Fullscreen HTML5Med Res (??MB)HTML4

Measure high temperatures with this: Thermocouple adaptor for DMMs How many times have you wondered how hot an object is? It might be the heatsink in your latest project, the inlet or exhaust manifold in your car or anything else that’s hot or cold. Now you need wonder no more with this thermocouple adaptor for digital multimeters. By RICK WALTERS This Thermocouple Adaptor for DMMs can use any of several readily available type K thermocouple probes. The probe is plugged into this adaptor which plugs directly into your digital multimeter. Any digital multimeter will be suitable whether it has a 3.5-digit (1999), a 4-digit or a 4.5-digit display. 32  Silicon Chip In essence, this Thermocouple Adaptor is a temperature to voltage converter. Its output is 0V at 0°C and this increases (or decreases for negative temperatures) at the rate of 10mV/°C. This means that the temperature can be read directly in degrees C on a digital multimeter that’s set to an appropriate DC voltage range. All Fig.1: the basic scheme for a thermocouple. It consists of two dissimilar metal wires joined together to form a measuring junction. The open end of the wires then becomes the reference junction. you have to remember is to divide the reading in millivolts by 10. The thermocouple probe you choose will depend on the tem­ perature you want to measure and how much you want to pay. You can pick up a low cost bare-wire thermo- Fig.2: the complete circuit for the Thermocouple Adapter. The ambient temperature is sensed by REF1 and this produces a compensating voltage which is added to the thermocouple’s output. This output is amplified by IC1 which then drives the meter (DMM). ZD1 provides the reference voltage for pin 2 of IC1, while VR2 is used for calibration. Table 1 couple which will cover the range from -40°C to 250°C or you can go for a more expensive high-temperature (type K) probe capable of measuring from -50°C to 600°C. “What’s a thermocouple?” you may ask. Basically, a thermo­couple consists of two wires which are of dissimilar metals (in this case Chromel and Alumel). The wires are connected at one end, which becomes the measuring junction while at the other end the wires are connected to a reference junction. Confused? Fig.1 shows the basic scheme. The measuring junc­tion is placed on the object whose temperature we want to meas­ure. We then use a meter circuit to measure the voltage developed across the reference junction which is normally at ambient tem­perature (ie, room temperature). This voltage will be proportion­al to the temperature difference between the measuring junction and the reference junction. This voltage effect is known as the Seebeck coefficient and is about 40.6µV/°C for a type K thermocouple. Note that the change in output voltage per °C is only approximately linear over a small temperature range (see Table 1). As you can see from Table 1, we are dealing with very small voltages here. This means that we need a high-gain circuit and we must take precautions to ensure that no spurious voltages are introduced into it. The adaptor described here covers the temperature range from -50°C to around +600°C with a reasonable accuracy of a few degrees at the extremes. Reference junction temperature In the laboratory, a reference junction can be held con­stant at 0°C using an ice bath but that’s not practical for a portable instrument. Instead, in this circuit, the reference junction floats at the ambient temperature. This means that we need to have some way of compensating for ambient temperature variations in order to obtain accurate readings. The way around this problem is to use another temperature sensor to generate a voltage that’s proportional to the ambient temperature. This compensating voltage is then added to the thermocouple output and this effectively nulls out any effect from ambient temperature changes. If the am- Chromel & Alumel: What Are They? We’ve mentioned that a type-K thermocouple uses wires of Chromel and Alumel but what are they? You might guess that they are alloys and you’d be right. Chromel is an alloy of chromium and nickel which is commonly used in heating elements, while Alumel is an alloy of aluminium, manganese, silicon