[contact Aren van Waarde about his project]

Background

The Bride of Zen, designed by Nelson Pass and described in The Audio Amateur, 1994, is a very simple line-level preamplifier which contains just one medium-power MOSFET per channel. It is meant to be used together with the Zen (or Son of Zen) single-ended MOSFET power amp. A complete description of the BOZ can be found at http://www.passlabs.com.Mr.Pass has given a very lucid account of the design philosophy, construction and alignment (the only alignment consists of adjusting the quiescent current which should be 40 mA per channel at a supply voltage of +60V). The lay-out for a printed circuit board can be downloaded from Passlabs.

Construction

Since the PC board is single-sided with wide tracks and easy to construct I decided to produce it myself. I made a screen dump of the lay-out (scale 1:1) on a HP Deskjet printer. Using a thick sewing needle, I perforated the screen dump at the positions of the component wires. Subsequently, I fixed the screen dump with sello-tape on the copper side of epoxy material for PC boards. (The copper should be cleaned of oxide with a household abrasive and of fat with an organic solvent). I marked all component holes on the PC board using a felt-tipped pen (Ed ding 400). After the ink had dried, I removed the screen dump and I drew all tracks by hand, using the same pen. Then I etched the PC board with ferric chloride. After rinsing and drying I drilled the holes (1, 1.5 and 3 mm size) using an electric machine in a vertical standard.

Initially I used both newly-bought components and parts from my junk box. This was not a good idea as you will read below, but hey, I am Dutch and I've been to Scotland! Thanks to the very clear instructions of Mr.. Pass, the prototype worked immediately. P102 has to be adjusted in such a way that the voltage drop over R108 (in the absence of an input signal) is 4 Volts. That's all!

The BOZ seen from the back. At the front panel: power indicator LED, right and left channel volume
pots, input selector. The RCA jacks at the back have now been replaced by better quality types.

Listening tests and modifications

Since my Zen power amp has not yet been finished, I evaluated the BOZ preamp in my current audio system. This consists of: Rotel RP850 turntable on a heavy concrete slab with Linn Trak MC cartridge, Van Willenswaard AT802 moving-coil preamplifier, Marantz CD-80 CD-player, Nakamichi BX2 cassette deck, Sony ST-S370 FM tuner, Audio Synthesis PAS-01 passive preamplifier, Van Medevoort PA222 power amplifier, Rogers Domestic Monitor transmission line speakers, home-made interlinks and Van den Hul Clearwater speaker cable. The PAS-01 was taken out of the system and replaced by the BOZ.

First I listened to some tapes on the Nakamichi. These sounded good. Driving the power amp via a passive preamp has always been somewhat of a problem for the Nak; the gain provided by the BOZ made the tapes sound better, more authoritative. Then I listened to a few vinyl records. The sound was rather "forward" in the bad sense of the term; there seemed to be a tonal imbalance, the midrange was emphasized with respect to the treble and the bass. One got the impression that the midrange drivers in my speakers were too loud! Moreover, individual voices and instruments were not well-focused, and the treble was rather sizzling and metallic in character. Of course I was very disappointed.

After a day of mind searching I guessed that the coupling capacitors in the prototype might be the culprit. I had used East German MKT from my junk box (the big yellow caps that you will see in the photographs). The only reason why I had used them was that they were immediately available and they fitted on the PC board. Moreover, they were within 5% of the specified value as measured with a capacitance meter. But of course that did not mean that they were of audio grade quality...

I decided to replace the coupling caps with polypropylene types. For the input capacitor, I used a type from Intertechnik (Audyn-Cap, 400V, 5%). For the output capacitor, I bought a type from Monacor (MKP, 250V, 5%). The main reason for using the Monacor was that it fitted the PC board; Solen or Intertechnik capacitors would have been too large. The Monacor caps were marked "Made in Germany" and they looked suspiciously like those made by Siemens.

After having replaced the caps, I resumed my listening tests. The difference was enormous! Of course I had hoped for an improvement, but I had not expected that it would be so large. The soundstage was now both wide and deep. Individual voices and instruments were well focused. After a while I switched from vinyl records to CD.

Once more I was disappointed. In loud passages there was a nasty distortion which sounded like clipping of the gain stage. Lowering the volume of the sound did not cure this problem. If you study the schematic of the BOZ, you will notice that the volume control is AF TER the gain stage. Therefore, adjusting the volume control does not help to avoid clipping.

Since overload occurred mainly on CD and much less on phono or tape, I guessed that a relatively high output of the CD player might be causing trouble (unfortunately I don't have an AC millivoltmeter so I could not actually check this). I tried to solve the problem by mounting a 470 k resistor in series with the CD input. A voltage divider is formed by this resistor and R107, resulting in a greater than 5-fold attenuation of the input signal to the BOZ. Later on, I also mounted 220k resistors in series with the phono and tuner inputs. Although overload was less noticeable with these signal sources, these resistors resulted in a sonic improvement. Nelson Pass admitted via E-mail that overload could occur in some cases and that the inclusion of series resistors was a good idea.

After this modification, my listening tests were satisfying. Overload did not occur any more. It came as a great surprise that the BOZ is so revealing. In some CDs which I thought I knew quite well I heard new details (novel melody lines in complex music for church organ; subtle variations in touch in a piano sonata; individual instruments in the back of the orchestra). I thought the BOZ should have been worse than the passive preamplifier, but in practice this was not the case. Apparently the extra gain and/or extra buffering produced by a single gain stage may be advantageous (I know that this sounds like heresy, but what can I say?).

