MAKING  THERMOCOUPLES FOR  AERIAL AMMETERS

Ray Robinson VK2NO

I was working on an WSNo.11(Aust), which is the Australian version of the Wireless Set Number 11, and attempting to load the transmitter into an aerial. There was very little aerial current. I tried various sized resistors, and finally tried a short circuit. It loaded beautifully into this, but as I was tuning up, the response was so quick, that the aerial ammeter hit full scale, and then flopped back to zero. I had burnt it out!



This type of meter is a plug in design, and is used in several radios, the WSNo.11, the WSNo.101, and the FS6. The meter comes in two values, both are the same physical size, and can be easily unplugged and a different one plugged in. The 1 Amp size is used in the WSNo.11 HP (High Power) and the FS6. The 350 milliamp size is used in the WSNo.11 LP (Low Power), the WSNo.101 and the Aerial Coupling Unit Type C (for the WSNo.11). The WSNo.101 is merely the Australian version of the WSNo.1, and the FS6 (Field Set 6) is the high power version of it. So I retrieved my box of spare aerial meters, and checked them. I had 28 meters in the box. All were faulty. I decided that I should fix one or two.

FAULTS

I sorted the meters into the 1 Amp and 350 mA sizes. I further sorted them into the different manufacturers. Some were labeled AWA, the manufacturer of the radios. AWA is the Amalgamated Wireless (Australasia) Ltd., the biggest Australian radio maker of that period. It was soon apparent that those meters were identical to the other meter manufacturers, so I concluded that they were not AWA manufactured at all, just made to an order and labeled AWA. This reduced it to four manufacturers. They were VANE, EMMCO, PATON, and an unnamed Model 505 (in a stepped sided case).

 

Picture: Different Meters (left to right, VANE, EMMCO, PATON, Model 505, top row 1 Amp, bottom row 350 mA)

I had assumed that they would all have burnt out thermocouples, but that was only partially true. Each type also had a unique fault, which affected the actual meter movement. I tested all the meter movements, to find the good ones. The VANE type was the best. Only one EMMCO  had an open circuit movement. The PATON type used a die cast support for the scale and pivot, and some had suffered pot metal disease, the scales were buckled, and the meter pivots had distorted or broken off. The Model 505 type, suffered from corroded and open circuit hairsprings.



Picture: Pot Metal Disease


Table of the fault types and the number of faulty meters.
-------------------------------------------------------------------------------------------------
TYPE           FAULT                                              350 mA                      1 Amp
------------------------------------------------------------------------------------------------- 
VANE          O/C thermocouple                               3                                   1
-------------------------------------------------------------------------------------------------
EMMCO     O/C thermocouple                                3                                   3
                     O/C movement                                     1                                   0
-------------------------------------------------------------------------------------------------
 PATON       O/C thermocouple                               3                                   3
                      pot metal disease                                 2                                   4
-------------------------------------------------------------------------------------------------
Model 505   O/C thermocouple                                1                                   1
                     corroded hairsprings                            2                                   1
-------------------------------------------------------------------------------------------------

MOVEMENT

Firstly, I cleared the bench, and removed all dirt and iron filings, a magnet being useful for this. I then put down a white cloth, and set up a bright light and used a head mounted magnifying lens. I disassembled the meters, and sorted them into the good meter movements. Most had loose glass, some had broken or missing adjuster screws, and two had broken cases. This had let dirt, iron filings, and moisture enter and affect the movement. I cleaned the cases and glass, glued the glass back in, used good cases, and swapped adjuster screws to the good cases. I used clear nail polish to glue the glass back in. Then I connected a bench power supply, with a 4K7 series resistor, with leads with alligator clips, to test each movement. I varied the voltage, to swing the needle up and down, looking for smooth swings, without sticking or jamming. I put the good movements aside, cleaned the sticking ones, and put the faulty ones in a box for spares.

