Topping PA5 Review (Amplifier)

[Matias]

We all seem to be tippy-toeing around this so what the heck, I'll take the plunge.

Hey @JohnYang1997 : are you using the TPA325X in the PA5? If not, what? Enquiring minds want to know!

(John: based on this darned review, I just purchased a PA5 AND an E50 this weekend, so be nice: I'm your customer! Please don't make me void my brand new warranty for the sake of science and curious ASR readers! )

He would have told already if he was allowed, and the marketing pics also don't tell, plus they had the trouble to cover the actual chip, so no way we will know... Unless a brave soul decides to open his/her unit, uncover the chip and take a picture. Of course losing waranty in the process.

 

[TheWalkman]

He would have told already if he was allowed, and the marketing pics also don't tell, plus they had the trouble to cover the actual chip, so no way we will know... Unless a brave soul decides to open his/her unit, uncover the chip and take a picture. Of course losing waranty in the process.

Matias,

Thanks, but with all due respect, I'd prefer we allow Mr. Yang a chance to speak for himself.

Cheers

 

[Zek]

.. Unless a brave soul decides to open his/her unit, uncover the chip and take a picture.

... and you will see a chip with a deleted label or enclosed in a capsule ...

 

[DanielT]

NAD 3240PE: 40 WRMS/ch w/2 ch driven.
I checked around and the amp will do ~500W peak stereo@2Ω for a short duration.
"With long 200-millisecond tone-bursts, representing the full duration of musical notes and chords, the 7240PE produces an impressive 100 watts per channel."

I crunched some numbers and if I am correct it can sustain 500w peak@2Ω/ch w/2 ch driven for ~10.9mS and then the power supply caps are drained to about ~0.8 of 1 RC time constant and the amp runs out of steam.
Don't quote me but if the transistors are in 100% saturation mode and the thermistors inline with the +/- rails are not resistive then we get that number.
If the transistors and thermistors are resistive then the time for duration for sustaining the 500w/ch @2Ω increases.

Aha aha.Thanks for your calculation. But is it good or bad then?
That is, if you translate the numbers into reality. Some song where there really is up and down, bangs, spouts in the trumpet and so on.
How would that one handle such a song with, say, 88 dB sensitive speakers, normal listening volume in a normal-sized living room. Would it then be driven into the clipping at transient peaks?

I just took an amplifier at random. Please do not say that it is good because then I will be eager to buy one and test. He he.

 

[sarumbear]

There's plenty of amplifiers specifically designed to run at full rated average power and still handle any peak you'd like to throw at them. A large number of commutating rail amplifiers built through the 80s, 90s and into the noughties.

Continuous average power is the standard. Everyone had their opportunity to argue for, and prosecute for a different standard when the FTC asked for comments. Did you put forward an alternative and more representative testing regime for them to consider?

Bob Carver argued this back in the 70s. His Phase Linear power amps was relatively limited in RMS but high power on peaks.

 

[PeteL]

He would have told already if he was allowed, and the marketing pics also don't tell, plus they had the trouble to cover the actual chip, so no way we will know... Unless a brave soul decides to open his/her unit, uncover the chip and take a picture. Of course losing waranty in the process.

I tend to agree that at 3 bills and change, it's only a matter of days or a couple week that some youtuber will not care about warranty an shoot a video of him popping up this chip shield/cover. He'll gather more money in clicks monetizing than the value of the amp anyway. And then everybody will know. I'm not sure about this strategy. Anybody who really want to copy will get a unit and dismantle it and we'll figure out most, not all of it, but will still probably fall short in measured performance. For the rest of us well knowing the chipset is not about competing Topping, it's of pure unbeneficial interest and Topping might be the RockStar of this customer base. It wouldn't hurt them to give them the Candy since the secret will travel anyway.

 

[Doodski]

Aha aha.Thanks for your calculation. But is it good or bad then?
That is, if you translate the numbers into reality. Some song where there really is up and down, bangs, spouts in the trumpet and so on.
How would that one handle such a song with, say, 88 dB sensitive speakers, normal listening volume in a normal-sized living room. Would it then be driven into the clipping at transient peaks?

I just took an amplifier at random. Please do not say that it is good because then I will be eager to buy one and test. He he.

From the reading that I did apparently many musical notes can be sustained within that 200mS @ 100W/ch time frame.
I think if it is cannon fire in the music that the time duration for sustaining that peak power is too short for that loud long canon noise.
In a average size room and with your parameters it should be OK.

