boXem Arthur 4222/E1 Amplifier Review

[bkdc]

.

3. For good cooling, you need
- high thermal conductivity
- low thermal mass
- as much surface as you can
A big block of aluminum doesn't tick one of these.

Thermal mass is not a bad thing. It acts like a heat sink to store the excess heat and take it away from the circuit board.
What you want is low thermal lag, but in general items with larger thermal mass also have higher thermal lag. You can remedy this with lots and lots of surface area (fins)

So a big block of metal designed well (lots of cooling fins) would help. If it didn’t, Noctua wouldn’t be successful selling cooling fins/fans for computer CPUs

 

[KMO]

So a big block of metal designed well (lots of cooling fins) would help. If it didn’t, Noctua wouldn’t be successful selling cooling fins/fans for computer CPUs

Computer coolers are very much the opposite of "big blocks of metal". They're extremely high surface area, and minimal structurally-sound mass for that area. Only real blocky mass is in the heat pipes.

 

[bkdc]

Computer coolers are very much the opposite of "big blocks of metal". They're extremely high surface area, and minimal structurally-sound mass for that area. Only real blocky mass is in the heat pipes.

Think of the big block of metal that sits directly on the CPU. Of course surface area to mass ratio makes a difference, but the thick pipes conduct the heat to the fins.

It’s simple physics. Even without fins, a big steel case weighing 10 kilograms will keep the board cooler than a tiny case weighing 500 grams As the heat is transferred to the case.

Here, we need a structural case (thick and heavy metal) with fins which are much thicker than on a CPU cooler. The concept is the same.. just not as efficient as a CPU cooler. But you have a huge heat sink in a heavy case as long as there is reasonable thermal conductivity between the amplifier board and the case.

If this didn’t work, why do all those Class A amplifiers weigh a ton? Thank goodness for class D. But you are still generating quite a bit of heat if you’re running up against the amplifier’s specified limits.

 

[KMO]

Think of the big block of metal that sits directly on the CPU. Of course surface area to mass ratio makes a difference, but the thick pipes conduct the heat to the fins.

It’s simple physics. Even without fins, a big steel case weighing 10 kilograms will keep the board cooler than a tiny case weighing 500 grams As the heat is transferred to the case.

Only until the whole thing reaches thermal equilibrium. Once the more massive case has absorbed all the extra energy, you'll have a big hot heatsink and the core will be at the same temperature as it would be with a small hot heatsink.

Mass just increases the time it takes to reach that equilibrium.

Now, in some applications (including audio!), relying on not having long-term maximum power, so getting away with "buffering" heat in a massive heatsink may make sense. That's exactly why you can get away with just metal cases with class D amps, and not lots of fins. They don't expect to run at 200W+ continuous, so they can dissipate energy surges over a longer time.

But a high-performance PC with a Noctua cooler absolutely is intended to run continuously at max power, so the thermal mass is pointless. We are fine with reaching equilibrium immediately, as we know our equilibrium point is a good temperature, cos we've got so many fins (and a variable speed fan). The pipes are thick simply to have high thermal conductivity, minimising the temperature gradient between the fins and the CPU. They're not there to act as a "buffer".

 

[KMO]

Just to make clear what I'm talking about - it's this sort of thing:

Nothing there is about "mass". It's thermal conductivity across the heat spreader and up the pipes (hence copper), then then area on the fins.

 

[Doodski]

Just to make clear what I'm talking about - it's this sort of thing:

View attachment 236632

Nothing there is about "mass". It's thermal conductivity across the heat spreader and up the pipes (hence copper), then then area on the fins.

Heat pipes where used in class A/AB for some years in the late 70's and early 80's. They worked very well in that application. Pipe diameter is I am going to guess was ~20mm/0.787" The thing is that the pipes where larger diameter.
Here is a 40W/ch class A/AB integrated from Technics.

 

[Matias]

If one wants sustained high power with heat dissipation and even a cooler there was the Vera Audio P400/1000 reviewed here too.

Vera Audio P400/1000 Review (Amplifier)

This is a review and detailed measurements of the Vera Audio P400/1000 Hypex NC500 based amplifier. It was sent to me by the company and costs 26 000 NOK ex VAT (US $2934). The case is milled from solid aluminum and has a nice quality and look to it. It is also quite heavy for a class D amp...

www.audiosciencereview.com

 

[boXem]

If one wants sustained high power with heat dissipation and even a cooler there was the Vera Audio P400/1000 reviewed here too.

Vera Audio P400/1000 Review (Amplifier)

This is a review and detailed measurements of the Vera Audio P400/1000 Hypex NC500 based amplifier. It was sent to me by the company and costs 26 000 NOK ex VAT (US $2934). The case is milled from solid aluminum and has a nice quality and look to it. It is also quite heavy for a class D amp...

www.audiosciencereview.com

Heat pipe + cooler indeed, for the Hypex NC500 OEM. Their iteration of the same amplifier with Purifi 1ET400A looses all the cooling system.
Not because the guys at Vera Audio are not competent, but because the Purifi modules are superbly efficient.