I tried several methods to remove the dirt and iron fillings from the meter movement gap. Blowing through the gap removed some, but left sticky or magnetic material in place. Pulling some thin strips of paper through removed some more. Using stainless steel (non magnetic) very thin tweezers allowed me to grab some iron filings. I discovered that a "Dissecting Kit" was useful, as it contained 3 tweezers, 3 scalpels, 2 scissors, one spike, and spare blades, in a soft zip closed satchel. This is available from Macquarie University for A$24 or 12 UKP. The best method proved to be laser printer labels, cut into thin strips, and poked into the meter gap. I used each strip once only, and threw it away or cut the end off, and kept on poking around until the gap was clear, and there were no particles on the sticky strip. This grabbed all foreign matter easily and quickly, and was less damaging than poking around or blowing. I just had to be careful not to touch the hairspring and stick to it. I now had 18 good movements, all they needed was a thermocouple. I should mention that I am not an instrument fitter and have no training in this field at all.

THERMOCOUPLE

I read up on thermocouples. These thermocouple meters were made in 1940, so it should be simple to duplicate, and I decided I would attempt to make one. It was more difficult than I thought.

Deciding on the correct type of thermocouple, and what materials to use, required some guesswork. All the information I could find, addressed the use of thermocouples to measure temperature. The classic approach, is to use two thermocouples, one to measure the desired temperature, and the other as a reference to reduce the effect of variations in ambient temperature. I was only interested in the temperature change, absolute temperature was not required, so I only needed one thermocouple. Thermocouples function by using two different metals touching or joined to each other. If wire is used, and the ends of the two wires are welded together, a voltage at the other ends of the wires is produced, and is dependent on the temperature of the junction. Several standard combinations of different metals are used, depending on the application, the temperature range, the cost, and the hostile atmosphere. In this application, ambient temperature, air, and non magnetic properties were required. Finding a supplier of the wire proved to be a problem. I eventually discovered OMEGA ENGINEERING, Atlanta, Georgia, <http://omega.com> in the United States would supply small quantities, various wire diameters, various wire metals, with web based mail order, and at reasonable prices. I ordered two 50 foot reels of ALUMEL and CHROMEL, diameter 0.13 mm (0.005 in), to make an "ANSI standard K" type junction. They were $14 for each reel, but there was an additional $55 shipping charge! I paid by Credit Card and the wire arrived a week later.

HEATER

To continue the classic approach, the thermocouple is placed where you wish to measure the temperature, which is often in a boiler, or in a hot gas, or in a chemical solution. This was no use to me, I wanted to heat the thermocouple with aerial current, so I needed a heater resistor, in contact with the thermocouple. I could find no information on this application. I knew wire wound resistors were made of nichrome wire, and I had a box full of them, so one of these should be suitable. I measured the remains of several burnt out thermocouples in the meters, of which half the thermocouple was usually intact, and also the length of the heater and the diameter of the wire. They were between 1 to 2 ohms in resistance, about 7 mm long, and about 0.125 mm in diameter. This should enable me to select some wire for the heating part of the thermocouple. I destroyed a wire wound potentiometer to provide the heater for the 1 Amp thermocouple. It was a 1K ohm potentiometer of about 10 watts dissipation. I destroyed a 5K resistor, 20 watt dissipation for the 350 mA thermocouple heater. Now I just had to join them together.

 

