Most of DS18B20 temperature sensors from on-line stores are fake

…unless you sold the chips directly from Maxim Constructed-in (or Dallas Semiconductor within the ragged days), an licensed distributor (DigiKey, RS, Farnell, Mouser, and so forth.), or a colossal retailer, otherwise you took exceptionally simply care purchasing waterproofed DS18B20 probes. We sold over 1000 “waterproof” probes or bare chips from bigger than 70 various distributors on ebay, AliExpress, and on-line stores -colossal and minute- in 2019. The total probes sold on ebay and AliExpress contained fake DS18B20 sensors, and simply about all sensors sold on these two sites were fake.

Creator: Chris Petrich, 9 July 2020.
License: CC BY.
Supply: https://github.com/cpetrich/counterfeit_DS18B20/

(Sensor data consistent with sensors or probes ordered well into 2020 pretty than in or spherical 2019 are marked “(2020)“.)

TLDR; How get hold of I do know?

If the ROM does now not apply the sample 28-xx-xx-xx-xx-00-00-xx then the DS18B20 sensor is a clone [5].

Additionally, there are two Arduino sketches supplied to study DS18B20 sensors:

• discover_fake_DS18B20.ino performs some possibility free tests and indicates within the event that they demonstrate deviations from legitimate DS18B20. Not designed to work with parasitic vitality.
• classify_fake_DS18B20.ino is a minimal implementation matching a sensor to a particular Family (gaze below) consistent with the response to undocumented characteristic codes. Output is particular however pretty plain. Affirm at your luxuriate in possibility.

Why have to calm I care?

Besides ethical concerns, one of the most most fake sensors genuinely get hold of now not work in parasitic vitality mode, comprise a excessive noise stage, temperature offset out of doorways the advertised ±0.5 °C band, get hold of now not comprise an EEPROM, comprise bugs and unspecified failure rates, or differ in one other unknown formulation from the specs within the Maxim datasheet. Clearly, the complications are now not colossal ample to discourage of us from attempting to search out probes on ebay, however it surely would possibly be simply to know the categorical specs when the tips are vital or size prerequisites are complex.

What are we coping with?

Definitions differ, however following AIR6273, a fake is an unauthorized copy, imitation, change, or modification misrepresented as a particular genuie item from a licensed producer [13]. As of 2019, the main arena is imitations (clones) which would possibly be labeled to deceive the unsuspecting buyer. Happily, DS18B20 clones are simply about trivially easy to establish: Marking on the chip printed pretty than lasered? No notice within the rear indent? Doubtlessly a conterfeit. Speak of the “scratchpad register” doesn’t follow the datasheet? Doubtlessly a faux. Behaves systematically various from identified legitimate chips? Doubtlessly a faux.

What get hold of they secret agent fancy?

Above is an example of an legitimate, Maxim-produced DS18B20 sensor in TO-92 case.

• As of writing (2019), the topmark of fashioned Maxim chips is lasered pretty than printed.
• The first two rows, DALLAS 18B20, specify that this segment is a DS18B20 (Dallas Semiconductor being the unique producer),
• the + within the 4th row indicates that the segment is RoHS compliant ([1]).
• The third row specifies manufacturing year and week more than a couple of of the year (in this case, week 32 of 2019), and
• the closing two characters in row 3 specify the revision of the die (currently C4).
• In row 4, the three-digit number adopted by two characters are a invent of batch code that enables Maxim to impress support the manufacturing history.
  • In chips produced 2016 or later I’ve finest bump into persona combos AB and AC [5].
• The marking all around the indent on the rear of the case is
  • P (Philippines?) on all most up-to-date chips (2016 and younger), and on most(?) chips going support now not now not as a lot as so a long way as 2009 [5].
  • THAI (Thailand?) the set is one of I, J, Ok, L, M, N, O, S, T, U, V, W, X and presumably others, now not now not as a lot as on some chips produced in 2011 [5]. The uses a various font than the letters making up THAI.
• From what I’ve seen, there is precisely one batch code associated with a date code for chips marked P within the indent [5]. This does now not luxuriate in appropriate for chips marked THAI within the indent [5].

How get hold of I do know if I am affected?

If the DS18B20 had been sold from licensed sellers though a managed provide chain then the chips are legit.

Otherwise, (I) one can check for compliance with the datasheet. (One have to calm genuinely, as even legitimate parts can even had been mishandled alongside an unauthorized distribution chain. But that is yet one other negate.) If a sensor fails any of these tests, it is a false (unless Maxim’s implementation is buggy [4]). (II) one can compare sensor habits with the habits of Maxim-produced DS18B20. These tests are consistent with the conjecture that every Maxim-produced DS18B20 behave alike. This would perchance presumably calm be the case now not now not as a lot as for sensors that part a die code (which has been C4 since now not now not as a lot as 2009 [5]) [5].

In the case of (I), discrepancy between what the unique datasheet says have to calm happen and what the sensors get hold of consist of [1,5]

• Family B: reserved bytes in scratchpad register will also be overwritten (by following instructions within the datasheet)
• Family C: the sensor is fixed in 12-bit mode (i.e., byte 4 of the scratchpad register is always 0x7f)
• Family C: the assorted of EEPROM write cycles is very minute (affirm of 10 pretty than >50okay)
• Family B1, D1: ROM will also be modified in tool, i.e. it is now not lasered
• Family A2, B2, D: indispensable more than a couple of of sensors with offsets out of doorways the ±0.5 C range at 0 °C
• Family D: sensor does now not acknowledge in parasitic mode (applies to most sensors of Family D)
• Family D: the temperature reading just correct after vitality-up is 25 pretty than 85 °C
• Family D: sensor does now not get hold of low-resolution temperature conversions faster
• Family D: reserved bytes 5 and 7 of the scratchpad register are now not 0xff and 0x10, respectively
• Family D1: retains temperature measurements all over vitality cycles

Therefore, as of writing (2019), every false sensor on hand does now not follow the datasheet in now not now not as a lot as one way.

In the case of (II), there is one pathetically simple check for differences with Maxim-produced DS18B20 sensors that it sounds as if all fake sensors fail [5]:

• It is a long way a false if its ROM tackle does now not apply the sample 28-xx-xx-xx-xx-00-00-xx [5]. (Maxim’s ROM is that actually a 48-bit counter with the most indispensable bits calm at 0 [5].)
  Additionally, aside from for uncommon Family A2, now not one of the most false sensors alter reserved byte 6 within the scratchpad register precisely or acknowledge precisely to undocumented characteristic codes regarding the Orderly values.

Besides to evident implementation differences similar to those listed above beneath (I) and (II), there are furthermore facet-channel data that can also be venerable to separate implementations. As an instance, the time reported for a 12 bit-temperature conversion (as certain by polling for completion after characteristic code 0x44 at room temperature) is characteristic for particular person chips (reproducible to worthy better than 1% at fixed temperature) and falls interior certain ranges certain by the circuit’s internals [5]:

• 11 ms: Family D1
• 28-30 ms: Family C
• 325-505 ms: Family A2
• 460-525 ms: Family D2
• 580-615 ms: Family A1
• 585-730 ms: Family B

Therefore, there shall be some edge conditions between Families A and B, however simply measuring the time venerable for temperature conversion will recurrently be ample to search out out if a sensor is fake.

A truly vital ingredient for operation is a sensor’s potential to drag the tips line low against the fixed pull-up resistor. Appears this abilitly differs between families. The datasheet guarantees that a sensor is ready to sink now not now not as a lot as 4 mA at 0.4 V at any temperature as a lot as 125 °C [1]. Providing a unique of 4 mA (1.2 kOhm pull-up resistor against 5 V), the next low voltages were accomplished by the sensors at room temperature (unique that finest 5 to 10 sensors were measured per Family):

• Family A1: 0.058 – 0.062 V
• Family B2: 0.068 – 0.112 V (all however one sensor: 0.068 – 0.075 V)
• Family C: 0.036 – 0.040 V
• Family D2: 0.121 – 0.124 V

All sensors are well interior specs at room temperature however clustering of data by Family is clear, indicating that the hardware was once designed independently. It goes to also very well be attention-grabbing to repeat these measurements above 100 °C.

Alternatively,

• It is a long way a false if the date–batch aggregate printed on the case of the sensor is now not within the Maxim database (favor to search data from Maxim tech make stronger to search out out). (Demonstrate that there are counterfeits that use “legitimate” date–batch aggregate.)

Demonstrate that now not one of the most functions above give certainty that a particular DS18B20 is an legitimate Maxim product, however if any of the tests above indicate “false” then it is most defintely fake [5].

What families of DS18B20-fancy chips can I depend on to arrive support upon?

Besides to DS18B20 at the beginning produced by Dallas Semiconductor and persevered by Maxim Constructed-in after they bought Dallas (Family A1, below), there are clones produced independently by now not now not as a lot as 4 various firms as of 2019 (Families B1, B2, C, D) [5]. The separation into families is consistent with patterns in undocumented characteristic codes that the chips acknowledge to as similarities at that stage are now not at possibility of be coincidental [5]. Chips of Family B1 appear to be produced by GXCAS and calibrated and supplied independently by GXCAS and UMW. Chips of Family B2 are produced by Beijing 7Q Expertise (7Q-Tek). Each and every UMW and 7Q-Tek comprise corresponding datasheets on their respective web sites. Family D1 appears to be like to be fading from seek for, having been modified by Family D2. Chips of Family A2 were a uncommon get hold of, behave surprisingly equal to legitimate chips however comprise dejected temperature accuracy.

