Loading…
TO-92 · DS18B20
Active1-Wire Digital Temperature Sensor
DS18B20
Maxim Integrated · TO-92 · 1-Wire Digital Temperature Sensor
Price on Request
Get a competitive quote within 24 hours. Volume discounts available. Shenzhen stock ships in 1-2 business days.
Overview
The DS18B20 is Maxim's (now Analog Devices) programmable-resolution 1-Wire digital thermometer and the de facto standard for distributed temperature sensing. The block diagram is deliberately simple: a temperature sensor, an 8-bit CRC generator, 64-bit laser-trimmed ROM, nonvolatile alarm registers and scratchpad memory, all connected to a single 1-Wire port. Because every device carries a globally unique 64-bit serial code — 8-bit family code, 48-bit serial number and 8-bit CRC — a practically unlimited number of DS18B20 devices can hang on one data line plus ground, each addressed individually by the bus master.
Temperature measurement covers -55 °C to +125 °C (-67 °F to +257 °F) with ±0.5 °C accuracy from -10 °C to +85 °C. Resolution is user selectable at 9, 10, 11 or 12 bits, giving increments of 0.5 °C, 0.25 °C, 0.125 °C and 0.0625 °C and maximum conversion times of 93.75 ms, 187.5 ms, 375 ms and 750 ms respectively — the power-up default is 12 bits. Conversion is initiated with the Convert T [44h] command and the result lands in the 2-byte temperature register as a 16-bit sign-extended two's complement value. The power-on reset value of that register is +85 °C, which is why an unconverted first read always looks like a hot sensor. Nonvolatile TH and TL registers provide an alarm window, and the Alarm Search command lets the master poll only devices that are out of limits — a very efficient way to monitor hundreds of points.
Power can be supplied conventionally on VDD at 3.0 V to 5.5 V, or drawn from the bus in parasite-power mode: the DQ pull-up charges an internal capacitor (CPP) while the bus is high, and that capacitor powers the die while the bus is low. Parasite mode needs a strong pull-up — a MOSFET rather than a passive resistor — during the whole conversion and any EEPROM write (tWR = 10 ms), because the device sinks up to 4.0 mA at 0.4 V. Maxim also warns against parasite power above +100 °C, where leakage currents can prevent communication. A master that does not know how its sensors are powered can issue Skip ROM [CCh] followed by Read Power Supply [B4h]: parasite-powered devices pull the bus low during the read slot. Packages are 3-pin TO-92 (top mark 18B20), 8-pin SO and 8-pin µSOP, and the part is software compatible with the DS1822.
Counterfeit exposure on this part is extreme — the DS18B20 is arguably the most cloned sensor in the industry, and the clones are good enough to answer a 1-Wire reset pulse. Documented failure modes on fakes include incorrect behaviour in parasite-power mode, wrong resolution-bit handling, missing or non-unique ROM codes, degraded accuracy at temperature extremes and different conversion timing. Verification is straightforward if you know what to test: read the 64-bit ROM and confirm the family code is 0x28 and that codes are unique across the lot; force 9-bit mode and confirm the conversion completes in about 93.75 ms rather than 750 ms; check that the low-order bits are masked exactly as the selected resolution requires; exercise parasite power with a strong pull-up; and sweep a sample from 0 °C to 100 °C against a reference. Physically, genuine parts show crisp laser marking, correct package mass and a date code traceable to Maxim/ADI. We stock ADI-channel DS18B20+ in TO-92 and DS18B20Z in SO-8; MOQ is 2,000 pcs on tape-and-reel for the TO-92 (Maxim's standard T&R quantity) or 50 pcs loose, with 1-3 day dispatch.
Key Features
- -55 °C to +125 °C with ±0.5 °C accuracy from -10 °C to +85 °C
- User-selectable 9- to 12-bit resolution (0.5 °C down to 0.0625 °C)
- Unique 64-bit serial code allows multidrop networks on one data line
- Runs from 3.0 V to 5.5 V on VDD or on parasitic power from the bus
- Nonvolatile TH/TL alarm registers plus an Alarm Search command
Typical Applications
- HVAC environmental control and building temperature zoning
- Equipment, machinery and process temperature monitoring
- Cold-chain, brewery, aquarium and terrarium logging
- 3D printer hotend and heated-bed sensing with long cable runs
- Consumer thermostats, thermometers and thermal safety cutouts
Specifications
- Temperature Range
- -55 to 125 °C
- Temperature Accuracy
- ±0.5 (-10 to 85 °C) °C
- Resolution
- 9 / 10 / 11 / 12-bit (0.5 to 0.0625 °C)
- Conversion Time
- 93.75 (9-bit) to 750 (12-bit) ms
- Interface
- 1-Wire, Unique 64-Bit Serial Code
- Supply Voltage
- 3.0 to 5.5 V
- Powering Modes
- External VDD or Parasite Power from DQ
- Alarm Function
- Nonvolatile TH / TL Trigger Registers
- Pins
- 3
- RoHS
- Compliant
Procurement Tips
Buy DS18B20 only with an electrical verification report attached — a ROM family-code check plus a 9-bit conversion-timing test catches the overwhelming majority of clones for a few minutes of bench time per lot.
