The cosmetic condition of this set was poor with some rust on the chassis and cabinet. The bottom is missing. I will have to cut a piece of metal to fit the bottom. I replaced the volume control which I had to destroy in order to remove its corrosion-welded knob. I removed much of the obvious chassis rust by scrubbing with plastic scrubbers and even some light sanding. The cabinet was also cleaned but will need a repaint. I will have the paint computer matched. The front panel markings are engraved which should allow restoration of the labels.
As usual, I used a bit of deoxit on all the tube sockets and controls. I reformed the electrolytic with an external variable power supply. The set's power supply choke had been replaced earlier in its life, but the choke's lead wires were simply spliced to existing wires. I corrected the poor wiring job. The set has a 0.1 MFD capacitor connected across the AC line. As expected, it was quite leaky. I replaced it with a smaller value cap (0.05 MFD) rated for direct power line connection. The 820 ohm resistor feeding the rectifier plate has been replaced at some time in the sets life. The replacement was poorly done, not only was there a cold solder joint, the resistor lead wire was loose on one end. I corrected the poor soldering job after verifying that the resistor was the correct value and in the correct location. I also found a 1000 ohm resistor in the screen circuit of the 12SA7 with the same problems and resoldered. The missing pilot lamp was replaced with a #47.
I slowly powered up the set with my isolated variac while watching current draw. This set only requires 24 watts, most of which goes to power the series filaments. Since the tubes and the pilot light began to glow, I knew that all the filaments were intact. By tuning the dial and changing the bandswitch to short wave, I was able to faintly pick up a couple of short wave stations. The broadcast band was dead. A quick check with a frequency counter showed appropriate readings from the oscillator section of the variable cap on all ranges except the broadcast band. There the counter read a consistent 5 MHz or so with only slight variations when changing the tuning dial. Suspected a problem with the broadcast band oscillator coil or bandswitch. Found that one of the wires leading fron the bandswitch to another coil had been bent enough to contact a broadcast oscillator coil terminal and short it out. Moving the wire solved that problem.
An "injected fault"
The set was operating but could only pick up strong stations. Tweaking the converter RF trimmers at the high end of the bands helped a little. Decided to do an alignment check of the 455KHz IF's. The primary and secondary of the first IF peaked sharply as did the primary of the 2nd IF. However the secondary (output side) of the second IF did not peak at all. Replacing the somewhat leaky AVC cap did not help. Removing the IF shield can did not reveal any obvious faults. Voltage checks on the 12H6 detector also did not reveal obvious faults. However, I noticed a wire leading from the BFO circuit to the 12H6 plate. The schematic showed no connection, only a dotted line capacitor which I assumed was a "gimmick", two insulated wires twisted together to act as a low value capacitor for BFO injection. I cut the wire. The sensitivity of the set immediately improved. I then proceeded to peak the IF secondary to its normal sharp setting. The set's sensitivity improved greatly to its normal level. I am assuming that a former owner, possibly in an attempt to improve BFO injection or probably seeing two twisted but not connected wires, directly wired to the detector tube and the set was blamed for poor sensitivity. I replaced the connection with a proper "gimmick".
Solving the "hot chassis" problem
This set is AC-DC with floating ground. Six capacitors, three of which are 0.1 MFD, connect the chassis to one side the power line. I replaced those with modern caps since all were leaky. (see the old caps in front of the chassis.) However, even new caps will create a shock hazard since the effective value of all the caps in parallel was about 0.33 MFD resulting in a leakage current of about 15 milliamps. I tried reducing the value of the 0.1 caps to 0.05. On testing, by connecting the ground lead, the GFCI on the AC line to the bench still tripped. This was no surprise since a GFCI will trip when it detects leakage of about 4 to 6 milliamps. On reversing the plug, the GFCI held but, as expected, tripped when I turned the set off via the power switch on the volume control. I don't like to operate sets that cannot pass the GFCI test. One solution might have been a proper polarized or 3-wire grounded line cord using an external power strip for turning the set off. However, I am a stickler for safety.
Solution - isolation transformer
I located a small 35 VoltAmpere isolation transformer in my "boxe de junque". I first thought about mounting it on top of the chassis but the size was small enough to comfortably fit under the chassis. I re-used one existing mounting bolt hole and drilled a second one. While I will only rarely modify a set, especially an "antique", in this case I felt it was fully warranted. The chassis is quite spacious and has plenty of room for such a safety improvement. By mounting the transformer under the chassis, the chassis top retains its appearance.
Comparing the NC-33 to the Hallicrafters S-38
The circuit of the NC-33 is similar to the first version of the Hallicrafters S-38. Both were intended for the same low-priced market. The National appears to be designed a bit better with some better components than the Hallicrafters. For example, the alignment trimmers are better quality and the set includes a power supply filter choke. It uses larger variable capacitors with better mechanics. The IF transformers are better quality. The NC-33 floating ground with those 6 caps between one side of the power line and the metal chassis is better than the early S-38 which had one side of the power line directly switched to the chassis and a cap between the chassis and the cabinet. I recommend operating either set with an isolation transformer. The NC-33 was more expensive, priced at $65.95 while the S-38 was priced at just under $50. Overall, I consider the NC-33 a better quality set than the S-38. However, the more compact Loewe-designed S-38 is usually considered to be the better-looking set.
What's A Tannhauser? - The Nordmende -Sterling Tannhauser USA was the previous item "on the bench".