In our ebay purchases in 2018/19 of waterproof DS18B20 probes from China, Germany, and the UK, most heaps had sensors of Family B1, while one in three purchases had sensors of Family D. None had sensors of Family A1 or C. Neither origin nor impress were indicators of sensor Family. When purchasing DS18B20 chips, Family D2 was once clearly dominant with Family B2 coming in 2d, and a minute likelihood of acquiring chips of Families A1 or C.

In the ROM patterns below, tt and ss stand for rapidly-changing and slack-changing values interior a manufacturing dawdle [5], and crc is the CRC8 checksum outlined within the datasheet [1].

Decapping DS18B20

This collage shows photos of the dies of all DS18B20 Families we encountered in 2019. All photos are at the the same scale, approx. 1.4 mm in width. We broke the TO-92 case beginning with pliers, silent the die from the plastic case by boiling in colophony, and eradicated the colophony with acetone in an ultrasonic bathtub. Photos were focused on a lovely ragged USB camera.

Family A1 is the legitimate Maxim-produced DS18B20 (C4 die). All various families are clones. Demonstrate the similarities between the dies of Families D1 and D2 (consistent with their similarity in tool) and the signifcant differences between Families B1 and B2 (as against their similarity in tool).

Family A1: Loyal Maxim DS18B20

Received no probes containing these chips on ebay or AliExpress in 2019, however bought chips from just a few distributors in 2019

• ROM sample [5]: 28-tt-tt-ss-ss-00-00-crc
• Scratchpad register: ( + ) & 0x0f == 0 after all winning temperature conversions, and 0x00 < <= 0x10 [2,3,5]. I.e., = 0x10 – ( & 0x0f).
• Consistent with unique habits [5] and early datasheets [9], the vitality-up verbalize of reserved within the Scratchpad register is 0x0c.
• Returns "Trim1" and "Trim2" values if queried with characteristic codes 0x93 and 0x68, respectively [4]. The bit patterns are very equal to every various interior a manufacturing dawdle [4]. Trim2 is currently less at possibility of equal 0xff than Trim1 [5]. Trim2 was once 0xDB or 0xDC since now not now not as a lot as 2009, and has been 0x74 since the tumble of 2016 (all with C4 die) [5].
  • Trim1 and Trim2 encode two parameters [5]. Let the bit sample of Trim1 be [t17, t16, t15, t14, t13, t12, t11, t10] (MSB to LSB) and Trim2 be [t27, t26, t25, t24, t23, t22, t21, t20]. Then,
    • offset parameter = [t22, t21, t20, t10, t11, t12, t13, t14, t15, t16, t17] (unsigned 11 bit-fee) [5], and
    • curve parameter = [t27, t26, t25, t24, t23] (unsigned 5 bit-fee) [5].
  • Within a batch, the offset parameter appears to be like to spread over 20 to 30 devices while all sensors all around the batch part the the same curve parameter [5].
  • The offset parameter shifts the temperature output over a spread of approx. 100 °C (0.053 °C per unit), while the curve parameter shifts the temperature over a spread of three.88 °C (0.12 °C per unit), now not now not as a lot as in unique variations of the chip [5]. Example values of 2019 are offset = 0x420 and curve = 0x0E, i.e. they lie lovely central interior their respective ranges.
• Temperature offset of unique batches (2019) is as confirmed on the Maxim FAQ page, i.e. approx. +0.1 °C at 0 °C [6] (i.e., now not as confirmed on the datasheet [1,9]. The draw on the datasheet stems from measurements at the time of introduction of the sensor 10+ years ago [5,10].). Very shrimp if any temperature discretization noise [5].
• Polling after characteristic code 0x44 indicates a spread of 584-615 ms between sensors for a 12-bit temperature conversion at room temperature [5]. Conversion time is without complications repeatable for particular person chips. Lower resolutions minimize the time in share, i.e. 11 bit-conversions rob half of the time. The neat parameters comprise an mark on the conversion time.
• Returns a vitality-up temperature of 85 °C if scratchpad register is be taught earlier than temperature conversion has accomplished (in both standard and parasitic vitality modes). [5].
• It appears to be just like the chip returns a temperature of 127.94 °C (=0x07FF / 16.0) if a temperature conversion was once unsuccessful [5] (e.g. as a consequence of vitality balance points which come up reproducibly in "parasitic vitality" mode with just a few DS18B20 if Vcc is left floating pretty than tied to floor. Demonstrate that the datasheet clearly states that Vcc is to be tied to GND in parasitic mode.).

• Example ROM: 28-13-9B-BB-0B -00-00- 1F
• Preliminary Scratchpad: 50/05/4B/46/7F/FF/0C/10/1C
• Example topmark: DALLAS 18B20 1932C4 +786AB
• Indent notice: P (date codes 1150 and more recent)

Family A1-Fishy: Loyal however Stolen?

