Summary of Arduino Auto-Ranging Ohmmeter: Divider Measurement and Accuracy Limits
This article explains building an Arduino-based ohmmeter using a voltage divider and ADC. It emphasizes that resistance is converted to voltage for measurement, with auto-ranging optimizing the ADC's usable range rather than improving accuracy. The guide warns against measuring live components and stresses calibration against known resistors due to errors from reference tolerance and VREF stability.
Parts used in the Arduino Ohmmeter:
- Arduino UNO R3 board
- ATmega328P microcontroller
- Analog-to-Digital Converter (ADC)
- Analog input pins A0 through A5
- Known reference resistors
- Unknown resistor under test
- Voltage reference (VREF)
- Switching mechanism for auto-ranging
- OLED display (referenced for companion voltmeter)
A divider-based Arduino ohmmeter does not measure resistance directly. It measures a voltage that depends on how an unknown resistor splits a known reference, then digitises that voltage with the board ADC. Auto-ranging only changes which reference resistor is switched in, so the reading lands in a usable part of the ADC range instead of collapsing toward an endpoint. Treat it as an educational measurement method for sorting and sanity-checking resistors, not as a calibrated bench instrument.
Why resistance has to become a voltage first
On the classic UNO R3, analogRead() uses a 10-bit ADC and A0 through A5 are usable analog inputs, per the official Arduino reference. The ATmega328P datasheet adds two points that shape any ohmmeter design: the conversion is single-ended and referenced to the selected VREF, and the ADC input discussion recommends a source impedance of roughly 10 kilohms or less while explaining sample-and-hold charging limitations. The ADC can only report where an input voltage sits relative to a reference, so to read resistance you must first convert it into a voltage, which is exactly what a divider does.
With a known resistor and an unknown resistor forming a divider, the voltage at the midpoint is a fraction of the reference set by the ratio between them, and the ADC digitises that fraction. When the two resistors are close in value the fraction sits near the middle of the range and small changes in the unknown produce a measurable change in the reading. When the unknown is much larger or much smaller than the known resistor, the fraction crowds one end of the range and the reading becomes insensitive.
What auto-ranging actually changes
A single fixed reference resistor behaves well only over a limited band. Auto-ranging switches among several reference resistors, choosing the one that keeps the divider voltage away from the endpoints. That is a resolution and usability decision, not an accuracy decision: accuracy is still governed by the tolerance of the reference resistors and the stability of VREF. Work out which ranges your own board can actually switch in, and measure each reference resistor instead of trusting its label.
Diagnostic table: reading behaviour and plausible causes
| Observation | Plausible cause | What to check |
|---|---|---|
| Reading stuck near zero | Unknown much smaller than the selected reference, so the divider sits near the low rail | Whether range selection has actually chosen a smaller reference resistor |
| Reading pinned near full scale | Unknown much larger than the selected reference | Whether the logic is selecting the next larger reference resistor |
| Digit noise and drift | High source impedance from large divider resistors, long leads or poor contact | Source impedance against the ~10 kilohm recommendation and lead length |
| Consistent offset versus a known resistor | Reference resistor tolerance or an uncalibrated scale and offset | Measured value of each reference resistor and the calibration constants |
| Range never changes | Selection logic or the switching itself is not operating | Which reference is physically in circuit at each step |
An ordered evaluation process
- Work only with disconnected, unpowered components. Never clip an ohmmeter across a live circuit, a charged capacitor or anything with stored energy.
- Confirm which reference resistors the design can actually switch in, and measure each one rather than trusting its label.
- Verify the ADC reference arrangement, because a 10-bit result is only meaningful against a known VREF.
- Check that the divider presents a source impedance at or below the recommended figure; otherwise expect the sample-and-hold to lag.
- Calibrate against at least two known resistors spanning the range of interest, deriving a scale and offset rather than a single-point correction.
- Compare results with a trusted reference meter and record the spread, not just one reading.
- Decide whether the observed spread is acceptable for the job, since sorting parts usually tolerates far more error than a value you intend to publish.
For a companion look at how a divider and an ADC reference interact in a voltage-measurement context, see the OLED voltmeter guide, and for a related resistance-measurement discussion see the milliohm meter article.
Where the numbers stop being trustworthy
The number of digits on a display says nothing about accuracy. In this method the dominant error terms are reference-resistor tolerance, VREF stability and endpoint behaviour of the ADC. It is also worth remembering that absolute maximum ratings in a datasheet are stress limits, not guaranteed operating conditions, so they cannot be used to justify driving the input harder.
Reader questions
Does auto-ranging make a cheap ohmmeter accurate?
No. Auto-ranging improves use of the ADC range so readings stay sensitive. Accuracy still depends on the reference resistors, VREF and calibration.
Can I measure a resistor still soldered into a circuit?
Not reliably, and not safely. The method assumes a disconnected, unpowered component; parallel paths and stored energy both invalidate it.
Why does the same resistor read differently on two ranges?
Usually a mix of reference tolerance, endpoint insensitivity and source impedance. Calibrating each range separately is the honest fix.
Is 10-bit resolution enough?
It sets the finest step size relative to VREF. Whether that is enough depends on the accuracy you actually need, not on the bit count alone.
Documentation: Arduino analogRead() reference (official reference-en source); Microchip ATmega328P datasheet, 7810D automotive edition.
Documentation only: the cited references cover ADC resolution, input mapping, conversion referencing and source-impedance guidance. No bench testing or calibration was performed, so treat readings as indicative until checked against a known reference.
- Does auto-ranging make a cheap ohmmeter accurate?
No, auto-ranging only improves the use of the ADC range so readings stay sensitive, while accuracy still depends on reference resistors and VREF. - Can I measure a resistor still soldered into a circuit?
No, you cannot measure reliably or safely because parallel paths and stored energy invalidate the method which assumes disconnected components. - Why does the same resistor read differently on two ranges?
This is usually caused by a mix of reference tolerance, endpoint insensitivity, and source impedance requiring separate calibration for each range. - Is 10-bit resolution enough for this project?
It sets the finest step size relative to VREF, but whether it is enough depends on the specific accuracy you need rather than the bit count alone. - What happens when the unknown resistor is much larger than the known resistor?
The fraction crowds one end of the range making the reading insensitive and potentially pinned near full scale. - How should I handle the source impedance for best results?
You must ensure the divider presents a source impedance at or below roughly 10 kilohms to prevent sample-and-hold lag. - Why do I need to calibrate against at least two known resistors?
Calibrating against two points allows you to derive a scale and offset rather than relying on a single-point correction. - What determines the dominant error terms in this measurement method?
The dominant errors come from reference-resistor tolerance, VREF stability, and the endpoint behavior of the ADC.
