I'm at the point where I need to seek help. I've fixed many problems with this unit and hope this is the last. I bought it cheap and I think it was someone else's donor unit. For example I found levers bent deep inside the mechanism. I have no idea how that could have happened. Problems: It switches by itself from forward play to reverse play after a few seconds. Fast fwd and fast rwd work for a few seconds, then the tape drive stops, with or without a tape loaded. By swapping modules from a known-good unit I have narrowed the problem down to the main board. Loading the motor by pressing on the motor pulley did not change the problem. The various related switches on the main board have been removed, cleaned and/or replaced. I replaced all large capacitors in the Logic section with not-known good caps from my donor unit. No effect. Q120 looked discolored. Switched with a not-known good Q120. No effect. By this time I had the unit down to the minimum parts for testing the motor direction. Equalizer and head cables desoldered to avoid damaging the cables. I tried pressing on the components with a toothpick, one by one, while testing. No effect. I put the unit in the freezer. While it was still cold and with power applied, the motor turned and reversed by itself as soon as power was connected but no switches or Play activated. This indicated the problem was temperature related. After the unit warmed up, I reflowed some resistor and capacitor smd pads. No improvement. The motor still reversed after a few seconds in fwd play. Fast rwd and fast fwd stopped after a few seconds. The exact time to failure varied a little. I only reflowed a few of the pads because I don't have a sharp soldering iron to easily get at all the pads. But I got the idea to heat transistors with the soldering iron, one by one, and testing. When Q121 was hot, it cured the problem. Q121 is in the Logic section and associated with auto reverse. This improvement with heat was consistent with general freezing making the problem worse. I replaced Q121 with same transistor but not-known good, from my donor unit. No effect on the problem. I continued testing Logic section smd components with heat applied and, accidentally, power still connected. D106, D111 and D120 caused more random motor direction and caused the motor to make a buzzing sound. These diodes are related to fast fwd, fast rew and auto reverse. Before I replace these diodes with ones from the donor unit... Does it make sense they are the problem? Did applying heat to them with power connected ruin them or just activate them? Could all 3 diodes be bad at the same time? Is it likely both Q121's are bad? I don't know enough to test the diodes in-circuit. I don't have voltage specs for testing the transistors. And I'm concerned about damaging things while testing. The alternative to solving this problem is to replace the main board. And I don't have a working spare. Thanks in advance for advice.
That sensor CB with photo detector was replaced together with main PCB ? An autoreverse loop is tied to the reel sensor signal, either the sensor is not producing a signal or a circuit problem. Regarding the potential failure of diodes, while possible it's very unlikely so many active components failed at the same time. Given the Q121 does change behavior with increased temperature, that area is worth more investigation. It is to be kept in mind though this doesn't necessarily mean the transistor itself is the problem. Increasing the junction temperature in Q121 does 2 things: - decrease the Vbe (so the junction starts conducting earlier, like 0.5V instead of 0.6V); - current gain becomes higher for same collector current; Something I would check is C179, even replace it if convenient to do so. Another thing to check is if you have good Vcc (3V) at the end of R173 connected to collector of Q111. In regard to testing diodes and transistors in circuit, the most basic test that can be done is this: - on diodes measure between anode (red probe) and cathode - marked with a line (black probe) with meter on diode mode. You should get a reading of 0.6-0.7V voltage drop. Should read OL when measured backwards, but may read some value depending on what's in parallel with it. - on transistors you do the same things between BE and BC (NPN), while measuring EB and CB (PNP). You should get a similar 0.6-0.7V reading and both junctions should read about the same.
