Oscilloscope
Input Stage and Ranges
The chain
A signal entering the input BNC passes through, in order:
- Relay K4, which connects the input only when the oscilloscope is active. In generator-only modes the input is physically isolated.
- The programmable gain stage — CH1’s de-amplifier/amplifier potentiometer pack, redirected into the measurement path by relay K5.
- The converter node, where either the LM393 comparator (SAR path) or the ATmega’s internal ADC input digitises it.
There is no input attenuator on the board
This is the single most important fact about the input stage, because it means a 1× probe or a bare lead puts the full signal into a chain whose converter node saturates at ±2.5 V. A conventional bench scope’s 1 MΩ front end would divide that down for you. This one does not.
The calibration suite records the probe attenuation alongside every measurement for exactly this reason — the same reading means different things at 1× and 10×, and the instrument cannot detect which is attached.
Programmable gain
| Gain range | ×0.19 - ×3.03 |
| Combinations | 10,000 (two X9C wipers, 100 positions each) |
| Converter full scale | ±2.5 V at the converter node |
The gain stage is the same hardware that sets the generator’s output amplitude in CH1 Generator mode. That is the reason the oscilloscope cannot run at the same time as CH1’s synthesiser — see Why the Modes Exist.
Input range
The converter node saturates at ±2.5 V, so the usable input is that ceiling divided by the chain gain, then multiplied by whatever your probe divides by:
| Chain gain | At the BNC (1× probe) | Through a 10× probe |
|---|---|---|
| Minimum (×0.19) | ±13.2 V | ±132 V |
| Unity | ±2.5 V | ±25 V |
| Maximum (×3.03) | ±0.83 V | ±8.3 V |
Choosing a range
If you know the amplitude of your signal, pick the highest gain whose range still contains it. More gain means the signal occupies more of the converter’s ±2.5 V full scale, which directly improves the effective resolution of the reading.
If you do not know the amplitude, use Autoset. It searches from minimum gain upward, so the comparator input is never overdriven during the search. This ordering is deliberate: a search that started at maximum gain and worked down would slam the comparator on the first step every time it met a large signal.
If the trace is clipped flat, the signal exceeds the converter’s ±2.5 V node full scale at your current gain. Reduce gain, switch to a 10× probe, or reduce the signal.
If the trace is a thin band in the middle of the screen, you have gain in hand. Increase it — resolution is quantisation at the converter node divided by chain gain, so higher gain means finer effective steps referred to the input.
Resolution referred to the input
The two acquisition paths quantise differently at the converter node:
| Path | Step at converter node |
|---|---|
| 16-bit SAR | 76 µV |
| Internal 10-bit ADC | 4.88 mV |
Referred back to the input, those steps are divided by the chain gain and multiplied by the probe ratio. At unity chain gain the SAR’s step is 76 µV at the BNC, or approximately 0.76 mV at the tip of a 10× probe. Raising the gain reduces those numbers proportionally; lowering it increases them.
Protection
Protection here is isolation, not clamping:
- Relay isolation. K4 disconnects the input entirely whenever the scope is not active. A board sitting in generator mode is not exposed to whatever is attached to its input BNC.
- Your probe. A 10× probe is the only attenuation in the system, and on this instrument it is a functional part of the input stage rather than an optional accessory.
Updated · Full project report (PDF)