A laboratory data logger built around a microcontroller shows three separate problems

PHYSICS · Multi-part · Fermi University Physics · Electronics, Instrumentation & Communication · Digital electronics, embedded systems and data acquisition · Digital electronics, embedded systems and data acquisition

Electronics for Measurement, Ch The microcontroller as laboratory hardware

Problem

A laboratory data logger built around a microcontroller shows three separate problems, each explained by a different member of the team. ((a)) The photodiode channel is sampled by a software loop intended to run at

5\,\mathrm{kHz}

(

200\,\mu\mathrm{s}

nominal interval). Over

10\,000

samples the measured mean interval is exactly

200.0\,\mu\mathrm{s}

, yet the digitised waveform shows small residual scatter about the expected sine. One team member says: "the mean interval is exactly right, so the sampling must be uniform and the scatter must come from somewhere else." Say what is wrong with this claim. ((b)) The same channel is also serviced by a timer interrupt that starts an SD-card write, with a latency of

4\,\mu\mathrm{s}

, a critical section of

60\,\mu\mathrm{s}

while the buffer is written, and one competing interrupt that can delay it by

6\,\mu\mathrm{s}

. The record shows occasional gaps of up to

70\,\mu\mathrm{s}

timed to the SD-card writes, but the designer calculated an "average" response time of

10\,\mu\mathrm{s}

(latency plus a typical, not worst-case, critical section) and concluded this was safe against the

200\,\mu\mathrm{s}

sampling interval. Say what is wrong with this reasoning, give the correct worst-case figure, and say whether it still fits within the

200\,\mu\mathrm{s}

interval. ((c)) A humidity sensor is read over a

3.3\,\mathrm{V}

I2C bus with a

2.2\,\mathrm{k}\Omega

pull-up, clocked at

400\,\mathrm{kHz}

(maximum allowed 30 to 70 per cent rise time

0.30\,\mu\mathrm{s}

). With a short cable the bus capacitance is

100\,\mathrm{pF}

and readings are reliable. Extending the cable raises the bus capacitance to

350\,\mathrm{pF}

and the readings become unreliable. A team member says: "the bus rate is set in software, so the cable length cannot be the cause." Say what is wrong with this claim, and support it with the rise time

0.85RC

for both cable lengths.

Browse the problem library