Thank you for the assistance. I believe the sensor was replaced when I swapped in the motor from a different unit. I also tested by cold-soaking the soldering iron and cooling components. This only led to additional perplexing results. But of course putting components outside their rated temperature parameters will do that. After trying all sorts of things, and minimally assembled to play tapes, the unit played, ffwd and frwd perfectly, except one channel was dead when playing tapes. I traced that to an unsoldered terminal on the head's flat cable and fixed it. To my consternation, the reversing problem was back, and ffwd and frwd again stopped after a few seconds. This unit has worked properly before. Poking and prodding at components, and flexing the main board, have not exposed any problem. I'm unable to correlate any circumstance to when it works I thought about how the mechanism works. Forward play engages the strobe gear and its sensor. Probably this is used for speed regulation and for detection of forward tape movement. If the strobe wheel is detected as not turning, this is interpreted as the end-of-tape. So the motor is told to reverse. The strobe gear is not used in reverse play. Why not? Maybe this mechanism is derived from a single direction mechanism and for some reason it was too much trouble to include the strobe gear and the sensor. I think it means speed regulation is less precise in reverse play, which may be why these units often have more w&f in reverse play than forward play. Perhaps the clutch function in gear assembly A is to relieve tension on the tape while tape direction and end-of-tape decisions are made. Otherwise the motor could be held at a stop while powered, which would be bad for it, or the tape could be pulled off a cassette reel. So how is tape direction reversed, or end-of-tape sensed (for continuous play) at the end of reverse play? I don't know. Perhaps current to the stalling motor. Back to the reversing problem. I considered the sensor board may not have been installed for my last test. So the logic would assume an end-of-tape situation, and reverse the motor. Then I found one of the wires on the sensor board was loose, but the break was hidden under the insulation. Could this be intermittent? But the wire was to the leaf switch engaged when the Play button is engaged. So it should not be involved in the errant reversing and failures of ffwd and frwd. None of this explained why Play, ffwd and frwd worked briefly. Next I will try a different sensor board. I doubt this will help because it is not associated with the ffwd and frwd failures. I have another defective HS-J09 in transit. As I repair it I can swap components with this one to again isolate the problem. In the meantime I will proceed with the tests you described. I am however very wary of putting current through transistors in a way that damages them.
Sounds like sensor stopped working. The sensor board should send square wave like signals from reading the alternate black/white pattern on the wheel. I think it’s not used for speed control, just to check whether tape is running.
The strobe gear is used for reel rotation detection only. In some decks (which have such sensors on both reels) it can also be used to determine the tape length and approximate current tape position in minutes based on the difference in speed between the 2 reels. It is, however, never used for speed regulation for a couple of reasons: reel speed varies during the length of a tape, signal resolution is very coarse and also not correlated to capstan speed. For motor speed regulation there are 3 ways it's done: 1. based on motor current draw (higher draw, higher voltage to compensate); this is by far the most used system. 2. based on motor FG coil that reads the motor speed directly and uses it for negative feedback loop control. SONY TPS-L2 has this system. 3. based on capstan FG coi, that reads the capstan speed and uses it for negative feedback loop control. DD (both disc drive and direct drive) systems use this. 3rd system is the most stable, but it needs the motor to either be the same as capstan/flywheel (direct drive system) or to be coupled directly to the flywheel with very little movement between the 2 (disc drive). A belt drive mechanism won't work in this configuration because belt is elastic and the error signal's phase can't be matched with the motor drive. Regarding FWD and REV, the strobe gear is used for both despite it's just one. That's because when the tape is moving, both reels are rotating, so it's enough to read any of the 2 not both. Some walkmans use 2 strobe gears and those could detect if tape is not spooling properly on both sides and can stop the unit if only one of the reels is not rotating. But this is rare. Exactly, that's its only purpose and works on both FWD and REV as when tape stops at the end, both reels stop.
(this post is rife with errors. The strobe gear operates in both directions when a tape is loaded.) To explore this further... I just checked to make sure, and the strobe gear only turns in fwd Play and fast fwd winding. The main gear assembly A turns during all operations. Gear assembly A always turns at "tape speed". Tape speed (and capstan rotation) in Play has to be constant in both directions. The strobe gear engages through other gears to gear assembly A. Takeup reel speed during Play would indeed be at varying speeds, and I'm not sure how this is done. Perhaps this is the function of the clutch in gear assembly A: to allow slippage in the gearing. Gear C in gear assembly A is fixed with respect to the main gear A. Gear B is free to turn by itself, limited by friction by being pressed against the felt washer. This would allow the takeup reel to vary speed during Play to accommodate the varying effective diameter of the takeup reel. Takeup speed during Play is held in check by tape speed, which is why, if the pressure roller is dirty or glazed, the takeup reel can pull the tape past the head faster than Play speed. Which in turn causes the chipmunk sound that's a symptom of a dirty or glazed pressure roller. Fast rewind and fast forward can be done at constant takeup reel speed. The difference in pitch we hear during fast winding is because the freewheeling reel changes speed. So because the strobe gear has to turn at constant belt speed during Play, it could be used for speed regulation. I speculated that is so because when I have attempted to adjust tape speed on a few of these players, it appeared to me that as long as I was in the right ballpark of forward speed adjustment, the speed automatically centered on the correct speed. I don't have equipment to verify this with instrumentation. It turned out my tape speed problems were due to coating exfoliating from the tape, or dirty pressure rollers. But since the strobe is only used for forward direction tape travel, speed adjustment for reverse play would have to be entirely manual. Using the strobe to regulate tape speed presumably would require a timer in the circuitry, and I don't know if there is one.
