Class D by Bruno P

[cafe67]

https://hometheaterhifi.com/30th-anniversary/the-now/life-on-the-edge-a-personal-perspective-on-the-past-present-and-future-of-class-d-audio-amplifiers/

Edited October 25, 2024 by cafe67

[Guest]

Thank you for the excellent article. Bruno has a wicked sense of humour  in trying to slay a few dragons: vinyl, GaN and the superiority of separates! 

 

From a Bruno fan, who owns Kii Three's. 

[Guest]

A MUST read for anyone interested in Hi-fi amps, even though you may prefer traditional ones.  Bruno Putzeys explains why a well designed Class D amp is so good & why the early ones were poor. All backed by science...

 

[almikel]

great read - thanks,

 

I've been considering acquiring a new amp for my AE TD18s, and a UcD or NCore amp is high on the list of likely candidates.

 

Mike

Edited November 4, 2024 by almikel
additional info

[crisis]

I understand precious little of the technical content. I can only say how impressed I have been with my Primare I32 and current I35 as far as sound quality. When I decided on the I32 it was in the midst of some negativity towards class D. The reviews were almost all complimentary. The I35 reviews were more so and my experience suggest to me the technology is vindicated.

[billstevenson]

I have been comparing my recently acquired Marantz PM10, which is slated for use in our summer home, with my Conrad-Johnson ART 7A & ART 88 separates.  PM10 sounds very good in comparison to this Class A tube electronics and I would say completely validates what Bruno says.  There is no reason for anyone to doubt that class D has arrived and offers a very viable option for audiophiles everywhere.  There are differences of course.  The preamp section of the PM10 is no match for the ART 88, but the amp section rated at 200 wpc, but tested by several labs to offer over 600 wpc into either 4 or 8 ohms is a far cry from the 36 wpc of the ART 7A.  So depending of the room size and the speakers to be driven, the choice of either might be optimal depending on use.  In my case the PM10 will be driving a pair of B&W 802 D Diamonds that I found NIB.  The dealer I bought the PM10 from also sells B&W, but is not selling the speakers in this case.  He also sells other high end electronics such as Mark Levinson and McIntosh.  He told me, though, that in his opinion no amplifier in his experience worked better nor sounded better than the PM10 driving B&W 801 or 802 irrespective of the series. 

[alex87]

Great article. Thanks for linking it. 

[andyr]

 @billstevenson's comment - 'one set of speakers bi-wired' - made me curious about something; hopefully, someone can explain it all to me.

 

To be able to be bi-wired, a two-way spkr needs to have:

• a 'parallel' XO (not a 'series' XO)
• and the HP & LP sections of this XO need to be separate (they are joined by the jumpers, for single-wiring).

 

So what happens to the nominal impedance of a two-way spkr, when you bi-wire it?  If a spkr uses 8ohm drivers ... one would expect the result would be an 8ohm impedance rating.  However, the XO messes this up - hence, you can get a widely varying impedance across the spkr's frequency range.

 

But in the case of a two-way passive spkr ... does removing the jumpers and bi-wiring, change the impedance that the amp sees?  After all, both drivers - and their part of the XO network - are still connected back to the one amp ... just with separate wires.

 

And does the situation change when you bi-amp a passive spkr - as now, each amp only sees the one driver (plus its associated HP or LP filter)?

 

[billstevenson]

The amp does not care about anything except the total load that it has to drive.  If you are concerned why don't you measure the resistance of the speaker(s) in whatever configuration you want to drive?  As long as the nominal load is between 4 and 16 ohms you should be ok.  If the load is very wonky the protection circuit will shut down the amp.  When you first start listening stay at low volumes to make sure everything is ok.

[almikel]

Hi Andy,

 

On 17/03/2025 at 11:49 AM, andyr said:

So what happens to the nominal impedance of a two-way spkr, when you bi-wire it? 

The short answer - nothing ...

 

On 17/03/2025 at 11:49 AM, andyr said:

If a spkr uses 8ohm drivers ... one would expect the result would be an 8ohm impedance rating.  However, the XO messes this up - hence, you can get a widely varying impedance across the spkr's frequency range.

The Xover doesn't "mess this up"!

