Thank you for such great responses regarding our in depth article on the Acoustic Research AU Amplifier.  It is refreshing how many people care and wrote back.  The unit we restored for this issue is the NAD Series 20 Model 3020 Amplifier.  This unit sold at Gig Harbor Audio for $389.  Just to give a context, it cost about $250 to repair it in parts/labor, so this unit was a steal.  The stores in the Pacific Northwest that do restorations/repairs like this besides Gig Harbor Audio are Hawthorne Stereo and Northwest Audio Service in Seattle and Echo Audio in Portland.  This is a dying art in that the institutional knowledge of these technicians can't really be taught.  They are all getting up there in age and have had lifetimes of seeing products come out, getting to know the work arounds for many unavailable parts, and understand that each of these units is a piece of history and innovation more closely resembling an artistic sculpture. 

On the NAD amplifier both channels were inoperative.  Both input line AC slo-blow power supply fuses were blown.  A check of the principle output transistors showed that in each channel one had a PNP power transistor defective and the other channel a NPN power transistor defective.  Also found on one channel one driver transistor (2N6554) was also defective.  Further investigation revealed that a 680 ohm bias resistor was charred.  Still was in tolerance surprisingly but replaced just the same.  Checked signal and power supply electrolytic caps and replaced the four principle power supply filter caps.  Also found two electrolytic caps in the power amp signal path leaky, replaced same.  When replacing the power transistors also found a common problem in that the heat conductive compound which is smeared on both sides of the mica insulator was excessive.  This actually inhibits heat transfer as there should only be enough compound to provide a very thin even layer on both sides, just enough to fill the air gaps when the transistors are installed and torqued down.

Initially when the unit was first openedwe were not especially impressed with the build quality and were curious why these amps have such a good rep. Once repaired it could be seen that build quality aside the engineering employed makes up for any issues.  For example the 2N3055/MJ2955 power transistors are used in the output.  These transistors can handle 15 amps of current and dissipate 115 watts of power each.  The only real limiting factor is keeping them cool with a robust heatsink.  Of course finned heatsinks are expensive and this unit was designed for domestic use so the existing heatsink is adequate to meet the demand.  This amp is conservatively rated at about 20 watts RMS/Chan into 8 ohms.  With that kind of engineering the only thing we can figure is that someone needed to arc weld something using the speaker leads in order to blow it up.  

Tested for 8 hours under various loads for the release of the magic smoke, no problems.  In fact the amp barely got warm.  DC balance and idle current OK.  This amp cranks and puts out more db's than we could stand to listen to with Klipsh speakers.  This is only when pushing 1 to 10 watts!  It can get very loud with the right combination of speaker load.  Sound is clean and crisp and with no signal in is dead quiet at an idle. Impressive. Cleaned switches and controls with Deoxit.  Tested inputs and phono stage, all OK.  LED power output indicators working.  It sounds great.