and nickel. Temperature °C Thermocouple Output (mV) -50 -1.889 -25 -0.968 0 0 25 1.00 50 2.022 100 4.095 200 8.137 300 12.207 400 16.395 500 20.640 600 24.902 bient temperature goes up, so does the compensating voltage and vice versa. In other words, for a given input temperature at the meas­uring junction, the output voltage from the Thermocouple Adaptor remains constant, regardless of the ambient temperature. Circuit details Let’s now refer to Fig.2 for the full circuit details. The ambient temperature is sensed by REF1, an LM335Z solid-state temperature sensor. This device generates an output voltage of 10mV per K(elvin). Because 273.12K is equivalent to 0°C, its output will be (nominally) 2.7312V at 0°C and will vary by 10mV for each Celsius degree rise or fall. This voltage change is reduced to 40.6µV/C° (ie, the same as the Seebeck coefficient for the K-type thermocouple) by feed­ing the LM335Z’s output into a voltage divider. This divider consists of the 100kΩ, 390Ω and 12Ω resistors and its output is connected December 1998  33 nal, while the meter’s negative terminal is connected directly to a +1.25V voltage reference (ZD1). Therefore, the meter will only read zero when the op amp’s output is at +1.25V. The reason for tying the negative side of the meter to +1.25V is to allow temperatures below 0°C to be measured. If the meter had been tied to 0V (GND), it would be unable to read down to even 0°C, since the OP07 cannot swing all the way down 0V. For temperatures below zero, the thermocouple voltage goes negative and pin 3 of IC1 swings below 1.25V. As a result, the reading on the meter (your DMM) will be negative – which is what we want. Now what about that offset voltage on pin 2 of IC1? This is set by trimpot VR2 which forms part of a voltage divider network across ZD1 (the 1.25V reference). In practice, VR2 is used to adjust the offset voltage at pin 2 of IC1 so that pin 6 sits at 1.25V at 0°C or 1.45V at 20°C. The meter will then show the temperature directly, provided that the gain of IC1 is set to 246.3 (100mV/40.6µV). This gain is set by the 82kΩ, 15kΩ, 390Ω and 12Ω negative feedback resistors. The 3% tolerance on ZD1 won’t worry us, as we compensate for this when we set VR2. The 0.22µF capacitor across the feedback resistors rolls off the gain of IC1 above 7.5Hz. This is done to prevent any hum signals picked up by the thermocouple leads from overloading the circuit. Power for the circuit is derived from separate 9V and 1.5V batteries. The 9V battery powers most of the circuitry, including the positive supply rail to IC1. The 1.5V battery is included solely to provide the required negative supply rail to the op amp (the op amp won’t work without Fig.3: install the parts on the PC board and install the wiring as shown here. The external battery test points are optional – just leave them out if you don’t want them. in series with the negative lead of the thermocouple. As a result, the thermocouple’s output is automatically compensated for ambient temperature variations. We still have one small problem though. As stated, the LM335Z has an output of 2.71312V at 0°C, which means that the output from the voltage divider sits at 11.73mV when the ambient temperature is 0°C. This 11.73mV offset voltage appears on pin 3 of op amp stage IC1 and needs to be cancelled out so that the multimeter reads 0V when the probe is measuring 0°C. One way of doing this would be to feed an equal offset voltage into the inverting input (pin 2) of IC1. In practice, we actually do feed in an offset voltage but it’s a bit more compli­cated than that, as we shall see. Take another look at the circuit. As shown, the op amp’s output (pin 6) connects to the meter’s positive termi- Table 2: Resistor Colour Codes  No.   1   1   1   1   2   1   2   2 34  Silicon Chip Value 100kΩ 82kΩ 39kΩ 10kΩ 4.7kΩ 3.9kΩ 390Ω 12Ω 4-Band Code (1%) brown black yellow brown grey red orange brown orange white orange brown brown black orange brown yellow violet red brown orange white red brown orange white brown brown black red black brown 5-Band Code (1%) brown black black orange brown grey red black red brown orange white black red brown brown black black red brown yellow violet black brown brown orange white black brown