example based on experimentation with different gear I know that noise levels in the mid-100 uV range are good for my setup to not have any noticeable hiss so this would be absolutely fine for me.

Just as a fun exercise if you take the combined noise from the DAC and the amplifier and calculate an effective dynamic range for a power level of 5 W into 4 ohms you can see how much the DAC degrades performance. Combined dynamic range is 20 x log(sqrt(4 x 5) / 63 x 10^-6) = 97 dB, so we have lost about 10 dB of performance.

Michael

Yup. And I want to run it with Motu Ultralite Mk5 as well. However I do not ask my amp to do the 500 functions Motu can do so I expect higher performance. That low noise and low gain sounds great for my purposes so I don't have to wonder 'what if?'. For example I was thinking, do I spend $500 on Buckeye amp? But the March p122 had better noise, though it costed $100-150 more or some such. Well new answer, I pay $180 less instead and get the pa5 and get lower noise and gain.

Lower noise vs more power... Depends on what you want. Meta already struggles at high volume, and I just don't find loud music *enjoyable* even if it's not enough to cause hearing loss.

Only missing a clipping LED/LED colour. Every amplifier should have one, especially the ones with such a modest power rating.

Good suggestion.

 

[MaxBuck]

I could see this amp paired with BMR Philharmonic monitors and a streamer-dac as a killer $2,500 system. So much so that I desperately want to convince my wife to let me do it.

 

[BoredErica]

Okay here's an unrelated question. The KH 80 takes a whopping... 6.5s? To turn on. I'm not sure why it takes so long. For an amp like pa5, or the pa5 in particular, how long between turning on the amp and music playing out of the speakers as intended? <2s? 5s? 5+s?

My pa5 is somewhere in transit to hifigo warehouse or something, because apparently silver case version will take longer to arrive than the black case version.

 

[Walter]

For those commenting on the price, I had also hoped that this amp would come in at the $300 or lower price point, as I mentioned on the pre-review thread. However, that was just a hope--not an analysis of what it should cost. As Amir already said, this amp might well have been $50 cheaper if it had been released a year ago. By comparison, the JLE Sylph TPA3255 board is $160-$180 depending on opamps, and in the thread, @jlesterp states "Around 110USD should be set aside to get a decent chassis and accessories. 60USD for a decent power supply." So total cost of about $350 but you have to assemble it yourself--although you do probably get a nicer case for that price. It has been reported that Allo is eyeing a price of $400+ for their TPA3255 implementation. So it does not appear to me that there is any price gouging going on here. I'll agree that in Europe, if this amp is going to cost 349 Euros instead of USD $349, the Audiophonics NC122MP amp might be the better deal. In the USA, where the cheapest NC122MP amp I am aware of is the VTV at $499, the value is just not there. I'd choose either a PA5 or a Buckeye NC252MP for the same $499, (assuming I could wait until March for delivery).

 

[misterdog]

this amp might well have been $50 cheaper if it had been released a year ago.

Sheets of MDF have risen in price by 200% in the last year.
The price of steel has doubled.

 

[pma]

I'll agree that in Europe, if this amp is going to cost 349 Euros

https://www.audiophonics.fr/en/integrated-amplifier/topping-pa5-balanced-class-d-amplifier-2x125w-4-ohm-black-p-15847.html

Correct, PA5 costs €349 in Europe (with tax and duties) and Audiophonics NC122 amp costs €399 with tax.

 

[zanlation]

I just ordered my PA5 to replace a DIY TPA3255 TI sample amp. It will drive a pair of full range Mark Audio 7 pencils with a Topping D50s DAC. I'll move my TPA3255 Amp to the Cinema where I need the power. Great to see a mid range power amp perform so well in tests.

 

[Mulder]

@amirm Is it possible to do a 18-19 khz test to see if there are any IMD.

 

[DanielT]

@amirm Is it possible to do a 18-19 khz test to see if there are any IMD.

I have seen in this thread that several people are interested in it, so it would be very interesting if Amir conducts a test / measurement.