5196106130

5®

r t

The rear of the PM-350. Both tip-ring-sleeve and XLR input connectors are prov!ded, for both channels. Recessed switches are used for bridged-mode mono and clipping elimination.

has three sets of output voltage taps, and each set are taken to separate rcc-tificr/filter circuits delivering three pairs of supply voltagćs — nominally +/-25V, +/-50V and +/-80V respcctively. A spe-cial “commutating Circuit*! switches thesc supply voltagcs in or out, again in response to the incomjng audio signal.

Ah, you may say, so that’s the secret of the Carver — a rcally fancy audio-modulatcd power supply, driving what is prcsumably a fairly normal amplifier. No, not quite!

In fact this “fancy audio-modulated power supply” doesn*t just drive the power amplifier, it actually is the power amplifier. Or morę accuratcly, it*s the bccfy part of it.

Essentially, the combined effect of the audio nioduiated magnctic field and commutated supply rails is to produce a high power step-approximation version (i.e., a bit rough) of the input audio. What thcn happens is that a smali and highly linear output stage with lots of negative fcedback applied is used to “smooth up” this rough approximation, to produce the finał output signal. This amplifier docsn*t need to handle high power, because it^łs really only correct-ing the errors — and handling the higher frequcncies (where there isn*t much power).

So in a Carver amplifier the power supply and power amplifier are rcally integratcd into the one functional enti-ty. In some ways this secms to be the really innovativc aspect of Bob Carver*s contribution to the art.

What about the two channels of a stereo amp? With the power supply so much an integral part of the amplifier(s), you’d pcrhaps expcct two totally separate supplies for stereo. But as far as we*vc bcen able to determine, thafs not so.

Prcsumably Carver’s thinking bchind this is that with the vast bulk of stereo program materiał, most of the energy is essentially mono and common to both channels. It also tends to be concen-trated in the lower part of the spectrum. So a common “rough approximation”

Circuit would not be inappropriate, pro-vided that the separate smoothing and error correction amplificrs in each chan-nel are designed to cope with normal cohtent diffcrences between the two stereo signals.

Both the PM-175 and PM-350 ampli-fiers oomc in compact 19“ rack mount-ing cases, only 90mm (3.5”) high and 350mm deep overall. The PM-175 has a mass of only 8.6kg and the PM-350 only 9 5kg. Both are tastefully styled, with a dark grey matt finish.

The only Controls on the front of both models are a pair of volume Controls and a power switch. Apart from this there are four indicator LEDs showing respectively power, signal present, whether or not either channel has ent-ered clipping, and the presence of a fault condition. Both amplifiers are able to detcci the presence of either a short-circuited load, or a sustained overload condition, and will turn themsclves off to prevent damage to either the amplifier itselfor the loudspeakers.

At the rear of both models there are the loudspeaker terminals, two kinds of input connectors for each channel (both 6 5mm tip-ring-sleeve jacks and XLR-typc sockets, with activc balancing), and a mono bridging switch acccssible via a smali hole in the back panel. The last-named of these allows the phase of one channel to be reversed, so that the two can be used as a bridged-output mono amplifier.

The PM-350 also has a second switch accessiblc via a hole in the back panel, to enablc or disablc a hard-clipping eliminator Circuit. Whcn enablcd, this limits amplifier clipping to 3% THD, evcn w hen it is overdriven.

Both models feature slow start-up, and input muting during the switch-on period to prevent “thumps”.

The interior of both models is almost identical, with a large PCB in the centre biscctcd by a length of extrudcd heat-sink. This is used to dissipate power from the 2()-odd power transistors used for commutation and the output of each channePs error correction amplifier. In the PM-350 there is a smali DC motor fan used to puli air in through the side of the case and blow it over the heat-sink, but in the PM-175 this iś appar-ently not needed.

At the front of the case in each unit is the power transformer, with that in the PM-350 a littlc larger than the other,- but both surprisingly smali for the power levels involvcd. Alongside the transformer on one side are the main reservoir clectros, and on the other side a smali PCB with the reso-nating capacitors and other ciicuitry needed on the 240V primary side of the “magnetic field”. The triac is mounted on the case just near this smallcr PCB.

Weil then, how did the Carvcrs check out in practice? On the power output side, they ceriainly delivered an impres-sive wallop. The PM-175 delivcred its rated 175W per channel into 8 ohms with both channels drivcn, with only .05% THD at 1kHz and below — well below the quoted 0 5% maximum. And it produccd very near its rated 250W per channel into 4 ohm loads, again for 0.5% THD at 1kHz and below.

Similarly the PM-350 easily produced 390W per channel into 8 ohm loads, again with both channels driven, for only 0.15% THD at 1kHz and below. It even produced 500W per channel into 4 ohm loads, with both channels driven, or 1000W into a single 8 ohm load in bridged mono modę — for only 0.3% THD at IkHz and below. Very impres-sive!

What we did notice, however, was that the THD began to rise noticeably above about 3kHz, for continuous test signals próducing over 20W output per channel for the PM-175, and about 40W per channel for the PM-350. Looking at the output with a scope, you could sce “kinks” appear in the output at these power levels — but only on signals abovc 3kHz or so. For 20W continuous output into 8 ohm loads. for cxamplc, the PM-175 reached the rated 0.5% THD level at about 8.5kHz. We suspect that this is some pcculiarity of the commutation circuitry.

ELECTRONICS Australia. September 1987

I