More than 30W per chanel Class A transistor amps

[Howard Stone]

Any recommendations? Reliability and ease of service important.

[Trevor Wilson]

On 5/02/2019 8:27 am, Howard Stone wrote:
> Any recommendations? Reliability and ease of service important.
>

**Krell KSA50. 50 Watts Class A, 75 Watts Class A/B. Very simple
topology, easy to work on, built-in fan cooling, reliable.
Krell KSA100. 100 Watts Class A. Uncertain about Class A/B power, but
likely in excess of 150 Watts. Similar construction to the KSA50, so the
same comments apply.

If unmolested, both will almost certainly require new electros by now.
The KSA100 was a particularly impressive sounding thing.

Dunno why you want Class A. A decent, high bias Class A/B amp, up to a
couple of Watts will sound just fine.

--
Trevor Wilson
www.rageaudio.com.au

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

[Howard Stone]

Ah yes I checked your old posts and saw you have spoken highly of that Krell in the past. Noted.

Why class A? Well, why not?

I have a little second system with JR 149s, right now it's just got a Quad 303/34. I love those speakers! And I feel that with really a really top amp they could be quite fabulous.

But I don't want the trouble of valves . . . so Class A seems a good line of investigation.

[Trevor Wilson]

**I have always regarded Class A as the lazy designer's way of achieving
good sonic performance. Thing is this:

The idea of Class A is to eliminate crossover distortion, which MAY
exist in a poorly thought out Class A/B amp. Crossover distortion (where
is exists) is caused by the non-linearity of the output devices (valves,
BJTs MOSFETs, etc) which occurs at low(ish) currents. In the case of
BJTs (Bipolar Junction Transistors - or just transistors), bias current
needs to be set at around 20mA (PER OUTPUT DEVICE PAIR) to eliminate the
effects of the kink (or knee) in the amplification curve. For MOSFETs,
the figure is somewhere around an Amp or so (>1,000mA). Many BJT amp
designers set the bias figure higher than 20mA, thus giving rise to
allegedly superior performance. A bias figure that equates to a few
Watts Class A (as opposed to the more usual, 20 ~ 100 milliwatts) will
provide all the benefits of Class A, but with far more modest power
consumption.

Either way, old Krells are plentiful, well built, easy to work on,
properly cooled (using fans) and provide genuine Class A performance.
This is as opposed to many of Krell's competitors, who CLAIM Class A
performance, but don't actually deliver on that promise. Musical
Fidelity is one of the more notorious companies here.

--
Trevor Wilson
www.rageaudio.com.au

[Howard Stone]

Which amps have been designed to have the crossover benefits of class A without the consumption.

I've decided to take a Electrocompaniet ECI-2 on a trial, it's not pure class A but I'll give it a go. The JR149s are by no means big speakers, but they have a reputation of performing much better if driven by a powerful amp. If it doesn't work out, then I can see a Krell KSA 100 with the offer of a trial -- though I'm a bit concerned about preamps (I only have the Quad 34 and a passive) I also noticed a cheap Sugden A21i for sale, but it's a small amp in terms of output -- may work though.

[Peter Wieck]

On Thursday, February 7, 2019 at 3:05:59 PM UTC-5, Howard Stone wrote:

> I've decided to take a Electrocompaniet ECI-2 on a trial, it's not pure class A but I'll give it a go. The JR149s are by no means big speakers, but they have a reputation of performing much better if driven by a powerful amp. If it doesn't work out, then I can see a Krell KSA 100 with the offer of a trial -- though I'm a bit concerned about preamps (I only have the Quad 34 and a passive) I also noticed a cheap Sugden A21i for sale, but it's a small amp in terms of output -- may work though.

Just curious: Have you ever tried a basic Brute Force amp? Something over 200 wpc @ 8 ohms? Not Phase Linear, but something well made.

Peter Wieck
Melrose Park, PA

[Howard Stone]

No, they’re very rare in the UK and Europe I think. Import charges are probably prohibitive for importing from elsewhere.

[Peter Wieck]

On Friday, February 8, 2019 at 9:43:50 AM UTC-5, Howard Stone wrote:
> No, they’re very rare in the UK and Europe I think. Import charges are probably prohibitive for importing from elsewhere.

Would you like to? Surprisingly often, I come across them in my travels. I have my eye on a Citation 16 now that is in need of maintenance that I might be able to get for the hauling - not instantly, of course. But reasonably soon.

http://www.highfidelityreview.com/inside-out-harman-kardon-citation-16-amplifier.html

Were that to come to pass - I would part with it for the cost of shipping. Sometime around end-April, I expect.

[Howard Stone]

That’s a kind offer but I really want to focus on exploring class A amps at the moment.

[Trevor Wilson]

**Well, in truth, very few modern (say: >1970) SS amps exhibit
measurable crossover distortion. Which is to say that, since the
distortion is unmeasurable, it is, by inference, inaudible. To ensure
reduced crossover distortion, I would look for an amp that has elevated
bias currents. In the 1980s (and possibly later) Marantz released a
range called "Quarter A", where bias current was adjusted such that
Class A power reached 25% of maximum power output. In a 100 Watt amp,
that means the amplifier deliver the first 25 Watts in Class A. This
would usually amount to far more than the average listening levels for
normal humans.