Frequently Asked Questions
What is the difference between the DS18B20, DS18S20 and DS1822?
All three use the same 1-Wire protocol and the same 64-bit ROM scheme but with different family codes: 0x28 for DS18B20, 0x10 for DS18S20/DS1820 and 0x22 for DS1822. The DS18B20 offers 9- to 12-bit programmable resolution over -55 °C to +125 °C at ±0.5 °C. The DS18S20 is fixed at 9 bits with 0.5 °C increments but achieves ±0.5 °C across a narrower range and has a different temperature register format. The DS1822 is a cost-reduced version with ±2 °C accuracy and lower resolution options. The DS18B20 is software compatible with the DS1822, which makes it the usual upgrade path.
How many sensors can I put on one bus and how do I address them?
The datasheet describes the number of slaves as virtually unlimited; in practice you are limited by bus capacitance, timing and enumeration effort. Standard wiring is one 4.7 kΩ pull-up from DQ to VDD with all sensors in parallel. Each device's 64-bit ROM is read with a Search ROM [F0h] algorithm, after which you address individuals with Match ROM [55h] plus the code, or all of them at once with Skip ROM [CCh]. Long runs benefit from a lower pull-up value (2.2 kΩ to 3.3 kΩ), shielded cable and a reduced conversion rate.
How do I wire parasite power correctly?
Tie VDD to ground, connect DQ through a pull-up to a 3.0 V to 5.5 V rail, and provide a strong pull-up — typically a MOSFET switched on by the master, or a resistor of a few hundred ohms — during the entire Convert T window (up to 750 ms at 12 bits) and during any 10 ms EEPROM write. The device sinks up to 4.0 mA while holding the bus at 0.4 V, so a passive 4.7 kΩ pull-up will not work. Issue Read Power Supply [B4h] to discover which devices on the bus are parasite powered. Do not use parasite power above +100 °C.
Counterfeit DS18B20 parts are a known problem — how do I screen them?
Bench-test, do not trust the marking. Read the ROM and check that the family code is 0x28 and that every device in the lot has a unique code. Switch to 9-bit resolution and confirm the conversion completes in about 93.75 ms and that the low three bits of the temperature register are masked off. Exercise parasite power with a strong pull-up — many clones fail here. Sweep a sample from 0 °C to 100 °C against a calibrated reference and confirm ±0.5 °C between -10 °C and +85 °C. Physically check laser marking, package mass and date-code traceability. We run this screen on every lot we ship and can supply the test records.
Why is my first reading +85 °C, and how does resolution affect timing?
+85 °C is the power-on reset value of the temperature register, so a read before any Convert T command returns it. After that, resolution and conversion time trade off directly: 9 bits gives 0.5 °C steps in 93.75 ms, 10 bits 0.25 °C in 187.5 ms, 11 bits 0.125 °C in 375 ms and 12 bits 0.0625 °C in 750 ms. For a multipoint network, convert all sensors in parallel with a single Skip ROM broadcast, wait the full 750 ms, then read each scratchpad individually — that keeps the total cycle time at about 0.75 s regardless of sensor count.
What cable and probe options work for long or wet installations?
The TO-92 with three wires handles tens of metres of shielded cable; for longer runs use a lower pull-up, add a series resistor at the sensor and keep the bus away from mains and motor wiring. Stainless-steel potted probes built on the DS18B20 are widely available for immersion, soil and pipe-clamp use — confirm the probe is hermetically sealed, because a leaking potting compound lets water reach the die and produces slow, drifting readings. For submersion and outdoor duty, buy probes with a documented IP rating rather than potting the TO-92 yourself.
What is the MOQ and lead time?
MOQ is 2,000 pcs on tape-and-reel for DS18B20+ (TO-92) or DS18B20Z (SO-8), which is Maxim/ADI's standard reel quantity, or 50 pcs loose for prototyping and repair. Shenzhen stock ships in 1-3 business days; the ADI authorized channel is typically 8-16 weeks. Given the counterfeit volume in this part, ask for the lot code, channel documentation and our verification report with every shipment.
Alternatives & Equivalents
View full cross-reference table →Need DS18B20 pricing and availability?
Submit an RFQ — we reply within 24 hours with stock, pricing, lead time and packaging options.
Request a Quote →