From George Smart's Wiki
Introduction & Motivation
Having had a WellBrook Loop ALA1530 for some time, I accidentally transmitted into it during a moment of madness where I wasn't using the RX antenna connector. I had been very impressed with the performance of the loop on LF frequencies, such as on 472 kHz and 1.8 MHz, and given the cost of a replacement, I decided to attempt a repair!
The loop is built into a electrical conduit joint box and potted with black resin compound to keep moisture out of the electronics when the loop is mounted outside. However, this made it difficult to repair.
At a local club meet, I was speaking with a friend about my mistake with the loop, and Dave G7UVW who works in an X-Ray lab, offered to x-ray the failed part to assess the feasibility of a repair. You can see these x-rays below. From the x-ray images, it was apparent the device was made on veroboard, with several capacitors, ferrite transformers, and some active transistors.
Since I had nothing to lose, I decided to try and remove the resin compound from the device so I could get at the circuit-board. Various methods were tried, but the most successful was a combination of using a craft knife and pick, as well as warm (around 35C) MEK (methyl ethyl ketone) to dissolve the resin. Over time, we managed to get through, until we could see all of the connections.
As a result of being transmitted into, the transistors were blown to pieces, but the parts in place, and, we were able to read the transistor part numbers as being ZTX327.
In getting to the circuit board inside the resin, many of the wires to and from the transformers were broken, one of the ferrites was cracked, as was the circuit-board, probably due to my haste in getting the resin off. Many of remaining components were worse for wear.
At this point, it was apparent I would need to rebuild the circuit from scratch, and in order to do so, I would need a schematic, since some of the wiring connections were broken. Unable to find a schematic online which matched what I had, I managed to piece the circuit together. As it became obvious that the circuit was balanced, I was able to compare both halves of the circuit to fill in the missing links.
I decided upon using BN-73-302 ferrites based on the graphs for the material, and measurements of AL and frequency response. These are the same size and cover the frequency range quoted for the original antenna design. All the available ferrites in the range were tested.
Starting from the schematic I had worked out, I laid out a PCB as I was pretty sure I had the design correct. I added the PCB order onto an existing board order, and waited patently.
I made the circuit boards up when they arrived and tested the design which oscillated uncontrollably! This turned out to be due to the phase of the input transformer having positive feedback instead of negative. An easy fix was found, namely, swapping two wires over on T1 (see below).
Another small mistake was the holes for the BNC connector pins are slightly too small, and may need to be drilled out or pushed in carefully - this wasn't strictly my fault, since the footprint was the default for BNC connectors in my PCB package. If you do need to do this, be careful that the ground side of the connector is soldered to the top ground-plane on the PCB.
The PCB was designed for the original ZTX327 transistors. However, they are now obsolescent, and expensive, so I have been using MSPH10's, which do not fit directly into the PCB. The ZTX653 is also a recommended replacement for new designs. See the section on transistors below.
Finally, with the circuit working as predicted, I connected the loop and observed the functionality of the newly built board to be indistinguishable from the original. I was pleased with the result!
One thing I did notice is that the transistors get warm when in use - I suspect this is due to the high DC quiescent collector current, which keeps the intermodulation distortion low.
Click on images to enlarge. X-Rays provided by Dave G7UVW and Dave M0TAZ.
An early version of the PCB had C6 incorrectly placed next to T2. Version 1.1 of the PCB has this issue fixed. In practice, this made no noticeable difference, but will be correct on future versions of the board.
Bill of Materials
Components for the loop PCB.
|PCB||WellGood PCB||1||I may have spare boards Contact Me|
|C1, C2||82pF||2||Farnell 1694324|
|C3, C7||2.2uF||2||RS 870-8752|
|C4, C5, C10, C13||100nF||4||Farnell 2525325|
|C8, C9||330nF||2||Farnell 2525327|
|Q1, Q2||ZTX327 (see below)||2||eBay ZTX327 or ZTX653 Farnell 9525580|
|D1, D2, D3, D4||1N4148||4||Farnell 2677463|
|LP1, LP2||Loop Connection (1 metre)||2||---|
|T1, T2||BN-73-302 / Fair-Rite 2873000302 (see below)||1||Mouser 623-2873000302|
|FB1, FB2||TUB4/2/5-3B1 (or similar)||2||Farnell 273156|
Bias Tee Board
Recommended components for a Bias-T board.
|PCB||BiasT Board||1||With main loop PCB|
|CON101, CON102||RF-BNC||2||Farnell 1712350|
|C101, C102||2.2uF||2||RS 870-8752|
Cx* is connected cross the input terminals of the bias tee PCB.
