Adcom GFA-555 II vs. GFA-5500 Dynamic Range & Speaker control

[Rein E. Narma]

Here is my repost from RAHE (also sent to the good folks at Adcom) that
addresses my first experience with similarly spec'ed solid state amps
(220wpc) that sound hugely different to my ear. I was hoping someone
out there has already made some of these measurements and can shed some
light what is going on here (and hear).

I am hearing some pretty major differences in the
"power", "slam", "grip" and general dynamic contrast between the two
designs. The 5500's sound like they have a lot less power even though
they play loud and don't visibly clip (light the leds).

My questions (to Adcom):

* Is this part of the trade-off for the sweeter sound I get from the
5500.

* Does the 5500 have noticably/measureable less headroom into 6-8 ohms.

* Is the 5500 more sensitive to line voltage than the 555 II.

* Is the output impedance different enough to account this.

* Is the 5802 the better choice for getting the sweetness and the slam.

Here are some of my recent postings to rec.audio.high-end that describe
my system and what I am hearing.

Thanks,

Rein Narma
714-461-0424

______________________________________________________________________

Subject:
Adcom GFA-555 II vs 5500 or MosFET vs BiPolar experience and
opinion
Date:
24 Apr 1998 21:42:39 GMT
From:
"Rein E. Narma" <ren...@pacbell.net>
Organization:
Pacific Bell Internet Services
Newsgroups:
rec.audio.high-end

Interested in comparing Adcom GFA-555 II vs 5500 or MosFET vs
BiPolar experience and opinion. I am curious whether my experience
relates to all/many mosfet designs (Hafler 9505 etc).

System is: Pioneer Elite PD-59 >AA DDE 3.0 > DTI Pro 32 > AA DLC >
Adcom 555 II or 5500 > VMPS Super Tower III.

cables dyi: interconnects (mic cable) and 4ft CAT 05 UTP and/or 8ga
line cord. Connection config is either biamped pairs (555 or 5500)
of amps or a single 555 or 5500 amp driving both speakers.

I am hearing much larger differences (pro and con) between my 555 II
and new 5500 than I expected regardless of mono or biamping. The
amps have nearly identical power ratings and very similar specs.

The fundamental technology differences are: bipolar vs MOSFET
(Hexfet) output devices and much higher (50 wpc) class A bias on the
5500.

The 5500 is smoother and exhibits less grain especially on female
vocals. It also seems to provide less dynamic contrast, "slam" and
speaker control particularly at higher volume levels. Loud passages
and transients seem to have less power and impact when compared to
the 555 II in each config. (I am going to do some measurements of
instantaneous current and voltage levels on peaks.) The sound seems
almost compressed, louder but with less and less dynamic range as if
it were "soft clipping". All this is happening at 200w peak levels
but without exciting the "distortion alert" leds.

Having said the above I still prefer the 5500 smoothness over the 555
brute force for most of my listening. I would be curious about
pairing 555 IIs with 5500s in a biamp config.

Any of this sound familiar?

Thanks,

Rein

-------------------------------------------------------------------

Thanks Greg and to Randall who also responded.

I have done some more critical listening since my post, based on which I
will now have to turn in my "Amps-Sound-Alike" membership card. The
more I listen the more differences I hear. My 70's vintage EE didn't
cover power MOSFETs and my only experience has been in small signal RF
work. Also, I don’t recall my old Hafler DH-500 sounding that hugely
different from my Ampzilla and Dynaco 400. The limiting factor then may
have been the lower efficiency and dynamic range of the Dahlquist
DQ-10’s vs the VMPS. My next steps will be to do some single blind
listening using two different outputs from my AA DLC after carefully
level matching the outputs with a test CD steady state signal and
O'scope.

Back to the listening. I expected the slight differences in grain and
detail with 555 II being more etched and the 5500 softening vocal
sibilance and cymbal brush strokes. I even expected some softening of
the bass. When listening again I heard what sounds like a looser "grip"
and reduction in dynamics and contrast even at moderate levels. On
some recordings this adds ease and smoothness. On others, it adds a
layer of thick gauze over rim shots and guitar licks. This was not
subtle! Less air is being moved.

I already know that "distortion alert" LEDs don't react (or more
correctly have little or no peak hold circuitry ) to very short term
clipping on 555 and probably don't on 5500. If my listening tests
continue to show the differences I will see if I can measure major
differences in current capability, slew rate and short term dynamic
power. I also need to see if I am having some line voltage problems
with the pair of 5500s as they draw a few hundred watts more on average
perhaps resulting in some power supply sag and early onset of clipping.

The Adcom folks have been very helpful and generally agree that the two
amps sound very different. Nothing in the specs hint at the obvious
dynamic contrast and "power" differences I am experiencing.

More to follow when my measurements are complete.

· Line voltage behavior
· Internal power supply behavior
· Peak short term Power / Current
· Square wave response

Thanks again.

Rein

[mn...@pobox.com]

In article <354653...@pacbell.net>#1/2,
ren...@pacbell.net wrote:

> My next steps will be to do some single blind
> listening using two different outputs from my AA DLC after carefully
> level matching the outputs with a test CD steady state signal and
> O'scope.

A simple voltmeter will be more accurate than the O-scope but I'd be
very interested in the results of your blind testing. I believe there is a
difference with Mosfet amps. I haven't critically listened to the 5500
but am quite familiar with the 555 and the early Hafler amps and other
MOSFET designs.

> I already know that "distortion alert" LEDs don't react (or more
> correctly have little or no peak hold circuitry ) to very short term
> clipping on 555 and probably don't on 5500.

The clip indicator on the 555 will indeed respond but the human
eye may not catch it. The circuit simply measures voltage inside
the feedback loop and if it exceeds a threshold, the amp is
likely distorting excessively with the NFB generating a large
error signal and the LED is lit.

> If my listening tests
> continue to show the differences I will see if I can measure major
> differences in current capability, slew rate and short term dynamic
> power. I also need to see if I am having some line voltage problems
> with the pair of 5500s as they draw a few hundred watts more on average
> perhaps resulting in some power supply sag and early onset of clipping.

You will likley find some differences but you'd probably be jumping to
conclusions to try and correlate them with the sound. Output impedance, for
example, has been reasonably proven to not have much effect as long as it's
below 0.1 ohms or so. There have been some rather strong statements in this
N.G. regarding how an amp can't really damp a speaker much at all.

The higher idling current of the 5500 shouldn't really change things much
unless you have the amps on a really bad circuit. The peak line current
demands are far in excess of the idling current. At worse, you might be
losing a tiny fraction of a db of headroom.

> More to follow when my measurements are complete.

Please keep us informed (particularly with blind testing).

-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/ Now offering spam-free web-based newsreading

[Brock Hannibal]

mn...@pobox.com wrote:

> A simple voltmeter will be more accurate than the O-scope b

What makes you believe this? Modern digitizing oscilloscopes are
actually VERY good AC voltmeters and derive their measurments from
acquired data. Now if you were to claim that a very good AC voltmeter,
like a 4 1/2 digit true RMS meter is more accurate I would say yes, BUT
not for waveforms that differ substantially from sinusoidal such as
triangle, ramp or square. Remember the "crest factor" and bandwidth
specifications of even your better AC volmeters are not all that great,
whearas a scope can acquire quite accurately almost any waveform whose
components are well within it's bandwidth. Oh, yeah, MODERN
oscilloscopes often have analog bandwidths that exceed 1GHz and sample
rates of 2 to 4 Gs/second so averaging can be invoked to achieve very
high resolution on very small signals in the audio spectrum.

--
Brock Hannibal
Hardware Design Engineer
Tektronix, Inc.
Beaverton, Oregon

"Any ideas or opinions expressed here do not necessarily
reflect the ideas or opinions of my employer."

[Rein E. Narma]

I probably should have specified my test gear in the original post. I am
comfortable with my choice of test equipment and how to use it. I agree
with Brock in that my Tek 455 (old though it may be) is a fine piece of
test equipment and will be most appropriate for matching peak to peak
levels across the audio spectrum.

O'scope Tektronix 455 (50mhz)
Allison O-Scope II PC interface/Digital 22mhz Storage Oscope
with FFT audio spectrum analyzer

DVM Testmate 365 3 1/2 digit 200 mv

Signal
Sources Stereophile Test CD 1,2,3 Merit FP-150 function generator
various other test CD's with tone burst and impulse content.

The program which drives the Allison interface is able to calculate and
hold peak and RMS values. Overall I am less interest in absolute values
than differences between the amps to help explain what I am hearing.

I will post the complete test process once complete. In the meantime I
am suffering from a head cold that prevents me from moving forward with
the single blind listen portion.

[mn...@pobox.com]

In article <354A13...@mdhost.cse.tek.com>#1/1,

Brock Hannibal <bro...@mdhost.cse.tek.com> wrote:
>
> mn...@pobox.com wrote:
>
> > A simple voltmeter will be more accurate than the O-scope b
>
> What makes you believe this? Modern digitizing oscilloscopes are
> actually VERY good AC voltmeters and derive their measurments from
> acquired data. Now if you were to claim that a very good AC voltmeter,
> like a 4 1/2 digit true RMS meter is more accurate I would say yes, BUT
> not for waveforms that differ substantially from sinusoidal such as
> triangle, ramp or square. Remember the "crest factor" and bandwidth
> specifications of even your better AC volmeters are not all that great,
> whearas a scope can acquire quite accurately almost any waveform whose
> components are well within it's bandwidth. Oh, yeah, MODERN
> oscilloscopes often have analog bandwidths that exceed 1GHz and sample
> rates of 2 to 4 Gs/second so averaging can be invoked to achieve very
> high resolution on very small signals in the audio spectrum.

Well in this case we're not worried about absolute accuracy at all--only
*relative* accuracy. The context of the post was level matching using a sine
wave source. Resolution and repeatability are the primary concerns and, in
general, a voltmeter will do better at both compared to a scope.

Further, I doubt the person making the post has a high-end scope. In fact,
it's unlikely their scope even has accurate cursor measurement capability
when compared with a garden variety 3.5 digit DMM. Lots of things can affect
the repeatability of the scope measurement while the voltmeter is hard to
screw up.

Given the tools the average person has at home (versus someone who works for
an instrumentation company) I was just suggesting what is very likely the
more accurate (and certainly easier) tool for the task at hand.

What is it with people taking things out of context here?

Mike

[j...@southwind.net]

. Oh, yeah, MODERN
> > oscilloscopes often have analog bandwidths that exceed 1GHz and sample
> > rates of 2 to 4 Gs/second so averaging can be invoked to achieve very
> > high resolution on very small signals in the audio spectrum.

Tek's "modern" digital scopes will often not do what their classic
analog ones did...and besides the purchase price is nothing short of
crazy,and the user interface sucks.When my 2247 goes out,I have no idea
what I will replace it,except nothing Tek now makes.