I've owned this little 'scope for several years now; received it in a non-operational state (the tubes were glowing but nothing did appear on the CRT). Finally decided/got around to sorting it out the past few evenings. For what it is it is reasonably sophisticated, having a proper frequency compensated attenuator for the vertical channel with specified sensitivities. Bandwidth is 4 MHz on the lowest vertical sensitivity.
The horizontal amplifier is a bit odd. A triode cathodyne phase splitter provides the complementary sweep waveforms to drive the horizontal deflection plates. The signals from the cathodyne are capacitively coupled to the plates and the plates are DC biased so that (in the absence of a sweep waveform from the timebase circuit) the beam is deflected off screen to the left. The saw-tooth sweep waveforms at the deflection plates are DC-restored by (hollow-state) diode clamps. There are no horizontal position or horizontal size/gain controls. So, you won't see/get anything on the CRT if the timebase is dead, and dead my timebase was.
Starting at the trigger input circuit for the timebase I found the first dud component noticing that some tube biasing potentials didn't measure right. A 680k resistor measured open circuit. With the resistor replaced the respective pentode stage still didn't appear to bias properly and this was further found to be due to a dud/low emission pentode (half of a 9A8 triode/pentode dual). After some searching through my NOS Miniwatt stash I found a 6BL8, which is a direct 9A8 equivalent. Substituting the dud tube caused the timebase to spring to life, but the saw-tooth waveforms weren't appearing at the deflection plates and I still didn't have anything visible on the CRT.
Probing the cathodyne stage immediately identified it as being dud. It wasn't biased properly, the cathode DC potential with no signal input being only a few volts above ground rather than a couple of hundred or so. The issue was traced to the 15k plate load resistor which measured open circuit. After replacing the dud 15k resistor the cathodyne sprung to life and I had proper complementary saw-tooth sweep waveforms at the cathode and plate.
That was the timebase and horizontal amplifier sorted, but the resulting signal trace on the CRT was still mostly off screen and all wonky and out of focus. I computed the DC potentials throughout the CRT circuit from the schematic diagram and compared these to measurements with my DVM. The voltages were well out and the cause traced to a 68k resistor that measured over 250k. With the dud 68k replaced the CRT circuit voltages came up as predicted, the visible trace could then be adequately focused, but the horizontal deflection was still stuffed as most of the signal trace remained deflected off screen. The DC potential at one of the horizontal deflection plates (measured with a 100M 100:1 probe for accuracy as the DC resistance to ground here is ~4.7M) was a couple of hundred volts too low. I took a punt that the 220nF signal coupling capacitor was leaky, pulling that deflection plates DC potential down towards the cathodyne's plate potential. This was verified by disconnecting one end of the capacitor and then witnessing the deflection plate DC potential return to the +430V it was supposed to be at. Scrounging through my junk capacitor box eventually produced a 220nF 630V polyester, which was promptly soldered in to substitute the dud.
Bingo. Five dud components down and I now have a seemingly fully functionally GM5650. It's still far from restored though. Many of the resistors still measure somewhat high and a lot of the ancient big black capacitors look so badly deteriorated that it's a wonder that they haven't all gone leaky either.
I also had to repair the CRT itself. It's a Philips-made DG7-32 of which I have several Mullard-made spares. As is typical for these old tubes the glue bonding the pin base to the glass envelope had given out. I carefully extracted the CRT from the scope (so as to not bust the interconnecting wires inside the pin base) and then re-bonded the base with some 24-hr Araldite. Shortly after doing that is when I learnt that none of my Mullard DG7-32's could be used as a substitute for that Philips one. To avoid waiting for the Araldite to harden I tried to plug in one of my Mullard ones, but it became a snug fit inside the mu-metal shield well before the base pins could properly mate with the socket. Compared to the Mullard DG7-32's, the neck of the The Philips' is ~12mm longer!
Home
Help
Search
Login
Register
