Quality CD-Mechanisms are long gone - let us build one ourselves!

What I like about the old-school CD player is that every part of the system is dedicated to the task of playback. A computer based system is far more complicated, where audio data is transferred over the same lines as the other data from memory, mouse and screen. Not to mention the software processing that is taking place. But even in a multi-tasking environment it still seems to work, just to difficult me to fully understand at a nuts & bolts level.
 
Last edited:
@stocktrader
correct me if I am wrong, but the iphone always automatically upsamples to 48kHz, even with 44.1 material via airplay. This was at least what I found out when trying to find a suitable base station.
Using Macs professionally as an editor and sometimes sound designer for 20 years I was a late adopter of the ipad and iphones and surprised how many sensible features were gone in favour for a strange concept of convenience.

But it is an undisputable fact that mechanical quality was better in first generations of players AND computer drives. Sometimes some deign flaws but durable.
The first super drive (made by Pioneer as far as I remember) features a brushless motor with 4mm shaft and a metallic cone on a spring for precise centering of the CD/DVD.
It also has a head with two lenses/lasers dedicated for the task of reading CD / DVD data. I bet eyepatterns were great back then.

This is what this thread is about:
Building mechanically reliable mechanisms that will outlive today´s crap.

But please folks: Do not let us waste space and pages about the pros and cons of this Idea. Just about how to get there!
Anyone in this forum who has contact to companies to provide parts? Maybe someone is around who was working on the development of laser mechanisms?


As promised I will check the Toshiba OPH-32 head and A Nak´s HLPL 030020 / CA80255A . But it will take some time
 
It is about understanding how a system works before trying to make improvements. The cheapest drive works flawlessly, it reads the pits and if the disc is not to damaged it will correctly output the data to the DAC. To me a good looking eye pattern is a bad design goal. What you should be looking for is system that can read a wide variety of discs, including things like reduced reflectivity or pit geometry. Adding different read strategies could help here.
 
The cheapest drive works flawlessly, it reads the pits and if the disc is not to damaged it will correctly output the data to the DAC.
You do not put bad/cheap 50$ optics prone to optical interfernce with bad latitude in front of a good DSL that covers 15 stops or do you?
It is not the goal that flawless discs can be read.
Dirt and sratches or manufacturing errors are noise.
A clean eypattern will give a good signal to noise ratio.

Goal is optics and a mechanism that works flawlessy for decades and reads bad errors.
Optics that can be cleaned. That can even give the possibilty to replace a burned out diode like a stylus on a record player with simple alignment.
Mechanics that will not add to the growing pile of eletronic garbage.
Goal is to squeeze the best out of bad discs. Get good glass and precise optics and give error correction lesser work.
After this is -mechanically- achieved, adapt any read strategy you want.
 
Last edited:
I seem to recall seeing Airplay upsample to 48khz which would create some artifacts.
I am using the audio headphone jack and local files for playback so I believe it's 44.1khz for mp3 and apple lossless files

Me too, but I use wav and vlc + file explorer for playback.

Let´s go back to topic.
Any help from people who were working in development of CD-players welcome.
Again no patents to break but many questions.
Also any company working in the field of solid-state lasers and
laser optics are welcome to benefit to this thread!
Anyone out there with some definately broken 1rst 2nd generation lasers out there? Would be nice to take them apart and to scan them. Ipromisecd to look at my lasers and will do the next days, but I fear they might still work and should be saved...
 
Administrator
Joined 2004
Paid Member
Hi stocktrader200,
The data has to come from somewhere to reach the memory first. If that area on the disc can't be read without errors, then the data stuffed into the buffer isn't what was on the CD.

If you are reading data from some object, then you are at the mercy of how well the device is operating. What the "skip-proof" memory did was enable the mechanism to retry a failed read. This only works in the cases where the drive was jolted, or the read error was a temporary thing. Lint on the disc for example. If the defect is permanent, like a scratched disc for example, that data is lost to you. The better the alignment is with the mechanism, the fewer cases like this will happen.

The fact remains, you need a working mechanism to read media. The better the mechanism is at reading the information, the more tolerant it will be for minor defects. Remember, the data does not exist until it has been captured / created by whatever mechanical means you depend on.

It really is that simple.

-Chris
 
I3 and I11 to ITOP, Symmetry, Radial Noise, Push-Pull Magnitude, and Cross Talk can be extracted from the eye-pattern or the HF signal.
You can also test land/pit jitter for each land/pit length. The problem is that it is not easy to do and most of this cannot be done by just looking at the eye-pattern.
 
