tl922mod

There are ten modifications I have made to my Kenwood TL-922A. Unfortunately, in today's world I have to add the disclaimer that I make no guarantees for these modifications, nor am I liable if they cause damage or injury.

Any or all of the modifications can be performed, but note the following:

If the QSK EasyKey Modification is done then the Band Switch Arcing Modification is not needed.

If the Tuned Input Modification is going to be performed, the Ground 3-500Z Grid Pins modification should be done because the tuned input circuit component values are optimized for directly grounded grids.

In my TL-922A, all of the original bulbs are still working. If any of them fail, I will replace all of them with modern LED's. This should be as simple as choosing the right color and size, then adding a series resistor in one of the LED leads to limit the current.

Lastly, the modification I believe most important is to do either the Band Switch Arcing Modification or the QSK EasyKey Modification, as either of these modifications help prevent arcing of the band switch.

So have fun, and get that TL-922 updated!

Eliminating band switch arcing was the first problem that needed to be solved with my TL-992A. I had a bad band switch caused by arcing, and needed to resolve this before replacing the damaged band switch, otherwise it would surely happen again. If you need a replacement band switch, check eBay or Harbach Electronics.

When this problem was first noted, the amplifier was operated upside down with the bottom cover off to observe the arcing while operating on 80 meters. There was an occasional arc at the band switch when switching from receive to transmit. Kenwood made a lame attempt at arc suppression by installing an arc suppressor (little glass bulb), but that did little to help.

The solution was to replace the large DPDT Relay R1 with a 3PDT P&B relay with 5 amp contact rating and 120 VDC coil. Two of the poles were wired identically to the old R1 Relay's RF connections. The third pole was wired to be in series with the DC voltage to the Relay R2 coil. This ensures that Relay R1 is settled *before* the bias is removed by Relay R2, which enables the amplifier tubes.

A second benefit was that the new relay was much quieter than the original. It was secured to the chassis wall with double sided sticky foam rubber.

After the modification, no further arcing was seen over a lengthy observation period. A new band switch was bought from Kenwood (still available in 1992) and installed. I took camera pictures of the band switch before removal to aid in switch replacement.

Some hams with TL-992A amps have never had this problem. I assume it is because they were not running VOX or the radio driving the amplifier had an adequate delay of RF output after the key line was asserted.

The TL-922 was supplied with 10 meter band capability. Later, the FCC made it illegal to sell amplifiers that had 10 meter capability. Thus, the TL-922A was released. This version had the ability to operate on 10 meters, but a Kenwood supplied modification had to be done. Click here to view the Kenwood 10 meter modification,

One failure that can occur with a 3-500Z tube is a cathode to grid short. If this happens, the 105 VDC power supply will be shorted to ground through the RF chokes on both sockets. This failure can cause damage to the RF choke and 105 VDC power supply. To prevent this potential problem, simply remove the wire that connects to the Normally Closed contact of relay RL-2-1 and the cathode of D1. The bias voltage will still be provided by the R7 100K resistor, but a cathode to grid tube short will not cause damage.

The measured AC voltage of the 3-500Z tubes in the TL-922A as measured at the tube pins was a bit high in my amplifier, about 5.4 volts. Adding a CL-80 current limiter in series with each of the primary 0 volt lead of the filament transformer dropped the voltage to 5.1 volts in my amplifier. A side benefit is that the startup peak current is limited, helping to extend tube life. Mount the CL-80 away from nearby components as they get hot. The CL-80 is available from RF Parts or eBay.

One day, while operating the TL-922A, the Power On Off switch was moved to the Off position, but the amplifier stayed on! The switch contacts had welded themselves together. I had a spare switch and installed it, and decided to add a Step - Start circuit to the amplifier to limit the peak current on the power switch. The circuit is a pair of CL-70 current limiters in series with each of the 15 Amp fuses. A DPDT 12 Volt coil relay is wired so that when energized, the CL-70's are bypassed. The 12 Volts DC is supplied by a half wave rectifier circuit using the 8 VAC light bulb transformer tap.

I believe W7RY sells a Step Start circuit board. There may be other sources if you don't want to make your own.

Occasionally a 3-500Z tube may arc. Protection can be added to your TL-922 with a length of very thin wire acting as a fuse. At the bottom of the picture is an added ceramic insulator with 2 solder ring lugs. The wire is just visible between the two insulators. A wire of size #42 (.065mm diameter) in series with the B+ from the power supply to the plates of the 3-500Z tubes is used. One source for this thin wire could be a spare IF can from a transistor radio.

Meter protection diodes across both meters can help prevent a fault from destroying a meter movement. Replacement meters are very difficult to find. As seen in the picture, a pair of 1N5408 diodes are connected anode to cathode across each meter. Using solder ring lugs make the installation much easier.