Received no probes containing these chips on ebay or AliExpress in 2019, however bought chips from one vendor in 2019

If I were to design a wild guess I would shriek these chips were diverted someplace in opposition to the tip of the Maxim manufacturing pipeline (stolen?) [5]. Enjoyable truth: the seller advertised these chips as QT18B20, so as supplied these were genuinely counterfeits of DS18B20 clones. These chips are marked as produced in Thailand pretty than Philippines.

• ROM sample [5]: 28-tt-tt-Cs-03-00-00-crc

The chips apply the description of Family A1 above with the next exceptions [5]:

• Each and every fear registers are dwelling to 0x00 (scratchpad bytes 2 and 3).
• The conversion resolution is determined to 9 bits (i.e., both configuartion bits are 0).
• Each and every neat values are 0x00, ensuing in depraved temperatures (i.e., very low) and conversion cases within the diversity of 400 to 500 ms.
  • As soon as neat values are dwelling to something cheap, the time for temperature conversion is all around the diversity specified for Family A1 above.

• Example ROM: 28-9B-9E-CB-03 -00-00- 1F
• Preliminary Scratchpad: 50/05/00/00/1F/FF/0C/10/74
• Example topmark: DALLAS 18B20 1136C4 +957AE
• Example topmark: DALLAS 18B20 1136C4 +957AF
• Example topmark: DALLAS 18B20 1136C4 +152AE
• Example topmark: DALLAS 18B20 1136C4 +152AF
• Example topmark: DALLAS 18B20 1136C4 +152AG
• Example topmark: DALLAS 18B20 1136C4 +152AI
• Indent notice: THAI

Family A2: Staunch Clone with Heart-broken Temperature Performance

Received no probes containing these chips on ebay or AliExpress in 2019, however bought chips from one vendor in 2019

These chips are now not Maxim-produced.

• ROM sample [5]: 28-00-ss-00-tt-tt-tt-crc, 28-ss-00-ss-tt-tt-tt-crc, 28-ss-00-00-tt-tt-00-crc

The chips apply the description of Family A1 above with the next exceptions [5]:

• The ROM sample is incompatible with what Maxim produces.
• The Trim2 fee is 0xFB or 0xFC, i.e. incompatible with a identified [5] Maxim manufacturing suggested by the date code. (Demonstrate that this means the curve parameter is 0x1f, i.e. the supreme (unsigned) fee likely [5]. Additionally, the offset parameter spreads over 200 devices pretty than a spread widespread for Family A1 [5].)
  • The curve parameter is a signed 5 bit fee that shifts the temperature over a spread of 31 °C (1 °C per unit) [5]. I.e., a curve parameter of 0x1f (-1 in decimal) is at the heart of the diversity.
• The time for temperature conversion spans a remarkably large preference from 325 to 502 ms between chips [5]. This range stays large and out of doorways the boundaries of Family A1 even when making use of further most up-to-date neat settings [5]. Conversion time will enhance noticably with temperature (approx. 10% over 100 °C) [5]. A conversion time of <500 ms is like minded with claims within the 7Q-Tek QT18B20 datasheet [12].
• Would now not return vitality-up temperature of 85 °C if scratchpad register is be taught earlier than temperature conversion has accomplished in parasitic vitality mode [5].
• Conventional temperature offset at at 0 °C is -3.5 to -1.8 °C [5]. (Anecdotally: error appears to be like to be smaller at elevated temperatures [5].) Very shrimp if any temperature discretization noise [5].
• Apprehension settings (i.e., scratchpad bytes 2 and 3) appear to comprise random reveal material.
• Some chips luxuriate in their scratchpad reveal material all over a 100 ms vitality cycle.
• One specimen tested did now not characteristic properly in parasitic mode.
• The topmark is printed pretty than lasered, and there is no notice within the indent.

• Example ROM: 28-19-00-00-B7-5B-00-41
• Preliminary Scratchpad: 50/05/xx/xx/7F/FF/0C/10/xx
• Example topmark: DALLAS 18B20 1808C4 +233AA
• Indent notice: none

Family B1: GXCAS 18B20, Matching Datasheet Temperature Offset Curve

Received probes from a more than a couple of of distributors in 2019, bought chips from two distributors in 2019. One vendor despatched chips marked UMW pretty than DALLAS