Administrator
Joined 2004
Paid Member
Hi Mark,
You can also test land/pit jitter for each land/pit length. The problem is that it is not easy to do and most of this cannot be done by just looking at the eye-pattern.
Just looking at an eye pattern can give you a very good idea how things are. Poor pit shape does distort the waveform, and jitter causes the right hand zero crossings to occur at different times, widening that area of the trace. Pretty much anything is there to see. Now, if you need to place numbers on things, they make a Jitter meter, and most current digital 'scopes will also measure these things directly off the eye pattern. The more expensive DSO's have a host of applications to characterize the quality of an eye pattern too.

Keep in mind that CD/DVD eye patterns are just one small area in electronics that use an eye pattern. Every communications link is characterized by examining the eye pattern, and now also the constellation pattern. 256 quam being a challenging specification to adhere to.

In order to minimize errors in data coming off a DVD or CD, you optimize the eye pattern. Only when the eye pattern is severely stressed at the best point would you begin looking at the C1 flags. The current standards deemed acceptable for CD eye patterns are so low that the early machines wouldn't have worked. This is something that has everything to do with the improvements in servo and auto-adjusting servos in the current products. DSP doesn't even come into play yet. Memory only receives data after the DSP processes the EFM signal, not the RF signal (eye pattern). So, you are looking at things well after the fact. Johnny-come-lately as it were.

Remember, the eye pattern is how the signal comes off the CD / DVD. You can't get any closer to that data unless you wish to examine the pits under a microscope. Therefore, it is the eye pattern that is the predictor of signal quality. This is by definition. Poor quality eye pattern = busy DSP and C1 series flags. Eventually, the signal may degrade to the point where the C2 series flags are active permanently. The only conditions that are worse entail an eye pattern so poor that the spindle servo can't even lock to the disc. Worse yet? Sure, discs where the focus servo can't even find the surface of the disc. Pit shape can be so bad that the focus locks, but the E-F tracking diodes can't find the track reliably and may run off to either end of the head's travel.

-Chris
 
Take into consideration that BLER-E22-E32 and the other digital data are the results of the processing of the RF signal. It is a digital representation of the RF signal and one that really matters.

If I test a disc in several players, the player that has the lowest errors (BLER-E22-E32) will be the better player. You will not get this clear data from the eye-pattern.
 
Administrator
Joined 2004
Paid Member
Hi Mark,
Take into consideration that BLER-E22-E32 and the other digital data are the results of the processing of the RF signal.
But, it isn't the result of processing the RF signal. The only thing the DSP is sent is the EFM signal. While it is also an RF signal, it is a limited and processed signal that doesn't reflect what is going on except in the very broadest terms. The EFM signal is always a consistent strong signal with well defined edges. An eye pattern on the other hand has constantly changing levels with lower and variable, less defined edges. The two signals are only loosely related to each other. Take an oscilloscope and look at the two wave forms and you will see what I mean. The envelope of the eye pattern will show you the surface conditions of the disc. You can also see the actions of the tracking servos in the eye pattern envelope. This is really interesting stuff.

-Chris
 
Administrator
Joined 2004
Paid Member
Hi Mark,
There are some things to check at the input at times.
Processed by C1 and C2, no DSP here. Would you test an amp at its input?
Would you test a CD player from the EFM input and forward? Because that is what you are saying here. I don't want to draw a parallel between a CD player and turntable, but you wouldn't service a turntable without operating it without playing a record - would you? How can you possibly separate the digital areas from the analog areas and expect to learn anything about the entire operation?

Taking the information off the media is the most critical phase in playing a CD or DVD. Why then would you exclude the most problem prone areas from a test? I can guarantee you that servicing a player without looking at the mechanism and analog wave forms isn't going to tell you much at all. Even the sled drive waveform, as well as the tracking and focus drive wave forms can uncover faults that you couldn't find via any of your digital test points. All you would know is that there is a problem somewhere, the rest being a total guess.

You can tell so much from the eye pattern that you can even determine what faults a disc has when using a properly working player. Suffice it to say that unless you are looking at the very first points where the information you want is taken from the disc, you will never know what issues a player might have until it causes problems. By that time, the machine is no longer functioning. The folks who look at the RF (eye pattern) can see trouble coming in advance. If that isn't enough to show you where to look for answers, I don't know what will.

I may be mistaken, but I think you are looking at test points that are accepted for the purpose of determining if your product is okay or not. Just an error count instead of a qualitative standard. If your error counts are below a certain number - you're golden. One of the first comments you made referenced this as being an acceptable procedure. One wonders if this is an industry litmus test, or one that is internal to your company. What I can say with certainty is that a high level test (abstracted) isn't acceptable for the service industry for CD or DVD players. You need to satisfy the manufacturer that his player produces an acceptable eye pattern. The standard for which is completely up to the manufacturer of the unit in question. One thing is certain is that as the eye pattern improves, error rates also drop.

