|
Dennis Had of Cary talks with Jonathan Scull about the circuit topology of the SLM-200
A REPRINT FROM STEREOPHILE, May 1997, Vol. 20 No. 5
by Jonathan Scull
Jonathan Scull: What mode does the SLM-200's output stage run in, Dennis?
Dennis Had: In class-AB1, close to class-A. In fact, most listening will be done in class-A. We output-stage KT88s are biased with a negative fixed-bias voltage for low distortion and maximum output capabilities.
Scull: And the SLM-200s are normally supplied in triode mode?
Had: Right. As shipped from the factory, all SLM-200s are wired for triode with the screen grid of the KT88 connected to the plate through a 100 ohm resistor. Triode puts the KT88 in low plate-impedance mode--670 ohms. We've found that triode operation is the most linear and yields the most musically satisfying presentation.
Scull: The amps put out +1OOW in triode?
Had: Right. About 130Wpc before hard clipping sets in. Our dealers can switch the SLM-200s into 200W ultralinear mode if you think 100W is not adequate for your needs. The output transformer does comes equipped with ultralinear screen taps for 200W operation.
Scull: Let's talk about the transformer...
Had: Sure. The output transformer is the most important component in the amplifier. Those in the SLM-200 have been specifically designed by us for use in this amp. We've taken an approach not too dissimilar from that of the single-ended amplifiers we design and produce.
Scull: How so?
Had: The primary and secondary windings on portions of the output transformer are wound in a bifilar process with the two inductors interleaving 16 times. The bifilar wind--two conductors wound at the same time - yields the closest balance and coupling of any design currently used in vacuum tube output transformers. The EI laminates are silicone-impregnated hypersil steel, which contributes to the extremely low loss of the output transformer.
Scull: So it's similar to your single-ended transformers, but not quite the same?
Had: That's right--no air-gap.
Scull: How much overall negative feedback do you use?
Had: 3dB of feedback, and that's to improve speaker damping. The SLM-200's original engineering design was implemented with zero feedback, so you can be assured that the feedback circuit is not used to correct distortion and transient response.
Scull: Whew--you had me worried! Let's get to the circuit details...
Had: Sure. The input pre-drive stage is the first 6SN7 dual triode on the SLM200--the same as we use in our single-ended amplifiers. It's a single-ended, class-A, anode current-source gain stage. The input signal from the preamp or line source is direct-coupled to the grid of one section of the 6SN7 input. The anode of this section is direct-coupled to the cathode grid of the second stage. This takes the place of the conventional dropping coupling-resistor network found in conventional gain stages. It's called the anode current section and offers infinite resistance and the proper current and voltage to operate the gain stage.
A balanced drive signal is applied to the control grids of the KT88 output tubes from the 6SN7 driver tube. The driver is configured in a dual-differential network that operates in the following manner: The first section of the dual 6SN7 tube is direct-coupled to the input pre-driver stage. The drive signal is amplified through this first section in a class-A grid-driven circuit with the output signal inverted 180' out of phase at the anode, then coupled to one half of the push-pull bank of KT88s.
At the same time, the second half of the dual-triode 6SN7 is cathode-driven in a grounded-grid, noninverting, class-A gain stage coupled to the other half of the push-pull bank of KT88s. With this network the balance is a perfect plus-and-minus dual-drive signal that drives the final output KT88s in a balanced push-pull configuration.
Scull: How about the power supply?
Had: The power supply in the SLM-200 actually consists of five different supplies. The power transformer is designed to operate at a 150% continuous commercial service at the full rated 200W output level. As the SLM-200s are delivered from the factory in 100W triode mode, the transformer is at 300% duty service.
The high-voltage power-supply section is a full-wave center-tap configuration - not some cheap voltage doubler, as is used in many competitors' amps - using a Pi network with a filter choke. This high-voltage section feeds the final output KT88s.
Another full-wave center-tap supply feeds a Pi network filter-choke medium voltage for the input and drive 6SN7 tubes. And yet another supply provides the negative DC grid-bias circuit, once again a Pi network. There's an additional DC supply with two voltage regulators to supply DC to the filaments of the two 6SN7 input driver tubes. These regulated filament supplies prevent AC ripple from capacitively coupling to the electrodes in the 6SN7 gain stages.
Scull: I understand that fast overload recovery is a major priority in your design.
Had: That's right. A great deal of attention was paid to that. The ability of an amplifier to instantly recover from clipping is much more important than is commonly believed. In the arnplifier power wars, the mentality has been focused on higher and ever higher power output to solve the clipping problem. In reality, the most critical aspect of this is how fast a recovery an amplifier can achieve after overload. With the incredibly wide dynamic range of music, even 2000W of power is not enough.
Scull: Yes; but really... it's the first few watts that count, isn't it?
Had: Exactly... most music heard in an average home listening room requires only about 3W of power. It's on the transients of loud, low-frequency program material that tremendous signal voltages appear at the input of the amplifier. It's here that overload recovery ability is of critical concern.
The SLM-200 has the ability to handle transients and instantaneously recover from brief and even extended overloads. And it will overload symmetrically at any frequency in the audio bandpass. It will also yield faithful reproduction of extremely low frequencies at full output levels. Power transformer, power-supply regulation, output-transformer design, and careful shaping of the overall frequency-response curve all play very important parts in the ability of the SLM-200 to recover quickly when overloaded.
Interestingly, if one were to monitor the high-voltage rail at 525VDC on our SLM-200 during soft and then loud music passages, it would be found that there is no more than a volt or so change.
Scull: The SLM-2OOs seem very robust. l couldn't believe it when yow asked me to pull all the output tubes and turn the amps on to check if the arcing KT88s from that first batch had anything to do with the amps themselves.
Had: That's right. The amplifiers can also be operated with no load - without the speakers - with no damage to the amp, output transformer, or tubes.
Scull: Thanks for the rundown, Dennis. You oughta put this in writing and add it to the manual!
Had: I'll do that!
|
|
