Choosing an Oscilloscope: A Hobbyist's Guide

Buying your first oscilloscope is overwhelming. Every spec sounds important — bandwidth, sample rate, channels, memory depth, FFT, decoders. Every model claims to have just the right amount of each. The price ranges from $100 to $50,000 with no obvious justification for the differences.

This article is the field guide that cuts through the marketing. We assume you are a hobbyist, student, or working engineer with a budget under $2000. Most people fit there. Most people overpay anyway.

The specs that actually matter

Bandwidth

The frequency above which the scope visibly attenuates the signal. A 100 MHz scope can faithfully display signals up to about 100 MHz. Above that, edges look slower than they really are.

Rule of thumb: pick bandwidth at least 5× the highest frequency you intend to measure. Most embedded work involves signals under 50 MHz (UART, SPI, I2C, slow PWM). For that, 100 MHz bandwidth is plenty. For USB high-speed, gigabit Ethernet, or fast SerDes, you need 500 MHz to 2+ GHz, which is well above hobby budgets.

Sample rate

How many samples per second. Generally 4–5× the bandwidth (so a 100 MHz scope has 500 MS/s minimum sample rate, often 1 GS/s). Higher sample rate buys you more accurate edges and the ability to use FFT meaningfully.

Number of channels

Two channels is enough for most one-off measurements. Four channels matters when you debug signals that interact — clock, data, and chip-select on an SPI bus, for example.

Memory depth

How many samples the scope can store from one capture. Critical for capturing longer events at high sample rate. A 1 Mpoint scope at 1 GS/s can capture 1 ms of data; a 100 Mpoint scope can capture 100 ms.

This is the spec where cheap scopes cut corners. The memory depth determines whether you can see a full UART message at high resolution or just one byte of it.

Vertical resolution

Bits of ADC resolution. 8-bit is standard; 12-bit (Rigol's newer models, Siglent SDS2000X HD) is dramatically better for low-amplitude signals. For most digital work, 8-bit is fine; for power supply ripple analysis or analog work, 12-bit is a real upgrade.

Tier 1: $300–500

Rigol DS1054Z

The hobby scope that defined the category. 50 MHz, 4 channels, 1 GS/s, 12 Mpts memory. Around $350.

The selling point: a long-documented firmware mod unlocks the hardware to 100 MHz bandwidth and all the included options. Doubles the value with no risk.

Verdict: if your budget is under $400, this is the scope. Limitations: 8-bit ADC, slow refresh rate, no protocol decoders without options. None of these limitations matter for the first year of hobby work.

Siglent SDS1104X-E

Around $450. 100 MHz, 4 channels, 1 GS/s, 14 Mpts. Includes UART/I2C/SPI decoders. Faster refresh than the Rigol. The current standard pick if you can stretch the budget.

OWON SDS1102

Around $250. 100 MHz, 2 channels. The bargain pick. Functional for digital work; thinner feature set than the Rigol or Siglent.

Tier 2: $500–1500

Siglent SDS1202X-E

Around $500. 200 MHz, 2 channels, 1 GS/s, 14 Mpts. Significantly higher bandwidth than tier 1. The right pick if you anticipate working with USB full-speed (12 Mbps), fast SPI, or composite video.

Siglent SDS2104X Plus

Around $1500. 100 MHz upgradeable, 4 channels, 12-bit ADC, large display. The 12-bit ADC is the real upgrade: low-noise measurements that 8-bit scopes cannot make.

Rigol DHO914

Around $700. 125 MHz, 4 channels, 1.25 GS/s, 50 Mpts, 12-bit. Newer architecture; serious tier-2 contender.

Tier 3: used / bench

Used Tek and Keysight on eBay

$200–500. Tektronix TDS3000 series, Keysight DSO5000, Agilent 54622D. 200–500 MHz, often 4 channels, generally excellent build quality.

Risks: aging electrolytic caps (factor in $50–100 for recap if doing it carefully), no warranty, slower software than modern scopes. Reward: capabilities that cost $5000 new.

Skill required: read the listing carefully, check seller history, ask for known-good photos. Easy to be ripped off; equally easy to find genuine bargains.

PicoScope

$300–3000 USB-based scopes. No screen; runs on your laptop. The 5000 series at the high end has 16-bit modes that no traditional scope offers at the price. Software is excellent.

The trade-off: no standalone operation. If your laptop crashes, your measurement is gone. For benchtop work where the laptop is always there, PicoScope is a sleeper hit.

What about the ultra-cheap stuff?

Hantek and similar $100 USB scopes exist. Honest assessment: they work, the spec sheets are roughly accurate, and they are dramatically better than nothing. They are also dramatically worse than a $300 entry-level real scope. Skip them unless your budget is firmly under $200; if you can stretch to a Rigol DS1054Z, the experience is significantly better.

Probes matter as much as the scope

The 1× / 10× passive probes that come with most scopes are adequate but not great. The cheap ones bundled with $300 scopes are particularly prone to ringing on fast edges.

For more serious work:

• Genuine 10× probes ($30–100 per probe) — less ringing, better calibration. Worth it once you outgrow the basics.
• Active differential probes ($300–1500) — for floating measurements, mains voltages, power supply switching nodes. Specialised; only buy when you need.
• Current probes ($200–1500) — clamp around a wire to measure current without breaking the circuit. Niche but irreplaceable when you need it.
• Logic probes — many scopes accept a 16-channel logic probe pod (mixed-signal scope). Useful for digital protocol decode without a separate logic analyzer.

Features that sound important and are not

• FFT. Most scopes have it; almost no one uses it well. A dedicated spectrum analyzer is a separate tool for a reason.
• Built-in waveform generator. Useful occasionally; usually a separate signal generator is cheap and more capable.
• Touchscreen. Marketing feature on newer scopes. Knobs are faster for most operations.
• Network connectivity. Useful for remote control in production test; rarely used at the bench.

Frequently Asked Questions

Should I buy a 2-channel or 4-channel scope?

4-channel for the same price tier if available. Many real-world debugs need three signals visible simultaneously (clock, data, chip-select on SPI; address bus + data bus + control on parallel buses). 2-channel quickly becomes limiting.

Do I need 1 GHz bandwidth?Almost certainly not. Most embedded work fits comfortably in 100 MHz. 1 GHz scopes start at $5000 used and go up sharply.

Is a logic analyzer better than a mixed-signal scope?For pure digital protocol decode, yes — cheaper, more channels, better software. For mixed analog-digital debugging (slow analog signal triggered by digital event), the mixed-signal scope is the right tool. Most engineers eventually own both.

What is the upgrade path?Tier 1 scope for $350. Use it for 1–2 years; identify what bothers you (memory depth? bandwidth? channel count?). Buy the tier 2 scope that fixes those specific complaints. Tier 3 is rare territory for non-professional users.

What about open-source scopes like the Glasgow or Scopy?Glasgow Interface Explorer is more of a logic-analyzer-meets-protocol-tool. Pluto SDR can be used as a poor scope. ADALM2000 is a legitimate but limited educational scope. None are direct replacements for a benchtop scope; all have specific niches.