Hi all,
after a long time, something new (or rather old) on the bench: a very high-end Mark Levinson CD 37!
This transport is highly praised for its build quality and smooth audio output. However, even the best build has its weaknesses, and it's no different here: while the CD drive subchassis, the laser, the circuitry, mostly everything is in really good shape, the tray mechanism suffers from multiple issues.
A video of its behavior is shown here:
So this is what we are seeing:
- tray not moving out at all on pushing the STOP/TRAY button, immediately beginning to read the CD TOC again
- tray moving out just a little, still clinging on to the CD and reading / playing it
- tray moving out in an irregular motion and rather sluggish
The owner states that he sometimes had to push the STOP/TRAY button twenty times to make the drawer come out eventually.
My first guess was a weak rubber belt in the tray mechanism so we ordered one from https://nanocamp.de/?belt (manufacturer and model have to be picked each time). It seemed to be the only place that has belts from current production built to the specifications required. The player's owner quickly found that he would have a hard time to change the belt. It's not as straightforward as one might think.
I offered my help and here it is. Removed the cover and found that the tray can easily be pushed out with a finger, and it helps to complete the process that is triggered by the button. A clear indication that the little motor tasked with moving it all is a little overwhelmed.
Once the tray is out, the motor, rubber belt, and transmission becomes accessible:
The red O-ring type belt (round shape) is my prime suspect. After more than 30 years, it is expected that the belt has lost some of its grip, softness and friction. So replacing it is always a good thing to do.
There is some challenge hidden in replacing the belt though. Unlike many other players, this belt is not running on "open" wheels on both sides. The larger transmission wheel sits on an axle that spreads across the entire width of the chassis. There is no way to get the old belt out or a new one in without some disassembly.
Here's a different perspective of the tray with the location of the axle marked yellow:
- the motor rotates at a rather high speed
- the belt transfers the rotation to the larger wheel, thereby reducing the rotation speed
- the larger wheel has a small gear which connects to the timing belt
- the timing belt is attached to the tray on the left side and will take the tray with it in either direction
- a little unconventional is the use of two belts. We need to take care of only one though as the timing belt basically lasts forever
If you ever wondered why they use belts at all, or why they don't use timing belts everywhere, this is because the tray might be mistreated by the user by pulling / pushing it hard which might overwhelm mechanical components. The rubber belt provides friction up to a point and then begins to slip through. This point is what protects the rest of the mechanics. Good belts will slip through at a rather late point (~ high load) and transfer energy reliably below that load. This point will move down to lower loads over time as they age.
Tools needed
- PH1 screwdriver to open the case
- PH2 screwdriver for the screws holding the tray/drive chassis and the spindle bridge in
- FD4 flat-head screwdriver for the stainless steel rod endcaps
- 1/16" hex driver for the countersunk screws that connect the tray and the rod glider beneath
- high-temperature grease
- new belt
- Isopropyl alcohol
- some wipes and cotton buds
Precautions
Mark Levinson products are, as almost anybody knows, very valuable, each cost a fortune. So it's worthwhile considering all steps in the repair with time and care and not hurrying through just to get it working quickly.
For instance, parts you remove and put aside for the time of the repair should be placed somewhere safe, screws should be kept in groups to ensure you don't mix them up later. The place where you work should be dry and as far as possible free of dust. If you do the repair in a phase where the weather is cold and dry, which you may learn from getting buzzed all the time when touching metal surfaces, be sure to take ESD precautions! It's a good measure in general but particularly important if the ambient promotes static build-up and discharge.
The surface finish of this player is anodized Aluminum with sharp 90° edges which are very prone to abrasion. The slightest mishap will let silver Aluminum shine through and ruin the expression. I wouldn't claim the hardware needs to be touched only with white gloves but be aware that surfaces should be protected or covered from accidental harm, especially on the front side.
The device has a mains input, therefore, up to 240 Volts AC can be found inside. You need to know what you are doing while you work on the device while powered. This can be *very* harmful and you do this at your own risk. If you have doubts about whether you can handle this, better turn towards an expert before you hurt yourself!
The purpose of this article is to show that this is not everybody's simple task to do. Besides the high voltage, there are also delicate electronics inside which are virtually irreplaceable. If anything gets destroyed along the way of the repair, it will be a very hard time finding replacement components! You need to ensure that everything is kept in the original (or better) shape, and that you avoid even touching components that are not part of the issue.
Phase 1
To get further into this mechanism, the tray and drive assembly needs to come out. There are three screws holding it in place, one is near the motor at the top (the little black PH2 screw north of the motor in the picture above). Careful, all these screws have washers beneath them that are easily lost!
Also, be sure to mark the plugs where the drive/tray unit connects to the mainboard. There are two 4-pin connectors (for tray motion detection with LED transmit/receive units) of equal dimension, and another two 2-pin connectors, one for the tray motor, another for the CD spindle. They might easily be confused later and while the LED sensors might not cause critical issues if swapped accidentally, the motors might need totally different voltages and power. Better don't risk it and give them labels or unmistakable markings.
Two more PH2 screws near the front can be found beneath the tray which needs to be moved out of the way to expose them:
Once the drive (and its connections) is loose, you can pull it out towards the back carefully. The drawer should be fully retracted as if a CD was set and playing. That way, the part that reaches through the slot in the front panel is as short as possible. You need to do this very slowly and bit by bit as the slim tray front bezel will hit metal along the way. If you are too brutal, you will cause abrasion there that is hard to undo later.
With the chassis out, we can now get access to its side where the CD "bridge" can be removed. This makes it easier to deal with the tray later. Remove two PH2 screws with shake-proof washers on each side of the bridge, then you can lift it off. Don't lose the washers.
When moving the tray manually, I found that it has some resistance it probably shouldn't have. The grease on the two stainless steel rods is probably dried-up. To replace it thoroughly, the rods need to be pulled.
ML designers made this really easy, lucky me. There is one hex / blade screw per rod to be undone, and loose it is.
Once both are loose, they can be pushed in a little by the white plastic end caps, then lifted up at a shallow angle, and pulled out towards the back. In the first step, I just pulled them back far enough to have both ends loose so the tray can be flipped up:
However, you can also see that one of the "gliders" is connected to a timing belt and it feels wrong to leave it under this tension. The gliders can be disconnected from the tray with two 1/16" hex screws on top of the tray surface. This makes it way easier to handle things:
The rods have O-ring seals at each end to dampen tray motion and prevent it from smashing into the metal body of the chassis. They need to be detached at the front side in order to pull them entirely away from the gliders. As luck would have it, one broke immediately:
I replaced it with some from my stock that seemed to be a good fit. No big deal.
Next up: cleaning the guides riding on the steel rods from old grease. I used Isopropyl alcohol and cotton buds to get out as much as I could. Did this from both sides of the guide as it is too long for one bud to reach through completely:
The rods were treated with IPA as well and it looks like it was time for a good clean:
Now for that rubber belt: to my surprise, the axle bolt can be pushed out sideways rather easily, there is no end stop, retaining clip or anything to keep it in place, it just sits there.
So I went and pushed it in from the side where the gear sits. Gently at first to ensure I don't crush anything or slip off:
Once the tool has pushed the bolt in far enough beyond the white plastic gear, it should eventually come off.
Space is tight so it takes some wiggling to get the gear out:
Comparing the old (left) and new (right) belt. Interestingly, the new one has a square profile, not a round one like the original. It's made from Silicone and will hopefully last a long time.
Once the new belt is on, the gear can be put back in its position to translate motor rotation to the timing belt, and the axle rod pushed back in carefully. It needs to align precisely to travel until its end position.
I found that there is some skew between the motor wheel and the large white plasic gear. The motor is suspended by two screws in rubber sockets, and tightening them up a little lined up things pretty well. Sorry, forgot to take pictures of it!
Reassembly is really easy. Just push each rod into its front opening, then align it so it runs horizontally, and pull the white end cap (along with the entire rod) towards the end. The little white nipple helps guide it near the end position.
There is an alignment aid as it seems. In the tray chassis corner near the IEC connector, there is a large drilled hole in the metal, and a round thread beneath it at the bottom. Both should be on top of each other. I had to illuminate this from below to illustrate, hopefully it becomes clear:
Phase 2
Now, all set? Not quite as I needed to find out the hard way.
After reassembly, the big moment ended as a big disappointment: the tray wouldn't come out, and a whining sound from the tray motor was the only reaction to pushing the TRAY button - not a good sign. So that tray is still moving too hard and the new belt cannot transfer the energy required. Actually that was worse than the initial situation.
First attempt to resolve this was to clean the belt while it moves with another IPA-soaked cotton bud. Unfortunately, that didn't make things any better. It would still slip when the tray should open. The other direction was flawless though.
I re-centered the white wheel on the axle to ensure it won't rub against anything. Added a lighter oil to the grease on the stainless steel rods, moved it all back and forth a few times, greased the hook near the back where the tray engages to lift up the pickup chassis once the CD is in place, even measured the driver voltage during operation of that motor. It all seemed reasonably okay, only it still didn't want to open.
Along the way, I found that the white endcaps of the steel rods have a lot of play before the screws are tightened. So it struck me: if the rods do not run exactly parallel, that might add a *lot* of friction. So I tried to align them. It might have been an engineering feature that they have so much play eventually. But whatever I did, fixing the screws would move the rods just a tiny little bit to misalign them, no matter how often I tried. So the final solution was to leave one of the rods with an endcap that isn't fully tightened. The screw is in deep enough it will probably never loosen itself, and this way the rods will align with each motion automatically.
The reward was that the tray now moves in and out as it should! Job done! It leaves a bit of an aftertaste that the new belt is a bit weaker than expected despite being manufactured exactly for the CD37 and CD39 drives.
Another funny thing I discovered is a typical workbench thing: the tray would move in and stop half-way to go back out in my first tests after loosening one of the rods. Turns out it was my workbench light that confused the sensors in the player!
https://youtube.com/shorts/k9QL2Rg0_6c
Eventually I turned the main light off / down and retried, and ever since, it's all working!