brown orange white black black brown black red black gold brown Parts List 1 PC board, code 04111981, 56 x 47mm 1 thermocouple probe 1 plastic case, 83 x 54 x 28mm, Jaycar HB-6015 or equivalent 1 DPST switch (S1) 1 9V battery 1 battery clip to suit 1 1.5V AA cell 1 AA cell holder, Jaycar PH-9203 or equivalent 1 10kΩ multi-turn cermet trimpot (VR1) 1 2kΩ multi-turn cermet trimpot (VR2) 2 banana plugs 2 solder lugs to suit above 4 PC stakes 3 2.5mm x 6mm countersunk head bolts 3 2.5mm nuts 3 solder lugs to suit above 1 M3 x 6mm countersunk screw 1 3mm nut Semiconductors 1 OP07CN op amp, Farnell Cat. 690-624 (IC1) 1 LM335Z temperature sensor (REF1) 1 ZR423 1.25V reference diode, Farnell Cat. 703-412 (ZD1) Capacitors 2 10µF 16VW PC electrolytic 1 0.22µF MKT polyester The PC board assembly fits neatly into a small standard plastic case. Note the method of mounting the 0.22µF capacitor near the top of the board. a negative supply rail). Double-pole switch S1 switches the power on and off. Finally, the circuit includes provision to test the batter­ies under load without opening the case. This is done by connect­ing the wipers of switch S1 and the 0V rail to three 2.5mm bolts on the side of the case. When the power is switched on, you can easily check the V+ and V- voltages (with respect to GND) using a multimeter. Construction All the parts except for the switch, the meter plugs and the 9V battery are mounted on a small PC board. This is coded 04111981 and measures 56 x 47mm. Before installing any of the parts, check the board care­fully for etching defects by comparing it with the published pattern (Fig.4). It’s rare to find any problems but it doesn’t hurt to make sure. Fig.3 shows the parts layout on the PC board. Begin by installing PC stakes at all the external wiring points, then install the resistors. Check each value on your multimeter as you proceed (Table 2 shows the colour codes). Once these are in, the semiconductors and the trimpots can be installed. Make sure that the semiconductors are Resistors (0.25W, 1%) 1 100kΩ 2 4.7kΩ 1 82kΩ 1 3.9kΩ 1 39kΩ 2 390Ω 1 15kΩ 2 12Ω 1 10kΩ all oriented correctly and take care to ensure that the trimpots aren’t mixed up. VR1 has a value of 10kΩ while VR2 has a value of 2kΩ. Op amp IC1 should be directly soldered to the PC board. Do not use an IC socket for this device. The reason for this is that it’s best to minimise the number of dissimilar metal junctions, as each junction is, in theory, another thermocouple. The PC board assembly can now be completed by installing the capacitors and the battery snap connector. Note that all the capacitors must be December 1998  35 The thermocouple at left (DSE Cat. Q1439) is a simple wire type which covers the range from -40°C to 250°C. If you want to measure higher temperatures (up to 600°C), you will need a probe type thermocouple such as the one shown at right (Jaycar Cat. QM1282) – see panel. Note that you will have to cut the plug off your thermocouple, so that it can be directly wired to the PC board. mounted with their bodies flat against the PC board, as shown in Fig.3. This is done to provide clearance for the 9V battery. The 1.5V battery holder should be secured using a 3mm coun­ tersunk screw and nut. You will have to drill a hole through the centre of the holder and the PC board to fit this. Once the assembly is complete, cut the screw off level with the nut so that the battery can be fitted. Drilling the case Before you start drilling the plastic case, remove the flutes along both the long sides using a sharp chisel. This is necessary to get the PC board to fit. Next, drill holes in the plastic box for the thermocouple lead, the two banana plugs, the switch and the three 2.5mm screws for the battery test terminals. Don’t forget to fit a small solder lugs under each nut of the test terminals. Note that the banana plugs must be accurately spaced so that they can be plugged directly into the terminals of your DMM. The standard spacing is 3/4-inch (19mm). Mount the two banana plugs on the end of the plastic box and fit a large solder lug under each nut. This done, make the connections between the PC board and the lugs using tinned copper wire. The short lead to the negative banana plug can be