The trip down memory lane continues with the second Dick Smith Electronics boxed kit that I purchased from the CCARC Wyong Field Day. You could say, I’m on a roll when it comes to these kits …
This posts covers a kit from the “delightfully” pink category of “Test Kit”. In fact, when it came to kits, building your own test equipment was a popular thing to do. Part of the reason was that branded test equipment for diagnosing electronics (e.g. thermocouple measurements, ESR measurements, transistor gain tests, microamp measurements, etc.) could be very expensive or difficult to obtain. As a result, being able to build something yourself was seen as a way to save some money and obtain some valuable capabilities.
This kit is the K7120 Thermocouple Adaptor for DMMs, from the December 1998 edition of Silicon Chip. This is a fairly recent kit by comparison with the last, especially with the reference to digital multimeters which would have been a luxury in the early 90s. Despite the passing of time, some things don’t change – the size of the box, the duo-tone labelling, the serial numbering and the Australian assembly of the kits.
The kit itself uses a K-type thermocouple which covers -40 to 250 degrees Celsius and is intended to plug into a digital multimeter to read out the temperature, in essence a thermocouple to voltage adapter. The kit is supplied with the components, hardware, PCB, case, front panel label and temperature probe.
When I was young, I do recall seeing this particular kit on the shelf and thinking to myself that it would be nifty to have one. Being of an inquisitive mind, I thought it to be interesting to be able to measure the temperature of things, especially transistors, as I had built a number of kits where I wasn’t certain if they had enough heatsinking. Another use would be to check the fridge, freezer and oven just out of curiosity.
Looking at the papery-stuff, there’s a label for the fascia of the case (no drilling through the label this time), the assembly manual, the “standard” Guide to Kit Construction, quality control card and disclaimer. It seems the market survey questionnaire was not included in this package.
Again, when it comes to components, a complete jiffy/zippy box package is supplied, along with a retail package thermocouple, which shows that whenever possible, they would just “shove” a whole retail package item into the kit. But for other components which are better supplied “counted” or “cut from tape”, these are all packaged into separate categories in exact numbers. A bit of wire and solder is also supplied, with the PCB.
The PCB is a bit more modern as well, featuring some text in the copper layer and the Silicon Chip logo. The pads are all optimised for soldering as well, with a fairly simple pattern that is concentrated on one side, as the rest of the PCB which remains blank is designed for mounting the batteries used to power the device. Just like the other board, it is a single-sided board with no silkscreening or solder resist. If I’m not mis-remembering, some of the later Discovery series PCBs did have silkscreening and solder resist.
It would seem that the supplied assembly manual is a “first edition” this time around, but it’s just the classic stapled double-sided black-and-white A4 laser print that normally accompanies the kits. As Silicon Chip is still going, I don’t dare reproduce more than just the header of the guide, but it is very comprehensive as usual, including some specific information about calibrating this kit.
Construction of the PCB is relatively straightforward, as the circuit is fairly simple with many resistors, capacitors, two trimpots, a reference and a 5V linear regulator.
The circuit is partially constructed above, ready to accept the thermocouple leads. As the kit doesn’t include thermocouple sockets in its design, it is intended that the thermocouple leads be soldered directly to the board (or pegs) – I chose the former as I wanted to minimise dissimilar metal junctions as they can impact on the accuracy of the device. It’s the same reason the IC is not socketed.
From there, there is a fair amount of wiring necessary to complete the kit, as this kit requires two batteries – one AA and one 9V, making this a rather annoying piece of equipment to keep powered. To be sure it operates correctly, there are separate test point connections, and to switch the power on and off, a dual-pole switch is used to switch each battery separately. It’s almost needlessly complex – nowadays we might use USB power and that would be fine for feeding directly into the 5V portion of the circuit. For the negative part, perhaps a charge-pump based voltage inverter would be employed or a virtual ground established on the 5V supply for the same effect.
Regardless, soldering of this board was relatively straightforward and as the board was a bit younger, it didn’t seem to have the same oxidation issues. The main annoyances came when it came to drilling out the case to wire up all of the necessary interfaces.
Holes in the case are used to let the meter interface and thermocouple wire out of the case. No strain-relief or captive cable grommets are supplied, instead, leads pass through holes in the case and thus could be yanked or otherwise nicked by the edges if they aren’t smoothed. This brings another complication – aside from the batteries being inside the case requiring disassembly to access, the 9V cell is secured by a releasable cable tie that proves very difficult to release. Not a great design in my opinion.
Another rather stiff challenge is to try to connect the wires to this type of soldered banana plug – it takes a lot of heat and careful alignment to do properly. I can foresee many difficulties, as it took me several attempts to get enough heat into them. But nonetheless, the kit was completed.
This time, there were no rubber feet for the bottom of the case, so I made sure to countersink the screws properly. Three screws hold the PCB inside the case.
Unfortunately, drilling for the test terminals didn’t proceed as smoothly as I had liked – using a hand-held cordless drill makes getting accurate hole positioning a little more difficult. It’s these subtle imperfections that add character and make it clear that this was something I (and my amateurish hands) have built.
Of course, securing these items to the sides of the case requires careful chiselling and filing away of the ribs which adds additional work.
When I rested the kit, it appeared to work just fine, so I tried to calibrate it. While there were clear instructions on how it should be done, it assumes access to a trusted temperature reference or measuring device. In essence, you’d be calibrating this with your meter to another meter that has some tolerance, so hence lies the crux of the problem. In many cases, these types of test equipment kits don’t come with any accuracy guidance as that really depends on how it was calibrated, how your meter accuracy is over the range of voltages, the temperature swings that the unit is exposed to, etc. Many thermometers and thermocouple read-outs are only about +/- 1 to 2 degrees Celsius “guaranteed” as there is tolerance in the thermocouple and in the meter, so it’s hard to be sure of accuracy of the unit as it probably drifts with temperature, as does the connected multimeter.
As a result, after getting it to within 0.1 degrees Celsius, I packed the kit away, as I couldn’t see it being accurate compared to a direct thermocouple input on multimeters which support them. This is just another layer of potential error – but it could be useful to have if you have a voltage-only datalogger.
The thermocouple adapter for DMM kit was constructed successfully, but reveals one of the biggest problems with building test equipment for yourself – namely that of accuracy and calibration. It may be good for use with voltage-only data-loggers as an indication, but without a proper thermal source to calibrate against and a highly accurate multimeter, any calibration is a stab in the dark. In fact, even if you did have those, you’d probably still come undone with the drift of the circuit as the room temperature changes – without any specifications, it would need to be characterised to provide assurance of performance.
The design leaves something to be desired too – with no thermocouple socket, it’s basically permanently attached to the supplied thermocouple. The use of both AA and 9V batteries makes powering the unit somewhat complicated especially as the releasable cable tie for the 9V cell is often reluctant to release. Like other kits, quite a bit of wiring is involved and drilling/filing of the case is necessary to fit the hardware.
By the end of all of this, while it can work, it’s probably just easier to use a modern multimeter that has a purpose-designed thermocouple input.