A low-profile affair with a bridge-tied load.
Page 1 | Page 2 | Page 3 | Page 4
The rear panel was, for this layout, nearly as complicated as the main chassis assembly. First, a look at the fan and the AC inlet. The fan includes a small daughterboard to smooth and filter the DC noise the motor can generate, along with a diode to protect against an accidental reverse conenction of the incoming supply cable from the fan control relay. If either heatsink reaches 65C, the fan will engage and attempt to ventilate the chassis. Normally, it should never run:
Figure 11. The fan, its daughterboard, and the AC inlet.
The AC inlet is wrapped in silicone insulating tape, which is concealing a "poor-man's noise filter" in
the form of a parallel 100nF polyester capacitor and 1MΩ resistor in parallel across the mains
terminals. The Active (hot) lead feeds through a thermal circuit breaker, and the output feeds over to
the softstart board via a Molex connector. The fourth position is a neutral return for the master
power switch, which includes an internal 120V neon bulb for indication.
For audio I/O, I decided to replicate two tricks that I had used on my P101 amplifier. One was a repeat, and that was to mount the RCA inlet jack on a section of fiberglass, then have it protrude through an oversize hole in the rear panel. A 100pF capacitor allows high-frequency RF to be immediately shunted to earth via the chassis, but the jack is otherwise isolated. A section of unetched copper PCB offers precautionary shielding between the output jack wiring and the exposed connections of the inlet jack, as this can be a source of noise or even oscillation:
Figure 12. The muting relay assembly and inlet jack.
The other repeat effort, pioneered for the first time on my P101 with great success, was to use a recessed
speakerbox plate for the amplifier output and mount the muting relay on the standoffs that would
normally support a passive crossover assembly. The relay is mounted on a fiberglass plate, which is then
mounted and wired into the jack. Not only does this improve the appearance from the rear and reduce the
likelihood of the output jacks suffering abuse, it reduces the amount of gear that has to be somehow mounted
to the chassis directly. 20A power relays aren't required for this build, but they will almost certainly
be one of the last components to ever fail in this unit.
The panel was completed and ready for installation. The IEC inlet jack, thermal circuit breaker, and speaker plates were perimeter sealed with 15-minute epoxy. The epoxy is weak and cracks under moderate stress, but when placed around a firmly-secured object, it seals the joint and hinders the slight flexing, that would otherwise occur, improving the general "feel" of the unit. The fresh epoxy runs readily and will creep through any gaps, so it does need to be applied with due caution.
Figure 13. The completed rear panel.
With that done, it was time to finish assembling the unit. The amplifiers were installed with beryllium oxide ceramic thermal pads to isolate the output MOSFETs. The compensation diodes were altered for direct installation against the heatsink on small perfboard sections, and wired back to the respective amplifier boards. All wiring was completed -- this always takes more and longer than ever expected -- and the chassis was ready for the front panel and cover to be installed:
Figure 14. The completed amplifier, ready for the front panel and cover.
Of course, before running the completed unit, it has to be biased. Having access to a pair of Agilent 6-1/2 digit multimeters isn't a requirement, but it does help, especially when the setting procedure for one channel requires fine-tuning two amplifiers simultaneously:
Figure 15. Biasing the amplifiers for a channel.
Once the channels had stabilized, sine tests were performed:
Figure 16. Sine-wave tests of a channel.
Once everything was determined to work correctly, the last step was final assembly of the chassis.
Optimized for a 1024 viewing width ~//~ All content (c)2008 aaronv.net