I recently completed JVC RC-838 JW II restoration and wanted to share some of my experiences and impressions. This was a non-working unit, and it had a broken Pause button. The radio was tuning and had reception; the cassette tape door was stuck, per usual. Good news: the cassette deck mechanics functioned (!): Playback, C-60 tape FF and rewind within 90 sec (spec - this means the capstan belt and idlers were OK), and autostop worked (the reel-to-counter belt was intact). But… It had no sound from any source - tape, radio, Line In. My first treatment was the usual one: I sprayed DeoxIT F5, intended for slider-type potentiometers, into the volume, treble, bass, and balance controls and worked them back and forth. It made no difference. I also noticed that the sliders did not move easily or travel through their full range. In hindsight, that should have been my signal to stop right there. But that's what experience is for. Disassembly: When disassembling the unit, you can skip parts B and C; they aren't necessary for this job, as I learned the hard way. Go straight to section D, "Chassis Assembly." The pesky screw #19: It isn't visible without a flashlight, but it holds the chassis at the level-meter area and prevents the chassis from separating. Also, don't forget to disengage the dampening shaft from the door. Bad news - one tiny wiper came out from one of the sliders. Now I knew what was coming: I took extra precaution and made sure I didn’t lose anything that could come out of the sliders. If one of the wipers flies out into the parts abyss, it’s a goner, and with it goes the boombox. These wipers were cleaned, but not reformed. Rebuilding the ALPS sliders The copper wiper had been deformed by my initial "DeoxIT wiggling" and needed to be reformed. The treble potentiometer had to be removed from the PCB, opened, and repaired. While desoldering it, three more wipers—all mangled—fell out of the bass and balance potentiometers. Fun! The 4 perimeter pins in these pots (ground connection) were twisted at the factory. As a result, you need to straighten them carefully before the pot can be separated from the PCB. The balance pot is soldered in a very difficult spot on the Control PCB - it was a real challenge to get it out. These ALPS sliders are unobtanium, and there are no direct replacements, other than from a donor unit (not an inexpensive option today, and it offers you no guarantee that the donor unit will have them intact). I’d say that most of them will require a complete rebuild. If the carbon tracks are destroyed or the wipers are physically broken, there isn't much you can do unless you're a real wizard at electronics repair. When taking these sliders apart, expect there to be a ball bearing and a small spring inside. Don’t lose them. Take pictures and make sure you assemble them the same way. Also, the direction of the pins matters during installation, and treble and bass are installed on the PCB in the opposite direction. Take careful notes as everything matters. Repairing broken wipers The challenges involved in repairing these ALPS potentiometers are well documented. One of the best demonstrations I've seen is Teddy Hashee's YouTube video showing his work on an RC-M70. The base of these potentiometers is made from Delrin, a self-lubricating plastic. Unfortunately, almost nothing adheres well to Delrin, and most glues fail. If you're in the United States, there's another complication: the Japanese-made two-part adhesive Teddy used in his video cannot be shipped here. Here's an alternative option: micro-drill a hole approximately 1.75 mm deep in the center of the potentiometer base for each wiper. I then cut a roughly 2.5 mm-long stud from a sewing needle and glued the stud into the hole. I used JB Weld Cold Weld, the slow-curing 24-hour epoxy. Slow cure allows you time to properly position the wiper base (essential for proper contact!) and build a rectangular "rivet" head from epoxy over the wiper base, essentially reproducing the original factory attachment. It bonds to the metal stud and the copper wiper base. This is delicate work, and you have to match the microdrill (I used 0.28 mm pictured) and a sewing needle (0.25 mm) so it is a snug fit and has room for a drop of epoxy. Make sure the “rivet head” doesn’t exceed the height of the guide that separates wipers. Also, the drills are fragile, and if you try this method, I hope you don’t break them: the drill bit will get stuck inside the hole, and there is little space to drill another one. I broke 2, and in the process found an even better solution: drill through the base, expose the metal stud by 0.75 mm on both ends, and epoxy it from both sides. I did it on the Bass pot. Time will tell. All that fun came after I had to reform bent copper wipers and clean them from oxidation - they were almost black, so they could not pass the signal effectively. Also, clean the contact legs thoroughly. I used a carbon fiber pen - quick and effective. Cleaning the carbon tracks with Deoxit F5 is compulsory; I used D5 on the copper rail. There is some good news regarding the wiper