Audio Research D110b
A 1980s Behemoth
This 92Lb. “Beast”, dropped off by a client, had blown the speaker fuses in one channel and taken out the speaker it was powering. The D110b is capable of delivering 220 Watts per channel (RMS Continuous) and 440 Watts in bridged mono mode into 4 Ohms.
Made back in the early 80’s this amp is still sought by audiophiles for its headroom and sound quality. They are extremely well made and built to last. The output transistors exist dead center of the unit, sandwiched between the massive heat sink on one side and three layers of PC boards on the other. This whole assembly is screwed into a metal bracket and mounted to the chassis on four sides. The PC boards are mounted “Neve style” with the boards soldered to each other by pins. Getting to the main transistors required complete disassembly of the chassis and the PC board assembly. This allowed access to the screws holding the heat sink on and the transistors.
This particular unit had been worked on before, probably several times, evident by the broken welds on the tee nuts inside the chassis. There were several, and regular nuts were threaded in from the inside to compensate. The power supply and the other channel were working so the problem was isolated to one channel.
Starting at the output, the first two output transistors I put my probes on were reading open. These were at the outer edge of the assembly and the only ones reachable without a complete disassembly. Removal of the PCs is tedious and care must be taken not to destroy the pads and traces. There are nine boards connected by 10 or so pins in three layers. Each pin had to be completely free of solder before each board could be pried off. Then the pins had to be cleaned of residual solder so as not to push out a pad when reinstalling.
The remaining 14 transistors on that channel tested fine but the cause of the demise of the first two was yet unknown. Since the boards were now removed it was easy to test the individual components so I started looking for shorted diodes and open capacitors. I found the four large caps, sandwiched between the boards, to be reading low by more than 30%. These caps were long and skinny so as to fit between the boards. The ones available today are shorter and fatter so spacers had to be added and the interconnect pins had to be lengthened to fit them in. I still didn’t see this as a reason for the blown output transistors. Upon further inspection I noticed a 14ga ground wire running from the wire harness to the main board was only being held on by 2 or 3 strands. The rest had broken free. I redressed the connection but still wasn’t convinced this was the cause.
Finding no other faulty components I reassembled the unit. I connected it to a large “dummy load” and brought it up slowly on a variac. No smoke and no blown fuses. After getting the unit up to 120 VAC I let it idle for about 10 minutes. I then injected a 1KHz test signal and put scope across the dummy load. The signal showed clean on the CRT. With no signal, I read the DC offset at the output to be 0.0017 Volts (1.7mV).
In conclusion, I think the blown voice coil in the speaker could have been the cause of the original problem. The coil may have overheated and shorted causing the impedance to drop below 2 ohms. This unit was designed to deliver a lot of power very quickly to handle musical transients. It’s very possible it overpowered the speaker.
Number of Channels: Single channel
Total Power: 20W (5W setting)
Speaker Size: 1 x 10" Celestion V-type Speaker
Preamp Tubes: 3 x 12AX7
Power Tubes: 2 x EL34
EQ: 3-band EQ
Inputs: 2 x 1/4" (high, low)
Outputs: 1 x 1/4" (16 ohm internal), 2 x 1/4" (single 8 ohm or 2 x 16 ohm), 2 x /14" (single 4 ohm or 2 x 8 ohm), 1 x 1/4" (DI out)
Effects Loop: Yes
Power Source: Standard IEC AC cable
Weight: 32 lbs.