View Single Post
Old 30th Sep 2021, 10:18 pm   #1
unitelex
Hexode
 
Join Date: May 2005
Location: Greater Manchester, UK.
Posts: 492
Default Nakamichi 582 cassette deck -unexpected failure modes

Since it too quite some time to diagnose a few problems with this set, I would like to share my findings and hopefully others can save hours of debug time.

This is a nice high end cassette deck from 1979...1981
The mechanism is quite precise and stable.
The audio design seems to be very good.
The logic control board design is, well complex and somewhat mind boggling. It uses straightforward 4000 series logic only for latching the modes, the remainder of the logic is done by many transistors resistors and diodes, oh and filament bulbs.

Problem 1: Bias Oscillator
The particular example I repaired came in with overheating resistor R319 in the feed to the bias oscillator. The oscillator was drawing way too much current but still oscillating. Its waveform was badly out of shape (sorry don't have an image) but it looked like a half cycle reasonably correct for the specified 105KHz, but the other half cycle stretched and ramp-like. Transistors and resistors checked out OK, I feared the transformer might be at fault but it turned out to be one of the base-collector coupling capacitors, a small orange 4.7nF plastic film? This is a push pull design so both such capacitors were changed for good measure and the oscillator started working correctly. The frequency only needed a slight trim via the transformer slug and it gave a perfect sine wave at 105KHz.

Problem 2: No audio during play (Muted)
The second problem was a lack of audio, or should I say extremely quiet, as it could be heard slightly on max gain.
After some investigation it became apparent that the logic was muting the audio. This machine has extensive muting logic, clearly the manufacturer wanted to eliminate all pops, clicks and other noise during switching or any mode other than Play. The manual shows a diagram of the muting logic, but it became apparent that there is more to it than the manual describes. In fact a very likely failure mode of a filament lamp which illuminates the stop button in stop mode can be a cause of no audio in play mode. This is because the cold filament of the stop lamp (not illuminated during play) provides a pull up current for a logic signal that tells the mute logic that the systems is not in Stop mode. Without this filament the mute logic will mute the audio during play. This is neither obvious nor is it described in the manual, but it is probably the most likely failure mode if you think about it. A filament lamp illuminated at all times the machine is powered but in stop, many hours... This will fail sooner or later.
Changing this tiny filament lamp is awkward as the button PCB is on a cable with little slack and no connector. Getting to the back of the board is difficult even with the board unscrewed from the chassis. The "grain of rice" bulb is a low current type. The remaining good ones measured ~50 Ohm cold. I tried a similar sized bulb I had in my stock however it drew too much current (~20 Ohms cold filament) and the machine logic did not respond correctly. I think the driving transistor did not have enough gain to saturate. I added 33R in series and this sorted the problem. The bulb brightness then matched others in the set. It might be better to replace these with LEDs however one would need to add a parallel resistor to ensure the logic sees the high pullup condition correctly. The bubs are mounted in a plastic or rubber grommet (maybe it was rubber and became hard like plastic due to heat/time?) This crumbled when I removed the bulb so I re-fabricated something similar using heat shrink sleeving. See photo

Problem 3: LH channel clunk and VU meter hard over during source select or Dolby mode select switching
I now had audio during play however I noticed the LH channel had an issue when switching source/tape mode and when changing Dolby setting. There was a step transition clunk and meter hard over momentarily (LH channel only). There was also some strange buzzing. The scope confirmed a huge step change in voltage, and also the buzzing showed up as a rectangular wave. This was traced to originate before the source select but was not present on the first pre-amplifier so was confined to the Dolby circuit on the LH channel.
I feared it might be a failure of the uA7300 IC, possibly difficult to source or expensive... But since these are socketed a quick swap of LH and RH chips left the problem still on the LH channel so both ICs must be OK. Comparison of DC levels on each IC pin between LH and RH channels did not show any appreciable differences. It turned out to be caused by C145 33nF orange film cap, almost no capacitance (hmm, pattern emerging here.)

Problem 4 occasional distorted sound and partial muting (Both channels)
The set had been working well under long term test but I noticed some brief second or two of distorted sound on both channels when initially starting Play.
Then after extensive testing the distortion came back from time to time during play. What could cause erratic yet identical performance degradation on both channels? Power supplies were checked and found to be stable.
So the mute circuit was suspect again. I found the mute control signal, normally either -8 .... -10V (non mute) and +8 ... +10V (muted) was drifting from a few volts negative to a few volts positive. The freezer spray soon narrowed in on the sensitive area, Q424 which drives the mute signal. Cooling this transistor solved the problem temporarily until it returned to room temp. Warming it made the problem worse. The lower temperature results in lower gain. Clearly it was being driven to a marginal state. The transistor tested OK but was changed anyway similarly the preceding stages Q425 etc were tested. Resistors checked out OK.
I suspected some leakage elsewhere. The mute circuit is complex and links to so many other functions. I tested all parts I could find in the path with no obvious culprits. This is all on the logic board which does not have any of the orange polyester caps which I had now come to suspect. It does have a few small electrolytics which were changed for good measure. This did seem to improve the situation quite a bit. It was now several volts negative during play. I was not entirely satisfied though because more heat applied from the hot air gun around Q424 eventually brought the problem back. I tested all transistors in the mute function but found nothing obviously wrong. So I adjusted the threshold on Q424 base by reducing R483 from 10K to 5K which resulted in closer to -10V on the mute signal during Play. Following this change the mute circuit worked well regardless of ambient temperature. I left it at that.

Problem 5: scale illumination lamp failure
More tiny filament lamps are used to illuminate the cassette window and VU meters. These had failed and there was evidence of heat damage to the surrounding plastic. I replaced all these with some surface mount white light LEDs built up to fit the small aperture. A couple of such LEDs back to back arranged to shine sideways, in series with a resistor replaced a single filament bulb. Three of these were fabricated with lead wires and soldered to the narrow PCB.

I cleaned and lubricated the mechanicals and cleaned and demagnetised the head which brightened up the treble considerably.

Summary
I was surprised to find failures of several capacitors of the type normally considered very reliable - These orange types I assume to be plastic film. Failures of different values also. So I eventually continued to change all of the caps of this type in this set for good measure. There were 15 of them. Out of the 15, I found a third suspect making the failure ~20% This seems to be more than a random component failure. The failure mode is very low capacitance, a few pF or even O/C Suggests to me to be some internal corrosion of the foil or the lead to foil bonding. Not sure of the brand, maybe somebody recognises it - they are marked with what looks like "N" inside a circle. Photo attached

During the diagnosis I had changed a lot of small electrolytics, branded Shoei, mostly light blue a few orange coloured but later testing out of circuit showed these to be in good condition. I tested but did not replace the grey coloured electrolytics, these are non-polarized types and tested OK anyway.


So if I am presented with another of these to repair, I would first replace any failed lamps (especially the Stop button one) and then replace all the orange film caps. And check the mute signal voltage! Followed by the usual clean, lube and demag.
Attached Thumbnails
Click image for larger version

Name:	Orange fim caps and one capule bulb.jpg
Views:	429
Size:	100.0 KB
ID:	242448   Click image for larger version

Name:	Bias Oscillator.png
Views:	324
Size:	54.8 KB
ID:	242449   Click image for larger version

Name:	Mute transistor.png
Views:	284
Size:	53.4 KB
ID:	242450   Click image for larger version

Name:	Dolby circuit.png
Views:	265
Size:	29.5 KB
ID:	242451   Click image for larger version

Name:	Button PCB.jpg
Views:	249
Size:	32.4 KB
ID:	242452  

unitelex is offline