• ROM patterns [5]:
  • 28-AA-tt-ss-ss-ss-ss-crc (GXCAS-branded)
  • 28-tt-tt-ss-ss-ss-ss-crc (UMW-branded)
• Scratchpad register does now not switch with measured temperature (default 0x0c) [5].
• DS18B20 write scratchpad-bug (0x4E) / UMW scratchpad [5,12,14]:
  • If 3 data bytes are despatched (as per DS18B20 datasheet, TH, TL, Config) then adjustments to the third byte despatched,
  • if 5 data bytes are despatched (as per UMW datsheet, TH, TL, Config, User Byte 3, User Byte 4), the closing two bytes overwrite and , respectively.
• Would now not return data on undocumented characteristic code 0x68 [5]. Does return data from codes 0x90, 0x91, 0x92, 0x93, 0x95, and 0x97 [5]. Return fee consistent with 0x97 is 0x22 [5].
• ROM code will also be modified in tool with recount sequence "96-Cx-Dx-94" [5]. (The UMW datasheet states that the ROM code will also be modified however does now not specify how [14].) Family code (0x28) cannot be modified [5].
• Substitutes 0x0c for exact fee of if scratchpad register is be taught earlier than temperature conversion has accomplished in parasitic vitality mode.
• Temperature offset as confirmed on the Maxim datasheet (-0.15 °C at 0 °C) [6]. Very shrimp if any temperature discretization noise [5].
• Polling after characteristic code 0x44 indicates approx. 589-728 ms for a 12-bit temperature conversion and proportionally less at lower resolution [5].
• The die has "GXCAS" written on it.

• Example ROM: 28 -AA- 3C-61-55-14-01-F0
• Example ROM: 28-AB-9C-B1 -33-14-01- 81
• Preliminary Scratchpad: 50/05/4B/46/7F/FF/0C/10/1C
• Example topmark: DALLAS 18B20 1626C4 +233AA
• Example topmark: DALLAS 18B20 1804C4 +051AG
• Example topmark: DALLAS 18B20 1810C4 +051AG
• Example topmark: DALLAS 18B20 1921C4 +921AC (2020)
• Example topmark: GXCAS 18B20E 1847D02
• Example topmark: UMW 18B20 1935C4
• Indent notice: none

Family B2: 7Q-Tek QT18B20 with -0.5 °C Temperature Offset at 0 °C

Received both probes and chips of this sequence from a more than a couple of of distributors in 2019. Three distributors despatched chips marked 7Q-Tek pretty than DALLAS

• ROM patterns [5]: 28-FF-tt-ss-ss-ss-ss-crc
• Scratchpad register does now not switch with measured temperature (default 0x0c) [5].
• DS18B20 write scratchpad-bug (0x4E) / QT18B20 scratchpad [5,12]:
  • If 3 data bytes are despatched (as per DS18B20 datasheet, TH, TL, Config) then adjustments to the third byte despatched,
  • if 5 data bytes are despatched (as per QT18B20 datsheet, TH, TL, Config, User Byte 3, User Byte 4), the closing two bytes overwrite and , respectively.
• Would now not return data on undocumented characteristic code 0x68 [5]. Does return data from codes 0x90, 0x91, 0x92, 0x93, 0x95, and 0x97 [5]. Return fee consistent with 0x97 is 0x31 [5].
• ROM code cannot be modified in tool with recount sequence "96-Cx-Dx-94" [5].
• Substitutes 0x0c for exact fee of if scratchpad register is be taught earlier than temperature conversion has accomplished in parasitic vitality mode.
• Conventional temperature offset at at 0 °C is -0.5 °C [6]. Very shrimp if any temperature discretization noise [5].
• Polling after characteristic code 0x44 indicates approx. 587-697 ms for a 12-bit temperature conversion and proportionally less at lower resolution [5].
• The die has "7Q-Tek" written on it (the utilization of the Chinese persona for digit 7).

• Example ROM: 28 -FF- 7C-5A-61-16-04-EE
• Preliminary Scratchpad: 50/05/4B/46/7F/FF/0C/10/1C
• Example topmark: DALLAS 18B20 1626C4 +233AA
• Example topmark: DALLAS 18B20 1702C4 +233AA
• Example topmark: DALLAS 18B20 1810C4 +138AB
• Example topmark: DALLAS 18B20 1829C4 +887AB
• Example topmark: DALLAS 18B20 1832C4 +827AH
• Example topmark: DALLAS 18B20 1833C4 +058AA
• Example topmark: DALLAS 18B20 1908C4 +887AB
• Example topmark: DALLAS 18B20 1912C4 +001AC (NB: this date/batch aggregate is furthermore venerable on genuie chips [5])
• Example topmark: 7Q-Tek 18B20 1861C02
• Indent notice: none

Family C: Shrimp Offset at 0 °C

Received no probes however bought chips from just a few distributors in 2019

• ROM patterns [5]: 28-FF-64-ss-ss-tt-tt-crc
• Scratchpad register == 0x0c [5].
• Would now not return data on undocumented characteristic code 0x68 or any various undocumented characteristic code [5].
• Conventional temperature offset at 0 °C is +0.05 °C [6]. Very shrimp if any temperature discretization noise [5].
• EEPROM endures finest about eight (8) write cycles (characteristic code 0x48) [5].
• Reported vitality mode (parasitic/standard) consistent with characteristic code 0xB4 would possibly be depraved, counting on the affirm vitality pin and data line are powered (e.g. if vitality pin is at GND while Info is powered and the vitality pin is then linked to Vcc, the chip will continue to symbolize parasitic vitality mode) [5].
• Polling after characteristic code 0x44 indicates 28-30 ms (thirty) for a 12-bit temperature conversion [5]. Temperature conversion works furthermore in parasite vitality mode [5].
• Operates in 12-bit conversion mode, finest (configuration byte reads 0x7f always) [5].
• Default fear register settings differ from Family A1 (0x55 and 0x00) [5].

