A few months ago on a lark I bought an HP333A distortion analyzer. It was supposed to be complete and unmodified, but upon inspection I discovered that the Electronics Magic Board, the one with the light bulbs and unobtanium CDS cells was missing. Upon further inspection I discovered that someone had started replacing the meter sensitivity switch: it was all disconnected. I got the HP PDF for it and figured out where the half-dozen clipped wires were supposed to go. Well, almost figured out. Today I got down to business of resoldering all those wires back in place. I then followed the basic checkout procedure … and got some serious flakiness and weird behavior. Turns out one of the wires didn’t belong on that obvious tab at all, but on that one there—what the hell? The resistor there had been clipped! Resolder the wire and the resistor, and try again.
I’ll interject here to say I really like having a four-channel o-scope. I got a 5110 a while back and bough another 2-channel plugin for it to match the one. It hasn’t got the greatest bandwidth in the world, but it’s plenty for the HP333A. I can watch my test signal, the analyzer’s output signal, and look at it at various points along the way.
The HP333A uses magic to get the automatic Wien-bridge balancing to work. Some eldritch wizardry with incandescent panel lamps in the circuit over here shining onto CDS cells in the circuit over there, affecting the resonant frequency of the notch filter. When I saw that the circuit board for the lights and cells was missing, I practiced what Old Saxon I know and examined the circuit diagram to see if I could bodge the thing to make it work without them. All the cells do is change some resistances in the filters. I shall have to figure out what the equivalent resistances are supposed to be with the thing set to Manual. I taped up all the connectors with some of that icky black electrical tape.
I suspect that the thing had been used to donate the Magic Board and the Meter Range switch to another, perhaps more deserving instrument. The Frequency Range switch in this one is clean and crisp and shifts like a Volkswagen. In comparison, the Meter Range switch is grungy and worn and flaky and feels like the shifter in an old Lancia.
I was getting no output from the thing at all. It turns out that the wire from the input amplifier and the wire to the notch filter are not the same wire, but they both plug into the Magic Circuit Board. So I pulled them out of the tape and connected them together. Now it makes output.
The voltmeter function works and the sensitivity setting procedure appears to work: the HP333A and the HP331 signal generator agree on powers of ten. But nulling does not work. The filter acts like a band-pass, not a notch: instead of a minimum signal at the center frequency, that’s where I get a max. At least the signal generator and the distortion analyzer agree within reason about frequencies. Now I need to read again the Theory of Operation and try to determine exactly what removing the CDS cells did to the notch filter. Perhaps I can replace them with ordinary resistors. There will be math.
You might be thinking, why doesn’t this guy go after the Fleabay seller of the dud distortion analyzer? I thought about that, but I paid only eighty bucks for it, and poking around in it to sleuth out its workings and problems has been fun. Maybe when I get tired of it I’ll dump it on Fleabay with a full description of its problems. Or I could try to channel Mister Carlson and design a substitute Magic Board. It’s just a fancy analog optoisolator with hysteresis, right?