Greetings all,
Signed up on Audiokarma specifically to post in this thread. My background includes about 15 years of working professionally as a home audio (“stereo” to most people) repair technician in the San Francisco Bay area. My parents bought a TK-66 new in 1968 or ’69 (i’ve forgotten which). I’ve repaired it 4 times:
1) Once for no/weak FM in 1975 (when i was in high school).
2) Second time was a total overhaul when i worked professionally (Resistance Repair, Berkeley, California), in 1984. Noisy controls and switches was the main complaint.
3) Third time was January 2001: recap + 2SC458 elimination, to cure a return of noisy controls and switches not amenable to contact cleaning.
4) Fourth time has been these past few weeks (late Nov. through early Dec. 2017), for FM fades in and out. NOT frequency drift: sensitivity comes and goes, visible on the signal meter. (Same on AM but no one in that household uses AM).
I know this model and this particular unit very, very well. I’m hoping by sharing here in this thread, i can help someone out (whether it’s 65W200 or anyone else).
My first WAG is that there may be capacitors fading in value and breaking down from heat. The second guess is that some transistors in the front end and/or IF could be doing the same thing.
Second guess for the winnnnn! At least that’s what has happened twice (1975 and 2017) in the family’s unit. Details below.
It appears that the unit contains almost all Germanium transistors (2SA234B, 2SA354, 2SA355, 2SB54, 2SB89A, 2SC281B, 2SC317, 2SC458LG, 2SC458B, 2SC494, 2SC664, and 2SC734) and some Germanium rectifiers. Resolving the problems with those could be almost more trouble than it is worth, since re-engineering the bias networks to get replacement silicon devices to work properly could take quite a bit of time.
2SA234B, 2SA354, 2SA355, 2SB54, 2SB89A: Germanium PNP, yes.
2SC317, 2SC458LG, 2SC458B, 2SC494, 2SC664, and 2SC734: Silicon NPN. No rebiasing needed.
In the 2001 repair, i replaced all the 2SC350s and 2SC458s with 2SC2320s. No idea if those are available any longer. They were the shop’s go-to low noise high gain NPN silicon small signal audio transistors, making preamp sections quieter and happier all over the place.
Hopefully it’s already general knowledge here, but if it’s not: the Hitachi-made 2SC458 transistors (any or no suffix) in the square package with a bevel which were manufactured up through the mid-1970s
love to leak and/or get noisy. The later ones from Hitachi or anyone else in the standard TO-92 package are fine. Back in the day back in the business, we replaced those original Hitachi flat-bevel 458s on sight—all of them, all the time (on a full service). They’re high beta, so whatever replaces them needs to be appropriately high in beta as well. I don’t know why i didn’t do that on my parents’ TK-66 back in ’84.
Note: no one at the shop ever noticed a problem with Hitachi 2SC460 transistors in that same case style. There’s an orange one at the Q1 position of the I.F. sub board for the meter, muting, and multiplex switching, which is still going great in 2017 on the ’rents unit.
Hitachi didn’t seem to be having a good time back then, making reliable transistors. The first repair i ever did on this TK-66 was replacing Q6 2SA234B in December 1975, to cure no FM. Back then, not knowing better, i went with a generic ECG 160 R.F. germanium PNP transistor. Worked
great, and still does! The ECG 160 (and i would hope the NTE 160) has 4 leads: emitter, base, collector, and case—just like the 2SA234B.
In 1984, i looked into changing the generic out for a closer-to-original match. Well… the shop didn’t have any, and couldn’t reasonably get any—
germanium PNP transistors were already super-rare in 1984. I left the ECG 160 in place, and it was fine.
Note that AM is picked off after I.F. transistor Q5, thus the failure of Q6 only affected FM. This time around, both AM and FM were intermittently failing. I was fortunate catching Q5 in the act of having an open B-E junction, with far more than 0.3V across it. I was even more fortunate to have a NOS ECG 160 to install! Again, it worked great. If both AM and FM are fading or super weak or dead, odds are high that it’s one of Q3 to Q5 on the I.F. board. Q6 and Q7 are the same transistor later in that stage, for FM only. I normally don’t recommend generics when there are other options, but in this specific case, the ECG 160 (the 1975 incarnation and some from the 1980s) work
great. I would hope the NTE 160, still apparently available for around $4 or so each in December 2017, would be equally good.
Alignment is necessary for best performance. Fortunately for me, i was a professional who had a lot of practice aligning receivers, and was able to buy a Sound Technology 1000A FM signal generator, one of our shop’s (switching) test panels, and other accessories when different repair shops went out of business. Having the great fortune to own a Nakamichi T-100 Audio Analyzer, i have the needed distortion metering to complete the alignment setup. For AM this time around, i set the I.F. to 455 kHz with a frequency counter and a funky Heathkit SG-8 (which itself needed some repair before i could use it), then dialed in the R.F. with over-the-air stations.