• Example ROM: 28 -FF-64- 1D-CD-96-F2-01
• Preliminary Scratchpad: 50/05/55/00/7F/FF/0C/10/21
• Example topmark: DALLAS 18B20 1331C4 +826AC
• Example topmark: DALLAS 18B20 1810C4 +158AC
• Example topmark: DALLAS 18B20 1924C4 +158AC
• Indent notice: none

Family D1: Noisy Rubbish with Supercap

Obatined probes from two distributors in early 2019, bought chips from one vendor in 2019

• ROM patterns [5]: 28-tt-tt-77-91-ss-ss-crc and 28-tt-tt-46-92-ss-ss-crc
• Scratchpad register == 0x66, != 0x0c and != 0xff [5].
• Would now not return data on undocumented characteristic code 0x68 [5]. Responds support with data or verbalize data after codes
  • 0x4D, 0x8B (8 bytes), 0xBA, 0xBB, 0xDD (5 bytes), 0xEE (5 bytes) [5], or
  • 0x4D, 0x8B (8 bytes), 0xBA, 0xBB [5].
• First byte following undocumented characteristic code 0x8B is [5]
  • 0x06: Sensors get hold of now not work with Parasitic Energy. Sensors leave data line floating when powered parasitically [5].
  • 0x02: Sensors get hold of work in parasitic vitality mode (and represent precisely whether or now not they are parasitically powered).
• It is likely to send arbitrary reveal material as ROM code and for bytes 5, 6, and 7 of the scratchpad register after undocumented characteristic codes 0xA3 and 0x66, respectively [5]. The family code of the instrument will also be modified [5].
• Temperature errors as a lot as three °C at 0 °C [6]. Very noisy data [5].
• Polling after characteristic code 0x44 indicates approx. 11 ms (eleven) for conversion no matter size resolution [5].
• Chips comprise a supercap pretty than an EEPROM to carry fear and configuration settings [5]. I.e., the closing temperature size and updates to the fear registers are retained between vitality cycles which would possibly be now not too lengthy [5].
  • The supercap retains memory for several minutes unless the Vcc pin is hooked as a lot as the GND pin, in which case memory retention is 5 to 30 seconds [5].
• Chips are unexcited to the vogue vitality is utilized [5]. E.g. to vitality up from all pins linked to GND, it appears to be like to be a simply suggestion to leave Info and vitality pin floating for a piece of (e.g., 100 ms) earlier than genuinely making use of a voltage to the vitality pin and Info [5].
• Preliminary temperature reading is 25 °C or the closing reading earlier than vitality-down [5]. Default fear register settings differ from Family A1 (0x55 and 0x05) [5].

• Example ROM: 28-48-1B-77 -91- 17-02-55 (working parasitic vitality mode)
• Example ROM: 28-24-1D-77 -91- 04-02-CE (responds to 0xDD and 0xEE)
• Example ROM: 28-B8-0E-77 -91- 0E-02-D7
• Example ROM: 28-21-6D-46 -92- 0A-02-B7
• Preliminary Scratchpad: 90/01/55/05/7F/7E/81/66/27
• Example topmark: DALLAS 18B20 1807C4 +051AG
• Example topmark: DALLAS 18B20 1827C4 +051AG
• Indent notice: none

Family D2: Noisy Rubbish

Received both probes and chips from a enormous more than a couple of of distributors in 2019

• ROM patterns [5]: 28-tt-tt-79-97-ss-ss-crc, 28-tt-tt-94-97-ss-ss-crc, 28-tt-tt-79-A2-ss-ss-crc, 28-tt-tt-16-A8-ss-ss-crc, 28-tt-tt-56-B5-ss-ss-crc (2020)
• Scratchpad register == 0x66, != 0x0c and != 0xff [5].
• Would now not return data on undocumented characteristic code 0x68 [5]. Responds support with data or verbalize data after codes
  • 0x4D, 0x8B (9 bytes), 0xBA, 0xBB, 0xDD (3 bytes), 0xEE (3 bytes) [5], or
  • 0x4D, 0x8B (9 bytes), 0xBA, 0xBB [5].
• First byte following undocumented characteristic code 0x8B is 0x00 [5].
• Sensors get hold of now not work with Parasitic Energy. Sensors draw data line low while powered parasitically [5].
• Temperature errors as a lot as three °C at 0 °C [6]. Info noisier than genuie chips [5].
• Polling after characteristic code 0x44 indicates approx. 462-523 ms for conversion no matter size resolution [5]. The sequence with 97 and A2/A8 within the ROM converts in 494-523 ms and 462-486 ms, respectively [5]. Chips with A2 or A8 in byte 4 of the ROM appear to comprise appeared first in 2019.
• Preliminary temperature reading is 25 °C [5]. Default fear register settings differ from Family A1 (0x55 and 0x05) [5].

