When an engineer isn’t happy with something, they often try to “fix” it. Sometimes this ends up in hilarious failure where the said object becomes more broken than when they started. Sometimes it ends up in success, but most of the time it ends up somewhere in-between. As I noted in my review of the AN-LP1, the performance of the active antenna was pretty marginal for me, so I wondered if there was a way to improve it. To do this, we first have to open it up and take a peek inside.
As it turns out, tearing down Japanese made products is relatively enjoyable, especially because all of the screws that need undoing to open the case are helpfully marked with arrows. Lets start off with a simple one – what’s in the filter unit?
Four small Philips screws later, we get our answer – it appears to be a common mode choke based around wrapping the wires around a clip-on ferrite, packaged as a fancy plug-in unit. Interestingly, because of the special connector arrangement in use, the male connector is the same moulded plug as supplied with the antenna controller box.
Getting a little more serious now, lets find out what’s in the antenna unit itself. It’s the same story – four Philips screws and we get access to …
… a PCB without any shielding around it. Hmm. Okay, taking a closer look, we can see that the loop antenna is actually formed by a thin, black-silicone insulated wire which runs around the outside of the sun-shade like apparatus. The steel “springy” wire that keeps the shade open is not used for the antenna itself, possibly because of connection difficulties. The PCB itself contains a host of diodes and FETs, the main star being the array of Sony 1T369 variable capacitance diodes which help do the tuning.
On the other side, there is a small transformer, and a through-hole inductor.
Time for the antenna controller box to get opened up – this time, five screws needed to be undone.
We are greeted with a very interesting and mesmerizing design. The antenna cable reel actually connects to the main PCB through another PCB with concentric ring tracks. The reel itself features dual-contact wipers which wipe around the rings as the reel is wound/unwound. Four contacts are available, with five contacts on the plugs – so ground between the two plugs is common-ed. Such an arrangement seems elaborate and similar to the mechanism used in generic multimeters for range selection, but I suppose that is how you would get a signal out of a rotating spool of wire.
To ensure a smooth action and good contact, the board and the spindle are coated in a conductive lubricant. The cable itself seems to be somewhat waxed for easier passage, and the flakes of wax worn off from the cable can be seen collected near the entrance hole at the top. This is the sort of attention to detail that is missing in some cheaper products.
The main board itself has an Alps rotary switch, which is a quality item. There are also a large array of variable resistors (trimpots) used to set the tuning peaks for each of the bands. The DC to DC converter seems to be shielded under a metal can which is soldered down to the board – so I’m glad they did do this to reduce the amount of interference it would cause.
The underside has another can on it, although smaller, but otherwise, it is pretty bare. Putting it all back together was pretty simple, with nothing broken in the process. Definitely a sigh of relief, because I only just got it.
Because the signal from the antenna didn’t seem to be sufficient, I surmised the main cause is the small aperture of the loop. If we look at the antenna module PCB, it looks very simple, and removing it only requires de-soldering two wires. Ignoring the inductance and impedance of the loop for now, I wondered if we could get better performance simply by using a larger loop of wire.
In order to make sure the PCB wasn’t damaged in my experiments, I opted to wrap it in a layer of electrical tape, and then wrap the wires in the tape as well, to provide some strain relief. The last thing I want to do is lift the pad off the PCB.
I had a reel of 0.9mm “fence wire” style galvanized metal wire. I decided to make a square loop, as that provides more wire for the same diameter. I decided to go with a 1m x 1m square, so it has a diameter of sqrt(2) meters, and the loop will total 4m of wire, which is about 3 times more wire than the included round sunshade. To make the connections, I used a screw terminal connector block, as soldering to such galvanized wire is a pain in the butt. The wire itself wasn’t very stiff, so I had to resort to finding something to support the structure of the loop. In the end, the best thing I had was an insulated fibreglass pole antenna for 27Mhz CB, which was about the right length. The corners were electrical taped to the antenna to ensure it physically held its shape.
Initial results were encouraging. Powering it up and tuning into the BoM’s VMC radiofaxes seem to show marked improvement, with the reception performance now indistinguishable from my 15m longwire outside.
In the course of this fax, I swapped back and forth between the two inputs on the Icom IC-R75 and found no meaningful difference – both antennas were subject to fading and weak signals, but both performed equally well. I added a coloured bar on the side to show when the antennas were switched. This was a welcome change.
But sadly, this change was not without caveats. For one, I found that higher frequency tuning was not possible. The range selector for 4-7Mhz worked correctly, but above that, the selector settings for 10-20Mhz instead only pushed a range of 8Mhz to 13Mhz. This limited the usefulness somewhat.
It seems that this may have arisen because of a dependency of the board on the inductance of the loop. The service manual claims that tuning has to be done with a dummy 2.2uH inductance, so I suspect the loop needs to have the same inductance to make it work properly.
There could have been some way to improve on this, by adjusting the tuning voltage pots on the antenna controller, but I didn’t want to create further problems right now. Looking at some datasheets, I’m not convinced that this would open up the range that is necessary – for one, the DC-DC converter in the controller has a limited output voltage, and the variable capacitance diodes also have a limited operational voltage. This is already 17V at the 20Mhz setting, and the diode itself has a maximum operational voltage of 28V. An increase of 11V doesn’t really change the capacitance very much, as read off the graph:
Voltage Capacitance 1V 58pF 10V 10pF 17V 4pF 28V 2.8pF
As a result, I don’t think tinkering with the potentiometer on the antenna controller would have been wise. Regardless, maybe changing the inductor values on the antenna board would be more productive, but I didn’t want to break it just yet.
I was thinking – “wouldn’t it be nice to shove the board in a jiffy box with two 3.5mm connectors for each antenna pad, and just use any pre-made length of 3.5mm to 3.5mm cable as a loop?” Evidently, things are not so simple. Bummer. But a 1m x 1m loop indoors was able to work as well as a 15m longwire outdoors – that was an interesting result in itself. Maybe the cheaper Degen DE31 might be more amenable to this sort of modding, but I suspect we will come up towards the same barriers of the loop’s inductance causing some limitation in tuning range.
The motivation for the modification was pretty obvious, and upon tearing down the unit, the most obvious thing to do was to increase the loop aperture by connecting it to a larger loop of wire. While that was done, and the performance did improve markedly, there was a side effect which resulted in a limitation of tuning range, with the selectors for 4-7Mhz working as expected, and 10-20Mhz instead selected a range of about 8-13Mhz instead. Despite the presence of a service manual, and the possibility of re-tuning the tuning voltage to try and get a little more range, I suspect there would not be a full recovery of the range purely because of the limited voltage output of the DC-DC converter and the voltage limits of the varicap in use. As a result, I am a little disappointed, but I’ve restored the unit to the original configuration, unharmed for now.