>
> I've decided to take a Electrocompaniet ECI-2 on a trial, it's not pure class A but I'll give it a go.

**EC products are generally pretty decent. You should find it acceptable.

The JR149s are by no means big speakers, but they have a reputation of
performing much better if driven by a powerful amp. If it doesn't work
out, then I can see a Krell KSA 100 with the offer of a trial -- though
I'm a bit concerned about preamps (I only have the Quad 34 and a
passive) I also noticed a cheap Sugden A21i for sale, but it's a small
amp in terms of output -- may work though.
>

**If you require big power, avoid the Sugden.


--
Trevor Wilson
www.rageaudio.com.au

[Howard Stone]

The electrocompaniet was a disaster -- distortion, overheating, keeps cutting out. So back it goes!

[Howard Stone]

I'm now interested in buying an Accuphase amp, and I've found one which looks as though it could be just the job, an Accuphase P-266. But it could need serious restoration.

I want to ask a favour. I have some images of the insides of the amp here. Could someone look at them and see if there's anything obviously wrong, any obvious red flags?

https://imgur.com/a/FyV3UcK


There's a description of the P-266 here

https://audio-database.com/Accuphase-kensonic/amp/P-266-e.html

[Trevor Wilson]

**Two problems that I can see:

1) It is not Class A (though it is claimed to be). In fact, it is not
even remotely close to Class A. Personally, I would avoid any product
where the manufacturer has deliberately lied in the description of the
product. Look at the claimed power output and relate it to the power
consumption. The THEORETICAL efficiency of a push pull Class A amp is
50%. In the real world, the figure will be somewhere between 30% ~ 40%.
Accuphase lied. Liars should cannot be trusted.
2) It uses MOSFETs in a standard Class A/B configuration.

OTOH, Accuphase products are generally quite well built, using quality
parts. Obtaining service data is often challenging (impossible), as can
obtaining some spare parts.

Find an old Krell and rebuild it. MUCH better choice and offers real
Class A muscle. Even better, old Krells don't use MOSFETs and the really
old ones are fan cooled.

[Howard Stone]

You are great help, much appreciated.

[Trevor Wilson]

On 18/02/2019 11:09 pm, Howard Stone wrote:
> You are great help, much appreciated.
>

**No worries. And do seek out an old Krell (KSA100 would be my choice).
Parts and schematics are readily available for the old fan cooled
models. You'll find lots of people who have rebuilt them describe their
experiences on the net. Properly rebuilt, they are bullet-proof.
Personally, I would avoid the later, convection cooled models. Fan
cooled models are much better and the topology is less complicated.

[Howard Stone]

I've just bought a KSA 50, it was practically given away. It blew its owner's speakers and he just lost confidence with it -- I've shipped it directly to a good engineer in the UK to do a full restoration.

Anyway assuming everything works out, my plan is to feed the source directly into the power amp, with no pre. I only have one source -- a DAC (DPA Bigger Bit) which has an output of 1.96V.

If not I have two spare preamps, a Quad 34 the pre section from my Electrocompaniet ECI-2

Everyone says that I should try it with am Audio Research valve pre -- but how can this possibly be an improvement over no pre at all???

[Trevor Wilson]

**"Everyone says"? Not me. ARC preamps are HIGHLY variable in quality.
Some models are excellent (IOW: They introduce almost no colouration
into the sound) and some are highly coloured (distorted). Many listeners
like that kind of distortion. I am not one of those listeners. The
REALLY expensive ARC preamps (REF10, et al) are very good indeed. They
are also cripplingly expensive (for me). Models like the SP7 and SP9 are
(IMO) junk.

I don't know the Electrocompaniet, so I can't comment.
I am not a fan of the Quad 34. It was a pretty uninspiring preamp.

You could use no preamp at all, providing input/output impedances are
suitable and that should provide an excellent result.

Congrats on your purchase. Be prepared to pay reasonable Bucks (or Quid)
to have it serviced. Those old guys are worth the effort.

[Peter Wieck]

On Sunday, March 3, 2019 at 7:21:11 PM UTC-5, Howard Stone wrote:

> Everyone says that I should try it with am Audio Research valve pre -- but how can this possibly be an improvement over no pre at all???

It is called "Gain" and "Transient". For the sake of simplicity and discussion, let us agree on three things:

1. Whatever amplifier is in use, it is rated at its full RMS output against a 2V input at 1,000 hz.

2. It is a well-designed device using quality parts and assembled with great care.

3. The speakers in use are also good, well-made and rated at 90 dB @ 1 meter @ 1 watt. And rated at an 'average' of 50 watts.

Some basics: The transient associated with a snare-drum rimshot is approximately 120 dB. The transient associated with the onset of the organ Bombard pipes in the Saint-Saens organ symphony on a well-made recording very nearly, or actually peaks at 30 dB over average.

Most amplifiers are capable of momentary peaks and transients far above their rated RMS. Most speakers can handle transients and peaks far above their 'average' rating.