The Xover impedance won't vary as much as the driver impedance - generally the Xover will have a low but flattish impedance within the pass band, and rising impedance through the crossover region with high impedance in the stop band - the Xover does not contribute much to a widely varying impedance across the speakers' frequency range, other than to increase impedance in the stop band which is by design and desired!

 

As you're aware, the impedance of a cone/dome woofer/tweeter changes a lot through its frequency range.

Looking at the impedance curve of the Vifa P21 woofers I used to run (impedance is the lower curve)

 

You can see the typical impedance curve of any cone/dome driver - low impedance at DC, big peak at resonance, then steadily rising at higher frequencies, due to the inductance - a driver is a big coil/inductor after all.

 

The impedance of a 2nd order low pass filter at looks like this:

ie "flattish" in the passband, rising steeply through the Xover region

On 17/03/2025 at 11:49 AM, andyr said:

If a spkr uses 8ohm drivers ... one would expect the result would be an 8ohm impedance rating. 

Not necessarily - only if the impedance of the crossover is much higher than the impedance of the driver.

 

With the driver connected in parallel with the crossover, the impedance seen by the amplifier of the combined load of Xover in parallel with the driver will always be lower than the lowest impedance of either by the formula:

Zamp = (Zxover x Zdriver) / (Zxover + Zdriver)

 

More generically, parallel impedances add like this:

1/Ztotal = 1/Z1 + 1/Z2 + 1/Z3 +... etc

 

For easy maths, let's say at a particular frequency the Xover has 8 ohms impedance, and the driver has 8 ohms impedance:

Impedance seen by the amp Zamp = (8 x 8)/ (8+8) = 4 ohms

 

If the Xover was 100 ohms impedance at the same frequency:

Impedance seen by the amp Zamp = (100 x 8)/ (100+8) = 7.41 ohms - just lower than the lowest impedance of either.

 

Note that I've never tested the impedance of a passive Xover, so I have no clue on what their real world impedance values are other than my googling showing low/flattish impedance in the passband rising above that for a low pass filter (opposite for a high pass filter).

 

My point is that with a passive Xover connected in parallel with a driver, the load seen by the amp will always be lower than impedance of the driver or the Xover, hence likely lower than the nominal impedance of the driver.

 

On 17/03/2025 at 11:49 AM, andyr said:

But in the case of a two-way passive spkr ... does removing the jumpers and bi-wiring, change the impedance that the amp sees?  After all, both drivers - and their part of the XO network - are still connected back to the one amp ... just with separate wires.

Yes by a miniscule amount - Rod Elliot goes into more detail on his site https://sound-au.com/bi-amp.htm

 

On 17/03/2025 at 11:49 AM, andyr said:

And does the situation change when you bi-amp a passive spkr - as now, each amp only sees the one driver (plus its associated HP or LP filter)?

Yes by a miniscule amount, with the gains marginal at best - similarly Rod Elliot goes into more detail on his site.

 

Hi Bill,

On 17/03/2025 at 12:21 PM, billstevenson said:

The amp does not care about anything except the total load that it has to drive.  If you are concerned why don't you measure the resistance of the speaker(s) in whatever configuration you want to drive?  As long as the nominal load is between 4 and 16 ohms you should be ok.  If the load is very wonky the protection circuit will shut down the amp.  When you first start listening stay at low volumes to make sure everything is ok.

Andy was just asking theoretical questions  

On 17/03/2025 at 12:21 PM, billstevenson said:

why don't you measure the resistance of the speaker(s) in whatever configuration you want to drive? 

Measuring the DC resistance of a speaker is trivial - measuring the impedance of a speaker across its frequency range is more involved...noting that @andyr has lots of experience with speakers where the phase angle presents difficult loads to the amp.

On 17/03/2025 at 12:21 PM, billstevenson said:

If the load is very wonky the protection circuit will shut down the amp.

Assuming your amp has appropriate protections circuits - otherwise here in Australia we call that "smoke testing", and after the smoke gets out, we haven't quite worked out how to get that pesky smoke back in

 

cheers,

Mike

[billstevenson]

The PM10 circuitry is pretty sophisticated.  In fact this amp has developed a very solid reputation for reliability.  Keep in mind it was tested during development by Ken Ishiwata using B&W 802 D speakers, which have a very difficult impedance profile.  I am using mine with 802 D Diamonds and it handles them without difficulty, soft or loud.  This is a robust amplifier.