-Chris

Even hard drives begin with an RF eye pattern. The
 
I am not suggesting you should test a CD player from the EFM input and forward, but from C2 backwards.

I can also see that if you are servicing a CD player that you would check the eye-pattern as it would show a problem with the laser or optics.

You keep saying that you can see so much from the eye-pattern, please give some examples.
If you had a few discs and 10 players, would you be able to categorize the players with the use of the eye-pattern only? I could do this quite easily using the error counters.
 
Administrator
Joined 2004
Paid Member
Hi Mark,
I am not suggesting you should test a CD player from the EFM input and forward, but from C2 backwards.
I would say that you have the direction backwards. C2 says you have a (big) problem, but what is it at the root? Or, you might conclude that there are no problems at all, not having any C2 flags. That might cost millions of dollars on a production line and a hit to the company's reputation. You don't have enough information.
I can also see that if you are servicing a CD player that you would check the eye-pattern as it would show a problem with the laser or optics.
In addition to the laser and optics, you can check the condition of the motors, mechanical alignment of the sled the head rides on and the servos. I would normally check in succession, eye pattern, eye pattern envelope, tracking servo, focus servo, spindle motor servo. Checking these things will also tell you what condition the mechanical parts are in. This series of tests does not take that long to accomplish and you now have a wealth of information about the machine. You could easily then check E-F balance and the VCO free-run frequency.

Generally speaking, problems in the digital sections are rare, but would be easier to deal with if you know the rest of the machine is clear of faults. But why check all of this? Simply because customers will not tell you if they or someone else opened the machine and twisted any controls (they always seem to). Plus, machines tend to accumulate multiple problems because people don't bring things in until they stop working.

Say you had a bad DSP chip, surface mount of course. So you get the part, successfully install it and spark the machine up - only to find it has a dead laser head (because someone turned up the APC control and fried the laser). The DSP chip can easily be $70 or more, heads from known stock are the same money or more. Oops!
You keep saying that you can see so much from the eye-pattern, please give some examples.
I have, several times already. You can see focus problems, disc motor problems, disc motor speed sync problems, azimuth problems, tracking servo problems and sled motor problems. Lastly, you can easily see eye pattern strength problems. However that can be anything from a contaminated lens, low RF gain or weak laser output. OF course, the CD might have light silvering and the laser beam just keeps on going until it hits another object, but it isn't coming back! You can see also defective pit edges and surface reflectivity problems that also distort the wave form. Pinholes will put notches in the eye pattern envelope and cause a glitch in the tracking servo waveform, as would scratches (but you can hear the steady "tick" noise every time it comes around).

I'm not sure what you are looking for Mark, but this list isn't a conclusive one.
If you had a few discs and 10 players, would you be able to categorize the players with the use of the eye-pattern only? I could do this quite easily using the error counters.
Absolutely! And they can be graded more finely too. If I wanted to specifically grade their DSP and error correction, I would look at the flags for error correction if they were available as test points. Additionally, by doing the tests, you would end up with 10 properly set up machines as they could be adjusted, and so be operated fairly at their peak performance. That way you would be fairly grading the actual best case standings of those machines.

Far more expensive equipment is graded and ranked on their technical merits using eye pattern quality. Their operation is periodically checked using - eye patterns. This is even true of optical data paths.

The eye pattern is a basic measure of health for all data transfers. Even the USB channels are tested via their eye patterns with minimum standards achieved with the use of masks for go and no - go areas. The same masks exist (or existed) for CD player eye patterns and DVD eye patterns. A basic test for quality that the device is working within approved limits. These limit tests offer a very fast pass or fail test for commodity items (like CDs or DVDs, including their players).

I still work on CD Players, and I look forward to this as the performance of the end product depends upon my skill. If I could still get the parts I would also enjoy working on tape decks and reel to reel players. Of course working on preamps, amplifiers and tuners is also something that is usually enjoyable. If each device only had a form of error counters, there would be no enjoyment because you could never tweak the setup to the fine levels as is possible with the raw signals. I even enjoy servicing test equipment, and for the same reasons. Alignment is usually more arduous than anything else though. There is nothing like a few error counters in test equipment repair and cal for one simple reason. Those indicators will never give you absolute confidence that things are in act operating as they should be.

-Chris
 
http://geertlangereis.nl/Publications/Papers/JJAP2001.pdf

I want to know how to identify playback problems by looking at the eye-pattern only. I found this example of adding a second signal within the normal signal. The T5-T11 pits and lands are altered by adding small pits to the lands and small lands to the pits. On page 1712 you can see that this has almost no impact on the eye-pattern. After reading this paper I am left with one question. Does it add errors? And that could only be seen by looking at the digital error correction.
 
Last edited: