Philips CD 100 - mysterious CD player fault

[ebastler]

I have a Philips CD 100 which I bought in 1984 -- the first type of CD player sold in this part of the world, and this particular unit was my first CD player back then. Due to its history, the impressively over-engineered optomechanics, and its industrial design I am quite fond of it and would like to continue using it. But is has started to act up:

• It sporadically emits loud crackling noises overlaying the music, sometimes followed by total silence until the sound spontaneuously comes back.
• This always affects both channels at the same time, but the crackles are different/uncorrelated on both channels.
• Mechanical bumps don't seem to start or stop the noise.
• Once I had disabled the muting relays (see below), I realized that the noise can also occur while the player sits there idly (powered on, but not playing a CD). That's a relief, since it probably means that it is neither the laser nor the pickup/tracking circuit.
  

Due to the sporadic nature of the problem, I have not been able to take any measurements while it was occuring. It might well be a thermal issue which does not want to show up while the unit is open, but I might be imagining that... What I have done is rule out the usual suspects mentioned in various threads on the net:

• I have replaced all electrolytic caps. Not something I would normally do pre-emptively, but the CD 100 does run quite warm. No change.
• I have shorted out the mute relays. (Two reed relays which normally interrupt the audio output while no music is playing, to impress the listener with "zero noise".) No change.
• I reseated all socketed ICs, resoldered quite a few joints in the area which sees heat and thermal expansion (power transistors in the back), and inspected all joints under magnification.

At this point I'm hoping to find someone who has seen this particular problem in a CD 100 or other early Philips player before. Or someone who has a creative idea where in the signal chain an error with the behaviour described above might lie. The "same time on both channels but different noise", as well as the fact that it also occurs in idle mode, seem quite characteristic -- but I can't pinpoint the problem, and seem to be largely limited to trial-and-error since it is so sporadic.

Many thanks for your help!

[Runco990]

Have you pinned all the ground plate throughs?  Another common issue.  I had to do mine, except I actually had a pre-production unit!  Sold it well.....

You can tap on the boards to see if it helps any before you do it.

[LA7SJA]

Have you checked all the powerrails with a scope? I have seen noise coming from the power on the CD100. I have the complete service manual and a sett of IC's and alignment tools and a complete laser assembly (from my stay in Philip Eindhoven learning to service CD) som where in strorage, but I need som time to locate it.

LA7SJA

[ebastler]

Thank you both for the quick replies, Runco990 and LA7SJA!

Have you pinned all the ground plate throughs? 

The PCBs in my unit are both single-sided, with components and jumper wires on the top side and traces on the bottom side only. So there are no plated through-holes, nor non-plated holes with rivets. Did you have a different version, or did I misunderstand?

Have you checked all the powerrails with a scope? I have seen noise coming from the power on the CD100.

In the normal (non-noisy) state, the power rails look fine. But maybe I should lie in wait and try to catch them when the noise is appearing? Do you think a power brown-out could create the uncorrelated crackle on the two channels?

[Runco990]

Mine had a ground plane, but it WAS a pre-production model.  Didn't even have part numbers on the laser and such.  So you are good if you have a single sided board.

Does it play a CD-R?  If it does, it's NOT the laser.  Although if it plays redbook cd's fine but has some trouble with CD-R's... the laser could be getting tired.  Really only scoping the RF Eye pattern will tell you.  The player may need a touch up align on a scope.

I have ALSO observed the spindle motor being lazy.  This can cause scratching sounds in the audio as well.  I have fixed 1000's of philips based players... not MUCH goes wrong with them.

The CDM-0 has a brush motor and is why I sold mine.  The later one had a CDM-1 with the indestructible hall motor.  I have found that taking it apart and oiling it and then re-adjusting the spindle height properly can fix many "problems".  This is often the case with Revox and Studer players as well.  You may have to try turning the bearing screw 1/4 turn clockwise if you have a CDM-1 to test if the spindle has worn a divet into it.

I'm not sure what to tell you.... When I have these things in front of me, I can usually figure it out in minutes, mostly by LISTENING to the player for the unique sounds it makes, but on forums it's really tough.

[ebastler]

Does it play a CD-R?  If it does, it's NOT the laser.  [...]
I have ALSO observed the spindle motor being lazy.  [...]

As mentioned in the OP, I am confident that it is not the laser, tracking, or drive (fortunately!), since the noise also occurs while the player is sitting idle without playing a CD. (The mute relays are bridged, since I had suspected them to be flaky and cause the noise.) That would suggest that the problem lies further downstream, on the demodulator/decoder board, right?

As suggested by LA7SJA, I have looked at the power rails again. All at their nominal values, and nice and clean. And of course, now that the player sits on the bench, opened up and with the scope connected, the issue has not shown up over the past three hours... Can't provoke it by heating things up with a hairdryer, nor by tapping the boards and connectors. Sigh...

One thing I only noticed now -- probably unrelated to my issue, but it makes me a bit nervous: The SRAM connected to the SAA7020 is by far the hottest-running IC. While it won't burn my finger, it is unpleasantly hot to the touch. (That's with the player just sitting there, powered on but not playing a CD.) Is that normal?