If a pre-amp has no gain over the original signal source, which is a nominal 2V, the peaks and transients will not be expressed - unless the actual listening level is a minimum of 30 dB below the average.

Pre-amps are awful things, just like democracy. Except for the alternatives.

Peter Wieck
Melrose Part, PA

[Howard Stone]

Aha, this is an argument for active pre-amps. Passive pre-amps presumably all have no gain.

[Peter Wieck]

On Monday, March 4, 2019 at 9:59:57 AM UTC-5, Howard Stone wrote:

> Aha, this is an argument for active pre-amps. Passive pre-amps presumably all have no gain.

Correct. Passive pre-amps are no more than switches.

Peter Wieck
Melrose Park.

[Howard Stone]

So I guess that, given that my project is about exploring Class A with my speakers, that IF I buy a pre-amp, it too needs to be Class A.

And I suppose I find out whether there are any Class A DACs, because I think their analogue end must involve some gain.

[dpierce.ca...@gmail.com]

On Monday, March 4, 2019 at 11:06:40 AM UTC-5, Peter Wieck wrote:
> On Monday, March 4, 2019 at 9:59:57 AM UTC-5, Howard Stone wrote:
>
> > Aha, this is an argument for active pre-amps. Passive pre-amps presumably all have no gain.
>
> Correct. Passive pre-amps are no more than switches.

And noise and distortion generators.

--
+--------------------------------+
| Dick Pierce |
| Professional Audio Development |
| cartchunk.org |
+--------------------------------+

[Peter Wieck]

It's only money: http://www.avalondesign.com/pdfs/hr_AvalonLit_PRECOMPCHARTnew.pdf

Or, not so much:
https://www.aliexpress.com/item/Lusya-SY99A-class-A-Preamplifier-HiFi-Stereo-Preamp-Assembled-Board-beyond-NAC-152-J2C-MBL6010-amplifier/32917979538.html?spm=2114.search0302.3.201.2d9c189bOAYW0l&ws_ab_test=searchweb0_0,searchweb201602_0_10130_453_454_10618_536_10890_317_537_319_10059_10696_10084_10083_10547_10546_10887_10307_321_10548_322_10902_10065_10068_10103_10884_10545,searchweb201603_0,ppcSwitch_0&algo_pvid=86cc1e2d-782c-4973-9cb9-44eaefa96d67&algo_expid=86cc1e2d-782c-4973-9cb9-44eaefa96d67-25

[dpierce.ca...@gmail.com]

On Monday, March 4, 2019 at 11:32:18 AM UTC-5, Howard Stone wrote:
> On Monday, 4 March 2019 16:06:40 UTC, Peter Wieck wrote:
> > On Monday, March 4, 2019 at 9:59:57 AM UTC-5, Howard Stone wrote:
> >
> > > Aha, this is an argument for active pre-amps. Passive pre-amps presumably all have no gain.
> >
> > Correct. Passive pre-amps are no more than switches.
> >
> > Peter Wieck
> > Melrose Park.
>
> So I guess that, given that my project is about exploring Class A
> with my speakers, that IF I buy a pre-amp, it too needs to be Class A.

The chances you're going to come across a preamp whose active
'stages are NOT biased class-A is pretty small. I can think of none
of the top of my head. (Some earlier IC-based preamp may have used
op-amps that had a push-pull output stage that may have run class
AB, but, again, I can think of none).

> And I suppose I find out whether there are any Class A DACs,
> because I think their analogue end must involve some gain.

Same goes here as well: you're not likely to find a DAC output
stage that is NOT class-A biased.

NOTE: A "DAC" really consists of two fundamental parts: the
digital-to-analog stage, which, these days, will also
include the bulk of the anti-imaging filtering, done
in the digital domain: this part feeds an analog signal
to the line driver stage, which, again, is very unlikely
to be anything but class-A.

Now, the issue of whether an audio power amp is biased class-A
vs class-AB or even class-B (and, yes, even various switch-mode
implementations) is a design issue forced on the designer byu
considerations of efficiency and thus weight (less efficiency
means more copper and iron in the power supply) heat dissipation
(larger heat sinks, bigger chassis area for heat sinks), cost
and therefore success in the chosen market segment. When you're
talking about even, say, a modest output power of 50 watts/channel
continuous, there's a very large difference in whether you
implement it as a class-A or class AB, and even withing those,
a large range of demands on power, heat and the subsequent costs.

However, in a preamp, where the total power output requirements
are orders of magnitude less, these considerations become far
less of a design driver and thus do not force decisions between
biasing classes that are either engineering or economically
driven.

Putting it more simply, it's MUCH easier to design, manufacture
and support class A amplifiers dealing with low(er) signal levels
and substantially lower powers than it is to to the same where you
need LOTS of power.

Case in point: consider a sophisticated solid state preamp with,
maybe, 4 gain stages (RIAA, gain/balance, tone and output), using
the output stage as one instance, say it's rated at 2VRMS into
10 kOhm load: that's, lessee 0.02 uW (or, to make it even more
impressive sounding: 20 billionths of a watt). And let's say each
stage has roughly the same power output requirement. The difference
in heat dissipation and power supply requirements from running all
those stages at class A vs all of them at some class-AB is probably
on the order of the amount of power required to run the front panel
lights.