Because the take up gear has to run at a variable speed, there has to be somewhere in the system to allow slippage, either at the take up gear or some center gear. the strobe (Gear H) in attached diagram, is permanently connected with the take up gear, so it will spin even in the REW mode, as long as to there is a tape. When there’s no tape, the strobe doesn’t spin and REW fast or play cannot sustain, the mechanism will trigger reverse or stop (motor) by design. Unless you push the REW continuously. What you have observed can be overridden by press and hold the REW button.
I loaded a tape and tested again. The strobe wheel does indeed turn in all operations. So much of what I just wrote is garbage. With or without a tape loaded, forward Play still reverses after variable delays, but within a few seconds. Similarly ffwd and frwd stop after operating for a variable few seconds. I'll try swapping in a not-known-good sensor next.
A bad sensor makes sense as a culprit for the symptoms. The substitute not-known-good sensor had no effect, with a tape loaded. I don't want to try a sensor from a known-good unit because disassembling these players has too much risk of collateral damage. Another non-working unit should arrive in a few days and I will try the sensor from it.
Ok, I've done the R173 -> Q111 check, and the diode testing. I hope my annotation makes sense. R173 -> Q111 0.013v Play not engaged 2.7v with Play engaged D103 AC/CA, AC/CA: 0.6/0, 0.6/0 D104 AC/CA, AC/CA: 0.6/0, 0.6/2.7 D105 AC/CA, AC/CA: 0.6/0, 0.6/0 D106 AC/CA, AC/CA: 0.6/0, 0.6/0 D107 AC/CA: 0.6/0 D108 AC/CA: 0.6/0 D109 AC/CA: 0.6/0 D110 AC/CA: 0.6/0 *D111 AC/CA, AC/CA: - red probe on anodes, black probe on cathode: 0.6/motor runs with 0.6v reading - black probe on anodes, red probe on cathode: 0/0 *D112 AC/CA, AC/CA: - red probe on anodes, black probe on cathode: 0/motor runs with 0.6v reading - black probe on anodes, red probe on cathode: 0/0 D113 AC/CA, AC/CA: 0.6/2.9, 0.6/2.6 D114 A/AC, AC/A; AC/C, C/AC: 0.6/1.3; 0.6/1.3 D115 AC/CA, AC/CA: 0.6/3.0, 0.6/3.0 D116 AC/CA: led (LED seems ok) D117 on motor board?: D118 on motor board?: D111 and D112 are attached to the ffwd and frwd switches. I have no idea if testing D111 and D112 activation of the motor indicates a problem. Earlier, checking C179 was suggested. I had already replaced it with the C179 from a donor unit. No effect. But I haven't tested either. No obvious problem with it.
Results of testing the transistors: Q107 NPN, BE: 0.68, BC: 0.6 Q108 NPN, BE: 2.5, BC: 2.5 Q109 NPN, BE: 0.72, BC: 0.72 *Q110 NPN, BE: 0.46, BC: 0.64 Q111 PNP, EB: 2.4, CB: 2.4 Q112 PNP, EB: 2.8, CB: 2.8 Q113 NPN, BE: 0.74, BC: 0.74 Q114 NPN, BE: 0.74, BC: 0.74 Q115 NPN, BE: 0.25, BC: 0.26 Q116 NPN, BE: 0.26, BC: 0.26 Q117 NPN, BE: 0.25, BC: 0.25 *Q118 NPN, BE: 2.6, BC: 1.6 Q119 NPN, BE: 0.74, BC: 0.74 Q120 PNP, EB: 0.7, CB: 0.7 Q121 NPN, BE: 0.7, BC: 0.7 Q122 NPN, BE: 0.74, BC: 0.74 Q123 PNP, EB: 2.6, CB: 2.6
Q110 gives the same readings on the main board in my donor unit when it is powered up, as on the malfunctioning unit. But Q118 gives 2.9, 2.9 on the donor unit, instead of 2.6, 1.6 on the malfunctioning unit. So I'll try installing Q118 from the donor unit in the other one. We'll see what happens.
