Dubrow TS-452 C/U Sweep Signal Generator / Oscilloscope
This is a mechanical sweep frequency generator, sometimes known as a wobbulator that like the Triumph 830, incorporates an oscilloscope to enable it to be used as a stand alone instrument for peaking RF circuits. It was introduced in 1959, mine is serial number 21 so I presume it is very early. Quoting from the manual “….used to check the alignment of intermediate-frequency and radio-frequency circuits in the 5 to 100 megacycle range. It may be used for the maintenance of equipment, as a laboratory instrument, for experimental work or for production testing.”
It was shipped in a substantial, splashproof olive drab case of metal and balsa wood laminations. When I received it, it was stuck firmly to the supporting rubber blocks in the case, I had a real job getting it out since it is also heavy.
Here is the swept RF oscillator; a motor rotates a vane condenser that is in the RF oscillator causing a 2:1 frequency modulation of the generated RF signal.
Attached to the left hand end of the same motor shaft is a (sweep) generator that generates a sinusoidal frequency at the same rate as, and synchronised to, the RF frequency sweep or wobbulation. I took it apart because the rotation was unsteady and noisy. I found a rusted bearing, interestingly, there are no other signs of moisture damage other than the one bearing (that is on the vane capacitor end of the motor armature.)
Here is the sweep generator:
The generated sweep signal is applied to the X axis of the built in oscilloscope. Here are the oscilloscope controls:
Here is the nice RF attenuator that is used to accurately control the amplitude of swept RF being applied to the circuit on test.
Changing the sweep range of the oscillator involves plug-in coils, an arrangement that is probably more reliable than using a band switch and most certainly cheaper!
Here is the RF oscillator re-installed and the coil set that is mounted on a door in the cabinet:
Operation: As the frequency sweeps, the beam moves across the screen; the swept frequency is applied to the circuit on test, the (rectified) response of which is applied to the Y axis (a Y amplifier is incorporated) thus the passband and peak can readily be observed and adjusted, providing it is in the swept frequency range. Don’t laugh, this is not always straightforward! A surprisingly accurate wavemeter (I verified this while screwing around with the thing) is included that resonates at the frequency it is set at so that when the sweep frequency passes through the wavemeter resonance, it puts out a signal peak that is differentiated and amplified then applied to the CRT grid to blank the oscilloscope beam at the peak. The wavemeter may be adjusted until the passband peak of circuit under test is displayed at the blanked spot and the peak frequency read directly from the wavemeter dial; it looks like this:
This tells us that the actual peak frequency here is 9.7 MHz, not the expected 10.7 MHz, most likely due to capacitive loading of the circuit by both the applied signal connection and the output connection. I was simply creating a demonstration, doing this properly requires experience and care. The other way to use it is to set the wavemeter at the desired frequency peak and adjust the circuit under test until the peak appears at the blank spot, thereby peaking the circuit at the desired frequency.
The operation of this instrument takes a little thought, here is the functional block diagram, enjoy your headache! The only element I have not so far described in this account is the multivibrator. Since the motorised rotary capacitor vane has to be balanced unlike a normal tuning capacitor, it has a double set of vanes, diametrically opposed for mass-balance. This means that the frequency rises and falls twice per cycle thus there would be four traces per cycle, forward rise then fall and return rise then fall. The multivibrator is synchronised to the sweep generator (that generates the X sweep) and arranged to output a gate pulse of one quarter turn duration; this gate is applied to the RF oscillator via the oscillator AVC circuit turning the oscillator on during just one rise per cycle resulting in a clean and stable single peak display.
Here is the inside of the wavemeter showing the tuner:
Here is the dial:
Restoration: You may have noticed my post on electrolytic capacitors. If you are active or interested in reviving vintage equipment, please check that post out. This is one occasion where I did not take my own advice and killed the previously beautiful sealed power transformer. EVERY electrolytic was dry and only two of them surrendered to reforming. This is a shame, the electrolytics are nice octal plug in units. In addition to this, every paper-in-oil cap, except those in the CRT HV supply, were leaky too as well as a couple of the silver micas. I think this unit must have spent most of its life in storage. I have left all the caps in place, simply disconnecting them and installing modern replacements. I spotted a ratty USM-32 oscilloscope on eBay and knew I would be able to adapt the power transformer from it. A few bucks won it, fingers crossed that the transformer would be OK! It was (is). Ironically, the clean USM-32 I restored needed a re-wind! Amazingly, the mounting bolt layout is the same. The B+ voltage is high however, I saw that simply removing the reservior cap to create a choke input arrangement would work. The resulting B+ is within 2% of the specified value of 250 V. The choke is a nice potted unit that does not hum, not that it would be audible over the racket the sweep rotor makes! The C- has no choke so I increased the series resistor to suit. I needed an extra 6.3 V winding, a Radio shack 12.6 VCT was in my box, it proved possible to open it up and separate the centre tap giving me an extra 6.3 V winding. I wanted this since I was not happy with the B+ rectifier (6X5) heater being common with all other heaters and grounded. Though the 6X5 has a heater/cathode insulation rating of 450 V, I did not like it. In fact the 6X5 was destroyed, one plate was burnt badly! I suspect an insulation failure here may have wrought the demise of the power transformer; there were signs of a previous attempt at repair and I think the whoever was involved gave up on it. I used the extra winding to power the B+ 6X5 heater.
Here is the underside, you can see the additional heater transformer at the top left. The replacement electrolytics are obvious, the wiring has been re-configured so that the plug-in electrolytic capacitors sockets can be left in place for the sake of top-side appearance.
Here is a picture of the topside, the replacement power transformer is obvious. The valves and octal electrolytic capacitors are retained by a neat sprung plate.
Here it is with the plate removed, you can see the defective capacitors marked “NG” for No Good:
The oscillator frequency is some 30% low, I have not yet got to the bottom of this issue however, the accuracy of the instrument is dependent on the wavemeter which at the two frequencies I checked it at, was dead-on.
Here is the schematic: