Today I calibrated my recently purchased Hewlett Packard 5315A Universal Counter, the portable sibling of HP’s 5316B bench counter. My unit is in very nice shape and has option C to go to 1GHz, and the option 4 ovenized oscillator — this is particularly important as this portable unit has no external 10MHz reference input. (Though I may construct one.)
I used my Trimble Thunderbolt GPSDO as a very solid 10MHz reference. Before calibration, the 5315A clocked it at 9,999,984 MHz which isn’t bad for an old instrument. But I am picky, so I calibrated the counter per the service manual.
The first step was a nice long warm-up over a couple of hours.
Old instruments like this don’t have case-closed calibration — I had to remove the entire case and set the circuit board on the bench.This is careful work as high-voltage (mains 120V) are exposed at the back of the board when it is plugged in (as it must be on to be calibrated).
Next, I checked the voltage of the 5V regulator on the oscillator card. It was in spec (5.025V) but I adjusted it anyway – to 5.00019V according to my recently Keysight Calibrated (ouch$) 34401A DMM.
On to the fun part… I placed an oscilloscope probe on the reference oscillator point on the main board on channel 2 of my scope, while channel 1 was connected to and triggered by the Thunderbolt. If these two sine waves had the exact same frequency, they would both remain static on the ‘scope display (though not likely in the same phase sync). Of course the frequencies will never match exactly — the waveform on channel 2 will scroll left or right past the triggered wave on channel 1 depending on whether channel 2 is higher or lower frequency; the “sliding” speed is greater the farther they are apart. The objective is to tweak the frequency adjustment screw on the 5315’s oscillator so that the picture on the scope remains still (no sliding/scrolling). Of course this is harder than it sounds and definitely harder than tuning a guitar. Not only do fine movements of the screw make an appreciable difference on screen, but the slight pressure of holding the screwdriver into the adjustment screw slot stresses the device and changes the frequency. Ergo, removing the screwdriver results in a frequency change. The trick seems to be over-adjusting slightly, so that it is (more or less) in the desired position/frequency after withdrawing the screwdriver. I needed more than a few tries (especially since I am picky).
Once the oscillator was adjusted, the circuitry needed to go back into the case and run for another 30 minutes to stabilize and ensure that the adjustment is on target. (Temperature inside the case and stress of mounting the circuit board can change the frequency.). Again, a couple of tries…