 

[restorer-john]

Heat pipes where used in class A/AB for some years in the late 70's and early 80's. They worked very well in that application. Pipe diameter is I am going to guess was ~20mm/0.787" The thing is that the pipes where larger diameter.
Here is a 40W/ch class A/AB integrated from Technics.

But let's face it, a 40wpc Technics from the 1980s was barely 40wpc, and didn't do much more into lower impedances. The heat-pipe was more a 'space-age' selling point. It had to cost more than extruded aluminium. It's not like they didn't have already a lot of fresh air in their consumer grade Technics amplifiers...

Laptops in the 90s really brought the heat-pipe back to prominence.

 

[MaxBuck]

There are a variety of ways to achieve effective heat transfer. Mass of the associated hardware isn't one of the primary determinants of effectiveness.

 

[pogo]

This has nothing to do with the module itself, but how the manufacturers, e.g. boXem, ... implement it.
Performance depends on physical implementation and system-level circuitry/configuration (thermal system, power supply section, ...).

Here are the cooling measures in the NAD M23 & M33 and probably the reason why NAD allows over 800W at 2 ohms (1ET400A): Link

 

[RustyGates]

No, it's because he wants to sell amplifiers, that's all. He's attempted to rewrite, in various ways, the measurement bandwidth narrative to suit their agenda, but it won't work.

On the one hand, they'll tout wide bandwidth source material as being beneficial with something like the Mola Mola Tambaqui with this fabulous HF IMD response...

Stereophile Jan 2022

View attachment 236587


But on the other hand, they want to brick wall filter at 20kHz when it comes to amplification specifications. I can see why...

Is he just trying to sell amps if the 1ET400A CCIF IMD test beats out the AHB2? If engineering is about making the right compromises to suite a particular application, optimally, then B.Putzeys has made all the right ones...

It also shows with the Tambaqui. I remember when I heard that thing; a jaw to the floor experience and certainly was an audition that sticks in my mind. Of course he could've brick-walled it, like Chord, but from his words on the Stereophile review, the filter was entirely tuned by ear, but results in a slow roll-off (thus smoother step response) but wonky 90kHz BW THD+N, yet the circuit is designed such that no components above 22kHz affect the audible band in any detrimental way (it also has an analogue filter at the end of the chain). It's basically doing what NOS is "trying" to do, but correctly, without the insane IMD and nasty multi-tone measurements you can see with e.g. the Holo May when input 44.1 kHz. In any case, any components you see above 22kHz is completely irrelevant.

 

[theREALdotnet]

Just to make clear what I'm talking about - it's this sort of thing:

View attachment 236632

Nothing there is about "mass". It's thermal conductivity across the heat spreader and up the pipes (hence copper), then then area on the fins.

Meh, in the olden days we used variable heatsinks:

 

[DonH56]

Meh, in the olden days we used variable heatsinks:

View attachment 236692

LOL. I have an old four- or five-gang version that handles up to 2 kW that I used to tune my ham antenna.

 

[restorer-john]

I love it! Variable Heatsinks. Lol

 

[restorer-john]

LOL. I have an old four- or five-gang version that handles up to 2 kW that I used to tune my ham antenna.

You’ve got antenna made of ham? Only in America.

 

[DonH56]

You’ve got antenna made of ham? Only in America.

Used to, bacon now. But don't knock ham; remember, "for a chicken, it's all in a day's work, but for the pig it's a lifetime commitment".

 

[juliangst]

Why does the 1ET7040SA based Apollon amp claim so much high power output? I thought all Purifi amps should perform the same if they have a powerful enough power supply

 

[Matias]

Why does the 1ET7040SA based Apollon amp claim so much high power output? I thought all Purifi amps should perform the same if they have a powerful enough power supply

View attachment 236928

An error...

Edit: hmm they are using the higher voltage PSU, the Hypex SMPS1200A700 is 85V. So maybe it is correct then.

 

[pogo]

I thought all Purifi amps should perform the same if they have a powerful enough power supply

It does not only depend on the PSU, see also my post #194 & #271.

Here is a good example with the 1ET400A:
The Dali Kore have an impedance minimum of 3.3Ohms at 65Hz. Considering that these are usually driven with four bridged NAD M23, that's quite a feat for any Purifi module. In bridge mode this impedance is halved for the Purifis, i.e. 1.65ohm and seems to work well in real life (pictures of setup)

With appropriate implementation, the 1ET7040SA should also be able to do more than 950W @ 2ohm.