Note that on this version 1.0 board, that C6 is incorrectly positioned. It should be installed between the two transistor bases, across pins 3 and 5 on T1.
Winding and mounting transformer T1 is the trickiest part of the build. The making of both transformers is described in detail. You should set the variable resistor RV1 approximately half way before initially powering the board. For the remaining components, it is simply a process of soldering the parts in. The notes describe the recommended approach.
Start off by making some bifilar wire; you can twist two bits of enameled copper wire together. It is often helpful to use two different colours of wire, or use a permanent marker to colour one of the wires in the twisted pair. If you can't colour the wires, it's easy to sort them after winding using a Ohm-meter. I've used 0.25 mm wire as I had that available, but this not too critical, as long as all of the windings fit on the cores. The original had 0.2 mm wire.
The first job is to wind the loop (input) side of T1. Wind 7 turns of the bifilar wire around the BN-73-302 core, creating a centre-tapped point, as shown below. The numbers refer to the solder pads in the colour placement diagram.
Flip the core around and repeat the process to make the first of the secondary windings for T1. As with the primary, two of the wires are connected together to form the centre-tap and will connect to pin 4. The two remaining wires need to be crossed to prevent the circuit from oscillating.
At this point, it may be easier to mount T1 transformer to the PCB and add the two single turn windings on afterwards - with a binocular core, one turn is a 'U' loop through both holes. These last two windings are made from a single piece of wire (not bifilar), which go in through one side of the core, and out of the other. One of the windings goes from pin 1 to pin 6, and the other between pin 2 and pin 7, as shown.
At this point, you should have transformer T1 mounted correctly.
Using the same bifilar wire as you made for T1, wind 8 turns for the primary of transformer T2. Separate the two strands and create the centre-tap as before, shown below.
The secondary winding of T2 is 6 turns of standard (single core) enameled copper wire, as shown below.
Mount this to the PCB and transformer T2 is also complete.
The transistors used in the original design were ZTX327, and so this board is designed to fit those. These are now obsolete and are quite expensive to acquire. A ZTX653 has a matching pinout with suitable characteristics, and would be a direct drop-in replacement for the ZTX327. Preliminary tests were done with an MPSH10, but the pinout is different, and so require some tweaking to fit to the board. The original ZTX327 transistors had a hFE-min of 15. It is important not to have too high forward current gain (hFE) on the transistors, as this will disturb the correct biasing of the transistors. The pinouts are shown below.
|ZTX327 / ZTX653||MPSH10|
To mount the MPSH10 on the PCB, you should switch the base and emitter wires around. The ferrite bead on the base will prevent them shorting.
Note that the transistors will get warm when the circuit is in use - this is due to the high quiescent current through the transistor pair, which maintains a high intermodulation level.
The variable resistor should be set to approximately half way before initially powering the loop. Once you have built the board, you will need to balance the current through the two transistors. The easiest way to do this is to adjust RV1 to minimise the potential difference across C3. Ideally this should be 0 mV, or as close as possible, within a few milli-Volts.
Tacking two wires to the back of C3 allows you to easily adjust RV1. Remove once you are happy with the balance.
Remember that the transistors Q1 and Q2 will get warm, especially if they are not balanced. This heating greatly effects their HFE (gain), meaning their temperatures will effect the balance. You should adjust the balance slowly, allowing the devices to heat up/cool down, as you are adjusting. I found blowing gently on the PCB between adjustments helped the temperature to settle more quickly
The BNC connector is used to connect back to the shack. The ground side connects to the top copper. On the first batch of boards, the drill holes for the BNC connectors may be too small and thus require drilling out.
Capacitors C1, C2, and C6 are ceramic disc type capacitors. The remaining are poly-carbonate. This is especially important in any RF signal paths (C3 and C7), where ceramic disc capacitors are not of high enough quality.
The remaining parts
The remaining parts can be fitted to the board as standard.
The loop is powered by a bias-T, which injects DC along the coax. The original bias-T was not examined, but, we have used a 2.2uF capacitor such as C7 and the same 1.5mH inductor such as L1.
This project was made possible with the help of the following people.