The Kenwood TS-2000X amateur radio transceiver is a great radio because it covers HF, VHF, UHF and L-band all in one box. It is not perfect on all fronts but it does everything well, much like a swiss pocket knife. Like most radios, it has some flaws though. One bug I discovered is a severely distorted SSB transmission on models built prior to year 2006. I happen to own one of these. Here is a detailed description of what I found and what I did to cure it.

One day, I decided to try the L-Band function on my TS-2000X in preparation for the ARRL June VHF contest. To my surprise, the other party reported that I had a very distorted audio in SSB and that he could barely understand what I was saying. Everything was fine in Rx. This sounded like a major case of non-linearity to me.

Following that evening, I did some research on the internet and found hints of what might have been happening. Others reported a design flaw kept quiet by Kenwood. The flaw is that the two hybrid power modules used as Tx driver and Tx power amplifier were under-biased on the Vbb pin for rigs dating earlier than 2006. Since my TS-2K's serial number starts with a 6 (which means 2004), I knew that my unit might be part of the defective batch.

So I called up Kenwood USA. A service person (nice, I must say) answered but provided little detail on the actual problem, other than confirming that there was a fix available and that Kenwood would have to do it for me since it involved a re-calibration via software of some of the internal parameters. He provided no detail of the fix. The bad news was that I would have to pay for THEIR design flaw... The reasons were that they gave only a one year warranty extension to the original warranty for this fix; my rig is older than that. The other problem is that I was not the original owner of the rig. Who cares!

I decided that I would not miss the VHF contest a couple of weeks later, so I went ahead and opened up the hood on the TS-2KX to try to better assess the problem. I probed the Vbb DC biases on pin 3 of the two modules when in Transmit and saw what follows:

The Mitsubishi specsheets specify a bias of +8Vdc and +9Vdc, with an absolute maximum value of +9Vdc and +10Vdc respectively. So the modules were clearly under-biased and the values corresponded to the problem description I had read on the internet, but was it really the source of problem?

Next, I decided to probe the RF signal to hear from where the distortion originated. I used a 20:1 resistive RF probe, fed its signal into a 60dB power attenuator and on to a receiver capable of tuning to 1296MHz. I terminated the antenna connection into a dummy load. I then keyed up the 1296 transmitter and probed the RF lines at the input of IC2, between IC2 and IC3 and at the output of IC3. I quickly determined that IC3, the P.A. hybrid, was the source of severe distortion.

Disconnecting IC3, pin 3 from its PCB connection is a rather simple thing to perform; it is just a matter of carefully lifting off the pin while heating up with a soldering iron. So I just did that. I then set the voltage on an external power supply to +8Vdc and connected it up to pin 3. I saw a constant 300mA of current flowing. The absolute maximum current specified is 500mA so everything looked nominal. I then transmitted again and YES!!! Clean SSB signal. I then knew how to fix it.

The following were the facts that I had to deal with in my implementation of a fix:

• When the Vbb is applied, the IC3 module draws a constant 300mA. So the Vbb bias is switched on only when the radio is transmitting on the 1.2GHz band
• There is no +8Vdc available in the vicinity of the 1.2GHz board
• IC2, the driver hybrid module worked well as is.

The two proposed solutions are shown on the schematic on the right hand side. Both will produce a solid +8Vdc when in transmit mode. In both solutions, the old Vbb signal is used to switch the new bias. Simple and effective. Now remember that you have to first lift up pin 3 of IC3 from its PCB pad. If you do not like the DIP relay solution, just implement the LT1764A solution shown at the bottom of the schematic page. Note that I have not tested it, but I have simulated it in PSpice and I have every reason to believe that it will work. The Zener diode and the neighbor 1K resistor are probably optional, but they guarantee a complete shutoff in case the Vbb bias is not quite 0V in Rx mode. The picture to the right shows how I physically implemented the new bias circuit. I used a piece of plated-through-hole prototype PCB. The lower-left corner of the 1.2GHz board is pretty much empty, with only the ground plane covering the surface. The 7808 voltage regulator is mounted underneath the PCB, so that its tab can be screwed against the PCB ground plane using the PCB corner screw. The voltage regulator tab can be seen to the left of the proto PCB. I used thermal compound to better transfer the heat to the PCB and chassis. The voltage regulator pins support the PCB. Care has to be taken so that the overall assembly profile fits under the 1.2GHz compartment cover.

Click on the figure to enlarge it. Click on the figure to enlarge it.