Plus the fact that, in the context of audio amplifiers, the choices
over whether to run things class-A vs class-AB is ONLY applicable
to push-pull topologies. To flip this around, if you wanted to have
the choice of running things class-A vs class-AB in a preamp, you
would have had, first, to make the design choice that you're designing
a push-pull preamp from input to output.

Don't spend a lot of time worrying about whither the preamp you
might buy is class A or class AB: because you'd be spending a HUGE
amount of time searching for the class-AB preamp that you don't want
to buy.

[Trevor Wilson]

On 4/03/2019 11:45 pm, Peter Wieck wrote:
> On Sunday, March 3, 2019 at 7:21:11 PM UTC-5, Howard Stone wrote:
>
>> Everyone says that I should try it with am Audio Research valve pre -- but how can this possibly be an improvement over no pre at all???
>
> It is called "Gain" and "Transient". For the sake of simplicity and discussion, let us agree on three things:
>
> 1. Whatever amplifier is in use, it is rated at its full RMS output against a 2V input at 1,000 hz.

**Just a nit: There is no such thing as RMS power. It is either
continuous Watts, or just plain ole Watts. The moniker RMS can be
applied to Volts or Amps, but not to the product of the two.

>
> 2. It is a well-designed device using quality parts and assembled with great care.
>
> 3. The speakers in use are also good, well-made and rated at 90 dB @ 1 meter @ 1 watt. And rated at an 'average' of 50 watts.
>
> Some basics: The transient associated with a snare-drum rimshot is approximately 120 dB. The transient associated with the onset of the organ Bombard pipes in the Saint-Saens organ symphony on a well-made recording very nearly, or actually peaks at 30 dB over average.
>
> Most amplifiers are capable of momentary peaks and transients far above their rated RMS.

**Most amplifiers that employ crappy power supplies, yes. The Krell
KSA50 is not such an amplifier. It's CONTINUOUS power rating is very,
VERY close to it's transient power rating. The continuous power of a
KSA50 is around 75 Watts. The transient figure is something like 76
Watts. Had Krell decided to skimp on the power transformer, then the
difference between the transient figure and the continuous figure would
be far greater. The weird thing is the manufacturers make a big song and
dance about the size of the transient number vs. the continuous one,
when the bigger the difference, the crappier the amplifier.

Most speakers can handle transients and peaks far above their
'average' rating.

**Oh, absolutely. One need only examine the voice of a typical bass
driver and relate that to the power handling. Almost all domestic
speakers are rated to be connected to an amplifier of XX Watts and being
driven with music. Since music has, at worst a peak to average figure of
around 10dB, then one may assume that the average power applied to a
speaker is always somewhere less than 10% of the continuous power rating
of the amplifier.

>
> If a pre-amp has no gain over the original signal source, which is a nominal 2V, the peaks and transients will not be expressed - unless the actual listening level is a minimum of 30 dB below the average.

**Well that depends on the sensitivity of the amplifier.

>
> Pre-amps are awful things, just like democracy. Except for the alternatives.

**A decent preamp need not be expensive.

[Trevor Wilson]

On 5/03/2019 3:32 am, Howard Stone wrote:
> On Monday, 4 March 2019 16:06:40 UTC, Peter Wieck wrote:
>> On Monday, March 4, 2019 at 9:59:57 AM UTC-5, Howard Stone wrote:
>>
>>> Aha, this is an argument for active pre-amps. Passive pre-amps presumably all have no gain.
>>
>> Correct. Passive pre-amps are no more than switches.
>>
>> Peter Wieck
>> Melrose Park.
>
> So I guess that, given that my project is about exploring Class A with my speakers, that IF I buy a pre-amp, it too needs to be Class A.

**Almost all preamps operate in Class A.

>
> And I suppose I find out whether there are any Class A DACs, because I think their analogue end must involve some gain.

**That will depend on how any OP amps used happen to be configured. I
wouldn't lose sleep over it.

[Peter Wieck]

On Monday, March 4, 2019 at 2:33:50 PM UTC-5, Trevor Wilson wrote:

> **Most amplifiers that employ crappy power supplies, yes. The Krell
> KSA50 is not such an amplifier. It's CONTINUOUS power rating is very,
> VERY close to it's transient power rating. The continuous power of a
> KSA50 is around 75 Watts. The transient figure is something like 76
> Watts. Had Krell decided to skimp on the power transformer, then the
> difference between the transient figure and the continuous figure would
> be far greater. The weird thing is the manufacturers make a big song and
> dance about the size of the transient number vs. the continuous one,
> when the bigger the difference, the crappier the amplifier.

Back to those nits: I keep a power-amp that will make a continuous 175 watts @ 8 ohms all day and all night. However, due to its power-supply reserve, output devices and so forth, it will make a transient (as a function of value and time) up to 1,000 watts for a very brief moment, less for longer moments.

> Most speakers can handle transients and peaks far above their
> 'average' rating.
>
> **Oh, absolutely. One need only examine the voice of a typical bass
> driver and relate that to the power handling. Almost all domestic
> speakers are rated to be connected to an amplifier of XX Watts and being
> driven with music. Since music has, at worst a peak to average figure of
> around 10dB, then one may assume that the average power applied to a
> speaker is always somewhere less than 10% of the continuous power rating
> of the amplifier.

Um. Not so much. Back in the day, AR made a huge point on the power-handling capacity of their speakers, and used "magic" such as ferro-fluid and so forth for heat dissipation.

a) Their speakers were rated rather low as compared to their contemporaries.
b) But their transient capacity was rated massively higher.
c) A great deal of my signal has a P/A of 20 dB. Some few examples have a P/A of 30 dB. Gregorian chant, single instrument, single voice may have the P/A of 10 dB that you reference, and those people may never turn the volume past 9:00 - but not all of us.

> > If a pre-amp has no gain over the original signal source, which is a nominal 2V, the peaks and transients will not be expressed - unless the actual listening level is a minimum of 30 dB below the average.
>
> **Well that depends on the sensitivity of the amplifier.

No. That is a fact. If the pre will put out no more than 2V, then it will drive the amp to peak for some percentage of the time. But not past the peak, ever. So, if the P/A is 10 dB, that will be well within the capacity of the amp (and the speaker). A close-run thing at 20 dB, impossible at 30 dB. But, using the example of a transient, and 90 dB efficient speakers - a perfect example of "impossible" and why I picked that particular item.

> > Pre-amps are awful things, just like democracy. Except for the alternatives.
>
> **A decent preamp need not be expensive.

That is absolutely true. A decent vintage pre may be had for well under US$200, tube or SS, and may, for approximately the same again or less, be tweaked to as-good-as anything off the shelf today.

[Trevor Wilson]

On 5/03/2019 7:11 am, Peter Wieck wrote:
> On Monday, March 4, 2019 at 2:33:50 PM UTC-5, Trevor Wilson wrote:
>
>> **Most amplifiers that employ crappy power supplies, yes. The Krell
>> KSA50 is not such an amplifier. It's CONTINUOUS power rating is very,
>> VERY close to it's transient power rating. The continuous power of a
>> KSA50 is around 75 Watts. The transient figure is something like 76
>> Watts. Had Krell decided to skimp on the power transformer, then the
>> difference between the transient figure and the continuous figure would
>> be far greater. The weird thing is the manufacturers make a big song and
>> dance about the size of the transient number vs. the continuous one,
>> when the bigger the difference, the crappier the amplifier.
>
> Back to those nits: I keep a power-amp that will make a continuous 175 watts @ 8 ohms all day and all night. However, due to its power-supply reserve, output devices and so forth, it will make a transient (as a function of value and time) up to 1,000 watts for a very brief moment, less for longer moments.

**And mine, with it's 5.5kVA (continuously rated) power transformer will
maintain it's power for as long as the mains breakers hold out. It's
transient power is VERY close to it's continuous power. Load impedance
is irrelevant (above 0.5 Ohms anyway). If the load impedance dips to a
low enough figure, then it can maintain a a couple of kW for a long time
(till the breaker opens).

>
>> Most speakers can handle transients and peaks far above their
>> 'average' rating.
>>
>> **Oh, absolutely. One need only examine the voice of a typical bass
>> driver and relate that to the power handling. Almost all domestic
>> speakers are rated to be connected to an amplifier of XX Watts and being
>> driven with music. Since music has, at worst a peak to average figure of
>> around 10dB, then one may assume that the average power applied to a
>> speaker is always somewhere less than 10% of the continuous power rating
>> of the amplifier.
>
> Um. Not so much. Back in the day, AR made a huge point on the power-handling capacity of their speakers, and used "magic" such as ferro-fluid and so forth for heat dissipation.

**Ferro-fluid is used only in HF drivers and some mids and, except for
NEAR drivers, not in bass drivers. And it is the bass driver that limits
the maximum thermal limits of a speaker system. Except for some pro
drivers, few domestic bass drivers can cope with sine wave power levels
of more than 50 Watts or so.

>
> a) Their speakers were rated rather low as compared to their contemporaries.

**[ANECDOTE] Back when I was a young(er) man, I built a pair of Bailey
T-lines, using KEF drivers. The bass driver was the venerable KEF B-139.
Rated at a massive 30 Watts. This was back in the day when manufacturers
like KEF rated their speaker as if they were required to handle sine
waves. As a young-ish man (early 20s) I tended to throw the odd party
for my friends. Much alcohol was consumed and much rock music was played
at high levels. Over time, I gradually improved my amplification,
culminating in my prized, Marantz Model 500 power amp. Rated at 250
Watts/channel, I measured the Model 500 at 310 Watts @ 8 Ohms. It's
superb fan cooling ensured that it could maintain high power levels at 8
or 4 Ohm loads for long periods (yes I know that maximum PDiss occurs at
around 40%). After one particularly grueling party, one of the B-139
drivers failed. It would have been at the end of around 5 hours of near
clipping levels for the Model 500. A couple of days later, I stripped
the driver down and repaired the voice coil lead in wire. No big deal.
There was absolutely no sign of heat distress on the voice coil. I
firmly believe that it was (slightly) sub-standard manufacture. However,
the same cannot be said about the electrolytic capacitors in the
crossovers. These had been severely damaged and were replaced with
plastic film types. IME, AR have always used adequately sized voice
coils, capable of handling far more power than my KEF b-139 drivers.

> b) But their transient capacity was rated massively higher.

**No, it wasn't.

> c) A great deal of my signal has a P/A of 20 dB. Some few examples have a P/A of 30 dB. Gregorian chant, single instrument, single voice may have the P/A of 10 dB that you reference, and those people may never turn the volume past 9:00 - but not all of us.

**I did say: "Worst case".

>
>>> If a pre-amp has no gain over the original signal source, which is a nominal 2V, the peaks and transients will not be expressed - unless the actual listening level is a minimum of 30 dB below the average.
>>
>> **Well that depends on the sensitivity of the amplifier.
>
> No. That is a fact. If the pre will put out no more than 2V, then it will drive the amp to peak for some percentage of the time. But not past the peak, ever. So, if the P/A is 10 dB, that will be well within the capacity of the amp (and the speaker). A close-run thing at 20 dB, impossible at 30 dB. But, using the example of a transient, and 90 dB efficient speakers - a perfect example of "impossible" and why I picked that particular item.

**Again: It depends on the sensitivity of the amplifier.

>
>
>>> Pre-amps are awful things, just like democracy. Except for the alternatives.
>>
>> **A decent preamp need not be expensive.
>
> That is absolutely true. A decent vintage pre may be had for well under US$200, tube or SS, and may, for approximately the same again or less, be tweaked to as-good-as anything off the shelf today.

**Perhaps.

[Trevor Wilson]

On 5/03/2019 5:11 am, dpierce.ca...@gmail.com wrote:
> On Monday, March 4, 2019 at 11:32:18 AM UTC-5, Howard Stone wrote:
>> On Monday, 4 March 2019 16:06:40 UTC, Peter Wieck wrote:
>>> On Monday, March 4, 2019 at 9:59:57 AM UTC-5, Howard Stone wrote:
>>>
>>>> Aha, this is an argument for active pre-amps. Passive pre-amps presumably all have no gain.
>>>
>>> Correct. Passive pre-amps are no more than switches.
>>>
>>> Peter Wieck
>>> Melrose Park.
>>
>> So I guess that, given that my project is about exploring Class A
>> with my speakers, that IF I buy a pre-amp, it too needs to be Class A.
>
> The chances you're going to come across a preamp whose active
> 'stages are NOT biased class-A is pretty small. I can think of none
> of the top of my head. (Some earlier IC-based preamp may have used
> op-amps that had a push-pull output stage that may have run class
> AB, but, again, I can think of none).

**I'm not certain about that. Obviously, depending on the load
impedance, many OP based preamps may, in fact, be operating largely as
Class A/B amplifiers. A system popularised back when the 5532/4 was
released, was to force it's operation into Class A mode, by inserting a
current source into the feedback loop. I've never measured it, but I
assume that suggests it is a largely Class B output stage, but by using
the kludge, becomes Class A (load dependent, of course). Maybe I'll
measure it one day....

>
>> And I suppose I find out whether there are any Class A DACs,
>> because I think their analogue end must involve some gain.
>
> Same goes here as well: you're not likely to find a DAC output
> stage that is NOT class-A biased.

**Not that it matters either, I agree.

>
> NOTE: A "DAC" really consists of two fundamental parts: the
> digital-to-analog stage, which, these days, will also
> include the bulk of the anti-imaging filtering, done
> in the digital domain: this part feeds an analog signal
> to the line driver stage, which, again, is very unlikely
> to be anything but class-A.
>

>

> However, in a preamp, where the total power output requirements
> are orders of magnitude less, these considerations become far
> less of a design driver and thus do not force decisions between
> biasing classes that are either engineering or economically
> driven.
>
> Putting it more simply, it's MUCH easier to design, manufacture
> and support class A amplifiers dealing with low(er) signal levels
> and substantially lower powers than it is to to the same where you
> need LOTS of power.

**Quite so.

>
> Don't spend a lot of time worrying about whither the preamp you
> might buy is class A or class AB: because you'd be spending a HUGE
> amount of time searching for the class-AB preamp that you don't want
> to buy.

**Load depending, yes.

[dpierce.ca...@gmail.com]

On Monday, March 4, 2019 at 4:47:15 PM UTC-5, Trevor Wilson wrote:
> On 5/03/2019 7:11 am, Peter Wieck wrote:
>>
>> Um. Not so much. Back in the day, AR made a huge point on
>> the power-handling capacity of their speakers, and used
>> "magic" such as ferro-fluid and so forth for heat dissipation.

Back in "which" day? Ferrofluid was not commonly used until the
1980's, long after the original AR days, and more like into
the Teledyne/Acoustic Research era and later.