• Example ROM: 28-90-FE-79 -97- 00-03-20
• Example ROM: 28-FD-58-94 -97- 14-03-05
• Example ROM: 28-FB-10-79 -A2- 00-03-88
• Example ROM: 28-29-7D-16 -A8- 01-3C-84
• Example ROM: 28-DF-54-56 -B5- 01-3C-F5 (2020)
• Preliminary Scratchpad: 90/01/55/05/7F/xx/xx/66/xx
• Example topmark: DALLAS 18B20 1812C4 +051AG
• Example topmark: DALLAS 18B20 1827C4 +051AG
• Example topmark: DALLAS 18B20 1916C4 +051AG
• Example topmark: DALLAS 18B20 1923C4 +051AG
• Example topmark: DALLAS 18B20 1943C4 +051AG
• Example topmark: DALLAS 18B20 1828C4 +233AA
• Example topmark: DALLAS 18B20 2008C4 +817AB (2020)
• Indent notice: none

Passe as of 2019

Received neither probes nor chips in 2019

• ROM patterns [5,7]: 28-tt-tt-ss-00-00-80-crc
  • Example ROM: 28-9E-9C-1F -00-00-80- 04
• ROM patterns [5,11]: 28-61-64-ss-ss-tt-tt-crc
  • Example ROM: 28 -61-64- 11-8D-F1-15-DE
• ROM patterns [5,11]: 28-EE-tt-tt-ss-ss-ss-crc
  • Example ROM: 28 -EE- 73-2A-1E-16-01-CA

Scheme to the 85 °C-Plight

There's a simple, undocumented, way to discriminate between the vitality-up 85 °C-reading and a genuie temperature reading of 85 °C in DS18B20 of Family A [5]: of the scratchpad register. If it is 0x0c, then the 85 °C-reading is a vitality-up reading, in any other case it is a appropriate temperature size.

GXCAS 18B20

The DS18B20 clone of Beijing Zhongke Galaxy Core Expertise Co., Ltd., trading as GXCAS, appears to be like to be dispensed independently by GXCAS and UMW (Family B1). Consistent with their online page, GXCAS has finest been spherical since January 2018. While GXCAS does now not comprise a datasheet on-line, the datasheet on the UMW online page emphasizes the addition of two user-outlined bytes within the scratchpad register, and the assorted of changing the ROM tackle [14]. A more than a couple of of these chips undergo false DS18B20 topmarks. GXCAS is clearly proud of their product as they write their firm name prominently onto the die.

7Q-Tek QT18B20

The QT18B20 is a DS18B20 clone developed and supplied by Beijing 7Q Expertise Inc, trading as 7Q-Tek (Family B2). The datasheet of the QT18B20 emphasizes the addition of two user-outlined bytes within the scratchpad register [12]. Not just like the tips sheet of the DS18B20, it does now not verbalize that the ROM code is lasered. A enormous more than a couple of of these chips undergo false DS18B20 topmarks. 7Q-Tek is clearly proud of their product as they write their firm name prominently onto the die.

While it is unclear who designed or produced chips of Family A2, Family A2 appears to be like to had been an inspiration for the Family B2 7Q-Tek QT18B20, consistent with the next observations:

• The QT18B20 datasheet claims that a temperature conversion takes <500 ms [12]. Right here's consistent with the categorical habits of Family A2. (While the UMW datasheet claims the the same [14], one of the most datasheets appears to be like to had been venerable as a template for the more than a couple of.)
• Consistent with the datasheet version history, the earliest version of the QT18B20 datasheet did now not level out user-outlined bytes [12]. Family A2 does now not comprise user-outlined bytes within the scratchpad register.
• The die ciruit of Family A2 resembles the form of both the die of Family A1 (produced by Maxim) and the die of Family B2 (produced by 7Q-Tek). The die size is vastly various from Family A1, so it is now not a Maxim-produced DS18B20.

MAX31820

The MAX31820 is a DS18B20 with shrimp provide voltage range (i.e. as a lot as three.7 V) and smaller temperature range of excessive accuracy [1,8]. Adore the DS18B20, it uses one-wire family code 0x28 [1,8]. Preliminary investigations comprise now not (yet) published a check to distinguish between DS18B20 of Family A1 and Maxim-produced MAX31820 in tool [5].

