TS-820s Mods

The Kenwood TS-820s

"Dream rig" from 1977 joins the lineup!

Around 1976, the ham gear at my station was a stout pair of tubed boat-anchors of excellent repute: The Hallicrafters HT-32 and Hammurland SP-600. As boat anchors go, very good gear indeed.

Unfortunately they were also huge, heavy, hot - and due to their combination of tube components and age, somewhat unreliable - significant problems for a 13 year old ham with zero cash and zero test gear and little technical skill beyond the general background required for an extra-class ticket back in the day.

So when the TS-520 came out, I was an eager buyer. The purchase was a bit complicated... The HT-32 was sold to a local and that provided the down-payment for the new rig. A bank loan, co-signed by my father, brought the full cash needed to complete the transaction.

At that time I lived in Hutchinson, KS - about 250 miles WSW from Kansas City - where the nearest real ham store was located. Fortunately, Elmer Larry Smith, WB0SSB, was a Kansas City resident and after some discussion, Larry offered to serve as both the buyer on my behalf. And having family in the Hutchinson area, Larry even provided the transportation service! I still remember the day Larry arrived. I had just stepped into the modern rig era!

While the TS-520 was all the rig I could drive, the guys with a bit more spending money would opt for the TS-820. The big difference in the 520 vs. 820 was the IF shift feature - and of course, the 820 "S" model had a digital dial - a real cutting edge feature then. So happy with the 520 on the desk, the 820s was my dream rig from 1977.

Check out the Shack History tab for a full run down on how I got into radio and what the early days were like. Shown at left, my ham shack circa 1981 just before heading to college. To the right of the 520 is an Accukeyer - a Bencher paddle is under the light-blue dust cover bowl.

Luck of the Irish

With that background, I've been kicking around the option to pick up a TS-520 variant over the years, but not having done so, I was pushed into action when the opportunity to check out a friend's TS-820S prior to sale. The rig was from the closet inventory of WA0WOF's late husband. Kay's a traditional red-head of Irish descent and is still very active in the local DX club, XYL clubs and quite a few other activities. Laying hands on that rig was enough to push me into action. And in the process of checking out the rig, it brought back the old memories which ultimately led to the rig staying in my shack.

30-something Troubles

Electronics some 35 years or so old is not going to be without some trouble. And the same was to be said with this rig.

Initial problems discovered included soft finals and a digital display reading wildly varying frequencies (a common issue with the rig). Cosmetically, the face of the rig was fine - but the cabinet was quite rough and unfortunately, the top facing area of the darker green front cover trim is scuffed pretty badly. And from an accessory standpoint, it was a "naked" rig in that a mic, rear apron 9p plug and power cable were missing.

Initial check out of the rig for basic functionality was done with jumpers stuck onto the power connection. Later I picked up a very nice cord and a VFO knob with spinner from Bob Mansker KE5FTF. The CW filter was missing as well - and I tracked one down on the Yahoo Kenwood Hybrid's reflector.

Getting the digital display to work was not difficult as the problematic display tube was intact and functional. The assembly has a number of board interconnects which are prone to intermittent contact and require some TLC cleaning. Each time I did the dis-assembly and reassembly, the display would run for a few days. But eventually, it would start to show an erroneous display pattern prompting another round of cleaning. It's been solid now for 4 days and I'm hoping the last time did the job!

Unrelated to the rig was an operation location - the shack desk is full. Fortunately, there was enough room to put a shelf along the left side of the shack desk and that provided enough room for the 820, SP-230 speaker and AT-200. The same shelf accommodated my SB-200 as well.

Warning: 820 MODs BELOW!!!

If your outlook toward old gear is "restore as originally built," then you are not going to like what I have done below at all.

However, if you are into tweaks and mods to get something new, or better, or different out of your 820s, then have fun with the items that follow!

MOD: Bulb/LED Swap Part 1 - METER

While I love old gear, the yellow glow of illumination bulbs is not part of the appeal. I like the crisp clean appearance that LEDs provide. So my #1 mod for this rig was to swap the old yellow meter lights with a set of LED as shown below.

Normally I will use T5 sized LEDs pointing down above a meter but this rig had no available vertical clearance. But for this rig, I used some SMT white LEDs. I built a series of 5 connected as shown here - spaced so that they covered about 80% of the width of the meter width.

The LED strip was placed on top of the meter and adjusted for best light pattern on the meter. I was unsure initially how the lighting would look and so placed a bit of black tape along the railing at the top. A few drops of hot glue hold the LED to the tape making the entire assembly easy to remove in the event the results were not good.