Q118 is Auto Reverse Mute according to the manual. I guess it’s affected by, but not likely causing the issue. If the sensor board works, at the “@“ sign on the diagram there should be ac signal generated by the sensor. Sometimes these old Aiwa may develop invisible cracks on the circuit board, more likely than bad transistors. Good luck!
The problem with locating cracks in the traces is mainly a physical matter. The transistors and diodes are on one side of the board, and the resistors and capacitors are on the other side. The traces go back and forth through the board. And the problem is intermittent. It may be easiest just to find a replacement for the main board. Is there a place where I can short out the sensor to give a constant unmuted signal from the sensor?
I agree with @novrain2012 that it's more likely problem is caused by a cracked trace or open via rather than faulty transistors. Would add it's also really unlikely 2 transistors are faulty, both the original and the one from the donor (even though it's not known good). That test alone already suggests it's not the transistor that's the problem. Also it's possible fact that you heated Q121 actually had an effect on the board (which may be the real reason problem disappeared) - copper expanded and the open circuit closed. Cracks in the PCB can be tested by pushing (to the point of bending) on the PCB to see if you obtain different behavior. This way you can narrow down a specific area of the fault. Regarding the sensor, stated from the beginning that if the sensor was not swapped together with the PCB it can be excluded. It's again unlikely 2 sensors are bad, even if the replacement is not known good. Only way to emulate the sensor is generate that squarewave (shown in bottom right of picture of 1st post) with a function generator and inject it into the circuit.
I'll do another round of poking and prodding while activating the ffwd and frwd switches. Ffwd and frwd halt after a short time, and the time to failure quickly diminishes with repeated trials. I believe this is the same problem as the unwanted reversal in Play. And so far as I understand how the thing works, both problems may be caused by a failure in the sensor circuitry. The varying time to failure suggests to me the problem is related to temperature, and if so, a break would be located in or adjacent to a component that warms up. Probably my prior tests of heat and cold sinking were not informative because I was exceeding the performance limits for the components. I can try that again with more gentle temperature changes. An alternative is to give up and replace the main board. I examined the main board from my donor unit, to judge whether it would be easier to fix it up than repair this one. It had suffered considerable corrosion damage. I think I could fully populate it with good components and bridge one rotten trace. The adjustable tuners or potentiometers seem to be most damaged by corrosion. I worry about tuning them after replacements are installed. Maybe the service manual addresses adjusting them. But first, more work on the one we're discussing.
Just an update. I went over everything in the Logic section of the main board, poking and prodding and pressuring each component while doing tests without a tape loaded. No smoking guns. I reflowed every solder joint I could get at in the Logic section. No change. But with a tape loaded it ran fine. Play did not reverse by itself and ffwd and frwd did not stall. I thought it was fixed. Next day, with nothing further done on it, the problem was back as if it had never gone away. I swapped transistor Q118 from the donor unit. No change. I put the original sensor board back in. No change. I'm going to take a step back and reverify the problem is in the main board. Doing so awaits arrival of a unit I have on order, because it will have to be disassembled anyway.
I think it's fixed, and the issue was the motor. Its original motor was bad. The motor I replaced it with had an intermittent problem of randomly stopping or reversing direction. A third motor seems to have fixed it. I found this out by having the second motor connected but sitting loose on the table. With steady power applied it would not run smoothly. It would twist around as it reversed, or caught and released. Hopefully this completes the refurbishment of this HS-J09. Other than my donor unit, it had more things wrong with it than any player I've bought. I suspect someone else had tried to fix it. I found things like bent metal components deep in the mechanism. It's very satisfying to return it to full functionality. During this effort I learned a lot about things like checking and soldering smd components. I thank those who offered their expertise. You too have a hand in the second life of this unit.