geometry. You don't have to perfectly reproduce the original angle of the single and double wipers. Once the wipers are glued to the base and reassembled, they will self-form against the rail after several test runs. When reassembling sliders, apply a dab of lithium-based grease on the sides of the metal case. Test the resistance by connecting legs 2-3, 1-3 in treble and bass; it should change from 0 to 100kOhm. The balance slider is 20kOhm, and the L/R channel works in opposite directions - one increases to 20, the other decreases. Once everything got soldered back on, do another resistance check. Function and mode switches The next mandatory work is switch cleaning for RC-838 was 4-way Function Mode (Tape-Radio-Line In- Phono, S703), and Mono-Stereo-Expanded-Biphonic. They are identical. The bad news is the way both are integrated into the circuit design means each can cause issues. I cleaned both of them with Deoxit before desoldering and opening them up, but that didn’t help, practically never does. Both switches failed resistance check. When you open these, expect 2 ball bearings and 1 spring. Take note of the pin locations (4- 10-10-6) and the black plastic switch position - it matters - and you have to put them back in proper order. ALPS JAPAN print on the case aligns with the 6-pin group. Clean all internal contacts on both sides; use a piece of paper wet with Deoxit D5 (per Mend It Mark method) to clean the inside of the tiny wipers. You can see the dirty and cleaned switches in the pic below. I actually had to remove S703 twice, which I'll get to later. The second time around, I also applied DeoxIT L260D grease to the contacts. The second time it did the job. Clean each of the 26 pins that get soldered to the PCB. These sliding wiper contacts are extremely delicate and bend very easily. Once bent, they lose contact, and you'll have to carefully reform them to their original shape. All four legs need to match in both width and height. Here's an important reassembly tip: carefully match the positions of the wipers on the switch contacts with the pattern on the black plastic internal switch piece. Then very gently drop the black plastic piece into place. It should fall in easily and settle into position. Do not apply any force to seat it. If you do, you'll likely bend the wipers and create even more work for yourself. Once it's in place, move the switch back and forth gently to make sure nothing is catching. Broken Key restoration. The key removal required more mechanical disassembly than ever needed. The keys are glued to the metal levers, which limits cover replacement options. After kicking around a few ideas, I settled on creating a silicone mold of a Stop key and using an injection of JB Weld Cold Weld 2-part epoxy to “reform” the missing part into the remaining key body. The mended key came out fine: not museum quality, but it was a good reproduction and functionally strong. Tape Transport service was relatively straightforward, with a few sticky points. I had a problem with intermittent left-channel sound that took me several days to resolve. Yes, it was the switch S703! It is always easy and obvious when it is done. Electrical calibration, etc., needed to be addressed due to major differences for earlier L/LB and later JW II models. I can write it all up in Part II if there is any interest. Apologies for the long post, but success and failure are often in the details. Alex
Very nice write-up, Alex. I'm most curious about your injection molding of the stop key. Since you only mended the part that broke off, did you add any reinforcement? That's something I've not ever done but am curious of the process. Did you align the radio part or was it fine as is? Cheers!
Thank you, Norm! I overfilled the mold slightly with JB Weld, and after the insertion of the Pause key, I only removed the escaping excess of the epoxy from the top (the patterned surface) and sides, allowing for some overage on the bottom. That layer created an additional reinforcement. It is not a thick layer, but it goes deep enough from the breakpoint towards the fulcrum. I did not work on the radio tuning, as I rarely listen to it, and it appears to receive more stations than I can get on my Walkman. Alex
Well, you get kudos for a tough repair. From the picture, one would never guess the key was broken at some point. It's been a few months since I've aligned a radio, the last one was when I restored a Sony TFM-8000W, that was a full recap and alignment of all the bands. Would you mind sharing what products you use for the silicone mold? I'm still intrigued. Cheers!
Thanks again, Norm. I used a Silicone Mold Making Kit, 21OZ Fast Cure Translucent Silicone Rubber (link to Amazon where I bought it - no affiliation). Here are the Stop and Pause (pressed) keys up close. I could have used a better-suited paint to match precisely, but I did not want to fuss over it. The key visual reference for the mold making was the first row of the pattern. Alex
Very impressive indeed, Alex. Thanks for sharing. Gives me inspiration to try something like this in the future.
