When the first CD came out
There's a reason, while people are busy talking about the new formats and DVDs, to bring up stories of when the first CD came out. I think the sound of optical software has essentially unchanged since then.
At the beginning of the CD era, people often said, "This will be the end of audio reviewing," or "Now there won't be any difference in the sound of any manufacturer's player we use."
Nobody would say that these days, but the gap between our anticipation toward the optical software and its reality still remains wide. There are people who still listen to music only through analog records.
I also feel that the digital playback is still having a hard time reproducing the inner voices (alto and tenor in case of a mixed chorus in four parts, viola in the case of string quartet). It sounds thin. Some people say it depends on the recording, but I decided not to complain about recording a long time ago. There are good recordings and bad ones, but it's out of my territory. I don't know the details of each recording. If I thought the recording is bad, I just wouldn't listen to it again.
With that in mind, I've been working on the developing of a CD transport for over a decade. Somebody estimated my way of working as "80% of intuition and 20% of science", and that's probably right! I always felt the mechanical construction is the weakest link in the CD playback since the first CD appeared.
In case of the analog turntable, it is said to be ideal for the base to have 10 times the mass of the platter. This is a tough number. If the platter weighs 4lb, the baseboard that supports the platter should have 40lb of mass. How about the ratio of the mass between the diamond tip of a cartridge stylus and the arm?
80% of instinct
I'd been working on a mechanism with which the laser head can read the pit at 90 degrees straight through. My instinct, almost a belief, told me that the sound would be far improved if the lens could follow the pit without moving its angle. It was quite sometime after that people started talking about the effect of servo current and jitter.
Also, if we could have a large ratio between the mass of the moving parts (lens and coil) and that of the solid support, the tracing and tracking ability must be improved too. You can see how the tracking ability affects the sound by checking the popping noise when you play a scratched disc. If the noise is a loud "pop", you can't expect a good sound reproduction. When it makes a light "pi", the sound is pretty good. In the case of the analog disk, when the scratch noise sounds like "buh" or "btch", it is the proof of a bad transient characteristic. The transient characteristic can be replaced as "speed" in more modern terms. The same thing can be said about a DAC and a phono equalizer in relation to the "buh" and "pi".
CDs are made quite a bit rougher than you'd expect. For example, one disc can be off centered more at the inner groove and another can be off centered at the outer groove. Some of them are even more off centered at the middle! Some are more warped than others. The lens is busily changing its angle and focus to adjust to the off centering and warping. Setting the disc on a coin sized platter and holding it with a magnet or a stabilizer of a couple of hundred grams can't correct the warping at all. Even if there's no warping, we can't check the angling and unevenness of the platter surface from out side of the black box.
Problem with the head moving mechanism
There are a couple of different types of head moving mechanisms in conventional CD transports, one utilizes a couple of flat gears, another combines a warm gear and flat gears to reduce the speed, and some use a linear motor. However, in most of the cases, these mechanisms are set inside a black box and we can't see what's going on in there.
Some years ago, while working on the PiTracer, I was surprised by the tremendous amount of vibration when I tried to pull up the whole machine by holding the guide rail. The vibration of the lens (with the actuator coil) has such an incredible energy and I realized that we'd need a great chunk of mass to accommodate it.
Now, running off the track, everybody knows an analog LP starts playing from the outer groove, but it seems many are not aware that CD starts from inner groove. You can drive a car not knowing the mechanics of the engine or the steering wheel system, but for an audio hobbyist, this kind of knowledge can be crucial.
The realization of mechanism by "20% of scientific mind"
To prove the 80% of intuition, you have to make it into an actual form. You can't draw a schematic with an intuition, or choose the parts. The electrical parts are basically standardized and come with certain limitations. In other words, you don't need to wonder around too much there.
However, to design a mechanical part by yourself is a different story. Your past experiences can be an obstacle. We tend to be conservative when it comes to our own specialty. When somebody comes up with an idea, you can counter it immediately with 10 different reasons why it doesn't work! This happens with every profession and we can't give birth to something new without overcoming this phenomenon! That's a warning I gave to myself while designing PiTracer.