Grounding the 3-500Z sockets will improve the power gain of the amplifier. I used a piece of solid copper #12 wire. The choke and resistor at each socket can be left in place if desired, as they are effectively removed by grounding. The TL-922A tuned input coils may need adjustment if this change is made.

When an Anan 100B was first used at my station, the SWR to the TL-922A input was too high on certain bands and caused the output power of the 100B to reduce. Other radios had a higher SWR also, but still seemed to put out full power or close to it. Examining the SWR for the various radios versus the band used was confusing, as one radio would have a good SWR on a certain band and another not.

What the heck was going on? Changing the length of the Anan coax produced large changes in SWR on various bands. A Heathkit SB200 was substituted for the TL-922A and suddenly all issues were gone. Coax length didn't matter, and all radios had a good SWR on all bands.

One TL-922 web site I found indicated that you should operate your radio with a 2 meter length of coax to the TL-922A input. What??? Something was definitely wrong with the TL-922A and needed to be fixed.

The culprit is a design flaw with the Q of the tuned input circuits. It is too low for the 160 through 20 meter bands. Comparing the schematic of the SB-220 and L4B tuned input capacitor values revealed that the TL-922A capacitor values were about half of what they should be.

The 80 meter band was selected to test and verify that this indeed was the problem. An inductance meter was used to determine how many turns of wire needed to be removed from the Kenwood 80 meter tuned input coil with the adjustment core at the halfway point. The coil was reinstalled in the amplifier with temporary compression trimmer capacitors to determine the optimum capacitance for both capacitors on either side of the coil. Results were very encouraging, and all radios now had SWR values of 1.3:1 or less at the band edges, and 1.0:1 SWR at the band center. As hoped for changing the coax length now has no effect!

Changes were made to all of the tuned input circuits. The capacitor combination values shown below are suggested. You can use different combinations to get the desired capacitance. The tuned input circuit changes follow:

160 Meters: Change C49-1000 & C50-560 to 1700 total with a 1500 & 200. Change C47-560 & C48-1000 to 2000 total. Remove 5 turns from 160 L9 coil resulting in 22.5 turns. (10uh to 6.3uh) Tune L9 for lowest SWR at 1900. Q is 1.02.

80 Meters: Change C52-330 & C59-220 to 1377 total with a 1247 & 130. Change C51-330 & C57-220 to 1303 total with a 1247 & 56. Remove 4 turns from the 80 L10 coil resulting in 14.5 turns. (4uh to 2.3uh) Tune L10 for lowest SWR at 3750. Q is 1.7.

40 Meters: Change C54-470 to 1150 total with a 1000 & 150. Change C53-330 to 1120 total with a 560 & 560. Remove 3 turns from the 40 L11 coil resulting in 6.5 turns. (2.3uh to 1.05uh) Tune L11 for lowest SWR at 7150. Q is 2.6.

20 Meters: Change C56-220 to 225 total with 195 & 30. Change C55-150 to 100 total with a 100. Remove 1.5 turns from the 20 L12 coil. The last half turn is made by 45 degree angle of wire rather then perpendicular to the coil. (1.2uh to .62uh) Tune L12 for lowest SWR at 14175. Q is 1.

15 Meters: No change needed. Tune L13 for lowest SWR at 21225. Q is .8.

10 Meters: Change C60-120 to 90 total with a 47 & 43. Add 47 to output (tube cathode side). Tune L14 for lowest SWR at 28600. Q is .8.

After thinking about what other modifications would be nice to have in the amp, an idea came to mind. A single custom board could accomplish all of the following: Key line compatibility with newer solid state radios, electronic 3-500Z bias switching, QSK using a pair of RJ1A vacuum relays, and control of the front panel STBY and On Air lights. Relay R1 is replaced with two Jennings RJ1A relays as seen in the picture below on the right. Relay R2 is removed from the amplifier. The result is a very quiet amplifier when switching between Transmit and Receive.

The QSK EasyKey board was developed to do all of the above as well as provide the proper timing and sequencing for QSK (Full break in) CW. There are also jumpers that allow setting 4 different 3-500Z bias voltages. All components on the board are easy to find and available from a variety of sources.

In the picture below on the left you can see the QSK EasyKey board installed to the left of the filament transformer. The board is secured with a pair of nuts to the existing bolts of the filament transformer.

I have sent several of the boards to friends who have installed them not only in a TL-922 but also Drake L4B amps. If you are interested in a board it is available for sale at $50 populated plus shipping and comes with installation instructions. Send an email to wb8bfs@fuse.net for more information. Installing the board requires soldering skills and attention to detail.