The 500W @2Ω peak worst case scenario of 10.9mS duration is too short for most musical notes. But nonetheless it is still cool that it can do that.
I think these power envelope design amps are in demand so I would not expect a low price.

 

[misterdog]

Hey @JohnYang1997 : are you using the TPA325X in the PA5? If not, what? Enquiring minds want to know!

(John: based on this darned review, I just purchased a PA5 AND an E50 this weekend, so be nice: I'm your customer! Please don't make me void my brand new warranty for the sake of science and curious ASR readers! )

You will either like it in your room and your system and wish to keep it, or not like it and send it back.

 

[pjug]

The power ratings for IcePower are very clear as to what is peak, maximum continuous, etc. I am not sure how other manufacturers are, but there is no ambiguity here for this Class D amp. If you take a look at the specifications table for the 1200AS2 section 8 datasheet, we see that max continuous power of 1200w for 1 channel driven without thermal shutdown is 15sec (120vac mains). But for continuous output without thermal shutdown is only 280w. It’s much better for 240vac mains so a lot of this appears to be limited by the on board PSU. View attachment 169129
With the TPA3255 (assuming now, that’s what is inside the PA5) the specs depend on the rail voltage and the amount of THD one can tolerate. For 53.5v rail and 1% THD for a signal at 1kHz, TI states in the datasheet that the amp can provide 155w per channel. This assumes that adequate heatsinking is provided and that the PSU can provide the goods. I have done continuous sine wave testing in the TPA3255 with a good heatsink and sufficiently large 600w SMPS and can confirm that it will make 150w continuously. The heatsinking here is not trivial - using an CNC milled thermal transfer block to a large 6mm thick plate aluminum chassis floor. Those small finned heatsinks they bolt onto run of the mill TPA3255 amps from China will not dissipate 30w (strereo) continuously. But not to worry, the TPA3255 has built in thermal protection shutdown, and it will give a fault LED warning if it does so. It also tells you if it is clipping and also shut down if there is excessive clipping.
View attachment 169135
The PA5 has a PSU de-rated from the 53.5v max to 38v, so this will impact the max power. The way to estimate max voltage swing of a BTL amp is to take the rail voltage and multiply by 2x and subtract 8v (for 2x 4v dropout due to MOSFET output). This is 68v and matches what they state as max output swing. 68v / 2.83 is 24vrms output. Power is V^2/R so this is 72wrms. It appears that the 140w spec is for both channels into 8ohms. But 140w per channel into 4ohms.

So really, this is a 70w per channel amp for folks with 8ohm speakers.

That’s just the math behind the PSU voltage stated on the back panel.

Seems like it should be able to do a little more power from a 38V supply. The TI datasheet, Figure 8, shows about 80 watts into 4 ohms and about 150 watts into 4 ohms for a 38V supply and 1% THD+N. But the Topping spec is just 65W into 8 ohms and 125W into 4 ohms (1% THD+N). Amir's Peak Power measurement into 4 ohms is 128W.

I don't think that's a big deal, but I wonder what explains it. It does not use the TI part, or something else with the implementation and power supply?

 

[csepulv]

If topping releases a minidsp alternative, or a DAC with dsp, I will be happy too.

Or if I really get to make a wish, a Topping AVR

 

[misterdog]

If it were so easy to buy some components and bring such a well measuring product to market at such a reasonable price, then why have more not done so ?

Even if the basic engineering architecture were revealed by John Yang I suspect most could not replicate it at this price.

If people wish to rip one of these amps apart and claim it is flawed or they could do better, go for it.

But don't press someone who has spent months/years designing a product, to give all his hard work away, to be ripped apart by the wannabee internet engineering influencers.

Do it yourself - simples.

Or buy the product and be contented with the sound.

 

[misterdog]

Or if I really get to make a wish, a Topping AVR

Indeed the basic amp modules in active speakers or AVR with Topping DAC and Pre would be revolutionary.

 

[PeteL]

If it were so easy to buy some components and bring such a well measuring product to market at such a reasonable price, then why have more not done so ?

Even if the basic engineering architecture were revealed by John Yang I suspect most could not replicate it at this price.

If people wish to rip one of these amps apart and claim it is flawed or they could do better, go for it.

But don't press someone who has spent months/years designing a product, to give all his hard work away, to be ripped apart by the wannabee internet engineering influencers.

Do it yourself - simples.

Or buy the product and be contented with the sound.