> **Ferro-fluid is used only in HF drivers and some mids and,
> except for NEAR drivers, not in bass drivers

Quite correct. Some tweeters used it, very few midranges used and
not a single bass or wide-band driver of any commercial viability
used it. It's use or non-use is well correlated with the maximum
excursion capability of the driver: it's completely unsuitable
for drivers with large excursions (typically greater than 0.5mm)
because it literally is pumped out of the gap.

> And it is the bass driver that limits
> the maximum thermal limits of a speaker system.

For the most part, the actual data from the field contradicts this.
<Looking at data from many of the clients I have dealt with over
the years, power-related (either thermal or mechanical) speaker
failures by far are dominated by thermal failure of tweeters,
while the remainder are mechanical failures of woofers. For certain,
there are overlaps (e.g., mechanical tweeters failures and thermal
breakdown in woofers), but, based on field data across a very large
number of systems, these latter are very much in the minority.

> Except for some pro drivers, few domestic bass drivers can cope
> with sine wave power levels of more than 50 Watts or so.

It's not so simple a relation as that. For sure, one can say that
if the voice coil is subject to sufficient input such that you're
dissipating the vast majority of the input power as heat in the voice
coil, that's largely true.

But consider the following: all other things being equal, for what
might be the output voltage of the amplifier that would result in
50 watts being produced at, say, lower midrange frequencies would
produce MUCH less power (and thus thermal dissipation) at below
100Hz or say at 2000 Hz, where the impedance is substantially higher.

Especially on the low bass, where the impedance is dominated by the
resonant peak from the woofer, and combined with the fact that the
excursion of the woofer gores as the inverse square of frequency,
the thermal problems a secondary compared to the mechanical limits.

[Trevor Wilson]

On 6/03/2019 5:27 am, dpierce.ca...@gmail.com wrote:
> On Monday, March 4, 2019 at 4:47:15 PM UTC-5, Trevor Wilson wrote:
>> On 5/03/2019 7:11 am, Peter Wieck wrote:
>>>
>>> Um. Not so much. Back in the day, AR made a huge point on
>>> the power-handling capacity of their speakers, and used
>>> "magic" such as ferro-fluid and so forth for heat dissipation.
>
> Back in "which" day? Ferrofluid was not commonly used until the
> 1980's, long after the original AR days, and more like into
> the Teledyne/Acoustic Research era and later.
>
>> **Ferro-fluid is used only in HF drivers and some mids and,
>> except for NEAR drivers, not in bass drivers
>
> Quite correct. Some tweeters used it, very few midranges used and
> not a single bass or wide-band driver of any commercial viability
> used it. It's use or non-use is well correlated with the maximum
> excursion capability of the driver: it's completely unsuitable
> for drivers with large excursions (typically greater than 0.5mm)
> because it literally is pumped out of the gap.

**You should tell the guys from NEAR (New England Audio Resource) that.
I've been using their speakers since the early 1990s. They are
delightful things. Although I've only been using their smallest product
from that time (the NEAR 10M), which incorporates a 130mm bass driver, I
can assure you that it has an Xmax of considerably more than 0.5mm. NEAR
does/did manufacture a 200mm bass driver as well.

http://www.nearspeakers.com/magnetic-liquid-suspension.html

**No argument from me there.

[Howard Stone]

This may be of interest

https://www.pinkfishmedia.net/forum/threads/rebuild-of-classic-krell-ksa50-fan-heater.227491/

(Still not heard it yet -- the latest info is slightly worrying in fact, I got an email saying:

"There is a slight hum from the transformers which is pretty quiet most of the time but at peak TV time last night with some DC on the mains it did hum quite badly so you may want to use it with a DC blocker. To be clear, this is from the amp chassis itself and mechanical hum from the transformers and not hum through the speakers. It is silent from the speakers in background noise." )

[Peter Wieck]

50 hz current has that awful effect in some cases - and will (very slightly) de-rate the amp. Given that those appear to be toroidal transformers, there is no easy cure other than an isolation transformer at the head-end. And that will be a pretty massive item, 20 A (@ 120 VAC) anyway.

As it happens, if that amp is still set to US voltage (120 VAC), there are some very good 240:120 isos out there that will do nicely.

http://www.farnell.com/datasheets/2279306.pdf?_ga=2.187485032.1926575215.1557160509-1096436046.1551096751&_gac=1.184557019.1556715333.Cj0KCQjwh6XmBRDRARIsAKNInDFhpeJKNAxViDjwsTsi2VzMrHglXPmgALeX25Ld31b6Fo3bVrcTomoaAlPcEALw_wcB

Nor is there anything wrong with an iso on the system for general noise reduction.

On the other hand, I suggest that you not panic, but see how it goes in your environment.

[Trevor Wilson]

**Pretty normal with many toroidal transformers. They can become noisy
when subjected to excessive DC on the mains supply. It is location
dependent. You may not have a problem in your location. As suggested, a
DC blocker is a simple, inexpensive possible solution.

A few comments on the repair:

* Your tech appears to have done an excellent and thoroughly
professional job.
* See if you can arrange to have the fan suck air UP, rather than
blowing it down. Those idiots at Krell made a huge mistake with early
KSA50 amps, by forcing air down, rather than up. (Hot air rises, after
all). This insanity was rectified in later variants.

[Howard Stone]

Oh the other thing Jez said to me which maybe he of interest was this

“There are speaker fuses at the output which are supposed to be 4A, 5A was found to be fitted. Such high values are unlikely to offer any real protection (2A is more common as a speaker fuse... and does sound bad!) to the speakers but could cause a slight reduction in sound quality. I will be opening it up again to check bias and offset after a couple of days use so if you wish I can bypass them altogether. let me know.”

[Trevor Wilson]

On 7/05/2019 7:47 pm, Howard Stone wrote:
> Oh the other thing Jez said to me which maybe he of interest was this
>
> “There are speaker fuses at the output which are supposed to be 4A, 5A was found to be fitted. Such high values are unlikely to offer any real protection (2A is more common as a speaker fuse... and does sound bad!) to the speakers but could cause a slight reduction in sound quality. I will be opening it up again to check bias and offset after a couple of days use so if you wish I can bypass them altogether. let me know.”
>

**Fuses are a really bad idea on speaker lines, as a fuse is, in
reality, a non-linear resistor. Therefore, a fuse (any fuse) will
introduce distortion. The lower the value of the fuse, the higher the
distortion.

[Howard Stone]

Brilliant, thanks for your advice.

[Trevor Wilson]

On 8/05/2019 5:33 pm, Howard Stone wrote:
> Brilliant, thanks for your advice.
>

**I should add, that, if the fuses are bypassed, then another means of
speaker protection should be considered.

[Peter Wieck]

On Tuesday, May 7, 2019 at 5:44:53 PM UTC-4, Trevor Wilson wrote:

> **Fuses are a really bad idea on speaker lines, as a fuse is, in
> reality, a non-linear resistor. Therefore, a fuse (any fuse) will
> introduce distortion. The lower the value of the fuse, the higher the
> distortion.
>

Source: Bob Cordell "Designing Audio Power Amplifies"

Section '13.11 Fuse, Relay and Connector Distortion' page 268.

Conclusion:

"At 20 Hz, amplifier distortion due to the fuse is calculated to be 0.0033 %."

Me: And, of course no harmonics are affected or altered.

How much are your speakers worth that you would put them at risk over the silliness of fuse adding distortion at audible levels?

[Peter Wieck]

RANT WARNING!

Guys and gals, there is common sense (which isn't), and then, there is received wisdom (AKA: Religion) when it comes to audio. Typically, but not always, they are mutually exclusive. Also commonly practiced in this hobby/religion are several common fallacies, the most prominent of which are:

1. Leaping to conclusions: My cat is grey, therefore all cats are grey.
2. Begging the question: Given that all cats are grey, therefore my cat must be grey.
3. False premises: My cat is wet, therefore it rained last night.

What brought this on was Trevor's post on fuses, which more-or-less encompasses two 1 & 3 of the above, with a glancing blow to 2.

Some Math:

If Output Watts = V^2/Speaker Impedance
Then V = Square Root of output watts x Impedance

OR: If watts =30
Impedance = 8
Voltage = 15.5
Amps = watts/volts
OR: 2 A in this model

Which, of course, will vary as the actual impedance differs from the nominal impedance.

A 4 amp fuse will not even warm up under normal operating conditions - and well-designed speakers will be able to handle a 30-watt input all day and all night as long as there are no/limited DC components. Sources of DC components could be from clipping in an amp with limited protection circuits, or catastrophic failure in one-or-more internal components. At which point the fuse comes into play. Aren't >you< glad that it was?

Other problematic speaker-damage sources: A user chooses to send a pure sine wave into a speaker at full volume. The fuse WILL NOT protect the speaker in this case, and 30 watts is 30 watts = heat. Even the very best voice coils will not withstand a constant 30 watts at a single frequency. Similar damage may be caused by a feedback loop.

Point being that a fuse is a line of defense that when deployed protects >your< speakers better than _ANY_OTHER_OPTION_.

That fuses can add distortion at audible frequencies does not mean that they do add distortion.
That a fuse once added distortion does not mean that all fuses add distortion.
That the speaker is distorting does not mean that the fuse is the cause.

And, of course, I could get started on the many types of fuses out there - including an entire cohort based on audiophoolery.

[Howard Stone]

In fact, the Krell did develop a slight problem. The fan would become noisy, a sort of rattle noise. If you turned it on and off then it would stop, but still I thought it was best to get it changed. The new one is almost silent and, as Trevor recommended, it sucks up. Jez says that the amp runs slightly cooler as a result.

The amp is wonderful!

[Trevor Wilson]

On 27/08/2019 8:38 am, Howard Stone wrote:
> In fact, the Krell did develop a slight problem. The fan would become noisy, a sort of rattle noise. If you turned it on and off then it would stop, but still I thought it was best to get it changed. The new one is almost silent and, as Trevor recommended, it sucks up. Jez says that the amp runs slightly cooler as a result.
>
> The amp is wonderful!
>

**Good to hear. Literally.