Picture: Resistance Wire



WELDER

The thermocouples appeared to be welded together. There were some thermocouple welders available on the OMEGA web site for US$1800, but this was too expensive for me. I decided to make one. I decided that an ARC welder would melt the wires, but I would have to hold them in place. I designed a jig that had two terminals to hold the heater wire. There was another terminal for each of the thermocouple wires, and a tube to guide them and hold in place. The terminals holding the wires would be connected to a meter and a current source, to test the thermocouple in place. The two arc welding electrodes, mounted vertically,  one above and one below, would meet at the contact point, create an arc, and melt the wires together. I constructed the jig from junk box parts, with a knob to push the welding electrodes together, and springs to retract them. I worked out the leverage, so that the top electrode moved at three times the distance and speed as the bottom electrode. I dismembered two carbon zinc AA penlight batteries, and used the carbon rods and the crimped on metal cap. I used a ceramic insulator as the guides for the electrodes. Initially I used, a SCOPE soldering iron transformer to provide the current source, as it provides 3.3 volts AC at a current capable of heating the carbon element to red heat and melting solder.  This proved insufficient.  I used a bench power supply to provide 30 volts DC at 2 amps. This also proved insufficient.

 

Picture: Jig

 

Picture: Prototype Welder

It appeared as though arc welding was not suitable. Perhaps spot welding would be better. I looked at some designs on the web and most seemed oriented to spot welding sheets of metal or battery tabs together. I selected a simple design, using two regulator Integrated Circuits (IC), a Silicon Controlled Rectifier (SCR), with a convention transformer power supply. I used an old Golf Buggy battery charger that had burnt out, but the transformer was still good. I put aside four computer grade capacitors, giving 170,000 uF in total, each the size of a can of soft drink. I used an LM317 regulator IC to adjust the voltage to between 5 to 19 volts DC. I connected up the SCR to the carbon electrodes, with one capacitor and tried it out. The carbon electrodes produced a small spark, but did not get hot. I decided to change the electrodes to copper, so that all the energy was dissipated in one place, rather than in the length of the electrode. I used an old copper soldering iron tip. This now produced a satisfactorily loud spark. However, the wires I attempted to weld, were vaporised! There was too much energy, even when turning down the voltage regulator. I reduced the capacitor size, and eventually settled for three normal sized electrolytics, each 4700 uF at 25 volts. The SCR I used was too large, but as it worked, I did not change it. Eight volts seemed to work the best, and 12 volts melted the wires. However, I could not see the junction, it was difficult to arrange the wires, I often only got two welded together, and I could not get them out easily, or trim them to size. I discarded the jig.



Picture: SCR   



Picture: Welder

Maybe a simpler approach was required. I screwed a large copper lug to the top of the plastic box, held the wires in place with sticky labels, and applied the upper electrode by hand. The heater wire and the two thermocouple wires, were gently pulled through some fine grade sandpaper, to clean them first. I could feel a slight "slump" or "give" when a successful weld occurred. If I pressed too hard, there was not enough contact resistance, and no heat and weld happened. If I had the voltage too high, a spark came out the sides, indicating the wires had been vaporised. If the wires or electrodes were dirty, then there was no weld. With a few trials, I settled on a method that works well. After the weld, I cut the wires with the scalpel, and picked the finished thermocouple up with tweezers. I gave each wire a tug to test the welds. The welds and the wires are quite strong. The wires can be bent without breaking, but the weld does not like being bent.



Picture: Thermocouple and Tweezers (ignore light reflections)

MOUNTING

The thermocouples were basically in two physical forms. Most meters used a straight heater wire, with the two thermocouple wires attached in the centre, and coming away at 45 degrees, so the shape looks like the letter "K". The ANSI standard also uses the letter K, as a type of junction, but I am referring to the shape here. The EMMCO meters had a straight heater wire, with the thermocouple wires attached in the centre, but coming away on both sides at 90 degrees, thus looking like a cross. The thermocouples are mounted on the bakelite meter case rear panel, and soldered in place. Ordinary 60/40 (tin/lead) solder will not work, as it does not wet or stick to the nichrome heater wire. I purchased some flux core "silver" solder, from the local hardware store, and adjusted my variable temperature soldering iron to full, as approximately 480 degrees F is required. The physical construction of the meters, has the two plug in pins attached to the rear of the meter case, and two copper metal bars almost meeting in the middle. The heater wire is soldered between these. The solid bars and the big pins, act as heat sinks, to remove the heat from the heater wire, so that the needle can quickly return to zero, when the heater current is removed. They also suck away the heat from the soldering iron, making it a little tricky to solder. The ends of the thermocouple wires are soldered to two small posts, and the meter movement is connected to these. There is provision of a third and fourth post, so that a series resistor can be added, to reduce the meter deflection (for calibration). Some of the 350 mA meters had a length of the heater wire soldered to these posts, the length being set for the amount of resistance required. The PALEC and EMMCO meters also had heat sinks on the thermocouple wire pins.  I accidentally dropped a few thermocouples, and they disappeared into the flotsam on the floor, or into the hair on my arms or legs, as the hair was about the same size. Wearing light coloured long trousers and a long sleeved shirt or dust coat would be an advantage.



