- Bandwidth
- 100 MHz on UTD2102CEX+; 150 MHz on UTD2152S-E; 200 MHz on UTD2202CEX+
- Channels
- Two analog channels, with independent coupling settings
- Sampling rate
- 1 GSa/s maximum
- Memory depth
- 64 kpts
- Display
- 7 inch TFT LCD, WVGA 800 × 480
- Waveform recording
- Up to 1000 screens recorded and replayed
- Automatic measurement
- 34 waveform parameters, with manual, tracking and automatic cursors
- Frequency counter
- Six digits, on the trigger signal
- Input
- 1 MΩ in parallel with about 16 pF
- Body
- 306 × 138 × 124 mm, about 3 kg
An oscilloscope answers a question a multimeter cannot: not what the value is, but what the signal did between two instants. The instrument in this range is the bench version of that question - two channels, a screen large enough to compare them, and enough memory to keep what it captured.
Product overview
Three models share a body, a display and a control layout. They differ in bandwidth: 100, 150 and 200 MHz, with the same two channels, the same record length and the same maximum sampling rate.
The front panel is arranged in the conventional way: a column of menu keys beside the screen, then vertical, horizontal and trigger blocks to the right, with the two BNC inputs below. Anyone who has used a bench scope before can drive it without reading the manual, and the AUTO key covers the case of someone who has not.
The case is about the size of a large book and weighs around three kilograms. It has a handle and feet, which is what a bench instrument needs if it is shared between two stations.
How the oscilloscope captures a signal
The input passes through an attenuator and a buffer, and is then sampled by an analogue-to-digital converter at up to one gigasample per second. The samples go into a record memory, and the record is what the instrument displays, measures and stores.
Two numbers govern what can be seen. Bandwidth sets the fastest edge the instrument can reproduce: a signal with significant content above the bandwidth arrives attenuated and rounded, whatever the sampling rate. Sampling rate sets how finely the record is spaced in time. Memory ties the two together, because at a given sample rate the record covers a fixed amount of time - a longer record means a longer window at full resolution.
Triggering decides which part of the stream is kept. The trigger circuit watches for a defined condition - an edge, a pulse width, a slope - and the acquisition is stored around that point. The frequency counter works from the same trigger signal, which is why it gives a stable reading without a separate measurement setup.
The measurement engine works on the stored record rather than on the live signal. That is why a frozen acquisition can be measured, cursors placed and parameters read after the event, and why recording successive screens is useful: the analysis can wait until the fault has been found.
How to select the right model
Bandwidth is chosen from the fastest edge, not from the repetition rate. A rule of thumb is to have several times the bandwidth of the highest significant frequency in the signal, so that the edge is reproduced rather than rounded. Where the shape of the edge is the measurement, more bandwidth is always the safer choice.
Channel count is chosen from what has to be compared at once. Two channels cover input against output, gate against drain, or a voltage against a current derived from a shunt. A third and fourth channel change the instrument class and are not available in this body.
Memory is chosen from the time window that has to be captured at full resolution. Looking at a switching edge and looking at a start-up sequence are different requirements, and the second one is where a short record becomes visible.
Connectivity is chosen last. A bench that records results needs the USB host for a memory stick or the USB device port for a computer; a bench that sorts parts needs the pass and fail output, which is not stated for every model here.
Measurement limitations and selection considerations
A probe is part of the measurement, not an accessory. A switchable passive probe in the 10:1 position loads the circuit far less than in the 1:1 position, but it also attenuates the signal and, at high frequency, the probe’s own compensation has to be adjusted against the calibrator on the front panel. An uncompensated probe produces a square wave that looks rounded or peaked, and the error is easy to attribute to the circuit.
Bandwidth and sampling rate are not interchangeable. A 100 MHz instrument sampling at 1 GSa/s gives ten samples per cycle at its bandwidth limit, which is adequate for viewing but marginal for measuring rise time. Where the edge is the measurement, the rise time specification matters more than the bandwidth headline.
The record length limits slow signals. At a slow time base, the sample rate falls unless the memory is deep enough, and the trace can look continuous while actually being reconstructed from far fewer points than the display suggests.
Finally, the instrument measures voltage against time. Current, power and impedance are derived from that, using a shunt, a probe or a maths function, and each of those introduces its own error. A current waveform from a clamp is only as good as the clamp’s bandwidth and its stated transfer ratio.
Every value below is taken from source documentation. Where a parameter is not stated in the available material it is marked as information required rather than estimated.