Other Problems
Noisy controls
Like the original Hitachi 2SC458s, this machine originally had a lot of little Elna electrolytic capacitors, wrapped in white plastic. White Elnas were as notorious as 2SC458s at our shop: replace on sight! They
love to leak. In the TK-66, that leads to D.C. on the controls and switches, thus noise no matter how clean they are. Gray and other color plastic cover Elnas are not necessarily problematic. I didn’t 100% recap all electrolytics in 2001, but did get most of them.
Re-capping the electrolytics, changing out the preamp transistors, and cleaning the controls and switches should lead to a mostly-noise-free machine.
Power Amp Overheating
Another post (maybe the one Watthour linked to) disparaged the build quality of the TK-66. Well, in my opinion, yes and no. Some parts of it are funky, but keeping in mind that this was an
early solid state receiver, it’s not that bad. The chassis is solid, and i will tell you for sure that the heatsinking of the power transistors is so overbuilt that it can run a long time under conditions which would trash most receivers from the 1970s onward.
Why do i write that? Because one part of the poor build quality was the horrible cheap-*** trimpots Kenwood used. This gets annoying anywhere in the unit, and would-be tragic in the power amp.
Bias trimpots VR5 & 6 tend to open up, making this nominally Class B (maybe AB) amp try too hard to run Class A! This machine of my parents lives (to this day) mounted vertically in a custom-built cabinet. Throughout the early years up until the 1984 repair when i was stunned with how much current the receiver was drawing and how hot the power amp was running, our family took it for granted that the metal front panel would be almost too hot to touch, and there would be heat expansion/contraction metal
tink sounds as it warmed up and once it was shut off.
This is not normal! Replace the power amp bias trimpots with something decent! And then adjust as usual to just barely eliminate or minimize the crossover distortion notch. A properly-biased TK-66 runs pretty cool, with the outside of the output transistor cage being right about comfortably warm, above ambient.
Quirks and Other Notes
* Volume control tracking is bad. More typical of American receivers than later Japanese models, which often tracked very well.
* Rec. Out loading can throw off high frequency response of the entire receiver. If you’re testing frequency response and want accurate high frequency results, you may need to disconnect anything on the record output jacks.
* These receivers were partially hand-built. Weird things sometimes happen. This time around, i found a funky solder joint on the I.F. board, which mostly looked OK but upon closer inspection was not. Resoldering improved AM & FM sensitivity.
* Weirdness #2: a truly bizarre wire lead/flux bridge between preamp input pin 2 and B+ at C19, putting +18V on the right channel Mag/Tape/etc. input. Repairing this cured a right channel noise in all modes other than Tape Monitor. Chances anyone else will ever see this are vanishingly small; the point of reporting it here is to get you to be on the lookout for truly weird things like this.
* Even when nominally in mono, the multiplex circuit isn’t pure mono. Maybe this is a failure i’ve not found. In practice, it works as a form of blending, producing a smoother transition between full stereo and hard mono.
* Adjusting the AM signal meter level may throw off the FM meter, muting, and MPX mono R.F. switch point adjustments.
Mods
AM Pop Elimination Mod
Even with perfect caps, transistors, and wonderfully clean switches and controls, a standard TK-66 still produces a loud POP out of the speakers when switched into AM mode, at any volume setting. Those whacky Kenwood engineers threw a little extra filtering in for AM,
in the power amp. C32 & 33 charge up suddenly when S1-5r closes, creating the pop. The solution/mod: add a 10MΩ 1/4W resistor across each of contacts 1&2 and 7&8 of S1-5r. This pre-charges the capacitors, so they won’t pop. That high a resistance will not have any meaningful effect on frequency response in non-AM modes. Remember: a TK-66 is mid-fi at best.
FM MPX Light Mods
The FM multiplex circuit of the TK-66 is typical of its era: primitive. The Stereo light tends to flicker, which does nothing for its longevity. During the 2001 repair, i studied the circuit and decided it would be wise to double the values of C24 and C25, electrolytics filtering the base and emitter of FM stereo light switching transistor Q7. This was a worthy improvement, with no side effects.
Finally in the last few months of 2017, the original FM Stereo incandescent light bulb which had been working
extremely hard for a majority of its 37-38 years of life burned out. Already fed by half-wave filtered D.C., this was meant for an LED mod. It’s so easy it’s silly:
1) Change R34 in series with the power supply to pin 13 of the MPX board (Q7’s collector) from 150Ω to 470Ω (i used a 1/2W). Interestingly, the unit here did not have a resistor R34—it was a direct wire connection.
2) Hook up a standard red LED with proper polarity in place of the original incandescent lamp. I used solid wire for its stiffness, which allowed the old-timey red LED to point straight at the red plastic lens. This probably would have been sufficient, but just in case the receiver might get jostled, i put some heat shrink tubing (red of course) around the LED and the back end of the red plastic lens. With or without the heat shrink, the end result looks basically identical to the original (with less flicker thanks to the earlier mod).