Picture: Burnt out PALEC Thermocouple



Picture: Burnt out EMMCO Thermocouple 

TESTING

I used the other half of my bench power supply, set to 10 volts, but with the current limit set to zero. I used leads with alligator clips, to test each thermocouple. These were connected to the meter rear pins. I had other leads with alligator clips, connected to the meter movement. The meter movement was lying on the bench in a plastic jam jar lid. This supported the movement, and prevented it from being damaged or picking up dirt and metallic contaminates.  I used the power supply current limit knob to swing the current up to 350 mA (or 1 Amp as appropriate). If the needle went the wrong way, I reversed the leads. Most thermocouples were good. Some failed straight away. I tried to use some failed thermocouples, but they were impossible to solder after being welded, and difficult to reweld. It was easier to throw them away and make another.



Picture: Testing the New Thermocouple

 I also tried another method of heating the thermocouples. I decided that they did not necessarily require the heater actually welded to the junction, just the heat from it. So I wrapped the heater wire twice around a junction that a heater wire had broken off. It was easy to do with tweezers as the heater wire is quite strong.  This worked, but it was sensitive to the heater current polarity. I was using DC to heat the wire. By wrapping the heater around the junction, it usually made electrical contact with both the thermocouple wires, and the electrical voltage drop along the heater wire, was conducted to the meter movement. This made the meter reading inaccurate, as it was dependent on the applied DC current polarity, one direction increasing the reading, the other direction reducing the reading. This may not have been a problem when measuring AC aerial current, but I decided not to use this method. I also noticed the same effect on some successfully welded thermocouples, but on close inspection, the heater was not exactly at the junction weld, but a little back from it, and touching the two wires. These were also discarded.

CALIBRATION

The 1 Amp meters were the easiest to do. They were usually within 10% of the FSD (Full Scale Deflection), some as close as 1%. The 350 mA meters were harder to calibrate, most were within 20% but a few  were reading 100% low (a reading of 350 mA FSD for 700 mA through them). I had to put in a new thermocouple to get them closer to the correct value, as my manufacturing technique had some specification variation in the finished thermocouples. I found none that read high (over the FSD), so fitting an internal series resistor would not have helped. I concluded that I needed more temperature from the heater. I was using a smaller diameter heater wire for the 350 mA meters, but the same thermocouple wires as the 1 Amp meters. I was not prepared to buy another two reels of smaller diameter thermocouple wire, so I accepted the inaccuracy. I rationalised this by considering Aerial Current meters are a tuning aid, to achieve maximum aerial current, and the actual aerial current is not that important, provided it is the most you can achieve. Some transmitters only have a series light globe as a tuning indicator.

CONCLUSION

It is possible to repair thermocouple ammeters. The more sensitive ones are more difficult. I read with interest, an article about repairing aerial ammeters with a current transformer and diodes. This is also a good approach and easier to do. As an interesting aside, I could blow on the thermocouple, and watch the meter reading drop with the cooling, and then recover. Also when working in reduced light, I noticed the heater wire was visibly red hot, at about 1.5 Amps. It would burn out at about 1.7 Amps.

Welder Circuit

Copyright Ray Robinson

Back to the INDEX (more radios)