Many oscilloscopes lack a delayed-sweep function. Plus, because typical users only occasionally require this function, simpler, more adaptable methods are worthwhile. Cross-triggering two standard oscilloscopes produces both main and delayed-sweep displays. This arrangement also grants designers certain advantages over purchasing a specialized dual-trace oscilloscope with delayed sweep.
For one, it comes at much lower cost, particularly if two single-trace scopes are already available—including those with higher bandwidth than the dual-trace unit. Additionally, this double-beam system offers both main and delayed-sweep traces on every shot, an essential feature for single-shot work. In contrast, the dual-trace unit supplies only one of the two traces, alternately.
Also, the gate output, G, may easily be used as a delayed-pulse generator to trigger or probe a DUT at various times, with timing and resulting outputs clearly displayed. Finally, other delayed scopes can be triggered from the ramp of either unit to monitor additional circuit details. Modern oscilloscopes don't generally feature the required outputs.
Neither oscilloscope needs the delayed-sweep function, but both should contain triggered-sweep and input delay-lines for entire signal display. Furthermore, one should have a time-base ramp output, T, and a Z-input for beam-intensity modulation by gate output G (Fig. 1).
Alternatively, a second channel may be used to display G and identify the delayed-sweep interval. A third method would be to add the gate G to the input, VIN, so that G shows as a displacement on the display, using a simple resistive averaging network to form the sum.
A 10-MHz Philips PM3230 (main) and a 15-MHz Tektronix 515A (de-layed) were implemented for these tests. In this tandem delayed-sweep system, input VIN is connected to both 1-MΩ oscilloscope inputs (In). However, depending on the source impedance, one or both may be connected via a 10×, low-capacitance probe, or 50-Ω terminated cables.
First, the "Main" scope triggers, either by VIN or some suitable external trigger, and its trace displays VIN. Soon its synchronized ramp output, T, triggers the delayed scope via its external trigger input, E. The system's trigger-level control (LEV) determines the point on the ramp where triggering occurs.
Thus, arbitrary adjustment of this LEV knob controls the second display's start delay, which may be chosen over the entire main-display range. Duration of each scope's display is adjusted in the usual manner, using their time/division knobs (not shown).
The gate output, G, is simply the unblanking pulse generated within the delayed scope. This can readily be added to the main scope's unblanking pulse by connecting it to its Z input. If G is too strong, an ordinary 10× (or 5×) low-capacitance probe may be used for the connection, rather than a coaxial cable. The augmented main beam will accurately highlight the delayed-display interval, as shown on the PM3230 display.
This basic arrangement works very well. Yet, the delayed display exhibits jitter on fast transitions from the combination of the signal's own jitter, ramp T fluctuations, and trigger comparator at E. The latter two may be eliminated by triggering VIN at any constant phase of the 60-Hz power frequency. This procedure always eradicates power-supply ripple effects.
Further illustrating this point: If VIN is a fast pulse triggered at a slight offset frequency, like 61 Hz, a low-jitter pulse will actually meander to-and-fro on the delayed display, at the difference frequency of 1 Hz. Precisely at 60 Hz, the pulse freezes.
To examine a fast transition at higher repetition rates, the transition must be retriggered after the delay. Most commercial delayed-sweep units can do that. The optional passive averaging network in Figure 2 provides just such a delay-qualified trigger function. If ramp T and the VIN transition have equal polarity, then by tuning the LEV knob, E can reliably trigger the delayed scope on that transition—excluding all others. On the other hand, when they're of opposite polarities, one must be inverted before the adding network by an inversion transformer or amplifier.
This network was first tested using a 2-kHz squarewave for VIN to determine the optimum compensation capacitor values of 56 pF. The process is analogous to tuning a low-capacitance probe. To display the resultant E while it externally triggers the delayed oscilloscope, employ a 1-pF, 100-MΩ low-capacitance probe. If plugged into the Tek 515A input, or any other 1-MΩ scope, this 100× attenuation probe will show the VIN signal added to the ramp, while triggering on the desired transition at E. After plugging in VIN, the 515A will display any transition reliably and clearly, with no perceptible jitter.