🚀 go-pugleaf

  I own a Carver M400 (the infamous 'cube') and some time ago replaced
it with a tube amp, and more recently with a heavily tweaked Linn LK100.
More out of curiosity than tweaking, I made up a set of schematics for
the M400, to find out what all the hoopla was about. 
  One thing I can now be sure about, there was some very elegant design
work that went into these units, considering their cost, power output,
and weight. High-end they're not. The unit I have can deliver within
the 3db, full power from several Hz to about 40kHz. At very high
power levels, (overload and frequency extremes) there is power supply
hash that modulates the signal.
  Distortion is not bad (.1-.4 %),noise is very low, it overshoots
with an inductive load (as in a speaker) under a square wave.
Percussion and violin music don't quite sound right through this amp,
at least in my opinion. I honestly didn't care for the sound it made
once I had heard other units, but that's a matter of taste.
  The amp is a real pig to service. Only one channel is accessible (from
the component side), non of the boards have any component identifications
or any helpful silk-screening. In order to access anything else, you
must strip it apart. Faults in the other channel are impossible to 
trace with the amp running. It took me about one and a half hours to
remove the three circuit boards (2 amplifier outputs and power supply).
The boards are held in by the power transistor mounts, so you have
to remove all 6 power transistors and 3 smaller heat-sunk transistors to
access the amplifier board. Oh, and they are all piggy-back, so if you
have to get at the lowest (power supply), they all must come out.
You're guaranteed to curse Mr. Carver, despite any previous respect you had  
for him. You'll also get your hands and clothes coated with heat-sink
goop (which is poisonous as hell).
  This unit does not have tap changing transformers. It does however,
use a TRIAC in series with the transformer primary. This requires the 
use of a well designed transformer (low leakage inductance). The
TRIAC is controlled by feedback from the power supply voltage level,
and also by 'feedforward' using the output level going to the speakers.
This allows it to 'gun' the supply before it has already pooped out.
In effect, the power supply is regulated by the loop from the voltage  
sensing, through the TRIAC, power transformer, and rectifier circuit.
That maybe the 'theory' but in practice it didn't do that well. Under
no load, the power supply would deliver +- 75 volts DC. At full load
it would drop to +-60to65 volts DC. I ran a burst of 1 kHz (about
10 cycles) every 1/2 sec., at full output, and watched the recovery
of the power supply with a 'scope. All in all, it was fairly well behaved,
the output supplies would drop about 3-4 volts and recover in about 40-50
milliseconds. I had my doubts about the triac control. Watching the 
the switched input to the transformer primary, I saw that when the amp
was idling, the controller was feeding AC pulses of only one polarity to
the transformer. That would explain the rough 'humming' from the 
power transformer, and why these transformers sometimes 'flame out'.
There is an adjustment to set the operating point of the triac controller
and I tweaked it so that BOTH sides of the AC input were switched. The
transformer quieted down, but now the power supply was running +- 85 volts.
IMHO, they should have used an SCR around a bridge rectifier instead of
a triac, to get a symmetrical control of the 110V. Instead of the cheapy
'light dimmer' circuit (with phototransistor), Carver should have used
something a little more sophisticated, to get better control at low
phase angles. It basically had no control at the first 1/4 of the 60Hz
sine wave (and that's where you need it, 'cause that's when the power
supply caps charge up), and only pulled in towards the remaining 1/2.
At that point the loop gains start to drop because of the shape of the
waveform, and you get poor regulation, along with all the nasties of
a non-linear feedback loop. Maybe they should have just put in a heftier
power transformer, ditched the 'Magnetic Amplifier' (aka 'light dimmer')
lost a bit of profit, and saved a bit of their reputation.
  I suspect you can fry this unit by transients on the power line,
so it's probably worth putting MOV transient protectors on the 110v
circuit that feeds it. The unit will generate a fair bit of 'hash'
on the line, so it's also a good idea to put a line filter on the 110v
as well. On overloads, this amp can easily draw 5-10 amps peak, so
any wiring should be done properly. I would also replace the capacitor
that's stuck across the 110 V input with a U/L or CSA approved one,
so that there is no fire hazard. A Siemens # B 81121 polypropylene
type is recommended.
  The audio part of this circuit is quite complex. Considerable attention
has gone into load and power transistor protection. Judging from the
circuit, (3 series transistors between speaker and each rail) good
high power transistors were not available or cheap enough at that time.
The power transistors all use in-house numbers on them, so who knows
what they really are. In the front end of the amplifier is an op-amp,
a TL081CP. This is a jfet input opamp, with low distortion, full power
output to 100kHz, and 13 V/uS (unit gain). There are about 21 transistors
on each channel. Other than 4 small electrolytics on each channel, all
the capacitors are Mylar types (those scruffy looking green ones, with
a shape like a small bag of potato chips). For all the elaborate
protection schemes, they really bombed out on protecting the transistors
themselves. The voltage ratings on the transistors do not come up
to what is required for an amp running from a +- 70to80 volt supply.
I can't speak for the power output transistors, they all have funny
part numbers. The driver stages are rated at 70v (mje2955), 25v!!!(2N3904),
40v!(TIS93).
  The heatsinks are the case itself, but it's a very heavy gauge aluminum,
and takes quite a bit to heat up. If you run your amps wide open all the
time, or have compressed music with a limited dynamic range, these amps
will run too hot. The guy I bought it from used it for DJ work, they
ran a little muffin fan next to it, in order to keep it cool.
  My complaints about the amp were it's poor ability to clearly 'place'
instruments in the sound field, a washed out high frequency end, and
a faint buzzing sound that eminated from the power transformer. 
  It's good points were it's cheap!, light, 200-300 watts per channel,
mechanically rugged.
  And now for the best part..... While I was stress testing this
amp, with repetitive bursts of 400w of output (1 kHz, 5% duty cycle) it
ran fine for about 15 minutes. The case was warm (not at all hot), the
200W power resistor was too hot to touch, and then.....BBBRRRAAAAPPP!!
A big cloud of smoke shot out the back, and that channel went dead!
#$%^&*(@! I figured the transformer blew, so I stripped the entire
unit down (1 hour). The power supply was fine! Turned out that one
of the driver transistors failed (not surprising, knowing their voltage
ratings) and blew out all the following drivers, and then all 6 power
output transistors. Sheeit! And I was hoping to sell this thing to
some unsuspecting sucker! Probably a good thing it went now, or maybe
in a week or two I might have had no kneecaps.... After staring at the
not so little mound of fried parts for 15 minutes, I scooped the whole
mess, amp included, into the garbage can. That's what you get for
$300 and about 50 hours of tracing out a set of schematics from the 
boards. Live and learn..... If you want good stuff, buy only the good stuff.
  I'm told a lot of them failed with spectacular smoke or flame displays.
If you have any good stories about Carver amps, I'd be interested,
it seems everybody either hates or respects Mr. Carver. I've read a
few of his advertising blurbs, and they come across pretty poorly to
me, but then he also wrote up some stuff for the Oregon Triode Society
that was very sensible, and cut through a lot of the mystique and 
high-end bullshit. (those can be found in the earlier archives of
the ftp site for rec.audio.high-end....csd4.csd.uwm.edu)
  Oh, if anyone wants, I can give them a copy of the schematics of this
Carver M400 (accuracy not guaranteed!), 'cause they are no use to me
now. I can give you a lot of techno-babble about this amp if you want,
but I don't think this newsgroup is the place to do that.

-Paul

-- 
-----------------------------------------------------------------------------
Paul J Guy			work phone:519-885-1211 ext 6371  
paul@gaitlab1.uwaterloo.ca	home:519-744-5016   FAX:519-576-3090
pguy@healthy.uwaterloo.ca	..remember...bullshit baffles brains...