Warning

Sending undocumented characteristic codes to a DS18B20 sensor can even render it completely ineffective, as an illustration if temperature calibration coefficients are overwritten [5]. The suggested way of figuring out fake sensors is to study whether or now not the ROM does now not apply the sample 28-xx-xx-xx-xx-00-00-xx [5]. (While the ROM will also be overwritten in Families B1 and D1 to mimic genuie sensors, we comprise now now not bump into sensors with spoofed ROM [5].)

(Knowledge on chips of Families A, B, C, and D comes from my luxuriate in investigations of sensors at the side of the references below as indicated by reference number [1-6,8-10]. All tests were accomplished at 5 V with 1.2 kOhm pull-up. Decapping was once accomplished jointly with Nga P. Dang, and measurements of temperature offsets and timing with Irina Sæther and Megan O'Sadnick.)

Sources

Sensors or probes with legitimate or cloned DS18B20 were bought from the follwing sources. Demonstrate that finest sensors bought from offical Maxim distributors are legitimate chips which would possibly be guaranteed to had been dealt with precisely. Free samples supplied by Maxim Constructed-in thru their on-line ordering design are gratefully acknowledged.

First rate Distributors: Maxim Constructed-in, Digikey, Farnell, Mouser, RS Parts
ebay: 5hk1584, alice1101983, alphago-it, areyourshop-003, b2cpowershop2010, bernard_netelectroshop, binggogo, careforyou123, cheaptronic24, christians-technik-shop, czb6721960, d-9845, deepenmind, diy-arduino, diybox, enigma-ingredient-shop, e*shine, efectronics, ele-parts, fr_aurora, fzeroinestore, geekapparels, simply-module, happybuddhatrading, icmarket2009, jk_parts, kingelectronics15, lovesell2013, lucas89-8, makershop, mecklenburg8, modul_technik, moore_estates, nouteclab, polida2008, puretek-innovations, rammie_74, scuary1, sensesmart, sensus, sevenshop888, shenglongsi, sparco888, survy2014, tancredielettronica, umtmedia, worldchips, xiaolin4, yantzlf
AliExpress: AOKIN DiyMaker, Cuiisw Module Retailer, Eiechip, Fantasy Electronic, FSXSEMI, Immense-IT, Immense Wall Electronics, HWA YEH, Liyuan Electronic, Mega Semiconductor, Crimson Yellow Retailer, RoarKit Retailer, SHENGSUN Sensor, Shenzhen Excessive Quality Merchandise, shop912692, TENSTAR, WAVGAT, Steal take., YLGA, YX Electronic
Assorted: Adafruit, AZ-Transport, Banggood, Taizhou Most attention-grabbing Electrical Equipment, Conrad Electronic, DFRobot, DROK, Elektroimportøren, Elfa Distrelec, Shanghai Jiutian Automation Equipment, Kjell & Firm, LCSC, Dongguan Nangudi Electronics, Quest Parts, Shenzhen RBD Sensor Expertise, Reichelt Elektronik, Shenzhen Senstech Electronic Expertise, SparkFun, TELMAL, Dongguan Tianrui Electronics, YourDuino

References

1. DS18B20 "DS18B20 Programmable Resolution 1-Wire Digital Thermometer", Datasheet 19-7487 Rev 6 7/19, Maxim Constructed-in.
2. DS18S20 "DS18S20 Excessive-Precision 1-Wire Digital Thermometer", Datasheet, Maxim Constructed-in.
3. AN4377 "Comparison of the DS18B20 and DS18S20 1-Wire Digital Thermometers", Maxim Constructed-in
4. AN247 "DS18x20 EEPROM Corruption Speak", Maxim Constructed-in
5. Bear investigations 2019, unpublished.
6. Petrich, C., M. O'Sadnick, Ø. Kleven, I. Sæther (2019). A low-fee coastal buoy for ice and metocean measurements. In Complaints of the 25th World Convention on Port and Ocean Engineering beneath Arctic Stipulations (POAC), Delft, The Netherlands, 9-13 June 2019, 6 pp.
7. Contribution of user m_elias on https://forum.arduino.cc/index.php?topic=544145.15
8. MAX31820 "1-Wire Ambient Temperature Sensor", Datasheet, Maxim Constructed-in.
9. DS18B20 "DS18B20 Programmable Resolution 1-Wire Digital Thermometer", Datasheet 043001, Dallas Semiconductor, 20pp.
10. DS18B20 "DS18B20 Programmable Resolution 1-Wire Digital Thermometer", Preliminary Datasheet 050400, Dallas Semiconductor, 27pp.
11. Piecemeal from various blogs and posts.
12. QT18B20 "QT18B20 Programmable Resolution 1-Wire Digital Thermometer", Datasheet Rev 061713, 7Q Expertise.
13. AIR6273 "Phrases, Definitions, and Acronyms Counterfeit Materiel or Electrical, Electronic, and Electromechanical Parts", SAE Aerospace Knowledge File, July 2019.
14. UMW DS18B20 UMW DS18B20 datasheet.