left bare, while the longer lead to the positive plug should be sleeved with spaghetti tubing to prevent shorts. Next install the battery switch and connect it, following the wiring diagram of Fig.3. If you want the battery test fea­ture, run two leads from the switch to the positive and negative battery test terminals, plus a lead from the PC board to the earth terminal. We couldn’t find a socket to match the thermocouple’s plug, so it was removed and the leads soldered directly to the PC stakes. Now before you cut off the plug note that it is polarised and you will see “+” and “-” signs moulded into the plug housing. When you unscrew the plug you will find that it has red and yellow wires. The red wire is positive and should connect to PC stake close to pin 3 of IC1 while the yellow wire connects to the other PC stake. The thermocouple and meter positive stakes will have to be trimmed, to allow the battery to sit low enough for the lid to fit properly. Finally, complete the wiring by fitting the battery snap connector and running the leads to the battery holder. If you don’t like shoehorning all this into the plastic box we have specified, use a larger box. Dick Smith Electronics has a box (DSE Cat. No H-2874) It’s a tight squeeze when the 9V battery is installed but it all fits. The meter plugs must be spaced so that the unit can be plugged directly into a digital multimeter. 36  Silicon Chip OFF +9V 0V ON SILICON CHIP -1.5V TYPE K THERMOCOUPLE INTERFACE METER 2V which is 40mm high instead of 28mm and will give you lots more room (but at greater cost). Calibration This is the easiest part of the whole project. First, set your DMM to the 10V range and connect it across the two outer terminals of VR1. The best way to do this is to connect the positive meter lead to the end of the 100kΩ resistor that’s adjacent to REF1 and the negative lead to a convenient ground point. Now apply power and allow five minutes for the circuit to stabilise. This done, place an accurate thermo­ meter on REF1, allow it to stabilise and adjust VR1 until the meter reads 2.7312V + (temperature/100). For example, if the temperature is 23°, you adjust VR1 for a reading of 2.7312 + 0.23 = 2.9612V. Of course, if you have a 3.5-digit multimeter, the best you can do is a reading of 2.961V or 2.962V; the resolution that you can attain depends on the number of digits on your multimeter’s display. Now connect your DMM to the METER + and - terminals (ie, to the meter plugs), set it to the 2V range and adjust VR2 for a reading of 0.230V (230mV). This corresponds to a reading of 23.0° which is the same as the reading on the thermometer. And that’s all there is to it; the calibration procedure is complete. Fig.4: the full-size artworks for the front panel and PC board. Interpreting readings If you are using a 3.5-digit meter, the 2V range will cover temperatures from -50°C to 199°C. This should include most of the everyday temperatures you will want to measure. The 20V range will need to be selected to cover temperatures from 200-600°C. Choosing A Thermocouple As mentioned in the main body of the text, this project uses a type-K thermocouple. There are several units that are readily available and these are sold by Dick Smith Electronics (DSE) and by Jaycar. These are as follows: DSE Cat. Q1438: -50°C to 1200°C ($99.95) – probe type DSE Cat. Q1439: -40°C to 250°C ($19.95) – wire type Jaycar Cat. QM1282: -40°C to 750°C ($14.95) – probe type Note: this adaptor can only measure to just above 600°C due to circuit limitations. As stated earlier, you must convert the reading on the DMM to millivolts and then divide by 10 to get the temperature in °C. For example: (1) the meter reading is 4.73V. In this case, 4.73V = 4730mV and so the temperature is 4730/10 = 473°C. (2) the meter reading is 0.673V. This is equivalent to 673mV and so the temperature is 67.3°C. Finally, if you plan to use the adaptor to measure tempera­tures within a specific range (eg, 100-250°C), greater accuracy can be achieved by calibrating the unit at the mean temperature within this range (175°C for the example given). This involves subjecting the probe to this mean temperature and then adjusting VR2 to obtain the SC correct meter reading. December 1998  37