Since the laser pickup is something that vibrates, I gave a big mass to the part that accommodates it. At the same time, I wanted to minimize the vibration that is transmitted to the laser pickup from outside. These outside vibrations would be, 1) from the rotation of the disk, 2) from the head moving mechanism (except those that are transmitted through the floor and air). PiTracer has a transformer in a separate casing so the vibration of the transformer is a separate issue.
There are also lots of small metal parts on every audio component, bolts and nuts and other connectors. These small metal parts can be a real problem sometimes. (If you are a motorcycle enthusiast, you'd know there's a method, by attaching a weight at the tip of the handle bar, to cancel out the vibration of the engine by the resonance of the weight and the handle bar.) Those small parts can add resonance to create a complicated vibration mode affecting the sound as a result. So I'd like to minimize the overall amount of the parts too.
Now, let me describe each part.
Machined out from solid aluminum block (320W x 360D x 30H mm). The spindle motor, sled motor, warm gear, V guide track, motor drive amplifiers and output terminals are mounted to it.
2. Spindle motor
The spindle motor to rotate the disc is a coreless 9 pole motor. This motor has a small inertia and very efficient which is most suitable for CD transport that has to change the speed from 500 to 200 rotations in about a minute and also has to repeat stop and go in a micro scale.
This motor is directly attached to the baseboard that weighs almost 10kgm. By the way, the platter weighs about 90grm with a CD on, so the ratio of the mass of the rotating part and the baseboard is over 100 times which, I suppose, would be sufficient.
The platter is machined out of acrylic. Only the inner clamping area and the outer edge touches the disc. The disc would be clamped down at the center by an acrylic clamper. This makes the disc flattened out except the unevenness of the thickness of the disc itself. You can see the effect of this system by looking at the reflecting image on the rotating disc surface with which you won't be able to tell whether it's moving or not.
The reason I made the platter and the locking clamper by acrylic is to minimize the unnecessary inertia, to eliminate magnetic effect on the actuator coil, and most importantly, after the motor is fixed to the platform and the acrylic platter is pressured in, we can re-machine each platter one by one while it's rotating to ensure the flat plane of the surface. By this method, the aforementioned flickering of an image on the rotating disc is further minimized.
4. Head moving mechanism
This is the biggest feature of this transport. How to move the 1.5kgm of mass smoothly, and with a quick enough response without vibration? My 10 years were spent mostly to solve this issue.
In conclusion, I decided to drive the head with a combination of Kevlar thread (which doesn't stretch by time) and a drive shaft decelerated by a warm gear. This minimizes the already reduced vibration of driving motor to be transmitted to the head.
The driving motor itself is exactly the same one as the spindle motor. The high starting sensitivity (another feature of this motor) is the reason I chose this motor for this purpose. You can see the motor giving a shake to adjust to the off centering of a disc by the movement of the warm gear placed at the tip of the motor shaft.
5. Installation of the head unit
The head unit has to be installed to the exact location in relation to the platter. Once the guiding rails and the bearing wheels are placed and the smooth motion is confirmed, they are finally attached to the accurate location using a specifically designed device. The rails and the bearing wheels are made by Bishop-Wisecarver in the US. The surface of the rails is whetted by three different grades of whetstones to further ensure the smooth motion.
The way the thread is installed is decided after numerous experiments. It is angled at 90 degrees twice toward where it's attached to the head unit and pulled by a spring at the end. With this method, the drive of the sensitive motor is transmitted to the head unit efficiently and smoothly.
Going back to the original topic, what factor changes the resulting sound in case of a CD transport? In case of PiTracer, even the tension of the thread changes the sound! It is a similar difference when you change the tension of a belt-drive turntable. The only thing I can say for sure is that the entrance of the system is very important.
With a tight mechanism without much random resonance, you can get a rich reproduction of inner voices; the activity of viola is well reproduced. I can enjoy my most favorite piece, the beginning of Smetana's Strings Quartet, No. 1, B minor, "From My Life", with a same balance as in a concert.