Why are you convinced that knowing the tech would result in people claim it's flawed? Measurments is all you need to make claim like that, and there isn't much to rip appart. As you said, it's not because they will disclose their gross block architecture that someone can make it equal.
For this type of company it's been good marketing to say which main chip they use because that is exactly their market, the fan club that care about what's inside, most people don't care about that but it's not their core base.

 

[D700]

I concluded yes --- but I'm in a 10x10 ft room

5 watts on a set of LS50s is plenty loud when adjacent to your desk. I know from experience. I took SPL readings from dead center, front row Cleveland Orchestra one year and average SPL was 60-70, with occasional peaks into the 90s. I could easily replicate that at my desk with a little Nuforce Icon 20 watt thing.

This Topping Amp at 60 watts with LS50s will get you into high 90s continuous if my math is right...which no one should listen that loud...but would suggest it will handle transient peaks, at reasonable average levels, just fine. If you get into a larger room with greater distance from speaker to ear, then probably yeah, want something more.

I just want a matching form factor to my D/A30Pro or I'd have already pulled the trigger.

Sub management would be nice though.

 

[Papaya_X]

Usually, internally developed technologies are patented as it should be (The hardwork of the engineers need to be protected).
However, I don't think it is productive to hide the technology or the engineering approach/strategy behind a product.
If every scientific advances are open source (see scientific publications) there is a reason...
And I bet that the initial technology and engineering approach of such amps can originally be found in a electronic/audio scientific paper.

 

[TimF]

It will be fine.

A bit tiny, boring and limited in functionality, but fine. It's not the amp you want to impress someone, that's for sure. Hide it in a cupboard and pretend it's a McIntosh.

It looks modest. But so do I. Put it inside a box labeled Tiffany & Co., or one labeled Heckler & Koch GmbH, or one labeled Hermes. Soon these amps will be so small that they'll fit into the box you get when you buy a Rolex. So is the significance of this product that high audio performance per our (ASR) standards can be attained using a Class D chip that is reasonably priced and available as opposed to the current class of high performance class D amp modules (ICEpower, Hypex, Purifi) that are more expensive and are significantly larger in size? Some of the functions done on the circuit board of the Hypex and Purify modules are done in the Class D chip thereby reducing the size of the implementation. Again, I assume chips like the Texas Instrument Class D chip(s) are developed for broader market use beyond the consumer audio products.

 

[xrk971]

One area that a Class D amp will have a hard time beating a linear amp is if you look at the output trace on an oscilloscope. It may measure exceedingly low THD and noise levels in the audible band, but you will see a significant 625kHz (in the case of the PA5) sawtooth peak hash/fuzz of not insignificant amplitude (tens of mV) superimposed on the audio signal that leaks through the LC filter. You can’t hear it, but it’s definitely very visible on an O-scope. The linear Class AB or Class A amp however, will be smooth and clean. I try not to think about this but when I see it, it bothers me a bit. I then go back to my Class A amp with a clean trace and feel better knowing that even if I were a bat, I could not hear the carrier frequency. Weeks later, my memory of the hash/fuzz has faded and I’ll put the Class D TPA3255 amp back in rotation and say to myself, “Dang - That’s a quiet amp!”

Maybe that’s one more standard test or graphic Amirm can add to his usual tests. A high bandwidth O-scope screenshot of a 10kHz sine wave at say 8Vpp into 8ohms (1W). We will see the fuzz if it’s there.

 

[misterdog]

Might be more to that than meets the eye, I moved from Hypex NC400 Monos to Neurochrome class A/B and then to Benchmark AHB2 Class A/B because of subjective listening.

I also prefer my SMSL M400 to Gustard X16 I assume because of the AKM chip in the former.

Measurements don't reveal everything.

 

[Prana Ferox]

Yes, I really don't get the wailing and rending of garments that goes on every time Topping releases another compact device with TRS. Absolutely nothing inherently wrong with TRS.

Flashbacks to music production, perhaps. The 'standard' has been around for a while and is, let's say, loose, so there are a lot of cables and jacks there that are subtly the wrong size, and just generally a lot of sloppy and crackly cables out there. The connections don't retain well and, hopefully not a problem for hifi equipment, the contacts wear out after a couple thousand insertions.

For hi-fi folks it screams 'PA use' which for some is a dirty thing.

 

[nhs]

My speakers go down to 2.3 ohm, I am not sure if the PS5 is able to handle such low impedance?