A possible explanation for the better sound could be that the passive preamp with its 10k volume pot was a rather heavy load on the output stages of the signal sources. It is also possible that the output impedance of the PAS-01 was too high for the power amplifier. Moreover, internal wiring of the PAS-01 is unshielded and the left and right channel wires are not very far from each other. In contrast, the BOZ has shielded wiring and it is virtually a dual- mono construction; it even uses separate volume pots for each channel. Yet I am surprised, especially since the Audio Synthesis preamp has been very favorably reviewed in Britain (Hifi News & Record Review).

The BOZ seen from the front. The power transformer is mounted at some distance from
the PC board to avoid hum and noise. A mains filter was later included.

On to perfection

After these modifications the BOZ was very good; it seemed better than any other preamp I have ever possessed or made. However, it was not perfect. Its most glaring fault was a large transient during switch-on and switch-off. This could of course be cured with a mute circuit but this would bring a relay contact or an electronic switch within the signal path. If you turned down the volume before switch-on and switch-off, the transient was virtually inaudible. But of course somebody would forget this some time. With my low-efficiency Rogers speakers, the transient was not much of a problem, but with high-efficiency speakers such as Lowthers it would be a threat to the voice coils!

Although this is not drawn in the schematic, Nelson Pass suggests to include a mains filter in the power supply to suppress interference. In the initial prototype I did not use such a filter (I live in a rural area with a decent mains voltage so I thought it wasn't very necessary). However, the sound of the BOZ was so good that I wanted to find out if it could still be further improved. Also, I may take the preamp to the city for demonstrations to an audio club, and in urban areas the mains is usually much more polluted. So I bought a nice dual mains filter (see below) and I soldered it between the mains switch and the power transformer.

It turned out that the mains filter was a substantial improvement. The transient during switch-on and switch-off was now completely suppressed, even if the volume pot was kept at a level of three-o-clock. So my BOZ is now also acceptable to Lowther or Fostex owners. It seemed that the filter also improved imaging precision somewhat, individual instruments are now even more pinpointed on the stage.

Finally, I replaced the cheap RCA connectors on the back of the amplifier with high-quality insulated gold-plated connectors which ensure a better connection. Also, I glued three wooden bars between the front and back of the metal cabinet for cabinet stiffening and the suppression of torsion (I used a large and rather cheap steel cabinet which wasn't very stiff). Since then I have been happy listening all the time (I didn't know that my speakers could image so well!).

Printed circuit board of the Bride of Zen preamp. The yellow
coupling capacitors were later replaced by better MKP types.

ERRATUM
Mr. Colin Rutter, of Spectral Technologies PTY Ltd., Australia, wrote: "I have to inform you that the bad sound you experienced with the BOZ initially is due to the input signal being clipped asymmetrically as setting the bias current to 40 mA by measuring the voltage drop across the resistor may in fact not set the "zero point" correctly. A CRO (cathode ray oscilloscope) solves this problem or a bit of math. My Bride of Zen has its idle current set at 32 mA for symmetrical clipping. Using your method to attenuate the signal will only cause the top end to roll off from about 15000 Hz. Any large resistance in the input will form an RC filter with the capacitance of the MOSFET".

Aren has checked Colin's suggestions and he found that they were all correct. So Aren has lowered the bias of his BOZ (to about 29 mA, using an oscilloscope) and he removed all series resistors from the audio inputs. The sound of the BOZ is now even better than it was. May it play happily for many years to come.

Parts list
(These are the parts I used and where I found them)

R102, 202 22ohm metal oxide 1% 0.6W
R2,3 47ohm metal oxide 1% 0.6W
R106,206,108,208 100ohm metal oxide 1% 0.6W
R1 4K7 metal oxide 1% 0.6W
R103,203 10k metal oxide 1% 0.6W
R101,201 33k metal oxide 1% 0.6W
R105,205,107,207 100k metal oxide 1% 0.6W
R104,204 1k wire wound 5% 5W Vitrohm

Additional resistors:
Rcd,Rcd' 470k metal oxide 1% 0.6W
Rmc,Rmc',Rtun,Rtun' 220k metal oxide 1% 0.6W

C104,204 1uF Audyn-Cap (Intertechnik) MKP 5% 400V (available from Speaker & Co., Speakerland etc.)
C103,203 10uF Monacor MKP 5% 250V (available from Display Electronica & other DIY electronics shops)
C2,102,202 100uF/63V radial electrolytic, Roederstein
C1,3,4,101,102 1000uF/100V radial electrolytic, Sprague (difficult to obtain, I bought mine from Barend Hendriksen HF Elektronika)

P101,201 5k log mono volume pots (Omeg )*
P102,202 10k multi-turn cermet-type (Bourns) (for adjustment of quiescent current)

Q1 TIP29C Texas Instruments
Q101,201 IRF610 International
Rectifier D1...4 1N4004
Z1...6 10V zener Philips (BZX55 type)
D101,201 20V zener Philips (BZX55 type)

S1 2 x 4 pos Input selector switch (I used a Lorlin with 4 positions and 3 poles)

Transformer 220:2 x 30V 14 VA, both secondary windings in series Flat type for PC board mounting (Marschner 012242)

Mains filter Schaffner FN2060-1-06, also marked 9540A (1 A type, dual)

* Unfortunately I could not find a dual concentric 5 k log stereo pot which would allow adjustment of both volume and channel balance; therefore I used a single mono potentiometer for each channel. Although the Omeg pots are cheap, they function very well. In the future I may replace them by stepped attenuators.