[shakalnokturn]

Yes the SAA7020 do run hot in normal conditions current consumption and current variations at this IC are probably the highest of all, it's worth checking supply noise at the +5V decoupling capacitor for this one.

You may want to check the UNEC error flag at pin 36 while the output is crackly.

A bad RAM is not uncommon and could cause those symptoms.

[ebastler]

Yes the SAA7020 do run hot in normal conditions current consumption and current variations at this IC are probably the highest of all, it's worth checking supply noise at the +5V decoupling capacitor for this one.

You may want to check the UNEC error flag at pin 36 while the output is crackly.

A bad RAM is not uncommon and could cause those symptoms.

Thanks! It looks like the RAM (or, if I am unlucky, the SAA7020) may indeed be the culprit:

During further testing yesterday, the CD 100 did me the favor (?) of switching from crackling to permanently silent audio. At least I can measure stuff now...

• The eye pattern looks fine.
• The SAA7010 demodulator output looks plausible -- all clock signals as specified, and a busy stream of data on pin 27.
• But the subsequent SAA7020 always outputs data words with all 16 bits high, and always sets the UNEC flag high too.

Swapping the TMM2016 SRAM would be a possible next troubleshooting step, but I am a bit reluctant to rip it out. It is one of the few unsocketed chips. Ideally I would want some more evidence that the SRAM is the likely problem.

• I did look at the /OE and /WE signals from the SAA7020 to the SRAM, and they look plausible -- nicely alternating, never both low at the same time.
• The 5V supply looks fine at both chips -- 5.0 V DC, approx. 35 mV rms AC at the SAA7020 and 15 mV at the SRAM.

Any other idea what I could test to confirm that the SRAM is the likely culprit? Thanks for your help!

Edit: To partially answer my own question -- I should probably check the data lines during /WE, to confirm that a mix of 1s and 0s is written to the SRAM; then check during /OE to see whether all data lines can output a mix of 1s and 0s too. If the latter is not the case, it would clearly point to a faulty SRAM, I guess. If the former is not the case, a faulty SAA7020?

[shakalnokturn]

I don't think you need much more proof, it isn't a high speed device and should only run warm, the SAA7020 should be hotter than SRAM.

There are a variety of other references that you can use as substitutes, it would be worth investing in a socket if you do remove it.

If you need more proof check all the data lines with an oscilloscope (then address lines if there's nothing suspect on data) look for any too-low high level or too-high low level.

The TMM2016 datasheet states 50mA active current so if there's a RC filter with fusible resistor to the SRAM you could also measure current there just to make sure it is drawing 400mA instead of 50mA.

To me the heat and symptoms alone would justify replacement.

[ebastler]

The SRAM was indeed bad: It could not pull any of its data outputs high properly. So I finally stopped fretting, unsoldered the SRAM and put in a socket and a NEC D4016C I still had around. And voilà -- sound is back!

I'm keeping fingers crossed that the sporadic crackling won't return either. But it seems plausible that a problem at this stage of the signal chain could cause the observed symptoms (noise on both channels at once etc.). Hence I am optimistic that the SRAM was the root cause for the sporadic noise as well.

Thanks again to all of you for the advice and encouragement!

[Gyro]

You might want to consider some small  stick-on heatsinks. Those early Philips chipsets run pretty hot.

[fzabkar]

Mine had a ground plane, but it WAS a pre-production model.  Didn't even have part numbers on the laser and such.  So you are good if you have a single sided board.

I had this same problem on a Philips CD, although I can't remember the model. ISTR that it sometimes spun the disc in reverse, but ICBW.

[Runco990]

I have also had the RAM problem, but on a later model.  Once you mentioned that it did this when NOT reading a disc... well, it became pretty clear.   Information is critical.

Glad you fixed it.  These players are worth saving.  If it still plays a CD-R, it's a quick test if the laser is healthy.  Who would have thought that this ancient machine would be able to play much later disc technology when many newer players could not.

I have a crappy, scratched CD-R I found on a floor as a test disc.  If it plays THAT, it's healthy!!
Weeds out marginal pickups real quick before I even break out the meter and scope.

[ebastler]

I have also had the RAM problem, but on a later model.  Once you mentioned that it did this when NOT reading a disc... well, it became pretty clear.   Information is critical.

Which is why I wrote it in the original post. 

For the sake of future troubleshooters coming across this thread it's worth repeating that in an unmodified player one would not hear any noise caused by a malfuntioning error correction while no CD is playing. Normally the Kill (Mute) relays would suppress the noise in that state. But I had shorted out these reed contacts earlier in my troubleshooting attempts, because I was suspecting them to cause the noise.

And, come to think of it, I am not even sure now that the sporadic noises (while idle) could really have been due to the SRAM. The total loss of audio clearly was and is now fixed. But is the whole demodulation and error correction logic even active while no CD is playing, or does it get disabled via some control signal? I will take another look at the schematics and maybe do some more measurements -- mainly out of curiosity.