@worldwide137 Thank you, Norm. The feeling is mutual - your successful work inspires me too... Part II Tape Transport Service As you can see from the motor label, this unit was manufactured on January 23, 1981, which puts it close to the end of its production run. Compared with earlier models, this version uses Meta Perm PB/REC and SA erase heads, as well as a different type of plastic for some components. The tape transport is extremely well made: full metal construction, very few plastic parts, and remarkably precise tolerances. While working on this unit, I had the distinct impression that JVC’s engineers spared no expense in building this model. You can also clearly see the broken Pause key. Tape transport disassembled. The old (Molykotte?) grease, which was still in surprisingly good condition, was removed and replaced with fresh grease. All the rubber idlers appeared to be in good condition and did not warrant replacement. I cleaned and treated them with CaiKleen RBR Liquid. I also swapped the reels, which are identical parts, since the right reel gets considerably more use over time. This should help ensure long-term reliability. The original pinch roller cleaned up well, although it is a candidate for replacement in the future. By the way, the inner shaft diameter is 2.5 mm. The pinch roller dimensions are 13 × 8 × 2.5 mm (the original measures 12.8 mm). The tricky part was reassembling the pinch roller arm. It is an awkward assembly to work on, and it is best done with the capstan removed. The spring is very strong and must sit perfectly vertically on the shaft without binding. Otherwise, the pinch roller can skew the tape and cause it to wander up and down on the capstan. It took me several attempts to get it positioned just right. That’s what experience is for. Pinch roller pressure measured at 350 g. I replaced the reel-to-counter belt, which controls Auto Stop, as it was the loosest belt in the transport. I used a Russel Industries 5.2 SCY belt originally intended for the Sony Walkman WM-2. It was too tight for the Walkman but fit perfectly here. I kept the original capstan belt, as it was in surprisingly good condition for its age. To service the flywheel, you only have to remove two screws on the left. The screw on the right is spring-loaded and sets the flywheel clearance. In hindsight, I should have done that; doing so would have saved me some adjustment work later. An incorrect gap will result in higher W&F. The flywheel and capstan are massive and beautifully made, but I still had my doubts that this transport could achieve W&F of 0.07% WRMS. The pause mechanism—the part that still had the old grease—gave me the most trouble. It would not latch 100% of the time. The mechanism is extremely intricate. I took it apart, regreased it, and reassembled it several times until I finally got it working reliably. I studied the schematics, which indicated that I was missing a washer, although I don’t remember seeing one when I disassembled the mechanism. I fabricated a washer myself, which helped somewhat, but the mechanism still did not latch reliably. After numerous trial-and-error attempts, I figured out that the spring with the two widely spread legs was causing most of the trouble because it was not seating perfectly. I made a small jig from thick paper with a 2 mm hole to keep the spring from shifting when I placed the Pause linkage over it and secured everything with the e-clip. That worked. The paper jig, of course, tore off afterward. Eventually, I discovered another factor contributing to the problem: the key rod was not fully inserted into the bracket. This mechanism has extremely precise tolerances. A word of advice: if you don't absolutely have to take this linkage apart, don't. (Fast forward: when I got a proper service manual for JW II, I saw that there was no washer in this linkage, but there is one in L/LB.) I kept mine as I don't ever want to do that again. Transport assembled, and everything functioned well. The moment of truth… Chasing electrical gremlins The rewind and fast-forward were silky smooth and remarkably quiet. Really beautiful. Everything seemed to work, except there was no sound in the left channel in Stereo, although it was present in Mono. At least all the slider pots were working. At first, I suspected the Volume slider, but when I moved the chassis, the left channel came on intermittently. That was a useful clue. The PCB on this unit is unlike any I have seen before: there are many press-fit components mixed in with the soldered ones, and the traces are not easy to follow. It was a real challenge. With the free service manual for the L/LB model that I had, I couldn't clearly see the traces either. It was time to get the proper manual. The new, proper service manual revealed quite a few interesting things. There were many mid-production updates made to the JW II, including changes to the specifications. Curiously, the W&F specification changed from 0.07% to less than 0.2% WRMS, along with changes to the amplifier adjustment voltages. For signal tracing, I injected a 400 Hz sine wave from Line In. I was losing the signal at the base of transistor X103. The new service manual helped immensely in tracing the path from Line In to X103. I was losing the signal after C102, after it reached pin 21 of switch S703. Someone with more experience probably would have solved this in five seconds, but “it ain’t me.” It took me three days to find this “lost signal.” Moving the switch slightly from side to side while in the Line In position intermittently caused the sine wave to appear on the oscilloscope. Bingo! One more desoldering and service of switch S703, and the left channel was back. After that, it was time for the basic tune-up: Azimuth adjustment: 10 kHz HPR tape — check. Motor speed: HPR W&F and Speed (3150 Hz) — check. I was impressed by how stable the speed was on this boombox. Playback levels: The service manual recommends using a VTT-664 1 kHz tape with a reference flux level of 160 nWb/m (listed in some service manuals as 16 μWb/mm). I used an HPR equivalent. VR201 (R channel) was adjusted to 300 mV, while VR101 maxed out at 292 mV. This indicates some ageing of the components, but the difference is acceptable for now. At this point, I went back to point 1, readjusted the azimuth, and then repeated point 3. I did not adjust the recording current, bias frequency, erase and bias current, or the biphonic processor, as I am not planning to use those functions. Recording quality was acceptable as it was. Playback frequency response. It is neither spectacular nor terrible; it is within the specifications. Frankly, I expected slightly better performance from a boombox of this build quality. At spec. I think W&F could be improved a little with a new capstan belt and pinch roller. Time will tell. The Verdict: Overall, this is a much better-looking than performing boombox. It has good sound, although it is somewhat lacking in bass. However, this is not because of poor frequency response; the response is quite ample. Perhaps better 16 cm midrange speakers, or a different amplifier design, could have turned this into a much higher-performing model. But I will leave it be, as original as possible down to the last e-clip. It may be light on bass, but it is absolutely loaded with looks. For the conclusion, I am going to quote a well-known US Nakamichi tech, because I can't say it better: “There are many machines that impress me for construction quality or performance, but they don't always go hand in hand. Many machines are design masterpieces but have poor performance, or the opposite, superior performance with design that leaves me unimpressed.” The JVC RC-838 JW II is a clear example of such a machine. I would call it a design masterpiece with average performance.
