I didn't set out to acquire a 1605-A, but it was included with a pair of 716-C bridges I bought. The 1605-A sounded interesting because of two applications. First, it can measure remarkably low values of dissipation factor, something I'm interested in because of my audio interests. It's other feature is the ability to accurately measure the tracking of ganged variable capacitors and resistors, useful in both radio and audio. It can resolve .0001 radian of phase difference, which is enough for just about any application.
My 1605-A wasn't in the best of shape, as it didn't work, nor were the knobs all the same shade of gray. After obtaining a manual, I set to work. I'll say right up front that out of all the things I've fixed throughout my career, this was by far the most difficult.
The problem was that the instrument sort of worked, but I couldn't get through the whole calibration and balance on the more sensitive scales. Ultimately, I checked every single component in the thing to no avail. Someone had modified the inputs slightly, and I returned them to the balanced factory configuration. That wasn't it. some of the matched range resisters were a bit out of spec. I changed them. That wasn't it. I racked up close to 40 hours of troubleshooting time before finding the answer.
It's been too long to remember how I stumbled onto it, but the problem turned out to be a bad solder joint on the input of one of the voltage follower stages. The stage had so high an input impedance that signals coupled just fine across the capacitance of the open joint, and the stage appeared to be working normally. The small series capacitance did however, introduce a small phase shift compared to the normally working side. Enough to keep the instrument from balancing correctly on the more sensitive ranges. The board with the problem was somewhat difficult to remove, but as soon as the joint was touched up, the complete calibration procedure went through without a hitch.
Like the Z-Y bridge, this isn't an instrument you reach for every week or every month. It's extremely specialized and remains on the shelf waiting for just the right problem to solve. My only complaint is that it operates at only 4 discrete frequencies, 100, 1K, 10K, & 100Khz. The ultimate instrument would have a continuous range of up to a megahertz or so. Still, a remarkable piece of engineering for 1956, and I notice that it remained in production at least into the '70s, albeit as a solid state design.
C. Hoffman
last edit August 28, 2010