However, the SMT LED worked out very well indeed and so I added a few more small drops of the hot glue that just touches the top surface of the meter housing and that keeps the assembly in position very nicely.

The connecting wires are #30 wire-wrap lengths which are flexible and can be hidden on the meter's back side.

The original lighting was fed with 12V AC. While the LED can be used with a series resistor directly feeding from the 12V AC line, I did not like the flicker that I felt was present. I may have been imagining things but 30 hz may have been visually interfering with the shack's fluorescent lamps or something.

I added a separate diode and cap to provide a half-wave filtered DC supply to feed the LEDs and that eliminated the flicker. The parts are mounted directly to the terminal block as shown here. And the wire-wrap connection to the LED is tucked up on top of the meter, behind the original lighting grommets

With the rig assembled, the lighting looks great.

It's not as bright as it looks in the picture and with the series dropping resistor easily accessible with removal of the top cover, easily adjustable. That nice because I find it takes some time of "living with a rig" before the optimal brightness is found.

MOD: Bulb/LED Swap Part 2 - VFO DIAL

Looking at the rig a couple of days after making the meter LED mod, I was wondering why Kenwood did not put some kind of lighting on that VFO dial. Then again, the digital display is present and the dial is not the focus of attention.

Later, in looking through the service manual - I noticed that 3 light bulbs were listed - yet I had only found 2 on the meter! The logical place for the 3rd one is on the dial, as none of the other controls were subject to lighting. But having had the VFO out for lube, I am sure I had not seen a 3rd bulb. Maybe I missed it - at that time, I was focused on the VFO itself and not specifically possible lighting.

Subsequent inspection of the VFO showed that yes, there is a place to put the bulb. On this rig, however, the bulb, wires and grommet had been removed.

Fortunately, a T5 white LED and junk-box grommet made short work of the restoration. Again the wire-wrap wires and some heat-shrink were used for the feed. The T5 facing was roughed up a bit with fine sandpaper to improve the diffusion.

The LED was connected to the same PS connection used for the meter, and with a separate dropping resistor.

The LED alignment is just right without any special adjustment and with the bezel on the VFO, it illuminates the VFO dial cross-hair very nicely.

Both of these lighting mods are fully reversible, if needed.

MOD: 600 Hz CW Offset Part 1 - SIDETONE

My standard sidetone on the Yaesu rigs is about 600 hz. The Kenwood is setup for an 800 hz sidetone. And without an easy sidetone-spot feature to use, I thought I would be consistently tuning the 820 off about 200 hz. Which made me wonder if the 600 hz sidetone and filter centering could be moved down. Adjusting the CW sidetone seemed the easier of the two and it was my first target.

Looking at the schematic, the sidetone oscillator uses a phase delay RC line with 10K and 0.012 uF caps in cascade.

The resistors could be swapped for higher value ones, or the caps padded. The easier solution proved to be adding pad caps.

The PCB layout was easy to decode thanks to the service manual. The 4 cap locations are shown in the red squares at left.

Approximately 4 nF of pad is needed to shift the frequency down. In the junk box, I had some 3.9 nF caps in a 1206 SMT flavor which proved to be an excellent fit.

The final sidetone frequency measured at 592 hz following the modification.

MOD: 600 Hz CW Offset Part 2 - Filter Alignment

Using a sound-card driven AF spectrum analyzer and sweeping a tone over the antenna input, the 500 hz filter was indeed centered around 800 hz. Unfortunately, the filter response knee sits right under 600 hz that would give quite a high side response to a 600 hz note.

Plots of the various filter and modes are shown in the log plot below. For a high resolution plot, click the graph.

An easy way to solve this is just to turn the IF shift knob a bit and that would result in the filter being nicely centered. However, my goal here was to see if I could get the filter centered with the IF shift detent position centered. It seemed a bit of adjustment on the VCO oscillator board would do just that.

The process was experimental rather than measured. The steps were:

1. Adjust the CW Rx frequency to center the AF response by tweaking the USB trimmer. This will result in the filter response moved down about 200 hz. This step also shifts the USB AF response as a unavoidable side-effect.

2. Measure the USB frequency response on the AF SA. And adjust it as well to trim the LSB response to match the USB response.

3. Trim the CW TX frequency to be offset by 600 hz from the CW Rx frequency. I used a 2nd receiver to do this. Both rigs were zero-beat to a signal. Then the 820s was switched into CW transmit and the frequency adjusted until the transmitted tone was again zero beat to a 600 hz tone.

A full spread of filter response plots are shown in the following section.

ANALYSIS: Post 600 Hz Mod Filter Response

Final frequency response of the rig following the alignment for 600 hz CW operation is shown in the graphs below, linear scale this time.

The input signal strength of -102 dbm is about 3db less than what seemed to be the 7.000 MHz saturation point of the rig (without the AGC operating).

Notice the CW filter is nicely centered around the 600 hz center, and the USB/LSB filter responses overlap tightly.

Taking a closer look at the top end of the filter's response shows this graph - zoomed to the top 10 db of the filters responses.

-6db bandwidth points measured out as follows:

CW-narrow: 530 hz

CW-wide: 1350 hz

USB/LSB: 2000 hz

FSK: 615 hz

Notice that we have traded off some top end SSB response for the CW filter centering. However, given my operation preferences (almost all CW), that's a good choice. If I want more high end SSB rag chew AF sound, I can adjust the IF shift up a bit.

The ultimate attenuation as shown by the higher noise floor than what would be expected is marginal, but that is likely due to the noise level on the bench where the AF lines to the SA feed were unshielded - and is probably not an accurate assessment of the full filter skirt. In fact, separate VNA tests on the YG-88C alone show ultimate attenuation to be about -70 dBc.

IF SHIFT ADJUSTMENT RANGES

Adjustment of the IF shift knob in CW mode, with the 500 hz filter enabled, looks like this below. Notice the plot marked "centered" is nicely centered around the 600 hz target. Data for the other 3 lines is taken with the IF shift control turned "left 1 panel marking tick," then "right 1 panel marking tick," and finally "right 2 panel marking ticks."

Repeating the same process for the USB mode provides the response below.

The similar IF shift data for LSB shown below. Note it matches the plots for the USB data almost exactly.

And finally, if you have wondered how old school rocks match up with new age DSP filtering, we have this comparison using an outboard AF DSP filter. The combination of the 500 hz crystal filter will hold AGC pumping down and avoid some close-in IMD problems. And then the AF DSP can provide single signal selectivity in a crowded band condition. The two make a nice pair. That must be why the best rigs have rocks up front, and DSP on the back end!

MOD: CW Wide/Narrow Switching

By default, with the CW filter installed (and PCB jumper set), the rig will invoke the 500 Hz filter by default. I prefer to use a wider bandwidth if the band is not crowded and found myself flipping between USB and CW quite often.

The need for filter switching was addressed in the 530/830 rigs but users are left to work up a mod of their choice for the 820 series. In one of the Kenwood service bulletins they mention re-tasking the digital hold (DH) button and that's the path I took. It's documented here a bit more clearly as I found the instructions on the web as well as the KW bulletin a bit hard to follow.

This mod allows the wide (SSB filter) or narrow (500 hz filter) to be selected in the CW mode. With the DH switch in it's OUT position, the wider filter is selected. Pressed in, the 500 hz filter is selected. Operation on SSB/FSK is unaffected.

First step is to identify the termination point of the black and white wires that are already routed to the DH button.

The black one terminates onto this grounding terminal located behind the mode switch as shown at left.

The white wire terminates into one of the PCB connection points. I cut both wires and fished them out of their tie-wrap retaining.

A third wire is needed and added that (green color, in my case) as shown here.

The wire set leading to the IF board runs very near the mode switch and that was a convenient place to splice in. The color coding of the various wires is shown below as I had them initially connected to verify functionality.

The green/purple/gray wires come from the wire harness.

The gray wire is cut with the black/white wires from the DH switch inserted.

The newly added green wire from the DH switch is spliced into the existing green wire as well completing the connection.

After testing, the wires are heat shrink insulated and then tucked in under the wire harness as shown at left. This results in a very tidy appearance as none of the surgery is readily visible.

MOD: The Sounds of Silence

Modern rigs are largely silent in their T/R operation. And so by comparison, a relay switched rig like the TS-820 really makes a lot of noise when flipping operation modes as it moves between RX/TX. So in the interest of environmental serenity, mod we must!

The obvious target is the big relay in the center of the rig. Lots of contacts and the base is connected to the chassis, via the solid PCB.

The first step is to break the acoustic conduction between the relay board and the chassis. Here some 3/8 single side adhesive foam covers the bottom of the PCB. It serves both to deaden the sounding-board effect of the PCB itself, and serves to provide some standoff from the chassis.

