FAQ
TL;DR: For 0–50°C skin-difference sensing over 5 cm, a Pt1000 Wheatstone bridge gives ~1 mV/K; “use a Pt100/Pt1000 with small footprint and insulate.” [Elektroda, Enrico Koeck, post #21673186]
Why it matters: Clinicians and makers need stable, low-noise differential readings to spot small left–right skin temperature differences.
Quick-Facts
- Target use-case: two points ~5 cm apart on human skin, 0–50 °C difference range. [Elektroda, zeyad asaad, post #21673183]
- Bridge sensitivity: Approx. 1 mV/K with two Pt1000s in a Wheatstone bridge. [Elektroda, Enrico Koeck, post #21673186]
- Self‑heating: Typical ~0.5 K/mW for Pt1000; limit excitation or use a current source. [Elektroda, Enrico Koeck, post #21673188]
- Single-supply op-amp needs a mid-rail (V+/2) reference; zero reads ~V+/2, reducing small-signal headroom. [Elektroda, Enrico Koeck, post #21673191]
- Hundredths‑of‑kelvin resolution needs an instrumentation amplifier and high‑quality components. [Elektroda, Enrico Koeck, post #21673193]
Quick Facts
- Target use-case: two points ~5 cm apart on human skin, 0–50 °C difference range. [Elektroda, zeyad asaad, post #21673183]
- Bridge sensitivity: Approx. 1 mV/K with two Pt1000s in a Wheatstone bridge. [Elektroda, Enrico Koeck, post #21673186]
- Self‑heating: Typical ~0.5 K/mW for Pt1000; limit excitation or use a current source. [Elektroda, Enrico Koeck, post #21673188]
- Single-supply op-amp needs a mid-rail (V+/2) reference; zero reads ~V+/2, reducing small-signal headroom. [Elektroda, Enrico Koeck, post #21673191]
- Hundredths‑of‑kelvin resolution needs an instrumentation amplifier and high‑quality components. [Elektroda, Enrico Koeck, post #21673193]
How do I wire a 3‑pin LM335 for body measurements?
Use the LM335 per its datasheet pinout; the third pin trims output. The package, not a specific pin, contacts skin. However, for tiny skin temperature differences, LM335’s larger thermal mass is not ideal. Prefer small‑footprint RTDs (Pt100/Pt1000) placed directly on the skin and insulated from ambient. [Elektroda, Enrico Koeck, post #21673186]
What sensors work best for small left–right skin temperature differences?
Use two matched Pt100 or Pt1000 RTDs with very small footprints. Mount them directly on skin and thermally isolate them from air. This setup improves response time and reduces ambient drift when measuring subtle differential changes near room temperature. [Elektroda, Enrico Koeck, post #21673186]
How do I get a true differential output from two probes?
Wire the two RTDs into a Wheatstone bridge and read the bridge imbalance with a differential amplifier. Keep sensor cable lengths equal. Set gain with A = (R2 + R1)/R1, and polarity so output is positive when sensor 2 is warmer. Add an output low‑pass filter if needed. [Elektroda, Enrico Koeck, post #21673186]
Do I need to zero one probe every time?
For a narrow operating range, set the bridge zero with the trimmers once, or occasionally to counter part tolerances. Use equal‑length cables to minimize offset from wire resistance and its temperature coefficient. For automatic zeroing, add a microcontroller to supervise calibration. [Elektroda, Enrico Koeck, post #21673188]
How do I avoid cable resistance errors over 5 cm spacing?
Use identical cable types and keep lengths the same from the bridge to each sensor. Mismatched wire resistance and its temperature drift will create false imbalances that look like temperature differences. Good wiring symmetry reduces this error source. [Elektroda, Enrico Koeck, post #21673186]
What gain and filter values should I start with?
Set amplifier gain with A = (R2 + R1)/R1. Start modest, then increase while watching noise. Add a low‑pass at the output: fC = 1/(2πR3C1). For very slow changes, consider an active filter to avoid huge capacitors. The circuit also works without the filter. [Elektroda, Enrico Koeck, post #21673186]
Can I power the op‑amp from a single 9 V battery?
Yes, but create a mid‑rail at V+/2 as the reference for a differential stage. With zero temperature difference, output sits near 4.5 V, which makes tiny signals harder to resolve. Dual supplies avoid this offset issue but increase complexity. [Elektroda, Enrico Koeck, post #21673191]
Why are my PT100s reading hot?
Excess excitation current causes self‑heating. In one example, ~4.5 mA produced ~20 mW, warming the RTD by about 10 K, so it acted as a heater. Reduce sensor current and series resistance to keep power low, or readings will skew high. [Elektroda, Enrico Koeck, post #21673191]
How small a temperature difference can I realistically measure?
Hundredths of a kelvin are challenging with simple op‑amp stages. “You would need pretty high quality components and an instrumentation amplifier circuit.” Use precision parts, good PCB layout, and low‑noise amplification to reach that level. [Elektroda, Enrico Koeck, post #21673193]
What is a Wheatstone bridge, and why use it here?
A Wheatstone bridge is a four‑resistor network that converts small resistance differences into voltage. Using two RTDs in opposite arms yields a near‑zero common‑mode output and a signal proportional to temperature difference, about 1 mV/K for Pt1000. [Elektroda, Enrico Koeck, post #21673186]
What is an instrumentation amplifier in simple terms?
It’s a precision differential amplifier with high input impedance, tight gain accuracy, and strong common‑mode rejection. It resolves tiny differential signals—like hundredths of a kelvin—from noisy, drifting environments better than a basic op‑amp stage. [Elektroda, Enrico Koeck, post #21673193]
How do I minimize ambient air effects on the probes?
Use small sensors directly on the skin and insulate them from ambient air. Foam or tape insulators reduce airflow errors and improve stability when comparing adjacent body points. [Elektroda, Enrico Koeck, post #21673186]
How do I build a simple differential skin thermometer? (3 steps)
- Mount two small Pt1000s 5 cm apart on skin and insulate from air.
- Wire them into a Wheatstone bridge feeding a differential op‑amp; match cable lengths.
- Set zero with the pots, choose gain A = (R2+R1)/R1, and add a low‑pass filter. [Elektroda, Enrico Koeck, post #21673186]
Can I just buy a ready‑made differential skin thermometer?
Yes. Given the precision needed for hundredths‑of‑kelvin differences, consider commercial instruments instead of DIY. They integrate low‑noise amplification, filtering, and calibration workflows out of the box. [Elektroda, Enrico Koeck, post #21673193]
Is the LM335 suitable for tiny skin temperature differences?
Not ideal. Although usable, LM335’s larger thermal mass slows response for subtle skin gradients. A matched Pt100/Pt1000 pair with careful mounting and insulation performs better for small differential measurements near room temperature. [Elektroda, Enrico Koeck, post #21673186]