The board was replaced and secured with two slightly longer screws with built-in washers. A grommet was split and placed between the screw head and the PCB to avoid conduction from the PCB to the chassis via the screw.

Next, the relay itself is insulated with the same foam, cut to fit. The red tape shown here serves to keep the foam in place. Because of the foam thickness, it's easier to place the relay into the socket without the adjacent PCB connectors fitted - they slide onto the contacts easily after the relay is positioned.

Not shown in this picture, I was a bit worried that the bottom of the case would compress against the foam and push the relay into the socket causing a bit of PCB shear force relative to the tall electrolytics. To ensure a more even force spread across the board, a bit of foam was added spanning the two electrolytics giving this operator a bit more peace of mind.

The rig was much quieter with this relay treatment - but still, it was a lot louder than I had expected.

After a couple of days, it occurred to me that there must be another relay in the RF compartment and that it was unlikely that Kenwood would have routed the TX signal all the way to this central relay location. Sure enough, hidden under what I had assumed to be a connector enclosure was the T/R relay.

The relay is similar in configuration to the relay board relay insulated above, but the shield is a tight fit to the relay enclosure. The same insulation treatment would not work for this relay...

The relay is mounted directly to the chassis and to break that sound conduction path, some spacing was needed. My eventual solution was to insert a bit of a felt spacer as shown here. The socket was not screwed back into place relying instead in the shield to hold the relay into the socket - and the mounting holes for the socket as it protrudes through the chassis keep the relay in position from an X-Y basis.

For the relay sound damping, a thinner foam adhesive was used along 3 of the sides. There is enough vertical clearance to use the thicker foam on the top side. As well on the side that faces the rear of the rig. The covering tape on the thinner foam was left in place to help provide the shield with a slick surface making fitting of the shield easier.

Finally, the shield 's 3 sides were expanded slightly to allow for a more relaxed fit around the foam covered relay. Other than the foam corners exposed, the modified shield looks quiet OEM. Only your ears will notice the difference.

With the T/R relay modified, the rig was very quiet. It's not silent as a solid-state switched modern rig is, but the sound it makes is (for reference) much quieter than the toggle switch makes even if you are very slowly moving the SEND switch.

I did have a bit of trouble with the relay not switching after this modification. It turned out to be a dirty socket. Subsequent treatment of the socket and relay pins with contact cleaner eliminated the problem and the relay has been a reliable performer since.

Note: This mod (other than the deformation to the T/R shield) is reversible.

MOD: Adding the Ham Gadgets Pico Keyer

Once the rig was back together and working, the next step was to GET ON THE AIR! But alas, that was to be more problem that I had imagined... None of my modern high-tech keying goodies are setup to key the negative 65V signal that is present on the keying line.

I had messed around with some isolation ideas using an opto-isolator - but none of the ones in my junk box had enough headroom. They were 70V max parts and I thought they may fry in actual use. What to do?

Then I remembered I had one of Dale's Picokeyer kits in the drawer - something I had planned to incorporate into my home brew digital interface, and as that project was swapped for a Microham MKII, the kit sat unused. Until now.

It had the advantage of being an excellent keying module from a user "feel" standpoint. And it would key the rig .

Building the kit took about 15 minutes and required no instructions.

The keyer has quite a set of built in functions including my must-have three - an Iambic-B emulation, keyer weighting control and a memory for CQ with repeat.

The keyer worked great and after a few QSO, thinking about where to mount this, I remembered Dale was selling a small custom enclosure for the keyer. It turns out that the box is exactly the right size to fit under the rig, with a velco attachment. And a quick order for another Picokeyer kit (need another one for inventory) and an enclosure brought the set to my door a couple of days later.

MOD: Adding a Bandscope

The TS-820 has an IF out connection which provides an easy access point to it's post-IF filter feed. Unfortunately, that feed is not suitable for bandscope use. Kenwood describes in the SM-220 manual where to obtain a pre-filter tap.

In my 820, rather than adding the additional feed, I re-tasked the existing shielded wire away from it's original location, and connected it to the IF point as shown in the picture below. A bit of hot-glue holds the wire to the PCB and serves as a strain relief. The 47 pf cap is specified for coupling in the Kenwood manual.

By modern standards, the passband is lumpy however for the purpose of providing an indication of how the band looks, it will work just fine. Of course, there's not a CAT feed available from the rig so no fancy bandscope/tuning interaction can be had. I will pair this IF feed up with a dedicated Softrock IF Lite kit and run it into one of the available shack PC sound card inputs.